Health, Risk, and Adversity 9781845458713

Table of contents :
Contents
List of Figures
List of Tables
List of Boxes
Foreword. Framing Health, Risk, and Adversity
Introduction. Health, Risk, and Adversity: A Contextual View from Anthropology
PART I • HEALTH RISKS AND DISEASE IN TRANSITION
Understanding Health: Past and Present
1. Health Consequences of Social and Ecological Adversity Among Indigenous Siberian Populations: Biocultural and Evolutionary Interactions
2. A Multidisciplinary Approach to Understanding the Risk and Context of Emerging Primate-Borne Zoonoses
3. Viral Panic, Vulnerability, and the Next Pandemic
PART II • GENERATIONAL AND DEVELOPMENTAL CHANGE
Thinking about Health through Time and Across Generations
4. Adaptation, Health, and the Temporal Domain of Human Reproductive Physiology
5. Changes in Risk Factors for Breast Cancer in Migrant Women: An Intergenerational Comparison Among Bangladeshis in the United Kingdom
6. Family Structure and Child Growth in Sub-Saharan Africa: Assessing “Hidden Risk”
PART III • GENE EVOLUTION, ENVIRONMENT, AND HEALTH
Explaining Health Inequalities
7. The Developmental Origins of Health and Disease
8. Beyond the Gradient: An Integrative Anthropological Perspective on Social Stratification, Stress, and Health
9. The Slavery Hypothesis: An Evaluation of a Genetic-Deterministic Explanation for Hypertension Prevalence Rate Inequalities
Conclusion. Adversity, Risk, and Health: A View from Public Health
Contributors
Glossary
Index

Citation preview

Health, Risk, and Adversity

Studies of the Biosocial Society Series Editor: Catherine Panter-Brick, Professor of Anthropology, Durham University, UK The Biosocial Society is an international academic society engaged in fostering understanding of human biological and social diversity. It draws its membership from a wide range of academic disciplines, particularly those engaged in “boundary disciplines” at the intersection between the natural and social sciences, such as biocultural anthropology, medical sociology, demography, social medicine, the history of science and bioethics. The aim of this series is to promote interdisciplinary research on how biology and society interact to shape human experience and to serve as advanced texts for undergraduate and postgraduate students. Volume 1 Race, Ethnicity, and Nation Perspectives from Kinship and Genetics Edited by Peter Wade Volume 2 Health, Risk, and Adversity Edited by Catherine Panter-Brick and Agustín Fuentes Volume 3 Substitute Parents Biological and Social Perspectives on Alloparenting in Human Societies Edited by Gillian Bentley and Ruth Mace

Health, Risk, and Adversity • • • Edited by Catherine Panter-Brick and Agustín Fuentes

Berghahn Books New York • Oxford

First published in 2009 by Berghahn Books www.berghahnbooks.com ©2009 Catherine Panter-Brick and Agustín Fuentes

All rights reserved. Except for the quotation of short passages for the purposes of criticism and review, no part of this book may be reproduced in any form or by any means, electronic or mechanical, including photocopying, recording, or any information storage and retrieval system now known or to be invented, without written permission of the publisher.

Library of Congress Cataloging-in-Publication Data Health, risk, and adversity / edited by Catherine Panter-Brick and Agustín Fuentes. — 1st hardcover ed. p. ; cm. — (Studies of the Biosocial Society ; v. 2) Includes bibliographical references and index. ISBN 978-1-84545-455-5 (hardback : alk. paper) 1. Health risk assessment. 2. Social medicine. 3. Medical anthropology. I. Panter-Brick, Catherine, 1959– II. Fuentes, Agustín. III. Series. [DNLM: 1. Health Status Disparities. 2. Disease—ethnology. 3. Disease— etiology. 4. Risk Factors. WA 300.1 H434 2008] RA427.3.H428 2008 362.1—dc22 2008020720

British Library Cataloguing in Publication Data A catalogue record for this book is available from the British Library

Printed on acid-free paper. ISBN: 978-1-84545-455-5 hardback

Contents

• • • List of Figures List of Tables List of Boxes Foreword Framing Health, Risk, and Adversity Alan Goodman Introduction Health, Risk, and Adversity: A Contextual View from Anthropology Catherine Panter-Brick and Agustín Fuentes PART I • HEALTH RISKS AND DISEASE IN TRANSITION Understanding Health: Past and Present Charlotte Roberts 1. Health Consequences of Social and Ecological Adversity Among Indigenous Siberian Populations: Biocultural and Evolutionary Interactions William R. Leonard, J. Josh Snodgrass, and Mark V. Sorensen 2. A Multidisciplinary Approach to Understanding the Risk and Context of Emerging Primate-Borne Zoonoses Lisa Jones-Engel and Gregory Engel 3. Viral Panic, Vulnerability, and the Next Pandemic Ann Herring Appendix 1 Was the 1918 Pandemic Caused by a Bird Flu Virus? 88 Appendix 2 Applying the Syndemic Approach: Whooping Cough at York Factory 90

vii x xii xiii

1

13

26

52 78

vi

• Contents

PART II • GENERATIONAL AND DEVELOPMENTAL CHANGE Thinking about Health through Time and Across Generations Darna L. Dufour

101

4. Adaptation, Health, and the Temporal Domain of Human Reproductive Physiology Peter T. Ellison and Grazyna Jasienska 108 Appendix Building an Integrated Concept of Health

123

5. Changes in Risk Factors for Breast Cancer in Migrant Women: An Intergenerational Comparison Among Bangladeshis in the United Kingdom Alejandra Núñez-de la Mora and Gillian R. Bentley

129

Appendix Breast Cancer Risk Among South Asians: Heterogeneity, Trends, and Prevention 144

6. Family Structure and Child Growth in Sub-Saharan Africa: Assessing “Hidden Risk” Daniel W. Sellen Appendix Poor Growth and Risk of Death

150

167

PART III • GENE EVOLUTION, ENVIRONMENT, AND HEALTH Explaining Health Inequalities 175 William W. Dressler 7. The Developmental Origins of Health and Disease Keith Godfrey and Mark Hanson

185

Appendix The Developmental Origins of Health and Disease (DOHaD) concept—a journey from maps to mechanisms to societies 202

8. Beyond the Gradient: An Integrative Anthropological Perspective on Social Stratification, Stress, and Health Thomas McDade

209

9. The Slavery Hypothesis: An Evaluation of a Genetic-Deterministic Explanation for Hypertension Prevalence Rate Inequalities 236 Lorena Madrigal, Mwenza Blell, Ernesto Ruiz, and Flory Otárola-Durán Conclusion Adversity, Risk, and Health: A View from Public Health Martin White

256

Contributors

273

Glossary

278

Index

286

List of Figures

• • • Figure A A web of causation representing interrelated risk and protective factors, with the agent (a spider) responsible for their root Figure 1.1 Map of Siberia showing the geographic locations of the Buryat, Evenki, Ket, and Yakut Figure 1.2 Mean (+SEM) (a) BMI (kg/m2) and (b) percent body fat of indigenous Siberian men and women living in different size communities Figure 1.3 Relationship between basal metabolic rate (kcal/day) and fat-free mass (kg) in indigenous Siberian men and women Figure 1.4 Mean (+SEM) physical activity levels (TEE/BMR) of indigenous Siberian men and women living in different size communities Figure 1.5 Total cholesterol levels (mg/dL) by age group for men and women of three indigenous Siberian populations (Buryat, Evenki, and Yakut) compared to the US 5th and 50th centiles Figure 1.6 LDL-cholesterol levels (mg/dL) by age group for men and women of three indigenous Siberian populations (Buryat, Evenki, and Yakut) compared to the US 5th and 50th centiles Figure 1.7 Systolic blood levels (mmHg) by age group for men and women of three indigenous Siberian populations (Buryat, Evenki, and Yakut) compared to the US 50th centile Figure 1.8 Diastolic blood levels (mmHg) by age group for men and women of three indigenous Siberian populations (Buryat, Evenki, and Yakut) compared to the US 50th centile Figure 2.1 Risk Analysis Figure 2.2 Route that performing monkeys may take as they are captured from the wild, transferred to live animal markets and then to their owner’s compound and eventually into urban centers

3 31 34 35 36

37

38

39

39 59

70

viii

• List of Figures

Figure 3.1 “Live-poultry markets like this one in Hanoi speed the spread of the virus from farm to farm when vendors take leftover birds back home, along with any flu viruses they’ve picked up.” Figure 3.2 Influenza death rates under age 45 (without infants), 1911–1945, United States Figure 3.3 Tuberculosis death rates under age 45, 1900–1960, United States Figure 4.1 The temporal continuum of adaptive response Figure 4.2 Differences in salivary progesterone profiles observed among Lese women in the Ituri Forest of the Congo Figure 4.3 Differences in salivary estradiol among Polish women with differing levels of energy expenditure in habitual activity Figure 6.1 Growth status of young children, 0-3.5 years, sampled in a Datoga community, Tanzania, according to number of co-wives of the mother Figure 6.2 Association of growth indicators with household wealth and marital status among surviving children, 0-11 year old, sampled in a Datoga community, Tanzania (n=188) Figure 6.3 Relationships between environment, growth, and other health outcomes Figure 7.1 Coronary heart disease death rates, expressed as standardized mortality ratios (SMR), in 10,141 men and 5,585 women born in Hertfordshire, UK, according to birth weight Figure 7.2 A model to explain the epidemic of type 2 diabetes in urban India Figure 7.3 A conceptual framework for the developmental origins of health and disease Figure 8.1 Exemple of Blood spot collection Figure 8.2 Association between household economic status (tertiles, based on parental occupation) and log-transformed EBV antibody level (mean ± SE) in Samoan adolescents (10 to 20 years), controlling for age, sex, region of residence, BMI, and current infection. Figure 8.3 Association between westernization experience (tertiles) and EBV antibody level in Samoan adolescents, controlling for age, sex, region, BMI, current infection, and SE.

84 86 87 113 117 118

153

157 161

189 194 197 215

219

221

List of Figures

•

ix

Figure 8.4 Association between status incongruity and EBV antibody level in the capital city of Apia, Samoa, controlling for age, sex, region, BMI, current infection, and overall social status 223 Figure 8.5 Interaction between skin color and income in predicting systolic blood pressure in African-Americans in the United States (age 33 to 45 years), controlling for age, sex, BMI, smoking, and use of anti-hypertensive medication 226 Figure 9.1 Genetic Deterministic Framework of the Slavery Hypothesis Figure 9.2 Probabilistic Framework for Hypertension Risk Inequalities Figure 9.3 Taking blood pressure measures in Puerto Limón, Costa Rica Figure 9.4 Estimates of hypertension prevalence across different studies for people of African descent in: (a) the United States, (b) the United Kingdom, (c) Africa, (d) the Caribbean, and (e) South America Figure 9.5 Median hypertension prevalence rate by region including both genders, all definitions of hypertension Figure 9.6 A genetic map of sixteen Afro-Caribbean groups based on gene frequencies

237 238 241

243 246 249

Lists of Tables

• • • Table 1.1 Indicators of growth status in Evenki children measured between 1991 and 1995, during the “Soviet” and initial “Post-Soviet” periods Table 1.2 Adult nutritional status of Evenki men and women (>18 years) measured between 1991 and 1995, during the “Soviet” and initial “Post-Soviet” periods Table 1.3 Percent overweight and obese among indigenous Siberian populations Table 1.4 Body weight, basal metabolic rate (BMR), total energy expenditure (TEE), and physical activity levels (PAL) of adult men and women of native Siberian populations Table 1.5 Multiple regression analysis of the correlates of systolic blood pressure in native Siberians Table 1.6 Multiple regression analysis of the correlates of LDL cholesterol in native Siberians Table 1.7 Comparison of metabolic and cardiovascular health parameters for indigenous Siberians and Pima Indians Table 2.1 Characteristics of human populations that influence infectious agent transmission, attributes of the environment that influence crossspecies transmission, and characteristics of animal populations that influence infectious agent transmission Table 4.1 Mean morning salivary testosterone levels in men from four different populations subdivided by age. The populations differ in their average values over all ages and at young ages, but not in the oldest age category

32

32 33

36 40 41 44

65

122

Lists of Tables

•

xi

Table 5.1 Endocrine-Related Risk Factors for Breast Cancer Table 5.2 Groups in the Bangladeshi Migrant Study

131 135

Table 7.1 Mechanisms linking developmental influences with the risk of cardiovascular disease and type 2 diabetes in later life

193

Table 8.1 Measuring stress: Advantages and disadvantages of using physiological indicators

213

Table 9.1 Hypertension prevalence rates by region according to several definitions of hypertension, both genders Table 9.2 Hypertension Prevalence Rate by Gender and Region

244 247

List of Boxes • • • Chapter 1 Ethnographic Background of Select Siberian Populations Unresolved Issues/Future Directions for Research Summary Points

30 46 47

Chapter 2 The Risk of Pathogen Transmission Components of Risk Analysis

55 58

Chapter 5 Characteristics of Migrant Studies

130

Chapter 6 A Polygynous Community in sub-Saharan Africa Growth as an Indicator of Health, Risk, and Adversity

152 160

Chapter 7 Evolutionary and Theoretical Perspectives: Predictive Adaptive Responses

187

Chapter 9 Two Contrasting Views on Hypertension Rates in People of African Descent Summary Points

237 252

FOREWORD

• • • Framing Health, Risk, and Adversity Alan Goodman I think that the tendency of applied science is to magnify injustices until they become too intolerable to be borne, and the average man whom all the prophets and poets could not move, turns at last and extinguishes the evil at its source. —J.B.S. Haldane, 1923

A

fter completing a large-scale review of worldwide variation, James Tanner, the preeminent British expert in human growth and development, surmised that mean adult heights invariably increase with greater socioeconomic status. This association, he stated, is consistent for all historical periods and in many dozens of countries around the globe. To say it more grandly still, everywhere that Tanner looked, and he looked in a remarkably large and varied number of places, socioeconomic status was indelibly written into and onto the body. During the 1960s, when Tanner first summarized his findings, the links between socioeconomic conditions and biological outcomes were less well studied than they are today. Now, thankfully, there are burgeoning fields of social medicine and social epidemiology, devoted to understanding the social, economic, and political origins of illness and disease. As well, anthropologists have begun to reconsider how social and economic conditions inscribe themselves onto bodies. As the editors of this volume point out in the introduction, leading theorists and practitioners of social epidemiology have called for studies of the political-economic and social origins of disease,

xiv

• Alan Goodman

or as Nancy Krieger (2001) proposes, an eco-social theory of disease. In this new human biology, group height and health have their origins in social interactions and institutions. This insight about the social origins of biological difference and suffering is profound for human biologists who have gotten used to looking toward proximate physical and biological explanations, rather than distant social explanations. Yes, tracing back through Boas, anthropology and human biology have a long history of efforts to link social and economic conditions to biological outcomes. That said, the attention of human biologists and anthropologists has certainly been less sustained on the social and the economic and more sustained on proximate environmental factors such as temperature, oxygen availability, and energetic yields. It is not that this focus is wrong, in fact it is essential. However, it might also be insufficient. Acknowledging the eco-social origins of biological maladies is the easy part. The hard work is unraveling the myriad way that the eco-social becomes local and get under the skin. How is the eco-social embodied? While we are now aware of many strong associations between eco-social conditions and biological outcomes, we have little sense of the causal processes that underlie these associations. Moreover, as a number of authors in this volume note, social conditions are notoriously fickle and flexible: They make terribly unreliable variables. In 1992, Thomas Leatherman and I brought together a group of biological anthropologists and other social and biological scientists who were interested in rethinking biocultural intersections. Like the symposium upon which this book is based, ours was also funded by The Wenner-Gren Foundation for Anthropological Research. Then, we wished primarily to explore a way to move biological anthropology to focus more on the root political-economic causes of disease risk, adversity, and illness (Goodman and Leatherman 1998). The group helped to move the focus of inquiry from biophysical stressors to sociocultural stressors as central to health, from evolutionary adaptation to the adaptive constraints brought on by poverty and inequality, and from race as a risk group to racism. We highlighted, with some success, the political economy of health for biological anthropologists. What we did less well is to explore the mechanisms and processes by which the eco-social and the biological are linked. Since our symposium, a number of publications have begun to specify these processes and mechanisms. Health, Risk, and Adversity is a great leap forward, taking us further than we have ever been. The editors, Catherine Panter-Brick and Agustín Fuentes, have brilliantly organized Health, Risk, and Adversity. Each section of the book includes juxtaposed chapters from a medical or biological anthropologist, an epidemiologist, and a clinician. They maximize the opportunities for readers to make their own connections. More important still, there are ample summations to guide readers from all fields and all levels of expertise. In addition to an introduction by the editors, the volume includes section commentaries from eminent scientists.

Framing Health, Risk, and Adversity

•

xv

This book is special. First, it interrogates how scientific ideas and concepts are shaped by the intersections of scientific and popular discourses. For example, the editors (Introduction) and many of the authors detail how “risk” is both a categorization (being in a risk group) and a statistical notion (degrees of risk). Both ideas about risk are powerful discourses that shape causal thinking and the political economies of research and clinical practices. Yet, risks are statistically real as well. Second, the volume as a whole goes further than any I am aware of in explicating the specific mechanisms and processes by which the eco-social gets under the skin. How is it that adverse situations such as poor diets, exposure to infectious agents, and physiological stress lead to disease in some but not in others? The chapters outline many potential mechanisms that are worth further exploration. Reading this book reminded me many times of the famous parable about the political economy of health. The original is credited to the medical sociologist Irving Zola and has been modified to fit critical medical anthropology by Nancy ScheperHughes. As Lorena Madrigal and colleagues note in Chapter 9, the parable was modified by me and Thomas Leatherman to fit work in biological anthropology. It goes something like this: A group of anthropologists sitting on a river bank are startled by the screams for help of a drowning individual. Just as the first body floats by and eventually out of sight and sound, another and then many others float down the river. The anthropologists discuss and debate what is going on. One of them sets off to find someone with biomedical training. Another, remembering her old CPR course, dives heroically into the waters. She manages to pull out and resuscitate a few individuals before exhaustion overcomes her. None of the anthropologists wanders upstream to expose the systems of hierarchy that are throwing individuals into the river. Inequalities are written in myriad way onto and into human bodies. Inequalities have biological consequences and the biological consequences add up and may result not just in increased adversity, failing health, and shorter lives, but in the declining resistance of mothers and caretakers, thus upsetting household coping and the ability to resist further insults. Poverty is a viciously biocultural cycle. The chapters in this book provide grounded theory. They enrich our sense of the history and evolution of health and the contemporary political-economic factors that operate downstream and get under our skins. The interdisciplinary conversation has begun to turn a social human biology from a serious of thoughts and mediations into a program of research. It is time for anthropologists to work with each other and fellow social and medical scientists to take a deep and long look at how individuals and groups become vulnerable and how inequalities in health arise and become bioculturally cyclical. This volume gets us as far as we have ever been. The authors and editors are not only wandering gallantly into the stream of human risk and adversity, they are trudging up stream against powerful tides.

xvi

• Alan Goodman

References Goodman, A. and T. Leatherman, eds. 1998. Building a New Biocultural Synthesis. Ann Arbor: University of Michigan Press. Krieger, N. 2001. Theories for social epidemiology in the 21st century: an ecosocial perspective. International Journal of Epidemiology 30: 668–77.

INTRODUCTION

• • • Health, Risk, and Adversity A Contextual View from Anthropology Catherine Panter-Brick and Agustín Fuentes Health and Adversity Both anthropology and public health have focused extensively on the question of who suffers from poor health outcomes. Both fields are still developing ways to examine why and how health differentials emerge over the lifetime of individuals. This book aims to enhance understanding of outcomes and processes that govern relationships between health, risk, and adversity—to facilitate linkages between multiple levels of inquiry, into who or what brings about health disparities, as well as into how, when, and why differential health outcomes occur. As editors, our goal is to relate the risks of poor health to contexts of adversity, defined as social or physical environments that create hardship or affliction. “Health” and “risk” are words that have entered the everyday language of many contemporary societies (Beck 1992; Lupton and Petersen 1996; Douglas 1997; Lupton 1999; Boyne 2003). Why should these two terms be linked to “adversity”? Adversity “matters” to health because it is responsible for human suffering on a scale that demands both objective research and ethical intervention. Most adverse environments are due to human action or inaction; most sources of ill-health are liable to remedy or prevention; and most health differentials across human populations are demonstrably unnecessary, avoidable, and unjust (Whitehead 1992). Health is “a state of complete physical, mental and social well-being,” according to the World Health Organization. This definition is uncannily similar to that of happiness, but with an important difference: health is ratified by the United Nations

2

•

Catherine Panter-Brick and Agustín Fuentes

as a universal human right, happiness is not (Saracci 1997; Panter-Brick 2003; Ross 2006). Systematic disparities in health across populations are currently a high priority for the international public health agenda (WHO 2008). The focus is on major health inequalities across socioeconomic groups, and, where closely aligned with a human rights approach, on health inequities produced by social advantage or disadvantage (Braveman and Gruskin 2003; Casas-Zamora and Ibrahim 2004)—namely, the avoidable and unjust distribution of resources with respect to gaps between health-related needs and service provision (Carter-Porras 2002). Addressing inequities is a “vital mandate” in the public debate about health (Editorial, Lancet 2008; Marmot, 2007). Health disparities are unmistakable evidence of significant life adversity. They permeate human experience across a whole range of sociocultural contexts, ecological settings, and developmental trajectories. Such disparities are often conceptualized and evaluated under the rubric of risk—in this book, we examine experiences of risk associated with genetic and physiological makeup, environmental exposure to poor nutrition and disease, and social marginalization. Vulnerability, susceptibility, and adversity are key constituents in this investigation, whether pertaining to the issues of risk measurement, risk evaluation, or risk communication. Anthropology as Context Anthropology provides meaningful conceptual frameworks and diverse methodological tools for health research. Within this field, a comparative approach is used to examine health experiences across space and time. Anthropology as an academic discipline is currently differentiated between bio-evolutionary and socio-cultural fields, with medical anthropology in an interstitial position (McElroy and Townsend 2004). Thus, several conceptual frameworks are in existence. For biological anthropologists, health risks and health outcomes across populations are shaped by the expression of genetic inheritance and the relative fitness of individuals confronting environmental challenges over evolutionary time (Stinson et al. 2001). For social anthropologists, health risks are dimensions of experience as well as behavior, which derive meaning from a shared cultural model or competing social constructions of hegemonic and lay knowledge (Popay et al. 1998; Krieger 2001; Loch and Nichter 2002). A common ground is advocated by researchers promoting systematic ways of relating “biology” with “culture,” to critically evaluate health disparities within and between populations in terms of political economy and social structure (Goodman and Leatherman 1998; Hahn 1999; Helman 2001; Kaplan 2004; Dressler 2005; Leatherman, 2005; Dufour 2006). These approaches provide a powerful complement to epidemiological mapping of health risks in terms of “webs of causation” (a metaphor referring to the numerous, interrelated risk and protective factors related to disease), in providing a critical evaluation of the “spiders”—agents responsible for spinning the “web” in the first place (Krieger 1994). Health risks do not simply appear in a web of complex, interrelated factors; they are produced or reproduced by social agents, hegemonic structures, or

Health, Risk, and Adversity

•

3

ecological pressures. A “factorial analysis” of health risks without critical explanation has limited appeal (Parker and Harper 2006), even if it does pave the way for public health action. A better understanding is required of the processes through which health disparities emerge across space and time—an understanding of how health risks are spun (Figure A). Collaboration across Disciplines Few would disagree that effective collaboration across disciplines enables better understanding of the processes governing the emergence of health risks in adverse living conditions: “no single discipline has a monopoly of insight” (Parker and Harper 2006:1). However, what exactly does “collaboration” mean? How is it fostered in research or applied contexts (Lambert and McKevitt, 2002; Porter 2006)? White’s concluding chapter to this volume illustrates some of the difficulties of talking across disciplines or bridging methodological practices, but emphasises the value of interactions across applied and academic domains. The academic steps involved in health research are akin to working on a puzzle, where individual players initially focus on particular areas, find critical missing pieces, rectify faulty moves, and contribute as a team to decipher the overall picture. Cross-disciplinary collaboration is far more

Figure A. A web of causation representing interrelated risk and protective factors, with the agent (a spider) responsible for their root.

4

•

Catherine Panter-Brick and Agustín Fuentes

compelling than any single discipline for revealing specific links between health risks and adversity. Research involving a cross-disciplinary, cross-cultural study of health risks and adversity often involves collaboration between different institutions—academic, political, nongovernmental, advisory committees, and other professional agencies—on an international scale. Collaboration of this nature introduces additional complexities: financial constraints, linguistic barriers, negotiation of research permission and ethical issues, different expectations about fieldwork practices, and varied access to published literature or analytical software. Health research must thus accommodate a diversity of conceptual frameworks, a dialogue between research and intervention, and institutional collaborations on an international scale. Architecture of this Book This book is divided into three main sections: health risks and diseases in transition; generational and developmental change; gene evolution, environment, and health. Each section has three chapters, written by biological and medical anthropologists, epidemiologists, and/or clinicians. Three discussants, from the fields of archaeology, human biology, and sociocultural anthropology, provide introductory analysis and commentary to these sections. The concluding chapter, written by an expert in public health, provides an overview of the contributions, highlighting the strengths and weaknesses in the approaches exemplified in this book. This format serves as a thematic and a pedagogical tool. We refer readers to the commentaries (by Charlotte Roberts, Darna L. Dufour, and William W. Dressler) and the conclusion (by Martin White) for elegant expositions of take-home messages and critical evaluations of individual chapters. Our own emphasis, in this introduction, is twofold: we draw attention to how narratives permeate cultural discourse and scientific inquiry on health, risk, and adversity; and highlight conceptual approaches for evaluating people’s lives in the context of social, ecological, and evolutionary change. In our view, health research is at its best when it engages with both the discourse and the science evaluating health risks and adversity. Risk Narratives Risk narratives are story lines or discourses that weave together cultural and scientific accounts regarding the conceptualization and communication of health risks, or relative vulnerability to poor health, in personal lives, media reports, political directives, and expert accounts. Narratives play central roles in our understanding of adversity and risks to health: indeed, they are a core component and often a problematic aspect of health research and public health policy (Oaks and Harthorn 2003; Bennett and Calman 2005; Porter 2006). Writing about vulnerability to avian flu or influenza, Herring (Chapter 3) states the fact that “narratives have a powerful influence on

Health, Risk, and Adversity

•

5

public concern about health crises and health policy…it is important to identify and critique the narratives and moral lessons that run through scholarly and media discussions of epidemics.” A preoccupation with “emergent diseases” is a good example of a risk narrative that took on global significance when Western populations began to experience “risks of contagion,” of the kind that had long been suffered by indigenous people in their local contexts, for example multi-drug resistant tuberculosis (Hurtado et al. 2005; Farmer 1999). Indeed, Farmer took to task the public health policy, resource distribution, and “professional commentary on emerging infectious diseases,” as based on a “rhetoric” of risk for Western populations—to the detriment of “those who have been silently suffering with these diseases, often for generations” (1999, p. 57). We invite our readers to evaluate the ways in which risk narratives shape our understanding of the links between health, risk, and adversity. In this volume, narratives are explicitly associated with “hype” about primate-borne infections and human pandemics, epidemics, and smaller-scale crises, by Jones-Engel and Engel in Chapter 2 and by Herring in Chapter 3. They are insidiously linked to ethnic categorization and social profiling of hypertension, by Madrigal et al. in Chapter 9, and shown to guide investigation into the real or hidden risks to child health, by Sellen in Chapter 6. Risk models and multifactorial analysis may seem empirically grounded in objective scientific inquiry. Yet the representation of risks factors as a “web of causation” is bound in a cultural “story,” a discourse that structures the measurement of risk factors, the communication regarding genesis of health disparities, and the promotion of health interventions and public health policy. Even the explanatory frameworks invoked in this volume—such as evolutionary theory, developmental origins, or acculturation—are, to a degree, narratives that weave together complex elements of biological and cultural processes responsible for health. Risk narratives also take the form of social categorization. In the field of public health, “risk factors” are variables that expose or predispose individuals to ill-health: the assessment of risk usually proceeds from the statistical measurement of factors that affect health and well-being. The quantitative approach to risk is a matter of ordering reality in a calculable form, to help assign individuals into low-risk or highrisk categories (Dean 1999, pp. 143–44). Such “at-risk categorization” understandably permeates biomedical practice (e.g., decisions for hypertension treatment are guided by blood pressure thresholds demarcating risks of “health” and “disease”). We fall prey, however, to problematic generalizations, if a categorization of individuals leads to flawed policy or social marginalization (see Panter-Brick 2002 for homeless children, Parker 2006 for sex workers). An empirical, probabilistic statement is readily transformed into a normative, deterministic statement regardless of historical context. In this book, this is well illustrated by Madrigal et al. (Chapter 9), who show a deceptively simple “scientific hypothesis”—black skin as a risk factor for hypertension—to be deeply flawed both in terms of evidence and as a conceptual model. Because risk is both a statistical notion and a social categorization, anthropology makes

6

•

Catherine Panter-Brick and Agustín Fuentes

a significant contribution to analysis, by coalescing the scientific evaluation of health processes with a critical appraisal of health outcomes in social context. Conceptual Approaches The commentators in this book (Roberts, Dufour, and Dressler) review the material offered across chapters in ways that call attention to conceptual frameworks and contextual analyses of health risks and adversity. Roberts draws a parallel in the study of past and present human populations, between archaeological approaches, used to appraise diversity of skeletal or cultural archaeological material, and the evolutionary and biocultural perspectives developed in this book. Archaeologists also make use of risk analysis to appraise the relatively unknown processes through which health disparities develop over evolutionary time. Dufour usefully critiques the use of the term environment, which in the broadest sense encompasses all the living conditions of a population, including physical, social, and cultural context. We often talk about environments without specifying which features of the “environment” or “environmental change” are actually relevant to people’s lives. To reach a better understanding of health contexts, it is necessary to specify which features in our environment produce disparity in health outcomes. Dressler opens his commentary with a focus on the terms health inequalities and health disparities, vocabulary currently espoused in the UK and the USA respectively. He reviews how this area of research is enhanced by biocultural research in anthropology. He warns, however, that we must take the concept of “culture” seriously, culture as defined by “shared meaning” regarding belief and behavior. As editors, we think that health researchers could all go a step further, by showing a deeper appreciation of biology, behavior, ecology, and culture. This book contains clear examples of why health risk research requires careful attention to scientific enquiry and cultural discourse. Consider the contributions on obesity and cardiovascular health by Leonard et al. (Chapter 1), hypertension by Madrigal et al. (Chapter 9), breast cancer by Núñez-de-la-Mora and Bentley (Chapter 5), psychosocial stress by McDade (Chapter 8), and growth retardation by Sellen (Chapter 6). These chapters employ terms widely used in risk modelling and discourses about risk or adversity: “socioeconomic status,” “ethnic group,” “environment,” “modernization,” “lifestyle,” and socioeconomic “transition.” Such well-worn constructs serve in analyses of health across place and time, but more often than not they have no simple ethnographic or ecological basis. As McDade aptly shows, there are multiple dimensions of “status” in existence even on a relatively small island, with potential mismatch of social and economic experiences linked to heightened psychosocial stress. Leonard et al. explicitly demonstrate, with reference to populations in Siberia, that relationships between “lifestyle,” metabolism, and cardiovascular health are multifaceted, given the context of major geopolitical changes and socioeconomic transition. Our contributors model risks to health against a backdrop of evolutionary, developmental, ecological, or social change—it is change that facilitates comparative

Health, Risk, and Adversity

•

7

analyses of health and adversity. They draw upon a strong anthropological tradition of comparative and contextual analysis, in their examination of life trajectories over evolutionary time, during growth and development, and under ecological and social change. Professional outlook and definitions give rise to important issues in health research and practice. In Chapter 4, Ellison and Jasienska reflect upon the conceptualization of health in reproductive ecology, a field of interest to biomedical practitioners as well as anthropologists trained in evolutionary and ecological paradigms. If we are to understand the production of health differentials, we require a consistent usage of terminology, such as adaptation and adaptability, two fundamental constructs in human biology. Many textbooks, for example, envisage adaptation—defined as a beneficial response to environmental adversity—solely in biological terms, while others relate it to both biology and behavior. Understanding of adaptation is often dangerously tautological: evidence of adaptation is linked to “success” in overcoming a difficult environment, while success is all too easily attributed to “adaptation.” Assessment is made all the more difficult given the human capacity for niche construction (modifying and creating environments, Odling-Smee et al. 2003), which alters the patterns and types of ecological pressures on populations. The concept of adaptability—flexibility in the face of environmental adversity—has even greater significance for human life, since most life trajectories negotiate a great deal of change; but it is also more challenging to document in the case of humans. Adaptability involves complex, interacting responses, with costs and benefits manifested not just as immediate trade-offs, but over the lifetime of an individual and across several generations (Kuzawa 2005). Making judgements about trade-offs is difficult; a response is deemed adaptive only if benefits outweigh potential costs. This was evident in debates rejecting the “small but healthy” hypothesis (children have reduced body size under conditions of food scarcity, enhancing short-term survival but entailing longer-term risks to maternal-child health and work capacity [Martorell 2000; Norgan 2000]). The power of this evolutionary conceptual framework is nonetheless widely recognized, not just within anthropology but in other health-related fields. For example, Godfrey and Hanson are clinicians who review the significance of predictive adaptive responses in early human development, in engendering health disparities in later adulthood (Chapter 7). Their evaluation of physiological processes, contingent on a mismatch between uterine and later nutritional environments, as well as critical evaluation of the magnitude and timing of physiological trade-offs, provides us with a complex example of human adaptability. This moves us away from reductionistic models of genetic inheritance that envisage only linear impacts of natural selection, a static or tautological definition of adaptation, and a simple binary definition of health and disease. Godfrey and Hanson demonstrate the complexities of fetal development across multiple physiological and social contexts, to show that genetic makeups and environmental contexts interact over the course of a lifetime. Again, the adaptive or maladaptive nature of fetal responses is difficult to demonstrate (Jasienska

8

•

Catherine Panter-Brick and Agustín Fuentes

et al. 2006), while the mismatch between early and later “environments” remains illdefined. Clinical health research here draws closely on modern evolutionary theory to understand the processes by which humans and their environments (writ large) shape one another to produce critical health outcomes for mothers, babies, and later adults. Our contributors also highlight methodological challenges in health measurement across time, place, geography, generations, and families. They seek to identify the processes that generate poor maternal, fetal, infant, child, and adult health outcomes, responsible for associations between health risks and adversity. The tracking of migration and acculturation patterns, maternal physiology and behavior, and fetal health and the very specific social contexts in which growth and reproduction takes place requires not just integration of biology and culture but longitudinal assessment. Research that follows individuals across their lives, the lives of their parents or spouses, and their acculturation to novel “environments” demands a long-term investment of money, time, and energy. Throughout this book, there is a call for identifying measures in the short-term that can reliably predict health risks in the distant future, for example by using growth status to foresee critical threats to child health (Sellen, Chapter 6), and by using developmental markers to predict breast cancer (Núñez-de-la-Mora and Bentley, Chapter 5). There is also systematic reflection on when and how to intervene to bring about significant, population-wide improvements in well-being—for instance, to improve maternal-fetal interactions, which Godfrey and Hanson argue (Chapter 7) provide the most effective opportunity for large-scale public health intervention. Audience and Pedagogy We address this book to social science and public health researchers who, like us, are guided by the simple principle that “health matters.” Some authors have included a brief description of their own research trajectories, interests, and/or reasons for pursuing research in their field—a personal narrative to encourage reflections on the research process. The chapters include boxes (in text or chapter appendices) that summarize take-home messages, enhance discussion, expand on particular topics, present thematically related information, or illustrate central concepts. These boxes thus serve for quick reference, additional information, clarification, and, from the standpoint of pedagogy, as foci for discussion or further investigation. Terms and concepts requiring definition are featured in the glossary. We hope that this volume will engage both researchers and students, appealing in content and style. The chapters express multiple viewpoints, with perspectives that may differ or overlap—this is an important component of research endeavors. Such perspectives highlight our current state of knowledge and practice in a contested field, but they also demonstrate that research itself is an ongoing dialogue. We wish to thank the Wenner-Gren Foundation, the University of Durham and the University of Notre Dame Institute for Scholarship in the Liberal Arts for partial support of the original discussions leading to the essays in this book.

Health, Risk, and Adversity

•

9

References Beck, U. 1992. World Risk Society. Cambridge: Polity Press. Bennett, P. and K. Calman, eds. 2005. Risk Communication and Public Health. Oxford: Oxford University Press. Boyne, R. 2003. Risk. Buckingham: Open University Press. Braveman, P. and S. Gruskin. 2003. Defining equity in health. Journal of Epidemiology and Community Health 57: 254–58. Carter-Porras, O. and C. Baquet. 2002. What is a ‘health disparity’? Public Health Reports 117: 426–34. Casas-Zamora, J. and S. Ibrahim. 2004. Confronting health inequity: the global dimension. American Journal of Public Health 94(12). Dean, M. 1999. Risk, calculable and incalculable. In Risk and Sociocultural Theory: New Directions and Perspectives, ed. D. Lupton. Cambridge: Cambridge University Press, 131–59. Douglas, M. 1997. The depoliticization of risk. In Culture Matters: Essays in Honor of Aaron Wildavsky, eds. R. Ellis and M. Thompson. Boulder, CO: Westview Press, 121–32. Dressler, W. 2005. What’s cultural about biocultural research? Ethos 33(1): 20–45. Dufour, D.L. 2006. Biocultural approaches in human biology. American Journal of Human Biology 18(1): 1–9. Editorial, Lancet (2008). Social determinants of health: a call for papers. The Lancet 371(9627): 1812-1812. Farmer, P. 1999. Infections and Inequalities—The Modern Plagues. Berkeley: University of California Press. Goodman, A. and T. Leatherman, eds. 1998. Building a New Biocultural Synthesis: PoliticalEconomic Perspectives on Human Biology. Ann Arbor: University of Michigan Press. Hahn, R. 1999. Anthropology in Public Health: Bridging Differences in Culture and Society. Oxford: Oxford University Press. Helman, C. 2001. Culture, Health and Illness. London: Arnold. Hurtado, A., C. Lambourne, et al. 2005. Human rights, biomedical science, and infectious diseases among South American indigeneous groups.” Annual Review of Anthropology 34: 639–65. Jasienska, G., I. Thune, et al. 2006. Fatness at birth predicts adult susceptibility to ovarian suppression: an empirical test of the Predictive Adaptive Response hypothesis.” Proceedings of the National Academy of Sciences 103(34): 12759–62. Kaplan, G. 2004. What’s wrong with social epidemiology, and how can we make it better? Epidemiologic Reviews 26: 124–35. Krieger, N. 1994. Epidemiology and the web of causation: has anyone seen the spider? Social Science and Medicine 39(7): 887–903. ———. 2001. Theories for social epidemiology in the 21st century: an ecosocial perspective. International Journal of Epidemiology 30: 668–77. Kuzawa, C.W. 2005. Fetal origins of developmental plasticity: are fetal cues reliable predictors of future nutritional environments? American Journal of Human Biology 17: 5–21. Lambert, H. and C. McKevitt. 2002. Anthropology in health research: from qualitative methods to multidisciplinarity. British Medical Journal 325(July): 210–13. Leatherman, T. 2005. A space of vulnerability in poverty and health: political-ecology and biocultural analysis. Ethos 33(1): 46–70.

10

• Catherine Panter-Brick and Agustín Fuentes

Loch, M. and M. Nichter. 2002. Introduction: from documenting medical pluralism to critical interpretations of globalized health knowledge, policies, and practices. In New Horizons in Medical Anthropology: Essays in Honour of Charles Leslie, eds. M. Nichter and M. Lock. London: Routledge: 1–34. Lupton, D. 1999. Risk and Sociocultural Theory: New Directions and Perspectives. Cambridge: Cambridge University Press. Lupton, D. and A. Petersen, eds. 1996. The New Public Health: Health and Self in the age of Risk. London: Sage. Marmot, M. (2007). Achieving health equity: from root causes to fair outcomes. The Lancet 370(9593): 1153-1163. Martorell, R. 2000. Body size, adaptation and function. In Nutritional Anthropology: Biocultural Perspectives on Food and Nutrition, eds. A. Goodman, D.L. Dufour and G. Pelto. California: Mayfield Publishing Co.: 258–65. McElroy, A. and P. Townsend. 2004. Medical Anthropology in Ecological Perspective, Boulder, CO: Westview Press. Norgan, N. 2000. Long-term physiological and economic consequences of growth retardation in children and adolescents. Proceedings of the Nutrition Society 59: 245–65. Oaks, L. and B. Harthorn, eds. 2003. Introduction: Health and the Social and Cultural Construction of Risk. Westport, CT: Praeger Publishers. Popay, J., Williams, G.,Thomas, C., Gatrell, T. 1998. Theorising inequalities in health: the place of lay knowledge. Sociology of Health & Illness 20(5): 619–44. Odling-Smee, F., K. Laland, et al. 2003. Niche Construction: The Neglected Process in Evolution. Princeton: Princeton University Press. Panter-Brick, C. 2002. Street children, human rights and public health: a critique and future considerations. Annual Review of Anthropology 31: 147–71. ———. 2003. Achieving health for children. In Changing Childhoods: Local and Global, eds. H. Montgomery, R. Burr, and M. Woodhead. Milton Keynes: Open University Press: 93–139. Parker, M. 2006. Core groups and the transmission of HIV: learning from male sex workers. Journal of Biosocial Sciences 38: 117–31. Parker, M. and I. Harper. 2006. The anthropology of public health. Journal of Biosocial Science 38(1): 1–5. Porter, J. 2006. Epidemiological reflections of the contribution of anthropology to public health policy and practice. Journal of Biosocial Science 38: 133–44. Ross, N. 2006. Health, happiness, and higher levels of social organisation. Journal of Epidemiology and Community Health 59: 614. Saracci, R. 1997. The World Health Organization needs to reconsider its definition of health. British Medical Journal 314: 1409–10. Stinson, S., B. Bogin, et al. 2000. Human Biology: An Evolutionary and Biocultural Perspective. New York: Wiley-Liss. Whitehead, M. 1992. The concepts and principles of equity and health. International Journal of Health Services 22(3): 429–45. WHO (2008). Closing the gap in a generation: Health equity through action on the social determinants of health (Executive Summary). Geneva, World Health Organization.

PART I

• • • Health Risks and Disease in Transition

• • • Understanding Health Past and Present Charlotte Roberts It is hard for people to know if things are getting better or worse….there is the myth of progress, which claims that technological innovation makes our lives less grueling, healthier, more productive, and happier…[and] there is the myth of the good old days… —G. J. Armelagos (1998:59)

The Complementarity of Studying the Living and the Dead As a person who does not research health in living populations, I find that this book not only provides fascinating and incredibly useful insights into living populations— that is, how we might approach evaluating how and why people get sick—but also shows how relevant medical and biological anthropology are to our understanding of the origin and evolution of disease over long periods of time. Palaeopathology, biological anthropology, and medical anthropology complement each other nicely; while palaeopathology can provide a window on disease evolution over long periods of time and highlight the main reasons for the appearance of specific diseases (Roberts and Manchester 2005; Larsen 1997), biological/medical anthropology focused on living populations can provide a better context for the many factors responsible for disease occurrence in populations today (McElroy and Townsend 1996; Sargent and Johnson 1996). Approaches to understanding health that are broad, holistic, and interdisciplinary, that emphasize the health implications of interactions between humans and their physical and biological environments, are something for which we should strive in palaeopathology, biological anthropology, and medical anthropology. The anthropology of modern populations, like palaeopathology, considers people in many environmental and cultural settings ranging from isolated and marginalized locations to urban community settings. However, one common difference is that in

14

•

Charlotte Roberts

palaeopathology there can be a tendency for the study of disease for its own sake with no consideration of the factors causing the condition observed, and a concentration on individual skeletons rather than populations, often with an emphasis on just one geographical location (e.g., Anderson 1997). In medical and biological anthropology of modern populations, this approach is usually avoided. It is surprising how many of the approaches seen within the three sections of this volume (Health Risks and Diseases in Transition; Generational and Developmental Change; Gene Evolution, Environment, and Health), have been followed in palaeopathology. However, palaeopathologists still have much to learn from biological and medical anthropologists about the impact of the many variables determining health status in the living. Researchers in past human health tend to come from diverse backgrounds (anthropology, archaeology, medicine, dentistry, nursing, anatomy, public health), but, more often than not, they lack an appreciation of the real impact of health problems on people living in the past because their dataset consists of observations from skeletal remains, often with no consideration of the association of signs and symptoms with the disease changes they observe. Furthermore, until recently, methods of analysis have been limited, and it has not been possible even to attempt to answer some questions about past health with the analytical methods available (although times are changing—see Brown 2000 for a commentary on ancient DNA analysis). However, there has always been some attempt, more often in North America, to consider past health both in the context of the origin and evolution of disease, and the socioeconomic and political factors relevant to a disease’s appearance and maintenance (see, e.g., Walker and Hollimon 1989; Merrett and Pfeiffer 2000). While advocating this approach, it must be acknowledged that it is often not that easy to apply when studying a sample of skeletons from an archaeological site that are often fragmentary (making diagnosis of disease very difficult), and with no knowledge of whether that sample is representative of the health of the original living population from the region (see Waldron 1994, and Wood et al. 1992, for a detailed discussion of the problems of inferring health from skeletons). Additionally, for most, there are limited analytical methods (usually restricted to diagnosis of disease that affects only bones and teeth) and sociocultural contextual data may be incomplete, making a biocultural approach to palaeopathology difficult. However, we make the best of what we have (Roberts and Cox 2003), make recommendations for future work, and hope things will improve. There are similar problems with collecting and interpreting data in the fields of medical and biological anthropology, although anthropologists who focus on living populations are better placed to investigate health risks under adverse environments. Diseases and Transitions Health and disease are measures of the effectiveness with which human groups, combining health and cultural responses, adapt to their environments (Lieban 1973:1031).

Understanding Health

•

15

The theme of this section, “Health Risks and Disease in Transition,” has been tackled in palaeopathology by some scholars for some time now (see, e.g., Swedlund and Armelagos 1990). Our ancestors experienced a change in their ecological relationships with the advent of hunting and gathering, pastoralism and nomadism, settled agriculture, and urbanism and industrialization, all with their attendant health problems; the developing complexity of life is viewed as broadly detrimental to health (see, e.g., Cohen and Armelagos 1984; Cohen 1989; Roberts and Cox 2003; and Steckel and Rose 2002). However, Froment (2001) has highlighted the problems of inferring health from skeletal remains, as illustrated in more detail by Wood et al. (1992). For example, even thought skeletal remains of hunter-gatherers may appear “more healthy” than those of settled agriculturists, there is a possibility that they may have died from diseases that did not affect the skeleton, or they could have died before disease had chance to make its mark on bone. Froment further indicates that nonsedentary groups will be more at risk from exposure to specific epidemiological risks such as closer contact with wild animals and more violent deaths through hunting and other accidents. Nevertheless, it is suggested that the contexts of hunter-gatherer groups in the past and present are “radically different” (Froment 2001:259). Clearly, there are different health risks for people around the world that are influenced by a variety of environments and economies, and these risks will have changed through time. Armelagos (1998) describes three major epidemiological transitions: the transition to agriculture, the post–World War II development of antibiotics to treat infectious disease, and the reemergence of new diseases as a consequence of changes in our environment (including the mutation of pathogens to resist pharmaceutical remedies). However, transitions and change can be very gradual and slow, with people experiencing different “worlds” at the same time (something that is true for living populations). Clearly, though, the dominance of the degenerative diseases in patterns of morbidity and mortality today, certainly in developed parts of the world, is being affected by emerging and reemerging infections. The topic of health risks and diseases in transition has been interpreted broadly in this book, but the focus is on the risks of human populations to disease as their living environment changes through (mainly short) periods of time. Much has been written on this theme, which includes the impact of the movement of people to new environments (whether it be for trade, tourism, or to potentially gain a better life— see, e.g., Mascie-Taylor and Lasker 1988; Roberts et al. 1992; Wilson 1995) and the consequences of socioeconomic change on disease loads in human populations (see, e.g., Cohen 1989; Morse 1995). For example, one of the many factors in the rise of tuberculosis today is increased mobility of human populations; Davies (1995) claims that immigration is one of the single most important causes in increases of TB in most developed countries. Likewise, malaria has seen a global resurgence as a consequence of many factors, not least the impact of particular agricultural systems on the ability for mosquitoes to survive and reproduce (Brown 1997). Of course, humans have a great ability to adapt to changing circumstances. If adaptation means survival,

16

•

Charlotte Roberts

then the human body needs to develop coping mechanisms whether these are cultural or biological, including genetic (through natural selection over a long time) and physiological (over the person’s lifetime) mechanisms. Clearly, this phenomenon has occurred in the past and continues to occur today. The chapters in this section are wide ranging in their treatment; they consider: • cross species transmission of disease, particularly primate-borne zoonoses in Asia, and the consequences for human populations (Jones-Engel and Engel); • the possible global impact of avian influenza, its relevance to the 1918 influenza pandemic and the potential of cross-species transmission from bird to human (Herring)—an issue that is very topical; and • the impact of evolutionary (cold exposure) and adaptive (changing economy) responses of indigenous Siberians on cardiovascular disease rates (Leonard, Snodgrass, and Sorensen). The chapters cover diseases both of an infectious and of a chronic degenerative nature. The latter is illustrative of people living in both developed and (increasingly, more recently) developing countries as a result of greater exposure to a “lifestyle,” in its broadest sense, that is conducive to chronic degenerative diseases (as indicated by research by Eaton and Boyd 1999 and by the chapter by Leonard et al.); the former tends to be more common in developing countries due to a set of particular circumstances, not least of which is a poor living environment and problems with access to health education and care, such as antibiotic therapy (although, increasingly, infectious diseases are being seen in the developed world). Both are indicative of our changing world and highlight where we have made mistakes, what risk factors we should look out for in the future, what we might do to address those mistakes, and whether the measures taken will be effective. It is a sad fact that it is usually when the developed world encounters a health problem that serious action to combat it is taken. Each chapter in this section illustrates the complex evolutionary and adaptive relationship that humans have with their changing environments, clearly emphasizing that the changes humans create in their “environment” can have grave implications for their health and well-being. Herring (Chapter 3) clearly illustrates the impact of factory farming and developed trade networks on avian influenza transmission. All chapters also remind us that the epidemiology of disease is highly complex, and that to “conquer” these diseases multiple factors need to be taken into account with people from different disciplines needing to work together to achieve that aim (as illustrated well in Jones-Engel and Engel in Chapter 2). While the medical profession and public alike thought some diseases such as TB had all but disappeared in the 1980s (e.g., Smith 1988), some “conquered” diseases are reemerging and totally new pathogens have also evolved. The journal Emerging Infectious Diseases first appeared in 1995, which perhaps illustrates the increasing awareness of the problem of infec-

Understanding Health

•

17

tions. For example, we have seen tuberculosis and SARS (severe acute respiratory syndrome) spread from human to human via droplet infection through travel, trade, and contact around the world, the impact of developments in the food industry and increased eating out in restaurants on the frequency of food poisoning (e.g., in the UK), and the effect of increased, and inefficient, use of antibiotics to treat infection, leading to antibiotic resistance in some pathogens (e.g., MRSA or methicillin resistant Staphylococcus aureus in UK hospitals; see Barrett and O’Hara (2005) on the 25th anniversary of the Journal of Hospital Infections). Cross-species Transmission of Disease In these chapters a number of observations emerge and themes develop. Firstly, JonesEngel and Engel emphasize the need to look more closely at the links between humans and nonhumans in the occurrence of disease in human populations. In particular, we must focus on nonhuman primates and especially the enzootic simian retroviruses, which have been shown to cross the species barrier (enzootic meaning “affecting animals in a limited region”—in this case Asia). Immunologically, physiologically, genetically and behaviorally, human and nonhuman primates are very similar, and thus nonhuman primates are more likely to infect humans than are nonprimate species. Clearly, in some parts of the world contact between nonhuman primates and humans is more likely, but it is suggested that the risk of this contact on disease transmission to humans has not been fully appreciated. In the context of Asia and nonhuman primates, monkey temples attract tourists, there are animal markets, people have nonhuman primates as pets, they hunt them for food, they are kept in zoos, they are used for harvesting purposes (sometimes) and to prevent crop raiding, and of course we see them as performing animals. Not only, then, can tourists come into contact with them, but the people that work with them are very vulnerable. The potential for human and nonhuman primates to come into contact and contract infections through body fluids via inhalation, ingestion, and skin contact is clear. However, Jones-Engel and Engel stress that pathogens have a variety of characteristics that influence their transmission, that humans and nonhuman primates also have characteristics that impact pathogen transmission, and that natural and human-made environmental factors influence whether pathogens are transmitted. In archaeology and paleopathology, this focus on human-animal interaction in disease transmission has not been prominent; this is surprising considering the long list of potential infective organisms that Jones-Engel and Engel list. The study of zoonoses in archaeological animal remains, or diseases that are passed from animals to humans, has been very limited and there is a general lack of interest in this field in the archaeozoological community, with a few exceptions (e.g., Brothwell 1991). This is despite recognition that zoonoses must have had an impact on humans in the past (e.g., TB, anthrax, brucellosis), and specific factors, such as working with animals, would have predisposed people to contracting diseases from their animals. Some re-

18

•

Charlotte Roberts

search has identified TB in animal remains (e.g., Bathurst and Barta 2004), but this type of work is limited to date. Of course, as Jones-Engel and Engel point out, it is not only direct animal-to-human transmission of disease that can occur; diseases can also be transmitted across species barriers by vectors, such as insects, even if humans and animals do not come into contact. A classic example of this is the transmission of the plague by the rodent flea (Xenopsylla cheopsis) to humans (Park 1993). This disease has a long history and remains a problem today in people exposed to infected rodents and fleas. Herring’s detailed analysis of the possible impact of avian influenza on human populations worldwide links well with Jones-Engel and Engel’s research, and similarly considers cross-species transmission; the chapter focuses on a viral infection that originates in birds but can be transmitted to humans. Herring and Jones-Engel and Engel observe the evolution of a pathogen from one that affects an animal host to one that affects the human population, and then ultimately a situation where human-tohuman transmission of the pathogen is enabled. The geographic spread of infectious agents is of course made possible by travel of both humans (for business, pleasure, or to gain a better life) and other animals (by way of humans, for trade, or by migration, as in the case of birds). The world waits for the H5N1 virus to evolve genetically to allow efficient human-to-human transmission and cause a pandemic. Of course, once a disease enters the human population it spreads rapidly to many geographic locations because of travel opportunities. Herring “anchors” avian influenza to the 1918 influenza pandemic and reveals that the media hype about avian influenza is largely unfounded. By anchoring H5N1 to the 1918 pandemic the media enhanced panic around the world, even though there was no scientific basis for making the claims that they did. When the data are dissected, it becomes clear that the extensive and severe threat from avian influenza portrayed in the media could not and cannot be upheld, and that this is because of a lack of understanding of the virus’s interactions with other pathogens and of links to underlying social circumstances. Furthermore, because the outbreaks of avian influenza have been in Asia, the blame has been lain there, and specifically with Asian agricultural practices and poor farming families. As a consequence, there has been an inevitable stigma associated with Asian agricultural markets. Interestingly, though, in small-village chicken flocks of low density, the virus has difficulty surviving, and it is when it enters factory farms that the virus spreads rapidly by droplet transmission, exacerbated by global trade networks. Clearly, there is a situation where scientific data are being ignored at the expense of fuelling worries about a possible pandemic of avian influenza, and this extends to making unfounded links to Asia as being a place where contact should be stopped. In the UK some restaurants took chicken and eggs off their menus in 2005; even though consumption of chickens and eggs is not a method of transmission of avian flu to humans, it was also recommended that they should be cooked thoroughly. It is still a rare disease, with 241 cases reported by the WHO from 2003–6, and 141 deaths (data from www

Understanding Health

•

19

.who.int). In some respects the “fear” of avian flu can be likened to the fear of leprosy in Medieval Europe, but also to situations in developing countries such as India today, where the lack of understanding or appreciation of why the infection appears in a person or population leads to speculation and associated stigma. In Medieval Europe much of the attitude to leprosy stemmed from the Bible, although it is now believed that the disease was not described there (Hulse 1976). The stigma attached to leprosy, however, led to (correctly?) diagnosed people being segregated into leprosy hospitals because of the associated shame (Richards 1977). The chapters by Jones-Engel and Engel and by Herring emphasize the importance of the movement of people and animals, and the interaction of the two, in enabling the respective infections to flourish and be maintained in both the animal host and the human recipient. Of course the picture is not that simple, with a myriad of variables affecting the success of the infections in humans and nonhumans and making a cross-disciplinary approach necessary for assessing health risks. Interactions between Diseases Herring also notes that interactions between diseases should be considered when assessing health risk, because suffering from one disease may make a person more vulnerable to another; in fact, secondary infections played a major role in mortality in 20 percent of people suffering from the 1918 influenza pandemic. Leonard et al.’s chapter (Chapter 1) considers the relationship between different cardiovascular health problems in their Siberian study groups (reindeer hunters, fishermen, cattle herders, and cattle and horse herders). Data on obesity, weight, body-mass index, basal metabolic rate, energy expenditure, physical activity level, blood pressure, and lipid levels were documented. The researchers found that elevations in basal metabolic rates appear to contribute to low plasma lipid levels, but also to increased blood pressure in both males and females. However, obesity was greater in females because of reduced activity, a consequence of sociopolitical events in Soviet and Russian history having changed the way these people subsisted, but something also influenced by adaptation to environmental stressors. The interaction between socioeconomic and political changes and health can also be seen in the HIV that is the single most important risk factor today for the progression of dormant TB into full-blown clinical disease (Raviglione et al. 1995). Because the infection compromises the immune system, people are more likely to develop TB, and because TB is a disease of poverty and poor nutrition, the problem of contracting TB by immunocompromised HIV sufferers is exacerbated. Jones-Engel and Engel also note that primate populations are usually in very densely populated parts of the world, and often where people’s immune systems are compromised by HIV and tuberculosis. In fact, compromises in immune system strength may predispose to many diseases, including infection, and there are questions today about the strength of our immune systems and its relationship to our changing lifestyles and increases in allergic conditions (Hamilton 1998). Similarly,

20

•

Charlotte Roberts

immune strength in people who have never been exposed to specific pathogens links closely with the mobility of people and the pathogens they encounter. For example, people today travel miles to exotic countries where the climate, environment, food, and water may harbor very different pathogens to which they have no resistance; they may also take pathogens home with them. In the past this was also the case—witness the impact of European contact on the Americas from the fifteenth century AD when new infections were introduced (Larsen and Milner 1994). While mobility of people today is a focus for research exploring the impact of movement on health, scientific analyses of skeletal samples to assess the impact of travel on the origin and transmission of disease (and change in organism strains) has yet to be accomplished; the analytical methods exist, but to a certain extent the interpretative tools for the generated data need developing (see, e.g., Budd et al. 2004 for a review of tracing mobility in the past). Differential Targeting of Groups of People by Disease Another theme that comes through in all the chapters is that different groups of people may be differentially affected by disease and this relates to their relative exposures to pathogens, their immune response, and what particular lifestyle they encounter. Herring notes that the 1918 flu affected young adults, males, pregnant women, immigrant and economically disadvantaged communities, and people with lack of access to health care. This is in contrast to the assumptions that the virus was “democratic” in how it affected humans. As for H5N1, there appears to be a focus on the young (and poor), which suggests that avian influenza has affected the human population before, and that older adults are immune to the infection. Newly introduced diseases are meant to affect everybody, but clearly this is not the case here. Leonard et al’s chapter, which considers health and lifestyle change in indigenous Siberians, finds higher levels of obesity, and high body-mass index in women, which in turn are related to energy expenditure and activity patterns (male activity is higher, especially in the case of Evenki reindeer herders). Furthermore, children in the same groups showed higher rates of under-nutrition and -growth when the adults became obese through developing sedentism. Clearly, there is a need to carefully analyze the interactions of diseases and their distribution among the various age and sex groups, but also socioeconomic statuses, and to think closely about the potential exposure of different groups of people to specific organisms. Who does the hunting and herding? Who spends most time indoors? Who is the least active? What foods does each group have access to? Lifestyle Change and Mobility of Humans and Animals Herring’s and Jones-Engel and Engel’s chapters clearly show the relationship between human disease (avian flu and nonhuman primate transmitted zoonoses) and move-

Understanding Health

•

21

ment of both people and animals. People move much more frequently than ever before for business and pleasure (local and long distance) and for many other reasons, such as fleeing from war-torn areas or accessing urban centers for work. They also are responsible for the organization of the movement of animals, and those people who encounter nonhuman primates in their travels are susceptible to the simian retroviruses and other zoonoses. Important here is the consideration of the groups of people who travel; are they the rich who travel for business and pleasure or are they the poor who move away from dangerous areas of the world? Perhaps, more than ever before, more controls on the movement of animals in particular may help prevent the transmission of zoonotic disease—witness the 2001 foot-and-mouth outbreak in the UK and the controls on movement of animals then, and the controls put in place in Fife, Scotland (April 2006), to prevent the spread of avian flu. In all three chapters we see the impact of lifestyle change on disease occurrence: Jones-Engel and Engel have emphasized the impact of tourism on the possible increase in contact of humans and non-human primates and resulting zoonoses; parks, markets, nature reserves, monkey temples, zoos and sanctuaries, and “performing animal situations” predispose humans to contracting primate disease even though there is probably little awareness on the tourist’s part of the possible health dangers. The term lifestyle change of course encompasses that increase in travel, but this is also covered in both Herring’s and Leonard et al.’s chapters in different ways. Clearly, developments in farming methods and international trade have fuelled this problem, and those underlying factors need addressing. Leonard et al. discuss their data on cardiovascular disease in Siberia in relation to social and economic changes that have impacted dietary and activity patterns, and they touch on the trend by some to move into urban centers from a previously rural existence. The effect on health of living in urban and rural environments is of course relevant to all the chapters, and it emphasizes the impact of geographical location on what diseases may be present (and, in the case of markets, many people and their animals from different geographic locations may be present). It also highlights the need to consider health risks at the local and the regional levels. The “pathogen internet” (Jones-Engel and Engel) is alive and well. Evolution of, and Adaptation to, Disease Evolution of, and adaptation to, disease is also a key theme that comes through in the chapters. Leonard et al. rightly point out that we must consider both evolutionary/ adaptive and biocultural perspectives in understanding health risks and diseases in transition. Jones-Engel and Engel note that characteristics of enzoonotic organisms, human host populations, nonhuman (reservoir) populations, vector species, and the environment are constantly changing and evolving with human and nonhuman organisms that are concurrently adapting; this provides almost endless new possibilities for infectious disease transmission. Leonard et al. discuss an adaptive response to cold stress in Siberian groups that is seen in MtDNA mutations for greater cellular heat

22

•

Charlotte Roberts

production and increased metabolic rate; this is common in North Asian populations. The origins of these mutations appear to coincide with the initial expansion of human populations to arctic climates and adaptation to cold stress. Thus, both physiological and genetic adaptation is noted for their study groups. Herring’s treatment of avian flu also considers the evolution of the disease organism and the fact that we now wait for the avian flu virus to evolve such that human-to-human transmission is possible (which is inevitable), but she notes that this could take variable amounts of time. Concepts of Disease and Risk Concepts of disease, and a disease’s perceived risk, are highly relevant to how a population understands the disease, how it is contracted and transmitted, and what measures may be put in place to combat risk. These concepts have changed considerably over time as we have a developed medical knowledge, especially about the pathogens that cause disease. For example, TB in a Vietnamese refugee group in New York State was believed to be caused by hard manual labor, smoking, alcohol, and lack of sleep (Carey et al. 1997); and, in a group in Vietnam, worrying too much and being unhappy also contributed to their conceptions of disease (Long et al. 1999). Clearly, the treatment for TB that these groups would accept (and its degree of success) would very much be determined by their concepts of TB. Does the world’s population truly know what is necessary for the transmission of bird flu to humans? Do they know that primate disease can be transmitted to humans and how? And are native Siberian groups aware of the reasons why they have cardiovascular health problems? Their understanding of their health problems thus affects the relative risk they have in contracting these conditions. Herring usefully discusses the confusion in the population about what actually led to the 1918 flu and the Black Death, and develops this theme for avian flu. Clearly, the media’s perception of the risk of avian flu to the human population was incorrect and rooted in their misunderstanding of the effects of the 1918 flu pandemic. Concluding Remarks Many common threads in the three chapters of this section (covering health, risks, and diseases in transition) have been highlighted. I draw nine principal messages from them: • the need to take a holistic view of health and to consider all available types of evidence to try and explain the patterns seen; furthermore, a range of practitioners from a variety of disciplines are needed to assess and deal with health risks;

Understanding Health

•

23

• both evolutionary/adaptive and biocultural (including behavioral) mechanisms need to be considered, because both have their part to play in determining health status; • diseases interact with each other and any one may impact on the other’s appearance; • specific groups of people may be more at risk than others (depending on age group, biological sex, social status, occupation, etc.); • lifestyle change has a significant effect on health risk; • emerging infections will play a major part in the world’s health for some time to come, more likely for the poor, but the rich are certainly not immune; • we need to understand the mechanisms underlying the health problems experienced in order to be able to assess real risk; • in assessing risk, we should be able to model the probability that an adverse event will occur under a particular set of circumstances; this could help health policy development; • more education is needed to explain to all (including the media) why specific health problems occur; this may prevent attaching stigma to certain diseases and their consequences. Research on health involves evaluating disparities in health associated with the experience of risk and vulnerability, with susceptibility and adversity being key constituents. By considering all the possible variables that might impact on health, and looking at differences in experience of disease among population groups in a geographic sense, disparities can be highlighted, the reasons for those disparities explored, and recommendations can be made for future management. As the organizers of this conference noted, we need to understand the relationships between processes and outcomes in health problems in order to overcome ill health. Risk analysis can support rational decision making in the face of uncertain probability, but often the underlying processes through which risk disparities in health develop are unknown. If hazards to health are identified (i.e., the probability of the introduction or escape of pathogens or other hazards into or out of humans or animals), risk can be assessed, managed, and communicated to people. This is needed to enable policy development. References Anderson, T. 1997. A probable osteoma of the mandible from Northamptonshire, Great Britain, Journal of Paleopathology 9(2): 69–72. Armelagos, G.J. 1998. Health and disease in prehistoric populations in transition. In Understanding and Applying Medical Anthropology, ed. P.J. Brown. Mountain View, CA: Mayfield Publishing Co., 59–69.

24

•

Charlotte Roberts

Barrett, S.P. and C. O’Hara. (2005). 25 years on—the Journal of Hospital Infection and the ‘backfile’ project, Journal of Hospital Infection. Bathurst, R. and J.L. Barta. 2004. Molecular evidence of tuberculosis induced hypertrophic osteopathy in a 16th century Iroquian dog. Journal of Archaeological Science 31(7): 917–25. Brothwell, D. 1991. On the zoonoses and their relevance to palaeopathology. In Human Paleopathology: Current Syntheses and Future Options, eds. D. Ortner and A.C. Aufderheide. Washington, D.C.: Smithsonian Institution Press, 18–22. Brown, K. 2000. Ancient DNA applications in human osteoarchaeology. In Human Osteology in Archaeology and Forensic Science, eds. M. Cox and S. Mays. London: Greenwich Medical Media, 455–73. Brown, P.J. 1997. Culture and the global resurgence of malaria. In The Anthropology of Infectious Disease: International Perspectives, eds. M.C. Inhorn and P.J. Brown. Amsterdam: Gordon and Breach Publishers, 119–41. Budd, P., A.R. Millard, C. Chenery, S. Lucy, and C.A. Roberts. 2004. Investigating population movement by stable isotope analysis: a report from Britain. Antiquity 78: 127–41. Carey J., M. Oxtoby, L. Nguyen, V. Huynh, M. Morgan, and M. Jeffrey. 1997. Tuberculosis beliefs among recent Vietnamese refugees in New York State. Public Health Reports 112:66–72. Cohen, M.N. 1989. Health and the Rise of Civilization. London: Yale University Press. Cohen, M.N. and G.J. Armelagos, eds. 1984. Paleopathology at the Origins of Agriculture. London: Academic Press. Davies, P.D.O. 1995. Tuberculosis and migration. Journal of Royal College of Surgeons of London 29:113–18. Eaton, S.B. and S.B. Boyd. 1999. Hunter-gatherers and human health. In The Cambridge Encyclopedia of Hunters and Gatherers, eds. R.B. Lee and R. Daly. Cambridge: Cambridge University Press, 449–545. Froment, A. 2001. Evolutionary biology and health of hunter-gatherer populations. In HunterGatherers: An Interdisciplinary Perspective, eds. C. Panter-Brick, R.H. Layton, and P. Rowley-Conwy. Cambridge: Cambridge University Press, 239–66. Hamilton, G. 1998. Let them eat dirt. New Scientist July 18: 26–31. Hulse, E.V. 1976. The nature of biblical leprosy and the use of alternative terms in modern translations of the bible. Medical History 20(2): 203. Larsen, C.S. 1997. Bioarcheology: Interpreting Behaviour from the Human Skeleton. Cambridge: Cambridge University Press. Larsen, C.S. and G. Milner, eds. 1994. In the Wake of Contact: Biological Responses to Conquest. New York: Wiley-Liss. Lieban, R.W. 1973. Medical anthropology. In Handbook of Social and Cultural Anthropology, ed. J.J. Honigman. Chicago: Rand and McNally, 1031–72. Long, N.H., E. Johnasson, V.K. Diwan, and A. Winkvist. 1999. Different tuberculosis in men and women: beliefs from focus groups in Vietnam. Social Science and Medicine 49: 815–22. Mascie-Taylor, C.G.N. and G.W. Lasker. 1988. Biological Aspects of Human Migration. Cambridge: Cambridge University Press. McElroy, A. and P.K. Townsend. 1996. Medical Anthropology in Ecological Perspective. Boulder, CO: Westview Press.

Understanding Health

•

25

Merrett, D. and S. Pfeiffer. 2000. Maxillary sinusitis as an indicator of respiratory health in past populations. American Journal of Physical Anthropology 111: 301–18. Morse, S.S. 1995. Factors in the emergence of infectious diseases. Emerging Infectious Diseases 1(1): 7–15. Park, K. 1993. Black death. In The Cambridge World History of Human Disease, ed. K. Kiple. Cambridge: Cambridge University Press, 612–16. Raviglione, M.C., D.E. Snider, and A. Kochi. 1995. Global epidemiology of tuberculosis morbidity and mortality of a worldwide epidemic. Journal of the American Medical Association 273(2): 220–26. Richards, P. 1977. The Mediaeval Leper and His Northern Heirs. Cambridge: D.S. Brewer. Roberts, C.A. and M. Cox. 2003. Health and Disease in Britain: Prehistory to the Present Day. Stroud: Sutton Publishing. Roberts, C.A. and K. Manchester. 2005. The Archaeology of Disease. Stroud: Sutton Publishing. Roberts, D.F., N. Fujiki, and K. Torizuka. 1992. Isolation, Migration and Health. Cambridge: Cambridge University Press. Sargent, C.F. and T.M. Johnson. 1996. Medical Anthropology: Contemporary Theory and Method. Westport, CT: Praeger. Smith, E.R. 1988. The Retreat of Tuberculosis, 1850–1950. London: Croom Helm. Steckel, R.H. and J.C. Rose, eds. 2002. The Backbone of History: Health and Nutrition in the Western Hemisphere. Cambridge: Cambridge University Press. Swedlund, A.C. and G.J. Armelagos, eds. 1990. Disease in Transition: Anthropological and Epidemiological Approaches. New York: Bergin and Garvey. Waldron, T. 1994. Counting the Dead: The Epidemiology of Skeletal Populations. New York: Wiley. Walker, P.L. and S.E. Hollimon. 1989. Changes in osteoarthritis with the development of a maritime economy among southern Californian Indians. International Journal of Anthropology 4(3): 171–83. Wilson, M.E. 1995. Travel and the emergence of infectious disease. Emerging Infectious Diseases 1(2): 39–46. Wood, J.W., G.R. Milner, H.C. Harpending, and K.M. Weiss. 1992. The osteological paradox: problems of inferring prehistoric health from skeletal samples. Current Anthropology 33(4): 343–70.

• 1 • Health Consequences of Social and Ecological Adversity Among Indigenous Siberian Populations Biocultural and Evolutionary Interactions William R. Leonard, J. Josh Snodgrass, and Mark V. Sorensen

Anthropological Perspectives on Human Biological Variation and Health In studying human health and well-being, biological anthropologists differ from most biomedical scientists in that we draw explicitly on both evolutionary and biocultural models (Stinson et al. 2000). As anthropologists we are interested in understanding the origin and nature of biological variation as well as the proximate social, political, and economic determinants of variation in human health. Thus, we recognize that human biological variation in health is shaped by adaptive responses to stress and adversity in our evolutionary past, as well as by ongoing social and ecological challenges to our health in the modern world. Increasing rates of obesity and cardiovascular (CV) disease are among the most pervasive threats to human health throughout the world. Today we find that these problems are emerging in many traditional societies where they were virtually unknown less than a generation ago. This is true of the indigenous populations of the vast Siberian region of Russia (Leonard et al. 1996, 2002b; Snodgrass 2004; Sorensen et al. 2005). In their traditional subsistence lifestyle, these populations were relatively

Health Consequences of Social and Ecological Adversity

•

27

protected from obesity and associated metabolic diseases by virtue of their high levels of physical activity and daily energy expenditure (see Shephard and Rode 1996). Today, as these native Siberian groups adopt more sedentary lifestyles, overweight and obesity have emerged as growing problems (Snodgrass et al. 2006a). Yet, while many of the factors contributing to these health changes (e.g., changes in activity patterns and food availability) are the same as in other parts of the world, unique sociopolitical events in Soviet and Russian history have produced patterns that are distinct to indigenous Siberian groups. Since the early 1990s, we have been studying aspects of human biological variation and health change among indigenous Siberian populations. Indigenous Siberians face multiple, interacting sources of adversity and risk, including: 1) a severe, marginal environment characterized by extremes in temperature and low biological productivity; and 2) ongoing social, economic, and political changes that constrain adaptive options and threaten health and well-being. In our work, we have examined both ecological and social determinants of health in native Siberians, specifically addressing three major issues: 1) biological adaptations to arctic climes, 2) the health consequences associated with lifestyle “modernization,” and 3) the impact of postSoviet political-economic changes on health. As initially conceptualized, these three research domains were largely independent of one another. However, as our research has continued, it has become clear that these domains are interconnected. We have found that key adaptations to cold stress appear to play important roles in shaping how lifestyle changes influence health in native Siberians. In other words, we have seen that basic aspects of human biological variation have important implications for determining how lifestyle and environmental factors may influence health. This chapter will summarize some of the key findings from our Siberian work, highlighting the interplay between the evolutionary and biocultural domains. Our work has shown that the ecological and social dimensions of risk faced by native Siberians do not impact the biology and health of all segments of population in the same way. Indeed, whereas the initial post-Soviet period was associated with increased rates of obesity in many adult Siberians, children showed higher rates of undernutrition and growth stunting (Leonard et al. 2002b). Additionally, we have also seen that the health consequences of lifestyle change among native Siberians differ from those observed in other “modernizing” groups. Despite increasing rates of obesity, native Siberians continue to have low cholesterol and triglyceride levels, whereas high blood pressure (hypertension) has emerged as a major cardiovascular health problem. Variation in blood pressure and plasma lipid levels are jointly influenced by variation in body composition, lifestyle factors, and metabolism. Both blood pressure and lipid levels are positively associated with body weight and body fatness (Leonard et al. 2002a; Snodgrass 2004; Sorensen 2003). Conditions of marginalization and poverty that have emerged in post-Soviet Russia also contribute to elevated lipid levels (Sorensen et al. 2005) and blood pressure (Snodgrass 2004; Snodgrass et al. 2005b). In addition, as a means of adapting to

28

•

William R. Leonard, J. Josh Snodgrass, Mark V. Sorensen

their cold, marginal environments, native Siberians have elevated basal metabolic rates (BMR) (Leonard et al. 2002, 2005). These elevated BMRs exert a significant positive influence on blood pressure while having a depressing effect on cholesterol levels. These effects persist after controlling for the influence of body composition and lifestyle variables. These findings suggest that in the face of nutritional and lifestyle changes, elevated metabolic turnover in native Siberian populations may protect them against hyperlipidia while predisposing them to hypertension. The clustering of cardiovascular risk factors in native Siberians is, in many respects, the mirror image of what is observed among acculturating populations such as the Pima Indians of the southwestern US. With the Pima we see low BMRs associated with moderate to high cholesterol levels, very high rates of diabetes, and yet relatively low rates of hypertension (Ravussin 1995; Spraul et al. 1993; Weyer et al. 2000). The differences between the Pima and indigenous Siberians appear to be the product of the unique historical changes in lifestyle and socioeconomic status of the two groups and underlying differences in metabolism shaped by adaptation to different environmental stressors. Our research underscores the importance of linking biocultural and evolutionary perspectives in studying variation in human biology and health. Moreover, it highlights the critical roles that biological anthropologists can play in addressing major global health problems. Indigenous Siberians: Historical and Ethnographic Context Siberia spans over 13 million square kilometers and has one of the world’s lowest population densities, with a total native population of only about 1.3 million. The initial human settlement of most of Siberia occurred relatively recently; although some evidence suggests an early settlement of Arctic Siberia (c. 30,000 years ago; Pitulko et al. 2004), most studies point to a more recent date (less than 20,000 years ago) (Goebel 1999; Mote 1998). Prolonged contact between Russian and indigenous Siberian populations began in the late sixteenth century as Russian explorers and traders expanded eastward across the Siberian plain in search of animal pelts to procure for the burgeoning European fur market (Forsyth 1992; Slezkine 1994). With the emergence of the Soviet State, native Siberians experienced major transformations in their traditional ways of life. Starting in the 1930s, the Soviet government began the process of collectivizing the indigenous populations of the north. These reforms were designed to “modernize” native Siberian groups by drawing them directly under the control of the Soviet state (Slezkine 1994). With collectivization, animal herds (e.g., reindeer) were no longer held by individual families but were placed into communal herds that were controlled by state-run collectives. As a result, the native populations were forced to shift from their nomadic lifeways and give up key aspects of their traditional culture (e.g.,

Health Consequences of Social and Ecological Adversity

•

29

language, shamanism). This transition also resulted in the restructuring of subsistence roles. With the traditional nomadic lifeway, indigenous groups had an unstructured division of labor, with both men and women contributing to subsistence production. After collectivization, gender roles were much more clearly defined, with men being largely responsible for subsistence food production (herding, farming) while women were responsible largely for domestic tasks. The collapse of the USSR in 1991 created conditions of economic hardship and unprecedented declines in life expectancy throughout Russia and the former Soviet states (Leon et al. 1997; Notzon et al. 1998; McKee and Shkolnikov 2001). These reductions in life expectancy have been attributed to a number of factors, most notably increased alcohol consumption (Leon et al. 1997; McKee and Shkolnikov 2001), impoverishment and material deprivation (Bobak et al. 1998), social stress and lifestyle changes (Carlson 2000), and the deterioration of the health-care system (Field 1995). In Siberia, the fall of the Soviet Union resulted in the dismantling of many of the indigenous herding and farming cooperatives. Under the collective system, the Soviet government funded the transport of food and medical supplies into remote villages and herding brigades (Forsyth 1992; Hannigan 1991; Slezkine 1994). Following the collapse of the Soviet system, these shipments were greatly reduced or eliminated entirely, resulting in a return to traditional subsistence activities among many of the more remote populations (Fondahl 1997; Leonard et al. 2002a,b; Snodgrass 2004). Today in Siberia, we are seeing the emergence of greater heterogeneity in lifestyles. While many families have shifted away from traditional subsistence-based activities to move to larger villages and towns, many of these same families continue to herd, hunt, forage, and/or farm in order to supplement their diets. These social and economic changes continue to have important yet variable impacts on health among indigenous Siberian populations. Our research has documented increased mortality rates (Leonard et al. 1997), declining levels of childhood nutritional status (Leonard et al. 2002), and increased risks of CV disease (Snodgrass 2004; Sorensen 2003; Sorensen et al. 2005) among indigenous Siberians during the post-Soviet period. Over the last sixteen years, we have studied four different indigenous Siberian groups: 1) the Evenki reindeer herders, 2) the Ket fisherman, 3) the Buryat cattle herders, and 4) the Yakut cattle and horse herders. The geographic locations of these four populations are shown in Figure 1.1. Brief ethnographic descriptions of each group are presented in Box 1.1. These groups live in communities that span wide variation in lifestyles, ranging from small herding encampments (brigades) of 35–50 individuals, to villages of 400–600 residents, up to larger towns with population sizes of more than 1,000 individuals. Those living in traditional herding units continue to pursue largely a subsistence-based lifestyle. Those living in small collective villages and larger towns have greater access to market goods and the wage economy.

30

•

William R. Leonard, J. Josh Snodgrass, Mark V. Sorensen

• • • Ethnographic Background of Select Siberian Populations In our research in Siberia, we have worked with four indigenous populations: 1) the Evenki, 2) the Ket, 3) the Buryat, and 4) the Yakut. The Evenki are a Tungusicspeaking population of reindeer herders from the northern regions of the Siberian boreal forest (taiga) (Forsyth 1992). The Evenki population numbered approximately 30,000 at the last major census (Fondahl 1997). Additional information on the study population, as well as on the Evenki in general, can be found in Leonard and coworkers (1994, 1996, 2002a). The Ket are a central Siberian population structured around fishing. They are extremely small in number, and in the 1989 census numbered less than 1,200 (Fondahl 1997). The Ket are apparently a remnant of a considerably larger population, which was centered in the Yenisey valley at the time of initial Russian contact but which was subsequently decimated by epidemics of infectious disease (Forsyth 1992). The Ket language is unique and, based on available evidence, appears to be unrelated to any known languages. The Buryat are descendants of Mongol populations that settled in the meadowsteppe region around Lake Baikal at the boundary of the northern forest (Forsyth 1992). The Buryat language belongs to the Mongolic language family. At the time of initial Russian contact, the Buryat population was relatively large and increased substantially during the Russian and Soviet periods; at the last census, the Buryat population numbered over 400,000 (Fondahl 1997; Forsyth 1992). Most rural Buryat today subsist off the products of cattle, which are fed through locally cultivated crops (Humphrey and Sneath 1999). The Yakut (Sakha), members of the Turkic language family, number nearly 400,000 and are concentrated in northeastern Siberia (Fondahl 1997; Forsyth 1992; Snodgrass 2004). The Yakut traditionally practiced a complex and locally variable subsistence strategy that was largely dependent upon regional ecological conditions (Tokarev and Gurvich 1964). In remote parts of the taiga, the Yakut subsisted by hunting and fishing, while in the Lena River Valley the primary subsistence activity was transhumant pastoralism (primarily horse and cattle).

• • •

Health Consequences of Social and Ecological Adversity

•

31

Figure 1.1. Map of Siberia showing the geographic locations of the Buryat, Evenki, Ket, and Yakut.

Health Consequences of Social and Ecological Adversity Nutritional Consequences of the Post-Soviet Transition in the Evenki Social and economic changes in Russia appear to have strikingly different nutritional consequences for children and adults. This point is most evident when we examine our data from the Evenki, whom we studied from 1991 through 1995 during the initial phases of the post-Soviet transition, when Evenki reindeer herding cooperatives were being dismantled. During the years immediately following the fall of the Soviet Union, anthropometric indicators of childhood undernutrition dramatically increased in the Evenki. Table 1.1 shows the percentage of Evenki children under the age of 6 years who were classified as “stunted” (height-for-age Z-score < –2), “underweight” (weight-for-age Z-score < –2), or “wasted” (weight-for-height Z-score < –2) during the Soviet and post-Soviet periods. Rates of stunting (a measure of chronic, mild-to-moderate undernutrition) increased from 34 to 61 percent between 1991 and 1995. Similarly, the prevalence of underweight children more than doubled during this period, rising form 18 to 43 percent. Wasting, an index of acute undernutrition, also rose dramatically from 2 to 17 percent. These levels of childhood undernutrition are comparable to those seen among impoverished populations of the developing world (see de Onis et al. 2000). In contrast, among Evenki adults overweight and obesity are much more common than conditions of undernutrition. Based on body-mass indices (BMI), only 2 percent men and 5 percent of women were classified as “underweight” (BMI < 18.5

32

•

William R. Leonard, J. Josh Snodgrass, Mark V. Sorensen

Table 1.1. Indicators of growth status in Evenki children measured between 1991 and 1995, during the “Soviet” and initial “Post-Soviet” periods (adapted from Leonard et al. 2002b).

Measures Low Height-for-Age (“stunting”) Low Weight-for-Age (“underweight”) Low Weight-for-Height (“wasting”)

Soviet (n=101)

Post-Soviet (n=54)

34% 18% 2%

61%*** 43%*** 17%***

Differences between the “Soviet” and “Post-Soviet” groups are statistically significant at: ***P < 0.001 (Chi-Square Analyses).

kg/m2), whereas rates of overweight and obesity (BMI > 25 kg/m2) were 11 percent and 32 percent in men and women, respectively. As shown in Table 1.2, mean BMIs of adults remained relatively stable between 1991 and 1995 during the initial postSoviet transition. However, levels of body fatness significantly increased in both men and women during this time period. The reasons for the sharp declines in growth status and nutritional health for children likely stem from the increased isolation that Evenki experienced during the initial post-Soviet period. The shift away from the collectivized herding system coupled with the general economic decline in post-Soviet Russia resulted in less regular contact between the small Evenki settlements (villages, herding brigades) and larger urban centers in Siberia. Plane and helicopter transport into these remote areas were much less common during the early post-Soviet transition. Thus, medical supplies and nonlocal foodstuffs that were regularly brought in by plane or helicopter during the Soviet era, became more scarce after the herding collectives were dismantled. As a consequence, access to important high-quality weaning foods (e.g., condensed milk, cereals) and health care became more limited during this period. Table 1.2. Adult nutritional status of Evenki men and women (>18 years) measured between 1991 and 1995, during the “Soviet” and initial “Post-Soviet” periods.

Males

Females

Measures

Soviet (n=123)

Post-Soviet (n=46)

Soviet (n=67)

BMI (kg/m2) Body fat (%)

22.6+2.4 14.2+4.7

22.2+2.8 17.1+6.2**

24.0+4.6 29.9+5.7

Post-Soviet (n=123) 24.3+4.9 33.3+7.3***

Differences between the “Soviet” and “Post-Soviet” groups are statistically significant at: **P < 0.01, ***P < 0.001.

•

Health Consequences of Social and Ecological Adversity

33

The increasing levels of adiposity among adult Evenki between 1991 and 1995 was an unexpected result in light of the dramatic declines in children’s nutritional status. For the adults, it appears that while the quality of the diet declined, energy intake did not dramatically change. Rather, the more important change for many of the Evenki adults resulted from the shift away from reindeer herding and the adoption of a more sedentary lifestyle. This shift in activity levels helps to explain why percent body fatness significantly increased among Evenki adults during the initial postSoviet period, while BMIs remained constant. Reductions in activity levels and relative stability in energy intake resulted in higher levels of fatness (and reduced muscularity) without substantial increases in overall body mass. Reductions in energy expenditure and physical activity levels associated with modernization/urbanization of lifestyle are trends that are evident throughout native Siberian groups and appear to be contributing to the relatively high rates of overweight and obesity now seen in the region. The correlates and health consequences of these adult lifestyle changes are explored in the subsequent sections. Lifestyle Correlates of Overweight and Obesity in Native Siberians The emergence of “overweight” and “obesity” in adulthood is a problem now common among all native Siberian groups, not just the Evenki (Snodgrass et al. 2006a). Table 1.3 presents the prevalence rates of overweight and obese adults based on the BMI for the four indigenous groups we have studied. Rates of overweight and obesity are systematically and significantly higher in women than in men (36 percent of women are overweight and obese versus 25 percent of men; P < 0.01). Table 1.3. Percent overweight and obese among indigenous Siberian populations (adapted from Snodgrass et al. 2006).

Population

Sex

n

Overweighta

Obeseb

Evenki

M F M F M F M F M F

148 174 14 19 51 80 150 264 363 537

10 22 21 37 22 29 25 24 18 24

1 10 7 5 8 15 11 13 7 12

Ket Buryat Yakut Combined a b

BMI: 25.0 – 29.9 BMI > 30.0

34

•

William R. Leonard, J. Josh Snodgrass, Mark V. Sorensen

Urbanization of lifestyle is associated with greater body mass (BMI) and higher levels of adiposity in native Siberians. Figure 1.2a shows variation in the BMI by residence location. Women have systematically higher BMIs than men (24.5 vs. 23.4 kg/m2; P < 0.001), but the degree of urbanization has a stronger influence on men’s body mass, with those living in the towns having significantly higher BMIs than those living in the brigades or villages (P < 0.05). When we look at percent body fatness (Figure 1.2b), as estimated from the sum of four skinfolds, a slightly different picture emerges. In this case, town residence is associated with significantly greater adiposity in both men and women (P < 0.01 for both sexes). Thus, female BMIs do not vary significantly with residence location, while levels of body fatness do vary. Influence of Energy Expenditure on Risks of Overweight and Obesity The marked gender differences in body weight and obesity levels between men and women appear to be partly attributable to differences in energy expenditure and activity patterns. Men have significantly higher BMRs than women, a difference that in part, reflects differences in body size. In addition, both men and women have BMRs that are elevated above reference values. Figure 1.3 shows the relationship between BMR and fat-free mass (FFM; kg) in men and women, compared to reference

Figure 1.2. Mean (+SEM) (a) BMI (kg/m2) and (b) percent body fat of indigenous Siberian men and women living in different size communities. Town-dwelling men have significantly higher BMIs than their counterparts living in either the herding brigades or villages (P < 0.05). For body fatness, town-dwelling men and women both have significantly greater adiposity than their counterparts living in less urbanized settings (P < 0.01).

Health Consequences of Social and Ecological Adversity

•

35

Figure 1.3. Relationship between basal metabolic rate (kcal/day) and fat-free mass (kg) in indigenous Siberian men and women compared to estimated values from reference equations of Poehlman and Toth (1995). BMRs of Siberian men average 1,746 kcal/day (7,400 kJ), 16 percent higher than predicted values. Siberian women average 1,388 kcal/day (5,800 kJ), 19 percent higher.

norms compiled by Poehlman and Toth (1995). Men have average BMRs of 1,746 kcal/day (7,400 kJ), as compared to 1,388 kcal/day (5,800 kJ) in women (P < 0.01). Siberian men have metabolic rates that are, on average, 16 percent above predicted values, whereas Siberian women deviate by +19 percent (P < 0.01, for both sexes). Total energy expenditure (TEE; kcal/day) varies markedly by gender and level of urbanization. Men have significantly higher levels of TEE. This pattern is true for the sample as a whole, and is true for each of the groups that we have studied (see Table 1.4). Men’s daily energy expenditure averages 600–700 kcal/day (2,500–2,900 kJ) more than women’s (2,773 kcal/day [11,600 kJ] vs. 2,106 [8,810 kJ] kcal/day; P < 0.001). These differences are consistent across all the ethnic groups we have studied and reflect gender differences in both body weight and activity levels. Activity patterns also appear to be influenced by lifestyle differences. Table 1.4 shows variation in the Physical Activity Level (PAL) ratio—the ratio of TEE to BMR (FAO/WHO/UNU 1985; James and Schoefield 1990)—in men and women by residence location. Men show only small declines in PAL with urbanization. Brigade and village dwellers have similar PALs (about 1.74), and those of town residents are slightly lower (1.66). In contrast, women show significant and marked declines associated with residence location. Women living in the brigades and villages have moderate daily activity levels (PALs = 1.62–1.66), whereas those living in towns have significantly lower PALs, consistent with very sedentary lifestyles (1.4–1.5; P < 0.05).

36

•

William R. Leonard, J. Josh Snodgrass, Mark V. Sorensen

Table 1.4. Body weight, basal metabolic rate (BMR), total energy expenditure (TEE) and physical activity levels (PAL) of adult men and women of native Siberian populations (adapted from Leonard et al. 2005).

Population

Sex

Weight (kg)

BMR (kcal/d)

TEE (kcal/d)

PAL (TEE/BMR)

Evenki, herders

M F M F M F M F M F

61.3 50.6 57.5 51.7 56.4 63.7 62.3 50.1 72.2 65.2

1619 1363 1543 1278 1622 1346 1622 1233 1848 1533

2805 2211 2669 2101 2316 1664 2727 1860 3102 2298

1.74 1.62 1.73 1.67 1.55 1.23 1.69 1.51 1.68 1.50

Evenki, village Evenki, town Ket, village Yakut, town

More fine-grained analyses from our most recent work among the Yakut pro-

Figure 1.4. Mean (+SEM) physical activity levels (TEE/BMR) of indigenous Siberian men and women living in different size communities. Town-dwelling women have significantly lower (P < 0.05) daily activity levels than women living in either the herding brigades or small villages. Men show more modest declines in activity levels with urbanization.

Health Consequences of Social and Ecological Adversity

•

37

vides additional support for the link between energy expenditure and changes in subsistence patterns. In particular, Snodgrass et al. (2006b) have shown that greater participation in subsistence activities (e.g., hay cutting, fishing, hunting, foragers) is associated with higher levels of daily energy expenditure. Conversely, greater reliance on market food items is associated with lower levels of daily activity and energy expenditure. Overall, the patterns of variation in energy expenditure are consistent with those observed in body weight and body composition. Women have significantly lower levels of basal and total energy expenditure than men. Lifestyle urbanization also appears to be associated with reductions in physical activity levels; however, these declines are more dramatic in women than men. Cardiovascular Health Risks Plasma lipids. Despite the increasing levels of overweight and obesity in native Sibe-

rians, we have found that plasma lipid levels are generally quite low (Leonard et al. 1994, 2002a; Mosher 2002; Sorensen et al. 2005). Figures 1.5 and 1.6 compare total and low-density lipoprotein cholesterol (LDL-C) levels in the Evenki, Buryat, and Yakut to the US 50th centiles from the US NHANES III survey (NIH, 2002). Total cholesterol levels of all the Siberian groups fall well below the US median values. The Evenki and Buryat track at about the US 5th centile; whereas the Yakut fall about about the 25th centile. LDL cholesterol levels are also low and show a similar pattern of variation to total cholesterol. For the Evenki and Buryat, men have LDL levels that that approxi-

Figure 1.5. Total cholesterol levels (mg/dL) by age group for men and women of three indigenous Siberian populations (Buryat, Evenki, and Yakut) compared to the US 5th and 50th centiles.

38

•

William R. Leonard, J. Josh Snodgrass, Mark V. Sorensen

Figure 1.6. LDL-cholesterol levels (mg/dL) by age group for men and women of three indigenous Siberian populations (Buryat, Evenki, and Yakut) compared to the US 5th and 50th centiles.

mate the US 15th centile, whereas women approximate the 25th centile. Among the Yakut, both men and women have LDL levels that fall between the 25th and 50th US centiles. Blood pressure. In contrast to the plasma lipid levels, Siberian men and women have

blood pressure levels that fall above the US reference values. Figures 1.7 and 1.8 show systolic and diastolic blood pressure values of Evenki, Buryat, and Yakut men and women compared to the US 50th centiles (Drizd et al. 1986). With few exceptions, systolic and diastolic blood pressure levels fall at or above the US median values for both men and women. Mean blood pressures (systolic/diastolic) in men are 135/86 mmHg in the Buryat, 133/80 mmHg in the Yakut, and 126/81 mmHg in the Evenki. Among women, the averages are: 136/84 mmHg in the Buryat, 127/82 mmHg in the Evenki, and 120/75 mmHg in the Yakut. Overall, 57 percent of men and 47 percent of women have elevated blood pressure (i.e., systolic BP > 120 mmHg and diastolic BP > 80 mmHg), as defined by most recent NIH (2004) recommendations. Correlates of cardiovascular risk factors. The cluster of cardiovascular risk factors observed in native Siberian groups—rising rates of obesity with the persistence of low lipid levels but elevated blood pressure levels—are distinct from what we find in many other “modernizing” populations around the world (e.g., the Pima Indians; see Weyer et al. 2000). This may reflect the fact that the types of lifestyle changes observed in post-Soviet Russia are different from those typically seen among other acculturating groups. In addition, the differences also raise the question of whether

Health Consequences of Social and Ecological Adversity

•

39

Figure 1.7. Systolic blood levels (mmHg) by age group for men and women of three indigenous Siberian populations (Buryat, Evenki, and Yakut) compared to the US 50th centile.

Figure 1.8. Diastolic blood levels (mmHg) by age group for men and women of three indigenous Siberian populations (Buryat, Evenki, and Yakut) compared to the US 50th centile.

key physiological or genetic adaptations among native Siberians may be structuring the health consequences associated with lifestyle change. Recent work with US and Nigerian populations has shown that BMR exerts a positive influence on blood pressure after controlling for the influence of age, sex, and body composition (Luke et al.

40

•

William R. Leonard, J. Josh Snodgrass, Mark V. Sorensen

2004). These findings suggest that the elevated BMRs of native Siberian groups may partly contribute to their high blood pressure (Snodgrass 2004). To examine these issues, we used multiple regression analyses to explore the joint influence of anthropometric, lifestyle, and metabolic correlates of blood pressure and cholesterol levels. Table 1.5 shows the multiple regression model for systolic blood pressure (n=289 individuals). Note that neither residence location (urbanization) nor smoking status exerts a significant influence on blood pressure, whereas BMI and age are both positively associated with blood pressure levels. In addition, BMR exerts a significant positive influence on systolic blood pressure, even after controlling for the other covariates. We find that of all the variables entered into the model, BMR is the strongest predictor of blood pressure. The results are similar for diastolic blood pressure, with BMR exerting a positive effect after controlling for the same covariates. However, the magnitude of the effect is smaller (regression coefficient = 0.004; P = 0.11). Table 1.6 presents the results of a multiple regression analysis of the correlates of LDL cholesterol variation. In this model (n = 175 individuals), residence location has a significant negative influence, suggesting the more urbanized Siberians have lower LDL-C levels. As expected BMI is positively associated with LDL-C. After controlling for the lifestyle and age covariates, BMR shows a significant negative association with LDL-C. This implies that the high metabolic turnover in native Siberians have a protective effect with regard to plasma lipid levels. In the total cholesterol model, the coefficient for BMR is also negative, but the effect does not reach statistical significance. For a subsample of 154 individuals, food consumption data were also available, allowing us to explore the influence the following dietary correlates of blood pressure and lipid levels: total energy intake (kcal/day), protein intake (g/day), fat intake (g/ day), and alcohol intake (g/day). Of these dietary parameters, only energy intake was Table 1.5. Multiple regression analysis of the correlates of systolic blood pressure in native Siberians.

Variable Constant Sex (0=female; 1=male) Age (years) Smoking (0=no; 1=yes) Urbanization (0=village; 1=town) BMI (kg/m2) BMR (kcal/day) Model R2 = 0.19; P < 0.001; n = 289

Coefficient (SE)

P

73.84 (7.62) 4.00 (2.87) 0.31 (0.11) 1.34 (2.18) –4.29 (2.82) 0.83 (0.32) 0.016(0.005)

< 0.001 0.165 0.006 0.537 0.129 0.010 0.003

Health Consequences of Social and Ecological Adversity

•

41

Table 1.6. Multiple regression analysis of the correlates of LDL cholesterol in native Siberians.

Variable Constant Sex (0=female; 1=male) Age (years) Smoking (0=no; 1=yes) Urbanization (0=village; 1=town) BMI (kg/m2) BMR (kcal/day)

Coefficient (SE)

P

81.38 (16.27) 8.37(6.45) 0.32 (0.25) –0.96 (5.04) –15.06 (5.59) 1.83 (0.65) –0.03 (0.01)

< 0.001 0.196 0.216 0.850 0.008 0.006 0.013

R2 = 0.15; P < 0.001; n = 175

a significant predictor of systolic BP (regression coefficient = 0.003; P = 0.04) when included with the other independent variables from the model in Table 1.5. Dietary fat intake approached statistical significance (coefficient = 0.072; P = 0.069). With the inclusion of either total energy intake or fat intake into the model, age, BMI and BMR all remain significant predictors of BP (energy model R2 = 0.24; P < 0.0001; fat model: R2 = 0.23; P < 0.0001). For LDL-C, none of the dietary parameters were significant correlates when entered with the other independent variables from the model presented in Table 1.6. We have also explored the lifestyle correlates of cardiovascular risk factors in greater detail among the Yakut. Sorensen et al. (2005) found lower total cholesterol and LDL-C levels among Yakut living in more urbanized settings and consuming a more market-oriented diet. The poorer lipid profile of the more rural Yakut reflects the conditions of marginalization that have occurred in smaller, more isolated communities following the collapse of the Soviet Union and the breakup of the collective farm system. These conditions contributed to more limited food choices, poorer dietary quality, and higher fat consumption among the rural Yakut (Sorensen 2003; Sorensen et al. 2005). In addition, Snodgrass and colleagues (2005b) found that higher blood pressure was associated with lower income levels in both Yakut men and women after controlling for age and body composition. Together, these findings suggest that conditions of social and economic marginalization in post-Soviet Siberia had significant effects on the cardiovascular health of indigenous groups. Thus, the distinctive clustering of cardiovascular risks in native Siberians appears to reflect the interaction of unique socioeconomic changes with underlying metabolic adaptations.

42

•

William R. Leonard, J. Josh Snodgrass, Mark V. Sorensen

Discussion Native Siberians face the challenge of adjusting to the interacting risks posed by both their marginal physical environment and ongoing socioeconomic and lifestyle changes. The economic collapse in post-Soviet Russia that occurred during the mid1990s had profound health consequences for native Siberian groups. This transition was associated with marked declines in the nutritional status of young children, while adults showed largely the opposite effects—low rates of undernutrition and increased levels of body fatness. These divergent trends have produced a phenomenon that is now increasingly common among urbanizing populations in the developing world— the co-occurrence of adult overweight and obesity with high levels of childhood undernutrition (Caballero 2001; Popkin 2001, 2002; Cameron et al. 2006). The dramatic declines in growth status witnessed among the Evenki children underscore how sensitive children’s growth is to “environmental quality” (Stinson 1985). Such examples of “negative secular trends” (i.e., declines in childhood growth and reduced adult stature) are relatively uncommon in recent human history and tend to be associated with conditions of socioeconomic hardship and impoverishment similar to those recently observed in Russia. Tobias (1985, 1986), for example, demonstrated negative secular trends in stature among South African blacks during the Apartheid era. Similarly, Ellison and Kelly (2005) have recently documented substantial declines in the growth rates of British children living on the Channel Islands under German occupation during World War II. Among adults, the growing rates of overweight and obesity in the Evenki and other Siberian groups are partly attributable to the shift away from traditional, subsistence activities to a more settled, urbanized way of life. In their traditional subsistence lifestyle, indigenous Siberians and other arctic populations were relatively protected from obesity and associated metabolic diseases by virtue of their elevated BMRs and their high levels of physical activity combining to produce high levels of daily energy expenditure (Godin and Shephard 1973; Heinbecker 1931; Milan and Evanuk 1967; Rodahl 1952; Rode and Shephard 1971). However, with the transition from a “traditional” to a more “modern” lifestyle, substantial and rapid changes in health are evident, including increases in levels of body fatness, declines in aerobic capacity, and increased rates of cardiovascular diseases (Naya et al. 2002; Rode and Shephard 1984, 1994; Shephard and Rode 1996; Young et al. 1995). Our research indicates that important gender and lifestyle differences in energy expenditure and activity levels strongly contribute to the patterns of obesity and cardiovascular disease risks seen among native Siberians. Men have significantly higher levels of energy expenditure (both basal and total) than women. This difference appears to underlie the marked differences in obesity rates between men and women. Based on the data compiled in Table 1.3, rates of overweight and obesity are 44 percent higher in Siberian women than in men. The lower daily energy expenditure and physical activity levels seen in women partly reflect the marked gender division of labor that is seen in native Siberian

Health Consequences of Social and Ecological Adversity

•

43

groups. Physically demanding subsistence activities such as herding animals (Evenki and Buryat) and hay cutting (Yakut) are performed mostly by men, whereas women are generally responsible for cooking, child care, and other domestic activities. This sharp differentiation between male and female roles has its origins in the restructuring of subsistence activities that occurred with Soviet collectivization (Forsyth 1992; Fondahl 1998). These differences have become exaggerated with urbanization and the adoption of more sedentary lifestyles. Indeed, as noted in Figure 1.4, gender differences in physical activity levels are most dramatic among native Siberians living in the most urbanized settings (towns). Although lifestyle modernization contributes to declines in total energy expenditure and physical activity, basal metabolism in native Siberians remains elevated compared to both reference values and nonnative populations living in the same communities (Leonard et al. 2002c, 2005). We have also found that contrary to earlier work on native circumpolar populations (e.g., Rodhal 1952) these elevations cannot be attributed to high dietary protein consumption (see Snodgrass et al. 2005a; Leonard et al. 2005). Rather, they appear to reflect adaptations to severe climatic stress. As for other measures of CV health, plasma lipid levels of native Siberians remain low, whereas blood pressure is quite elevated. Variation in lipid levels and blood pressure among native Siberians is explained, in part, by increased levels of adiposity and by the conditions of socioeconomic marginalization that have emerged since the collapse of the Soviet Union. Elevated blood pressure is also associated with increased energy and fat consumption. However, none of the dietary parameters were associated with variation in lipid levels. Our recent work among the Yakut has provided some additional insights into the social and lifestyle correlates of CV risks. Sorensen et al. (2005) found that poorer dietary quality contributes to higher lipid levels in the more marginal, rural Yakut communities. For blood pressure, our data from the Yakut suggest that conditions of poverty (lower household income) are associated with higher blood pressure, even after adjusting for the effects of age and body composition (Snodgrass 2004; Snodgrass et al. 2005b). It is possible that the pervasive psychosocial stress resulting from the collapse of the Soviet system and the dismantling of farming and herding collectives has contributed to the elevations in blood pressure now observed throughout native Siberian groups. Such effects are consistent with a growing body of literature on how chronic stress represents an important axis through which lifestyle change and social marginalization influence blood pressure (Dressler 1999; Dressler and Bindon 2000; Dressler et al. 2005; Madrigal et al., this volume). Additional work is needed to examine the influence of psychosocial stress on changing in cardiovascular health of native Siberians. In addition to body composition, dietary and lifestyle factors, elevations in BMR also have important implications for cardiovascular health in acculturating Siberian groups. After controlling for the effects of body composition and selected lifestyle

44

•

William R. Leonard, J. Josh Snodgrass, Mark V. Sorensen

factors, we find that BMR is positively associated with blood pressure and negatively associated with lipid levels. Thus, it appears that increased BMR may protect against elevated plasma lipid levels while increasing the risks of hypertension. A number of recent population studies have identified potential pathways linking metabolic rates and cardiovascular risk factors. In particular, alterations in the activity of the sympathetic nervous system (SNS) are now thought to play an important role in the regulation of cardiovascular homeostasis and energy balance. The SNS helps to regulate arterial pressure by altering the degree of vasoconstriction in the blood vessels (Malpas et al. 2001). Increased SNS activity in humans has been shown to be associated with elevations in blood pressure (Huggert et al. 2004; Masuo et al. 1997; Saad et al, 1991; Ward et al. 1996). In addition, it also appears that elevated sympathetic activity contributes to increased energy expenditure (both BMR and TEE) in humans (Spraul et al. 1993; Tataranni et al. 1998). Luke and colleagues (2004) have argued that increased SNS activity is a likely mechanism for explaining the associations they have found between blood pressure and BMR reported in both US and Nigerian populations. Our findings are consistent with this interpretation and suggest that heightened SNS activity may be responsible for contributing to both the high BMRs and elevated blood pressures in native Siberian groups. In this context, the recent findings on cardiovascular health of the Pima Indians of Central Arizona provide an intriguing counterpoint to our Siberian example. As summarized in Table 1.7, research on Pima has shown that they have low BMRs, very high rates of obesity, very high rates of adult-onset (type 2) diabetes, and moderate to high plasma lipid levels, and yet they have relatively low rates of hypertension (Ravussin 1995; Spraul et al. 1993; Weyer et al. 2000). This clustering of metabolic and cardiovascular risk factors is, in many respects, the mirror image of what we see among indigenous Siberians and appears to be linked, in part, to reduced SNS activity. Indeed, the low SNS activity of the Pima is thought to be an important conTable 1.7. Comparison of metabolic and cardiovascular health parameters for indigenous Siberians and Pima Indians.

Parameter

Siberians

Pima

Metabolic Rate Obesity levels Blood pressure Lipids NIDDM SNS activity

Elevated Moderate Higher Low Low? ???

Depressed High Lower Moderate Very High Low

Pima data derived from: Ravussin (1995); Spraul et al. (1993); Tataranni et al. (1998) Weyer et al. (2000).

Health Consequences of Social and Ecological Adversity

•

45

tributor to their sluggish basal metabolism and their high rates of obesity (Ravussin 1995). Reduced SNS activity is also thought to contribute to the low prevalence of hypertension in the Pima (Tataranni et al. 1998; Weyer et al. 2000). The differences between the Siberians and the Pima underscore the point that lifestyle modernization can produce markedly different health outcomes. These differences appear to stem partly from the unique historical changes in lifestyle of the two groups and partly from underlying differences in metabolism that may have been shaped by distinct selective forces in the evolutionary pasts of both populations. Further work is necessary to understand how the different dimensions of lifestyle change are contributing to variation in cardiovascular health. In addition, more research is needed to determine whether there is a genetic basis for these differences in metabolic rate and, if so, what the central pathways might be for regulating metabolism and cardiovascular risk factors. We are currently beginning to explore the nature of the metabolic adaptations in native Siberians by examining how BMR varies in association with major genetic markers in the mitochondrial genome that have been identified by Wallace and colleagues (Mishmar et al. 2003; Ruiz-Pesini et al. 2004; Wallace 2005). If the model by Wallace and colleagues is correct, we should find that individuals with the key mutations for greater heat production should have increased BMRs over those who lack the mutations. Conclusions The biology and health of indigenous Siberian populations have been strongly shaped by both environmental and social stressors. Through their evolutionary history, native Siberians have adapted to their cold, marginal climate, in part, through increased metabolic heat production. They have BMRs that are 15 to 20 percent higher than expected for their size. Until recently, these elevated metabolic rates, coupled with high levels of physical activity, protected native Siberians against obesity and associated cardiovascular risk factors. Recent social, economic, and political changes in Russia have had dramatic, yet variable influences on the health of indigenous Siberians. The collapse of the Soviet Union in the early 1990s had divergent effects on Siberian children and adults. While children’s nutritional status sharply declined during this period, overweight and obesity became growing problems for adults. The increasing levels of obesity among Siberian adults appear to be linked to reductions in energy expenditure and activity levels associated with the transition from more “traditional” subsistence lifeways to a more sedentary “modern” lifestyle. In addition, the rates of overweight and obesity are significantly higher in Siberian women than men. This gender disparity is partly attributable to lower levels of activity and energy expenditure in women. Although obesity and increased body fatness are growing problems, adult Siberians continue to have low plasma lipid levels. In contrast, the major cardiovascular risk for Siberian adults is elevated blood pressure, particularly among men. This “cluster-

46

•

William R. Leonard, J. Josh Snodgrass, Mark V. Sorensen

ing” of cardiovascular risks is markedly different from that seen in other modernizing groups, such as the Pima Indians. This patterning of risks appears to be linked to the joint influences of body composition, lifestyle change, and elevated metabolism. Conditions of marginalization and impoverishment have significant influences on lipid and blood pressure levels. Additionally, increased BMR is associated with reduced plasma lipid levels and increased blood pressure even after adjusting for the influence of body composition. Further work is needed to better elucidate the mechanisms through which the social and metabolic parameters are shaping cardiovascular health. Nonetheless, these results underscore the importance of linking biocultural and evolutionary perspectives in understanding the health consequences of lifestyle change.

• • • Unresolved Issues/Future Research Directions 1. Additional research is needed to explore the mechanisms through which lifestyle change and marginalization in post-Soviet Russia may be influencing cardiovascular health outcomes. Future work should explicitly examine the role of psychosocial stress in promoting hypertension among indigenous Siberians. 2. Although it is clear that native Siberians display systematically elevated BMRs, the nature of those elevations has not yet been determined. Further work is necessary to identify the genetic and/or developmental factors responsible promoting metabolic adaptation in these groups. It is likely that different populations exploit different genetic and physiological pathways to increase metabolic rates. 3. The influence of metabolic rates on cardiovascular risk factors needs to be explored in a broader range of human populations. Additional research is also needed to elucidate the mechanisms responsible for linking variation in metabolic turnover with variation in blood pressure and lipid levels. 4. Follow-up research is needed to determine the trajectory of health changes in native Siberians in the face ongoing political and economic changes in Russia. It is currently unclear whether the increased levels of childhood undernutrition and growth stunting of the mid 1990s continue to persist. In addition, the long-term health consequences of the initial post-Soviet transition need to be examined in the cohort of children studied during the early-mid 1990s.

• • •

Health Consequences of Social and Ecological Adversity

•

• • • Summary Points 1. The initial post-Soviet transition had divergent effects on Siberian children and adults. While children’s nutritional status sharply declined during this period, overweight and obesity persisted as growing problems among adults. 2. Among adults, increasing levels of obesity appear to be linked to reductions in energy expenditure and activity levels associated with the transition from more “traditional” subsistence lifeways to a more sedentary “modern” lifestyle. The higher rates of overweight and obesity in women are attributable to their lower levels of activity and energy expenditure. 3. Basal metabolic rates of native Siberians remain elevated in comparison to reference values. This increased metabolic turnover appears to reflect an adaptation to their cold, marginal environment. 4. Despite increased levels of adiposity, plasma lipid levels of native Siberians remain low, whereas blood pressure is quite elevated. Variation in lipid levels and blood pressure reflects the interaction of body composition, lifestyle change, and metabolism. Conditions of marginalization and impoverishment in post-Soviet Russia contribute to elevations in both lipid levels and blood pressure. Additionally, after controlling for the effects of body composition and selected lifestyle factors, we find that BMR is positively associated with blood pressure and negatively associated with lipid levels. 5. The clustering of metabolic and cardiovascular risk factors in native Siberians is, in many respects, the mirror image of what we see in modernizing groups such as the Pima Indians. The differences between the Pima and indigenous Siberians appear to be the product of the unique historical changes in lifestyle and socioeconomic status of the two groups and underlying differences in metabolism shaped by adaptation to different environmental stressors. 6. By linking biocultural and adaptive perspectives, we are able to gain fresh insights into the origin and nature of variation in human health. The integrative perspective of biological anthropology is particularly powerful for addressing many of today’s growing global health problems.

• • •

47

48

•

William R. Leonard, J. Josh Snodgrass, Mark V. Sorensen

References Bobak, M., H. Pikhard, C. Hertzman, R. Rose, and M. Marmot. 1998. Socioeconomic factors, perceived control and self-reported health in Russia: a cross-sectional survey. Social Science & Medicine 47: 269–79. Caballero, B. 2001. Symposium: Obesity in developing countries: biological and ecological factors. Journal of Nutrition 133: 866S–99S. Cameron, N., N.G. Norgan, and G.T.H. Ellison, eds. 2005. Childhood Obesity: Contemporary Issues. Boca Raton, FL: CRC Press. Carlson, P. 2000. An Unhealthy Decade: A Sociological Study of the State of Public Health in Russia, 1990–1999. Stockholm: Almquist & Wiksell. de Onis, M., E.A. Frongillo, and M. Blossner. 2000. Is malnutrition declining? An analysis of changes in levels of child malnutrition since 1980. Bulletin of the World Health Organization 78: 1222–33. Dressler, W.W. 1999. Modernization, stress and blood pressure: new directions in research. Human Biology 71: 583–605. Dressler, W.W. and J.R. Bindon. 2000. The health consequences of cultural consonance: cultural dimensions of lifestyle, social support, and arterial blood pressure in an African American community. American Anthropologist 102: 244–60. Dressler, W.W., K.S. Oths, and C.C. Gravlee. 2005. Race and ethnicity in public health research: models to explain health disparities. Annual Review of Anthropology 34: 231–52. Drizd, T., A.L. Dannenberg, and A. Engel. 1986. Blood pressure levels in persons 18–74 years of age in 1976–80, and trends in blood pressure from 1960 to 1980 in the United States. Vital and Health Statistics. Series 11, no. 234. Washington, D.C.: U.S. Public Health Service. Ellison, G.H.T. and M. Kelly. 2005. Growth of Jersey schoolchildren during the 1940–1945 German occupation: comparison with schoolchildren on mainland Britain. Human Biology 77: 761–72. Field, M. 1995. Health crisis in the former Soviet Union: a report from the ‘post-war’ zone. Social Science & Medicine 41: 1469–78. Fondahl, G. 1997. Siberia: assimilation and its discontents. In New States, New Politics: Building the Post-Soviet Nations, ed. I. Bremmer, R. Taras. Cambridge: Cambridge University Press, 190–232. ———. 1998. Gaining Ground: Evenkis, Land and Reform in Southeaster Siberia. Needham Heights, MA: Allyn and Bacon. FAO/WHO/UNU (Food and Agriculture Organization/World Health Organization/United Nations University). 1985. Energy and Protein Requirements: Report of a Joint FAO/WHO/ UNU Expert Consultation. WHO Technical Report Series no. 724. Geneva: WHO. Forsyth, J. 1992. A History of the Peoples of Siberia: Russia’s Asian Colony 1581–1990. Cambridge: Cambridge University Press. Godin, G. and R.J. Shephard. 1973. Activity patterns of the Canadian Eskimo. In Polar Human Biology, eds. O.G. Edholm and E.K.E. Gunderson. Chichester: Heinemann, 193–215. Goebel, T. 1999. Pleistocene human colonization of Siberia and peopling of the Americas: an ecological approach. Evolutionary Anthropology 8: 208–27. Hannigan, J. 1991. Statistics on the Economic and Cultural Development of the Northern Aboriginal People of the USSR (from the period of 1980–1989). Ottawa: Bureau of Indian and Northern Affairs.

Health Consequences of Social and Ecological Adversity

•

49

Heinbecker, P. 1931. Further studies on the metabolism of Eskimos. Journal of Biological Chemistry 93: 327–36. Huggert, R.J., J. Burns, A.F. Mackintosh, and D.A.S.F. Mary. 2004. Sympathetic neural activation in nondiabetic metabolic syndrome and its further augmentation by hypertension. Hypertension 44: 847–52. Humphrey, C. and D. Sneath. 1999. The End of Nomadism? Society, State and the Environment in Inner Asia. Durham, NC: Duke University Press. James, W.P.T. and E.C. Schofield. 1990. Human Energy Requirements: A Manual for Planners and Nutritionists. Oxford: Oxford University Press. Leon, D.A., L. Chenet, V.M. Shkolnikov, S. Zakharov, and J. Shapiro. 1997. Huge variation in Russian mortality rates 1984–94: artefact, alcohol, or what? Lancet 350: 383–88. Leonard, W.R., M.H. Crawford, A.G. Comuzzie, and R.I. Sukernik. 1994. Correlates of low serum lipid levels among the Evenki herders of Siberia. American Journal of Human Biology 6: 329–38. Leonard, W.R., V.A. Galloway, E. Ivakine, L. Osipova, and M. Kazakovtseva. 2002a. Ecology, health and lifestyle change among the Evenki herders of Siberia. In Human Biology of Pastoral Populations, eds. W.R. Leonard and M.H. Crawford. Cambridge: Cambridge University Press, 209–35. Leonard, W.R., P.T. Katzmarzyk, and M.H. Crawford. 1996. Energetics and population ecology of Siberian herders. American Journal of Human Biology 8: 275–89. Leonard W.R., A. Keenleyside, and E. Ivakine. 1997. Recent fertility and mortality trends among aboriginal and nonaboriginal populations of central Siberia. Human Biology 69: 403–17. Leonard, W.R., J.J. Snodgrass, and M.V. Sorensen. 2005. Metabolic adaptation in indigenous Siberian populations. Annual Review of Anthropology 34: 451–71. Leonard, W.R., M.V. Sorensen, V.A. Galloway, G.J. Spencer GJ, et al. 2002c. Climatic influences on basal metabolic rates among circumpolar populations. American Journal of Human Biology 14: 609–20. Leonard, W.R., G.J. Spencer, V.A. Galloway, and L. Osipova. 2002b. Declining growth status of indigenous Siberian children in Post-Soviet Russia. Human Biology 74: 197–209. Luke, A., A. Adebowale, H. Kramer, T. Forrester, and R.S. Cooper. 2004. Association between blood pressure and resting energy expenditure independent of body size. Hypertension 43: 555–60. Levine, M., L. Duffy, D.C. Moore, and L.A. Matej.1995. Acclimation of a non-indigenous sub-Arctic population: seasonal variation in thyroid function in interior Alaska. Comparative Biochemistry and Physiology 111A: 209–214. Madrigal, L., E. Ruiz, M. Blell, and F. Otarola. 2006. The slavery hypothesis: an evaluation of the genetic-deterministic explanation for hypertension prevalence rate (HPR) inequalities. This volume. Malpas, S.C., B.L. Leonard, S-J. Guild S-J, J.V. Ringwood, et al. 2001. The sympathetic nervous system’s role in regulating blood pressure variability. IEEE Engineering in Medicine and Biology Magazine 20: 17–24. Masuo, K., H. Mikami, T. Ogihara, and M.L. Tuck. 1997. Sympathetic nerve hyperactivity precedes hyperinsulinemia and blood pressure elevation in a young, nonobese Japanese population. American Journal of Hypertension 10: 77–83. Milan, F.A. and E. Evonuk. 1967. Oxygen consumption and body temperatures of Eskimos during sleep. Journal of Applied Physiology 22: 565–67.

50

•

William R. Leonard, J. Josh Snodgrass, Mark V. Sorensen

Mishmar, D., E. Ruiz-Pesini, P. Golik, V. Macaulay, A.G. Clarke, et al. 2003. Natural selection shaped regional mtDNA variation in humans. Proceedings of the National Academy of Sciences 100: 171–76. McKee, M. and V. Shkolnikov. 2001. Understanding the toll of premature death among men in eastern Europe. British Medical Journal 323: 1051–55. Mosher, M.J. 2002. The genetic architecture of plasma lipids in the Buyat: an ecogenetic approach. Ph.D. thesis, University of Kansas. Mote, V.L. 1998. Siberia: Worlds Apart. Boulder, CO: Westview Press. NIH (National Institutes of Health). 2002. Third Report of the National Cholesterol Education Program (NCEP) Expert Panel on Detection, Evaluation and Treatment of High Cholesterol in Adults (adult treatment panel III): final report. Bethesda, MD: NIH. ———. 2005. The Seventh Report of the Joint National Committee on Prevention, Detection, Evaluation, and Treatment of High Blood Pressure. Bethesda, MD: NIH. Näyhä, S., P. Luoma, S. Lehtinen, T. Lehtimäki, et al. 2002. Disease patterns in Sámi and Finnish populations: An update. In Human Biology of Pastoral Populations, eds. W.R. Leonard and M.H. Crawford. Cambridge: Cambridge University Press, 236–50. Notzon, F.C., Y.M. Komarov, S.P. Ermakov, C.T. Sempos, et al. 1998. Causes of declining life expectancy in Russia. Journal of the American Medical Association 279: 793–800. Pitulko, V.V., P.A. Nikolsky, E.Y. Girya, A.E. Basilyan, et al. 2004. The Yana RHS site: humans in the Arctic before the last glacial maximum. Science 303: 52–56. Poehlman, E.T. and M.J. Toth. 1995. Mathematical ratios lead to spurious conclusions regarding age- and sex-related differences in resting metabolic rate. American Journal of Clinical Nutrition 61: 482–85. Popkin, B.M. 2001. The nutrition transition and obesity in the developing world. Journal of Nutrition 131: 871S–72S. ———. 2002. An overview on the nutrition transition and its health implications: the Bellagio meeting. Public Health Nutrition 5 (1A): 93–103. Ravussin, E. 1995. Low resting metabolic rate as a risk factor for weight gain: role of the sympathetic nervous system. International Journal of Obesity 19 (Suppl 7): S8–9. Rodahl, L.K. 1952. Basal metabolism of the Eskimos. Federation Proceedings 2: 130–37. Rode, A. and R.J. Shephard. 1971. Cardiorespirtatory fitness in an arctic community. Journal of Applied Physiology 31: 519–26. ———. 1984. Ten years of ‘civilization’: fitness of the Canadian Inuit. Journal of Applied Physiology 56: 1472–77. ———. 1994. Physiological consequences of acculturation: a 20-year study of fitness in an Inuit community. European Journal of Applied Physiology 69: 516–24. ———. 1995. Basal metabolic rate of Inuit. Am J Hum Biol 7: 723–29. Ruiz-Pesini, E., D. Mishmar, M. Brandon, V. Procaccio, and D.C. Wallace. 2004. Effects of purifying and adaptive selection on regional variation in human mtDNA. Science 303: 223–26. Saad, M.F., W.C. Knowler, D.J. Pettitt, R.G. Nelson, et al. 1990. Insulin and hypertension: Relationship between obesity and glucose intolerance in Pima Indians. Diabetes 39: 1430–35. Shephard, R.J. and A. Rode. 1996. Health Consequences of ‘Modernization’: Evidence from Circumpolar Peoples. Cambridge: Cambridge University Press. Slezkine, Y. 1994. Arctic Mirrors: Russia and the Small Peoples of the North. Ithaca: Cornell University Press Smals, A.G.H., H.A. Ross, and P.W.C. Kloppenborg. 1977. Seasonal variation in serum T3 and T4 levels in man. Journal Clinical Endocrinology and Metabolism 44: 998–1001.

Health Consequences of Social and Ecological Adversity

•

51

Snodgrass, J.J. 2004. Energetics, Health, and Economic Modernization in the Yakut (Sakha) of Siberia: A Biocultural Perspective on Lifestyle Change in a Circumpolar Population. Ph.D. thesis, Northwestern University. Snodgrass, J.J., W.R. Leonard, M.V. Sorensen, L.A. Tarskaia, et al. 2006a. The emergence of obesity among indigenous Siberians. American Journal of Physical Anthropology 25: 75–84. Snodgrass, J.J., W.R. Leonard, L.A. Tarskaia, V.P. Alekseev, and V.G. Krivoshapkin. 2005a. Basal metabolic rate in the Yakut (Sakha) of Siberia. American Journal of Human Biology 17: 155–72. Snodgrass, J.J., W.R. Leonard, L.A. Tarskaia, and D.A. Schoeller DA. 2006b. Total energy expenditure in the Yakut (Sakha) of Siberia as measured by the doubly labeled water method. American Journal of Clinical Nutrition 84: 798–806. Snodgrass, J.J., W.R. Leonard, L.A. Tarskaia, M.V. Sorensen, et al. 2005b. Health and economic modernization in the Yakut (Sakha) of Siberia. American Journal of Human Biology 17: 256 (abstract). Sorensen, M.V. 2003. Social and biological determinants of cardiovascular risk among rural and urban Yakut: impact of socioeconomic upheaval. Ph.D. thesis, Northwestern University. Sorensen, M.V., W.R. Leonard, J.J. Snodgrass, L.A. Tarskaya, et al. 2005. Health consequences of failed modernization: dietary and lifestyle determinant of cardiovascular risk factors in Yakutia. American Journal of Human Biology 17: 576–92. Spraul, M., E. Ravussin, A.M. Fontvielle, R. Rising, et al. 1993. Reduced sympathetic nervous activity: a potential mechanism predisposing to body weight gain. Journal of Clinical Investigation 92: 1730–35. Stinson, S. 1985. Sex differences in environmental sensitivity during growth and development. Yearbook of Physical Anthropology 28: 123–47. Stinson, S., B. Bogin, R. Huss-Ashmore, and D. O’Rourke, eds. 2000. Human Biology: An Evolutionary and Biocultural Perspective. New York: Liss. Tataranni, P.A., L. Christin, S. Snitker, G. Paolisso, and E. Ravussin. 1998. Pima Indian males have lower B-adrenergic sensitivity than Caucasian males. Journal of Clinical Endocrinology and Metabolism 83: 1260–163. Tobias, P.V. 1985. The negative secular trend. Journal of Human Evolution 14: 347–56. ———. 1986. Physical stature in disadvantaged communities—Johanessburg blacks have not grown taller this century. South African Journal of Science 82: 585–88. Tokarev, S.A. and I.S. Gurvich. 1964. The Yakut. In The Peoples of Siberia, ed. M.G. Levin and L.P. Potapov. Chicago: University of Chicago Press, 243–304. Wallace, D.C. 2005. A mitochondrial paradigm of metabolic and degenerative diseases, aging, and cancer: a dawn for evolutionary medicine. Annual Review of Genetics 39: 359–407. Ward, K.D., D. Sparrow, L. Landsberg, J.B. Young, et al. 1996. Influence of insulin, sympathetic nervous activity, and obesity on blood pressure: the Normative Aging Study. Journal of Hypertension 14: 301–8. Weyer, C., R.E. Pratley, S. Snitker, M. Spraul, et al. 2000. Ethnic differences in insulinemia and sympathetic tone as links between obesity and blood pressure. Hypertension 36: 531–37. Young, T.K. 1996. Obesity, central fat patterning, and their metabolic correlates among the Inuit of the Central Canadian Arctic. Human Biology 68: 245–63. Young, T.K., Y.P. Nikitin, E.V. Shubnikov, T.I. Astakhova, et al. 1995. Plasma lipids in two indigenous arctic populations with low risk for cardiovascular diseases. American Journal of Human Biology 7: 223–36.

• 2 • A Multidisciplinary Approach to Understanding the Risk and Context of Emerging Primate-Borne Zoonoses Lisa Jones-Engel and Gregory Engel

Introduction Infectious diseases have exerted a powerful influence on the evolution and development of human societies. Over the millennia, people have struggled to understand the ravages of pandemics, often in religious terms, attributing mass illness and death to wrathful deities or dissatisfied spirits (see Herring, Chapter 3). In the past two centuries, scientists have been able to show (most of the) immediate culprits to be microorganisms; yet too often we find ourselves in the position of reacting to epidemics, while insufficient resources have been used to proactively seek out new, heretofore unrecognized threats. Animal reservoirs in particular have been an important source for epidemics of human disease. But, of the vast number of microorganisms infecting the planet’s fauna, only a relatively few are capable of causing human disease, and a small minority of those have the potential to cause epidemics affecting human populations. What strategy should be taken, then, to protect human populations from newly emerging infectious diseases? The emergence of zoonotic diseases (diseases originating in animals that can be transmitted to humans) as significant human pathogens depends on characteristics of human societies (the organization of food production, housing, and means of

The Risk and Context of Emerging Primate-Borne Zoonoses

•

53

production), characteristics of nonhuman populations (including animal reservoirs and vector species), characteristics of the infectious agents that parasitize them, and characteristics of the environments in which they interact. Because these four factors are dynamic in nature, their complex interrelationships are also constantly changing, opening up new possibilities for infectious disease transmission. Over the past decade, our group’s research on bidirectional pathogen transmission between humans and nonhuman primates has characterized both the varied contexts of human-primate contact in Asia as well as the infectious agents that are transmitted. This research has provided an empirical foundation for viewing interspecies pathogen transmission between humans and primates, in particular, and, by extension, for understanding zoonotic transmission in general. Asia is a compelling setting for this research owing to the region’s faunal diversity, which includes a plethora of primate and non-primate species. These habitats also contain a diverse microbial fauna, including infectious agents previously unencountered by human populations. What’s more, Asia’s dynamic primate populations and the infectious agents they contain are situated amid the world’s densest human populations (including a growing number of people immunocompromised by HIV and tuberculosis) in a region bustling with regional and international commerce, and a popular destination for tourists from around the globe. Epidemics of SARS and influenza underscore the reality of the global spread of pathogens from Asia. Taken together, these factors can potentially facilitate the rapid dispersal of infectious agents from Asian primates to human populations the world over—a virtual “pathogen internet.” Human social organization provides a backdrop for analyzing the multiple contexts of human-nonhuman contact that, in turn, have the potential to lead to interspecies pathogen transmission. In Asia, interspecies contact between humans and primates occurs in a variety of contexts: animal markets, pet ownership, monkey forests, bushmeat hunting and consumption, performing primates, zoos, and the aids (e.g., the use of macaques to harvest coconuts) and obstacles (crop-raiding primates) to agricultural production. Diseases can also be transmitted across the species barrier via vectors such as insects, even when humans and primates do not come into close physical proximity. In this chapter we will use data from the contexts of humanprimate contact in Asia that we have identified to establish a framework that describes how infectious agents can be transmitted between humans and primates. Historically, it is likely that bushmeat hunting or scavenging provided the first context for human exposure to primate zoonoses. Bushmeat hunting is still practiced in parts of Asia today; however, it is but one of a number of contexts of humanprimate contact. Widespread domestication of primates is virtually nonexistent on any continent; however, in parts of Asia it is common to use pig-tailed macaques to harvest coconuts. Within this limited context of domestication there are ample opportunities for human exposure to infectious agents. The origins and extent of primate pet ownership is unclear, but today in many parts of rural Asia pet ownership can be seen as a byproduct, to a large extent, of bushmeat hunting, as orphaned infants

54

•

Lisa Jones-Engel and Gregory Engel

are taken in as pets. The long-term and often intimate physical association between pets and owners provides an ideal context for the transmission of infectious agents. Primates occupy important roles in Asian cultures and religions, and in many locations temples have become associated with populations of primates, most often macaques and langurs, that are provisioned and protected by the human communities. Depending on the local customs, human-primate contact can be a common phenomenon at “monkey temples” (religious sites that have become associated with populations of monkeys, and where monkeys are usually protected), some of which have evolved, over time, into international tourist destinations. Perhaps a modern variation on monkey temples are the parks and nature preserves where visitors can watch, and often interact with primates. Cross-species transmission of infectious agents within the context of monkey temples is ongoing. Although performing primates are encountered throughout the world, Asian cultures have perhaps the longest and most vibrant tradition of using primates for entertainment. It is important to consider the origin of performance primates when evaluating the risk of primate-to-human transmission of infectious agents. Animal markets are often the source for performing primates (and also an alternative source for pet primates). Markets bring together a variety of species, primate and non-primate, from diverse geographic origins, each with its potentially unique burden of infectious agents. Animals in these markets are typically maintained at high density and often in poor conditions, which may compromise immunity and facilitate disease transmission. Dissecting individual contexts has provided some perspective on the variables that govern infectious disease transmission between humans and primates. Transmission of infectious agents can be seen as a stochastic process governed by multiple variables. A collection of methods known as risk analysis has emerged to support rational decision making in the face of events of uncertain probability. Applied to the problem of interspecies pathogen transmission, risk analysis can be used to evaluate the infectious risks associated with human-primate contact in order to prioritize cost-effective management strategies. In the risk analysis framework, mathematical models are powerful tools that help to predict the emergence of zoonotic diseases into human populations and anthroponotic diseases into primate populations. Of course, models of cross-species disease transmission are ultimately only as accurate as the data used to inform them. This points to the crucial importance of a multidisciplinary approach to this type of research. In this chapter, we discuss how the integration of data on micro- and macro-level variables has allowed us to improve our ability to predict patterns of disease transmission in Asia. This paradigm for analyzing crossspecies transmission holds the promise of providing a firmer scientific basis for making policy decisions that have implications for both human and primate populations in the years to come. This chapter addresses the issue of emerging zoonotic diseases in general and focus on the specific example of diseases transmitted to humans from primates. We lay a theoretical basis for regarding zoonotic disease transmission and discuss some

The Risk and Context of Emerging Primate-Borne Zoonoses

•

55

of the multitude of risk factors that influence the transmission of infectious agents (see The Risk of Pathogen Transmission box below). This theoretical construct will be applied to specific infectious agents enzootic to primates within several contexts of human-primate contact. A principle focus will be on the human-primate interface and the ways in which the contexts of human-primate contact influences the risk of interspecies disease transmission. This focus stems, in part, from a frustration or concern with much of the previous research on primate zoonoses, specifically HIV/SIV. Since the mid-1990s, when it first became evident through laboratory-based research that HIV/SIV was initially transmitted to humans from primates, there has been an assumption in the scientific literature that bushmeat hunting accounted for the transmission events from primates to humans (Hahn et al. 2000; Peeters et al. 2002). However, there have been no integrated epidemiological, epizootiological, and/or behavioral studies conducted to support or disprove this notion. There are no data that conclusively prove that bushmeat hunting is the mechanism by which humans initially became exposed to HIV/SIV. While bushmeat hunting does present significant risks for exposure to primate body fluids, there are multiple other contexts in which humans and primates interact that could account for the initial transmission events.

• • • The Risk of Pathogen Transmission Why Study Cross-Species Disease Transmission? Risk Perception —Predict infectious threats to primate populations and learn how best to prevent human-to-primate disease transmission. —Predict when, where, how, and among whom primate-borne zoonoses will occur in order to prevent primate-to-human transmission. Conceptualizing Bidirectional Pathogen Transmission Risk Factors What are the likely contexts of transmission? How can transmission occur? What kinds of infectious agents can be transmitted?

• • •

56

•

Lisa Jones-Engel and Gregory Engel

Finally, citing findings from research performed over the past fifteen years, we will advocate a multidisciplinary approach to the prevention of emerging zoonoses that uses a paradigm known as Risk Analysis to inform interventions designed to protect public health (see box on page 58). Why Primate-Borne Zoonoses? The ongoing pandemic of human immunodeficiency virus (HIV) has cast a spotlight on the role of primates as potential reservoirs for emerging zoonotic diseases. Theoretical justification for studying interspecies transmission from primates to humans derives from the immunological, physiological, genetic, and behavioral similarities between humans and our closest phylogenetic relatives. It is these very similarities that have led to the widespread use of primates as models for the study of human diseases and which underlies a general rule (though one with important exceptions) that phylogeny predicts susceptibility to infection. Consequently, it is generally held that infectious agents endemic in primate species are more likely to be able to infect humans than infectious agents from non-primate species (Acha and Szyfres 1980). The theoretical possibility of primate-to-human and human-to-primate disease transmission is underscored by a long list of pathogens shown to cause disease in both humans and primates. Brack (1987) and Fiennes (1967) catalogue some 49 viruses, 4 mycoplasmas, 49 bacteria, 25 fungi, 16 protozoa, 27 nematodes, 13 trematodes, 11 cestodes and two mites known to cause infection in both humans and primates, and the list continues to grow. The risk that these pathogens will be transmitted between humans and primates is further enhanced when the populations occupy overlapping geographic areas. Macaques in Asia Macaques are the most widely distributed taxa of primate in the world. In southern Europe and northern Africa pockets of the genus’s most ancestral species Macaca sylvanus can be found at sea level, in urban areas and ranging through the Atlas Mountains. The remaining nineteen species are distributed throughout Central, South, and Southeast Asia, with M. fuscata on the islands of Japan marking the eastern boundary of the genus’s natural distribution. Several macaque species have lived commensally with humans; their ecological and behavioral plasticity enabled them to thrive for centuries in the dense population centers of South and Southeast Asia. Owing to their numbers and their hardiness, macaques have also been extensively used in laboratory research, and are therefore among the most studied primates. Macaque-human contact is a frequent phenomenon in a variety of contexts throughout their distribution. These qualities all contribute to making macaques an ideal subject for the study of primate-to-human disease transmission. Their numbers and their frequency of contact with humans make it possible to use statistical and epidemiological methods to analyze their behavior, their biology, and the prevalence of enzootic infectious

The Risk and Context of Emerging Primate-Borne Zoonoses

•

57

agents found in different populations and demographic subpopulations of macaques. Thus among primates, macaques are the most logical choice for studying cross-species disease transmission. Perhaps as a result of the emergence of HIV/SIV, which has its origins in Africa, the issue of cross-species transmission in Asia and South America has been largely ignored. However, several factors make South and Southeast Asia a logical setting to conduct research on pathogen transmission between primates and humans. Asia’s habitats contain a rich primate fauna. These habitats also contain a rich non-primate fauna; several areas have been designated by Conservation Intentional as biodiversity “hot spots.” It is likely that these habitats also contain a diverse microbial fauna, including infectious agents previously unencountered by human populations. At the same time, deforestation of pristine habitats and encroachment of human populations is bringing ever-increasing contact between wild primates and humans. These wild primates and their offspring come into contact with humans in a variety of contexts: pet markets, pet ownership, monkey forests, bushmeat hunting and consumption, circuses, zoos, and agriculture production. It is important to realize that these primate populations are not isolated from one another; they are dynamic, overlapping, and can intermix. Animals taken from the wild become pets. Pets are released or escape into the wild (Jones-Engel et al. 2005a). People take animals with them when they move from rural to urban areas or to another island or another country. Animal markets bring multiple species together in close quarters, increasing the likelihood of pathogen transmission among primate species and creating the possibility of new host/pathogen combinations (Karesh et al. 2005). As a result, primates previously isolated from human contact may provide a conduit by which infectious agents previously unencountered by humans can enter the human population. Modeling Cross-Species Transmission of Infectious Agents If we hope to learn about the conditions under which future primate-borne zoonoses are likely to emerge and use this information to take action to prevent their spread in human populations, it is critical to learn more about the what, when, where, whom, and how of pathogen transmission. This is a complex endeavor given the variety of species, pathogens, and contexts for interspecific contact between humans and primates. Researching the phenomenon of cross-species pathogen transmission calls for a different approach than that used to investigate the transmission of specific diseases. It requires the acquisition and integration of data from different disciplines. Serologic and molecular data are needed to determine the prevalence of pathogens in both human and primate populations. Data on primate behavior and primatehuman interactions are needed to describe the contexts in which primate-human contact occurs. Epidemiologic and epizootiologic analysis is required to integrate the two, providing a population perspective on cross-specific transmission and the basis for modeling the conditions under which cross-specific disease transmission is likely to occur. In taking steps to reduce the risk of cross-species transmission between

58

•

Lisa Jones-Engel and Gregory Engel

humans and primates, it is perhaps advisable first to understand how transmission occurs—that is, the situations, conditions and behaviors that lead to contact and transmission. Risk Analysis: A Method for Evaluating Risk One approach to this challenge is to develop risk analysis models that break the “process” of disease transmission into component parts (Travis et al. 2006). Processes and interactions that could lead to cross-species disease transmission are explicitly described as a hypothetical infection chain. Data from laboratory experiments and field experiments are used to estimate the probability of each component, expected natural variation, and margins of error. When data are unavailable, expert opinion provides a guideline for probability estimates. Risk analysis can be used both to estimate the risk of transmission as well as the likely efficacy of various risk-reduction strategies (Hueston and Walker 1993; Metcalfe et al. 1996). This information can improve the efficacy and cost-effectiveness of public health interventions.

• • • Components of Risk Analysis Predicting Emerging Infectious Diseases What is the likelihood in the future that disease transmission will occur? Between which populations? Between which species? What infectious agents? What is the likely morbidity/mortality? Early Warning System Risk Analysis Hazard identification: What is out there? Risk Assessment: How likely is it to occur? Risk Management: What do we do now? Risk Communication: Whom do we tell?

• • •

The Risk and Context of Emerging Primate-Borne Zoonoses

•

59

Risk analysis is a multidisciplinary process used to evaluate existing knowledge in order to prioritize risks associated with the spread of disease. A collection of methods known as risk analysis has emerged to support rational decision making in the face of events of uncertain probability (Kellar 1993; Leighton 2002). Applied to the problem of interspecies pathogen transmission, risk analysis can be used to evaluate the infectious risks associated with human-primate contact in order to prioritize costeffective management strategies (MGVP 2004; Travis et al. 2006; Jones-Engel et al. 2006b; Engel et al. 2006). Risk analysis is a blend of hazard identification, risk assessment, risk management, and risk communication (see Components of Risk Analysis box above). Risk in this case, is the measure of the probability of the introduction or escape of pathogens or other hazards into or out of the animals/people of concern. The hazard identification process seeks to establish which hazards are of concern and how they may be introduced. Risk assessment is the process of modeling the probability that an adverse event, such as introduction of an emerging pathogen into a naïve population, will occur under a given set of conditions. The goal of risk management is to reduce both the likelihood and implications of the introduction of the identified hazards into the population of concern. The involvement of all potentially affected parties in the overall process is the goal of risk communication. In the risk analysis framework, mathematical models are powerful tools that help to predict the emergence of zoonotic diseases into human populations and anthroponotic diseases into primate populations (Figure 2.1). Of course, models of cross-species disease transmission are ultimately only as accurate as the data used to inform them. This points to the crucial importance of a multidisciplinary/translational approach to this type of research. There is great potential in integrating data on micro- and macro-level variables to improve our ability to predict patterns of

Figure 2.1. Risk Analysis

60

•

Lisa Jones-Engel and Gregory Engel

disease transmission. For example, economists, epidemiologists, anthropologists, and land-use specialists bring to bear their training, perspectives, and data on the macrolevel factors that go into the development of realistic models. Collaborations with physicians, virologists, and veterinarians generate data on micro-level variables that help elucidate the factors that impact cross-species transmission. This paradigm for analyzing cross-species transmission holds the promise of providing a firmer scientific basis for making policy decisions that have implications for both human and primate populations in the years to come. Recently, conservation and management decisions concerning both captive and free-ranging nonhuman primates have been greatly impacted by emerging and re-emerging disease issues. Zoonotic pathogens such as HIV/SIV, Ebola virus, herpes B virus, tuberculosis, measles, and influenza can lead to emotion-based decision making rather than science-based policy development. Decisions made in the absence of informed research and without the input of all stakeholders can be disastrous. The incorporation of the risk analysis paradigm into research and policy development will help bring together scientists and policy makers to make better decisions for both human health and primate conservation (Decision Tree Writing Group 2006; Lonsdorf et al. 2006; Fuentes 2006). Risks of Primate-Borne Zoonoses Primate-Borne Infectious Agents Retroviruses. Several enzootic simian infectious agents have been shown to cross the

species barrier, with varying levels of risk to human populations. Of these, the best documented are the enzootic simian retroviruses. Primates can simultaneously harbor several retroviruses, including simian foamy virus (SFV), simian T-cell lymphotropic virus (STLV), simian type-D retrovirus (SRV), and simian immunodeficiency virus (SIV). Simian foamy viruses (SFV) are members of the taxonomic subfamily Spumavirinae. These retroviruses are found in many mammals, including cats, cows, rodents, and sea lions, as well as in several species of primates (Meiering and Linial 2001). To date, no endemic human foamy virus has been discovered. Research in primates suggests that although host animals mount an antibody response to SFV, the virus continues to be detected in the host throughout its life (Falcone et al. 1999; Murray et al. 2006). Although proviral DNA can be found in nearly every tissue, indicating infection, the virus replicates to a detectable level only in the oral mucosa. Replication at this site facilitates transfer to other hosts through saliva (Murray et al. 2006). In spite of the virus’s persistence in a variety of tissues, no pathogenic state or confirmed risk to human health has been associated with SFV infection (Meiering and Linial 2001; Switzer et al. 2004). This is in contrast to all other known retroviruses, which have been shown to be pathogenic in at least some hosts. Type-D simian retrovirus (SRV), a multi-serotype group of exogenous retroviruses, is known to cause epidemics of an AIDS-like syndrome in its enzootic hosts, Macaca spp. (Marx et al. 1985; Maul et al. 1986; Tsai et al. 1986). The finding of

The Risk and Context of Emerging Primate-Borne Zoonoses

•

61

SRV as the etiologic agent of simian AIDS (SAIDS) in 1985 was considered to be the first simian model for human AIDS (Daniel et al. 1985; Letvin et al. 1985). Recent reports of new SRV isolates from laboratory macaques and free-ranging Indian langurs (Grant et al. 2005a) and evidence of langur-to-macaque transmission (Grant et al. 2005b) indicate that even more serotypes may be present in a wider distribution of primate species. SRV and SFV are present in saliva and other body fluids of infected primates, suggesting that bites, scratches, and mucosal splashes with primate body fluids have the capacity to transmit infection (Lerche et al. 1986; Neumann-Haefelin et al. 1993; Falcone et al. 1999; Jones-Engel et al. 2007). Previous studies examining laboratory and zoo workers as well as bushmeat hunters in Africa and monkey temple workers in Indonesia who came into contact with primates, have clearly demonstrated that humans can be infected with SFV and SRV. (Heneine et al. 1998; Callahan et al. 1999; Lerche et al. 2001; Brooks et al. 2002; Switzer et al. 2004; Wolfe et al. 2004; Jones-Engel et al. 2005b; Jones-Engel et al. 2008). The fewer than forty documented instances of zoonotic transmission of SFV have provided no evidence for secondary transmission, suggesting that humans may be dead-end hosts for SFV (Callahan et al. 1999). The long-term risks associated with human infection with SFV are unknown. To date, no pathogenic outcomes have been documented in people who have been infected with SFV. However, it should noted that very few infected individuals have been detected and follow-up on these individuals has either been nonexistent or has been for less than ten years. Simian T-cell lymphotropic viruses (STLV-1 and STLV-2) are closely related to human T-cell lymphotrophic viruses (HTLV-1 and HTLV-2). STLV is hypothesized to be the progenitor of human T-cell lymphotrophic virus (HTLV), through multiple cross-species transmissions (Vandamme et al. 1998; Gessain and Mahieux 2000). The close phylogenetic relationship between these viruses suggests an ancient ancestry, and increasingly they are referred to as primate T-cell lymphotropic viruses (PTLV) (Slattery et al. 1999). In humans, infection with HTLV-1 and 2 are associated with the risk of T-cell leukemia, myelopathy, tropical spastic paraperesis, and other neurologic and inflammatory diseases (Yamashita et al. 1996). Increased surveillance and improved molecular diagnostic techniques have recently led to the identification of several presumed cross-species transmission events of PTLVs within the context of hunting, butchering, and keeping primates as pets in parts of Africa (Wolfe et al. 2005; Calattini et al. 2005). SIV is enzootic in several species of African primates (Peeters et al. 2002). Molecular and phylogenetic research suggests that SIVsmm enzootic in sooty mangabeys is the likely progenitor of HIV-2, and that SIVcpz, which is enzootic among chimpanzees in west central Africa, is the closest phylogentic relative of all the major variants of HIV-1 (Santiago et al. 2005; Nerrienet et al. 2005). SIVcpz appears to have emerged in chimpanzees through at least two cross-species transmission events that may have been related to chimpanzees hunting different species of monkeys

62

•

Lisa Jones-Engel and Gregory Engel

that share their home range. Molecular characterization of SIVcpz suggests that this virus is actually a recombination within the chimpanzee of two distinct strains of simian immunodeficiency virus, namely, SIV from red-capped mangabeys (SIVrcm) and from greater spot-nosed monkeys (SIVgsm) (Bailes et al. 2003). It has been theorized that, similarly, transmission of SIVcpz from chimpanzees to humans occurred through the process of hunting, in this case humans hunting chimpanzees (Hahn et al. 2000). While virological and molecular evolutionary studies have given us tremendous insight into the diversity of immunodeficiency viruses present in human and primate populations, studies designed to generate direct epidemiological evidence of transmission remain uncommon. Cercopithecine herpesvirus 1. Cercopithecine herpesvirus 1(CHV-1), also known as Herpes B virus, is a member of the taxonomic subfamily Alphaherpesviridae. In Asian macaques (Macaca spp), CHV-1 is a common, latent infection. While CHV-1 infection in primates is almost always benign, CHV-1 infection in humans causes a severe meningoencephalitis with a mortality rate approaching 70 percent (Huff and Barry 2003; Holmes et al.1995). Since first reported in the 1930s, approximately forty cases of CHV-1 have been diagnosed worldwide, all occurring in the United States, Great Britain, or Canada and exclusively among people who had contact, either direct or indirect, with laboratory macaques (Holmes et al. 1995; Hummeler et al. 1959; CDC-MMWR 1989; Holmes et al. 1990; CDC-MMWR 1998). Several routes of primate-to-human transmission have been implicated, most involving direct inoculation of tissue or fluid from an infected macaque. One case of human-to-human transmission of CHV-1 has been documented (CDC-MMWR 1987). No cases of CHV-1 infection have ever been documented in people exposed to free-ranging macaques, in spite of a long history of human/macaque commensalism in Asia. Rhesus cytomegalovirus. Rhesus cytomegalovirus (RhCMV), Cercopithecine herpesvirus 8, is a betaherpesvirus enzootic among several species of Old and New World primates (Vogel et al. 1994; Minamishima et al. 1971; Nigida et al. 1979; Ohtaki et al. 1986; Eizuru et al. 1989; Blewett et al. 2001, 2003). RhCMV is likely transmitted horizontally from mother to infant via breast milk and saliva, similar to identified modes in humans (Alford and Britt 1993). While in immunologically intact animals RhCMV infections are asymptomatic, immunosupressed animals are at increased risk of severe disease (King et al. 1983; Kaur et al. 2003). Human infection with RhCMV has yet to be reported, though RhCMV has been shown to infect in human cells in vitro (Alcendor et al. 1993). SV-40. Simian virus 40 (SV40) is a polyomavirus enzootic among rhesus macaques

(Macaca mulatta). SV40 is present in the genitourinary tract of infected macaques, and the risk of transmission is thought to be through ingestion of urine containing the virus (Shah and Morrison 1969). SV40 first became an object of public health

The Risk and Context of Emerging Primate-Borne Zoonoses

•

63

interest in the 1960s when it became apparent that millions of doses of polio vaccine, which was produced in tissue cultures of monkey kidney cells, were contaminated with SV40 (Shah and Nathanson 1976). SV40 is known to produce tumors when inoculated into newborn hamsters, and a number of investigators have reported the detection of SV40 genomic sequences in human tumors (Cicala et al. 1993). Epidemiological studies, however, have yet to establish a firm link between SV40 and human cancer (Bergsagel et al. 1992; Carbone et al. 1994; Krainer et al. 1995). Other agents. In addition to the list of enzootic infectious agents presented here, there are several other infectious agents that have been transmitted from primates to humans where the primates were merely the intermediate hosts, including the monkeypox virus, tuberculosis, and, most notably, Ebola (Arita et al. 1985; Hobson 2003; Walsh et al. 2005; Hankenson et al. 2003). Filoviruses, including Marburg and Ebola, have gained significant public attention since they were first identified in the late 1960s. Primates are not the natural reservoir for these viruses; this is evidenced by the rapid and often fatal progression of the disease witnessed in primates that are infected. Since the first documented emergence of these viruses about four decades ago, more than two thousand humans and untold numbers of primates have died (Rouquet et al. 2005). Despite significant resources and efforts from the World Health Organizations and the Centers for Disease Control and others, the ensuing epidemiological search for the natural reservoir for these viruses has yet to produce definitive results. Recent outbreaks of this deadly virus have been linked to humans coming into contact with primate carcasses (Le Guenno et al. 1995; Amblard et al. 1997; Georges et al. 1999; Leroy et al. 2004).

Patterns The extent to which non-vector-borne zoonotic pathogens infect human populations depends on three basic rate-limiting steps: 1) the transmission of the pathogens from primates to humans; 2) subsequent transmission of the pathogen from humans to other humans; and 3) the geographic spread of the infectious agent. An approach to the study of emerging infectious diseases must encompass these three steps (Morse 1995; Daszak et al. 2004). Primate-to-Human Transmission: Risk at the Human-Primate Interface The first step, primate-to-human transmission, is a sine qua non of zoonotic disease for non-vector-borne infectious agents. In order for zoonotic transmission to occur, a population of nonimmune hosts must come into contact with a reservoir population containing transmissible infectious agents. With some notable exceptions, naturally occurring primate populations are clustered in tropical and subtropical countries, and as a result, most human-primate contact occurs in these “habitat countries.” However, it should be noted that, in addition to free-ranging primates in habitat

64

•

Lisa Jones-Engel and Gregory Engel

countries, primates and primate tissues can also be found in other contexts (zoos, laboratories, pets, and performing monkeys) in nearly every country in the world. Transmission of infectious agents is a stochastic phenomenon (one governed by laws of probability) that depends in part on the amount of contact between reservoirs of infectious agents and potential new hosts for those infectious agents. Here we use the term “human-primate interface” to describe both the amount and kind of contact that occurs between populations of humans and primates. The human-primate interface denotes all humans who come into contact with primates, primate tissue, or other primate-derived fluids. The human-primate interface, furthermore, can be described in terms of both temporal and spatial dimensions and is dependent on: 1) the number and characteristics of humans in contact with primates; 2) the number and characteristics of primates in contact with humans; 3) the duration of contact; 4) the type of contact (i.e., respiratory, skin contact with body fluids, mucosal contact with body fluids, ingestion; and 5) the intensity of contact—that is, the amount of potentially infectious material that comes into contact with a point of entry. Human-to-Human Pathogen Transmission: Expanding the Risk Primate-to-human pathogen transmission is a necessary but not sufficient condition for the emergence of disease that impacts a large number of people. In fact, there are several examples of infectious agents for which primate-to-human transmission has been shown but for which there is no evidence that subsequent human-to-human transmission occurs (Switzer et al. 2004; Boneva et al.2007). Once primate-to-human transmission occurs, many variables contribute to its ability to cause disease in the human population. A substantial literature treats the question of which pathogen properties contribute to the wide dispersal of some pathogens while others remain confined to small populations. These variables can be encapsulated by a concept known as the basic reproductive ratio, or R0, which is the number of secondary cases in a population caused by a single case. An R0 greater than 1 is associated with the potential to cause epidemic disease. R0 has relevance not only to the spread of pathogens at a local level but also to the ability of disease outbreaks to infect large numbers of people beyond the immediate geographic area. Characteristics of societies such as population density, population mobility, infrastructure, health status, and availability and quality of health care and health surveillance all influence R0. It can be argued that in many areas in Asia where humans and primates coexist, these societal characteristics tend to facilitate the emergence of pathogens. Factors that Influence Emergence of Primate-Borne Zoonoses Whether or not a particular enzootic microorganism has the capacity to cause significant human illness and death depends on several factors, including characteristics of the microorganism itself, characteristics of human (host) populations, characteristics of nonhuman (reservoir) populations, characteristics of vector species (for vector-

The Risk and Context of Emerging Primate-Borne Zoonoses

•

65

mediated diseases), and characteristics of the environments in which they interact (Table 2.1). Because these factors are dynamic in nature, their complex interrelationships are also constantly changing, opening up new possibilities for infectious disease transmission. Characteristics of Infectious Agents that Influence Zoonotic Transmission Pathogens possess a variety of characteristics that influence their transmission, including virulence, latency, mutability, and ability to infect a variable host range. Table 2.1. Risk Factors that Influence Cross-Species Transmission

Characteristics of Human Populations that Influence Infectious Agent Transmission

Attitudes toward animals Geographic location Mobility (individual and group) Diet Sanitation Water use Use of domesticated animals Population size, density

Demographics Social structure of families and communities Economic and political structure of community Health status Vaccination status Disease burden Nutrition status Previous pathogen exposure Availability/quality of medical care

Characteristics of Animal Populations that Influence Infectious Agent Transmission

Population size (ability to sustain an enzootic infection) Social structure Habitat Diet (e.g., reliance on humans) Range

Habituation to humans Health status Nutritional status Disease burden Intragroup interactions

Attributes of the Environment that Influence Cross-Species Transmission

Flora Fauna Food supply Land use

Pollution Natural disruption Human disruption

66

•

Lisa Jones-Engel and Gregory Engel

Similarly, many characteristics of human populations impact pathogen transmission: geographic location, mobility, social structure, diet, use of domesticated animals, and access to healthcare, to name a few. Characteristics of nonhuman primates, including diet, range, and social structure, influence the likelihood that they will transmit pathogens to humans or be exposed to endemic human pathogens. Finally, environmental factors, both “natural” and “human-made,” influence pathogen transmission. For example, rainfall and drought patterns can alter the migratory patterns of humans and primates as well as the distribution of pathogen vectors such as mosquitoes. War, with its attendant environmental impact and mass influxes of population, can profoundly affect ecological systems and make bidirectional pathogen transmission more likely. The myriad variables that influence pathogen transmission suggest the inadequacy of a single academic discipline for studying these phenomena. Rather, a multidisciplinary approach involving collaborative efforts among anthropologists, primatologists, veterinarians, physicians, epidemiologists, and virologists is integral to understanding the complex interplay of factors that are involved in primate-tohuman primate pathogen transmission. Context of Interspecies Contact The Human-Primate Interface In order to understand the critical first step of primate-to-human pathogen transmission, it is important to appreciate the various contexts of human-primate contact and how the specific properties of each context influence pathogen transmission. Worldwide interspecies contact occurs in a variety of contexts: pet markets, pet ownership, monkey forests, bushmeat hunting and consumption, circuses, zoos, and the aids (e.g., the use of macaques to harvest coconuts) and obstacles (crop-raiding primates) to agricultural production. Diseases can also be transmitted across the species barrier via vectors such as insects, even when humans and primates do not come into close physical proximity (Singh et al. 2004; Jongwutiwes et al. 2004). Monkey Forests/Temples Monkey forests are a common phenomenon throughout South and Southeast Asia, where primates play an important role in culture. Temple monkeys exhibit extensive sympatry with human populations, and tend to reside around specific temple structures either in or around areas of human habitation. These situations can be in human cities, small local shrines and temples in villages and towns, or specific temple sites that double as sacred places and tourist locations. This pattern is strongly associated with cultures that practice Hinduism and/or Buddhism and their associated rituals and temple construction patterns (Aggimarangee 1992; Fuentes et al. 2005). While construction of these Hindu and Buddhist temples reflects a time depth of no more than one to four millennia (depending on location), human and macaque sympatry and overlap in these areas has a much greater temporal depth. As deforestation

The Risk and Context of Emerging Primate-Borne Zoonoses

•

67

has increased, monkey temples have become refuges for populations of free-ranging primates. These primates subsist at least in part on the food and flower offerings left by temple worshippers. Over time, some of these temples have evolved into well known tourist destinations known as “monkey forests,” which offer domestic and international visitors the opportunity to observe and interact with the macaques. These temples of Southeast Asia present a particularly important setting in which to conduct research on cross-species pathogen transmission between primates and humans. Extensive, unregulated and often intimate contact between humans and primates occurs at these sites (Wheatley 1998; Engel et al. 2002; Fuentes and Gamerl 2005; Engel et al. 2006; Fuentes 2006; Jones-Engel et al. 2007, 2006a,b). Monkey forests differ widely in the amount and kinds of human-primate contact that occur in and around them. This depends on a multiplicity of variables including the species of primate, the density of the primate and human populations, the geographic accessibility of the monkey forest, food provisioning policies, availability of natural forage, the role the monkey forest plays in the economic life of the community, local and governmental regulations, and the cultural beliefs surrounding the primates, among other variable. At some sites, human-primate contact is actively encouraged, usually in the context of tourism, and almost always involves food rewards. Primates may climb onto the head and shoulder of visitors and may thus expose visitors to primate body fluids including saliva, urine, and feces that have the potential to infect humans via contact with their mucus membranes. In some monkey forests it is also not uncommon for aggressive behaviors to lead to primate bites and scratches, which have the potential to introduce primate-borne pathogens transcutaneously. Important aspects of monkey forests that bear on zoonotic transmission include multiple contacts of a population of primates with multiple humans over an extended period of time. In addition to physical contact, face-to-face proximity also frequently occurs in this context, providing the possibility of the transmission of infectious agents via respiratory droplets. A third possible route includes the transmission of disease through contact with contaminated objects (fomites). Finally, in the communities that surround the monkey forests humans and primates may share water sources, creating the possibility of water-borne transmission of enzootic primate pathogens. Primate Hunting and Consumption Primate hunting worldwide continues to take a large toll on primate populations and is an important context of interspecies contact. Primates are hunted using a variety of techniques, including guns, blowguns, snares, and traps. Once the primates are trapped, hunters may come into physical contact with primate body fluids including blood, mucous, urine, and tissue. The process of hunting and transporting and preparing primates for consumption or ceremonial or medicinal uses may bring a variety of people into contact. Still others may use or consume these primate-derived substances. Depending on the actual activity and individual, contact can be frequent and intense. Humans may come into contact with a variety of primate species, providing

68

•

Lisa Jones-Engel and Gregory Engel

the potential for different infectious agents to come into contact with each other and recombine. For the most part, however, the duration of contact with an individual primate is short. Primate fluids containing infectious agents have the opportunity to enter the human host via breaks in skin, through mucous membranes, or by ingestion. Importantly, because wild primates are more common in remote, rural areas where habitat can sustain them, bushmeat consumption is largely a phenomenon of rural areas that may be characterized by lower human population densities and less developed infrastructure. Thus, although bushmeat does appear on the tables of people living in distant lands, the bulk of the exposure involving bushmeat occurs near primate habitat. This serves to decrease the likelihood of human-to-human spread subsequent to primate-to-human transmission. Parks and Nature Reserves Parks or reserves constitute another context of human-primate contact. In many respects parks and nature reserves are similar to monkey temples in terms of the human-primate dynamics that occur in this context. Parks and nature reserves are globally more widely distributed than monkey temples, and include areas where there are no native primate populations. The amount and kind of interspecies contact at these parks varies widely, from frequent contact to virtually none (Adams et al. 2001; Lonsdorf et al. 2006; Gillespie et al. 2005). Pets From the perspective of interspecies pathogen transmission, the practice of owning pet primates bears particular examination. Previous research suggests that a variety of infectious agents may be transmitted from pet NHPs to humans as well as from humans to pet NHPs (Jones-Engel et al. 2001, 2004; Mack and Noble 1970; Ostrowski et al. 1998; Peeters et al. 2002). Exposure to pathogens can take place at several junctures during the life of a primate destined to be a pet: during the hunting or trapping process, while being transported, at market, and in the home of a pet owner. Pet primates live in close proximity with humans and may come into frequent physical contact over extended periods of time with their owners as well as with other members of the community. Pet markets bring multiple animal species together in close quarters, increasing the likelihood of pathogen transmission among primate and non-primate species and creating the possibility of new host-pathogen combinations (Schillaci et al. 2005, 2006; Karesh et al. 2005). Pathogens infecting pet primates may in turn be introduced into wild primate populations when wild primates come into contact with pets or when escaped pets come into contact with their wild counterparts (Jones-Engel et al. 2005a). Seen in this light, data on pet ownership may be crucial to understanding an important link in a chain of pathogen transmission that connects human beings the world over to wild primate populations. Two recent epidemics illustrate the prescience of this concern. The recent emergence of SARS in

The Risk and Context of Emerging Primate-Borne Zoonoses

•

69

Southeast Asia and monkeypox in the northern midwest of the United States (but note that monkeypox infection in humans is not linked to contact with primates) hint at the potential for the pet trade to link pathogen reservoirs in wild animals with human populations. Primate pet ownership provides a variety of contexts for bidirectional pathogen transmission. Urban Performance Monkeys Although performing primates are encountered throughout the world, Asian cultures have perhaps the longest and most vibrant tradition of using primates for entertainment (Thiereny 1987). Research in Java, Indonesia, suggests that performance primates are commonly found in larger metropolitan areas (Schillaci et al. 2005). The primates are either purchased by their owners from local animal markets or, in some cases, are bred by the owners. Training of these primates typically takes four to six months, after which the primates perform multiple shows six days a week in parks or along the streets in metropolitan areas. There is physical contact between the performance primates and audience members. Typically, the owners live in close, communal compounds, often with more than one primate. Extensive physical contact between the owner and primate occurs during the socializing, training, and performance periods. As previously described, the ecology of performance monkeys is distinct from that of primates that come into contact with humans in other contexts, such as temple monkeys, laboratory primates, and primates consumed as bushmeat (Schillaci et al. 2005). The performance monkey “niche” influences the amount and kinds of contact these primates have with humans and, in turn, influences the likelihood of zoonotic transmission occurring in this context. In regarding the risk of primate-tohuman transmission of infectious agents, it is important to consider the origin of performing primates. Figure 2.2 illustrates the typical route that performing monkeys may take as they are captured from the wild, transferred to live animal markets, and then brought to their owner’s compound and eventually into the urban centers. The movement of performing monkeys through live animal markets is significant because animal markets bring together a variety of species, primate and non-primate, from diverse geographic origins, each with its burden of infectious agents (Schillaci et al. 2006). The conditions in which the monkeys are maintained in the markets may compromise their immunity and increase their risk of acquiring novel infections. This, in turn, may have implications for the animals’ owners and “audience.” Markets Live primate/animal markets are a common phenomenon in Southeast Asia and may pose a particular risk for the emergence and dissemination of primate-borne pathogens (Karesh et al. 2005). Several factors contribute to this risk: 1) the animals in these markets are often kept in crowded enclosures, increasing the likelihood of

70

•

Lisa Jones-Engel and Gregory Engel

Figure 2.2. Route that performing monkeys may take as they are captured from the wild, transferred to live animal markets and then to their owner’s compound and eventually into urban centers.

pathogen transmission; 2) animals of different species and different geographic origins are brought together in a small space, making interspecific pathogen transmission possible; 3) primate/animal markets are places of high human density, increasing human-primate interaction and the possibility of human-primate pathogen transmission; 4) animals in pet markets are often poorly kept and thus more vulnerable to disease; and 5) primates acquired at these markets undergo a variety of fates, including

The Risk and Context of Emerging Primate-Borne Zoonoses

•

71

export to other regions or even other continents, and thus may act as vectors for the introduction of novel pathogens into human populations. Seen in this light, markets are potentially key contributors to the global spread of zoonotic and anthropozoonotic disease. Learning about pathogens reservoirs as well as the dynamics of pathogen transmission in markets is important to understanding these risks and ultimately how to control them. Zoo/Sanctuaries Zoos and animal sanctuaries in Asia are typically not bound by the restrictions on contact between visitors and animals that are observed in the United States. In fact, it is quite common for visitors to touch and feed animals in zoos/sanctuaries, particularly primates (Agoramoorthy and Hsu 2005). Intimate contact of this type can provide yet another context for the transmission of pathogens between primates and humans. Wild Primates Surveys of wild primate populations to determine prevalence of enzootric primate pathogens are important because they provide information on the primate “pathogen reservoir” and thus identify infectious agents to which humans may be exposed. In addition, learning about exposures to endemic human pathogens among wild primates can provide some idea of the extent of pathogen flow between the two populations. For example, the limited data available on bushmeat hunting in Africa suggest that primates typically make up a fairly low percentage of the yearly bushmeat haul, and data indicate that despite making up a low percentage of the animals hunted, primate populations are rapidly dwindling in Africa. A bushmeat hunter, when he processes his kill, will have an intense exposure to primate bodily fluids. These primate bodily fluids can come into contact with the hunter’s skin and mucosal surfaces. Slight nicks and cuts to the hunter during the butchering process are common. Conclusions A growing literature suggests that cross-species transmission of infectious agents occurs between humans and several NHP species in a variety of contexts and in diverse geographic areas (Jones-Engel et al. 2006b, 2005a,b, 2004, 2001, 2008; Switzer et al. 2004; Wolfe et al. 2004; Huff and Barry 2003; Engel et al. 2002; Lerche et al. 2001). Indeed, wherever humans and primates come into contact, the potential for cross-species transmission exists. Whether cross-species transmission occurs depends on a number of factors, including the prevalence of infectious agents present in the primate reservoirs, the contexts of interspecies contact, and the frequency with which contact occurs (Engel et al. 2006). Zoonotic diseases are likely to continue to constitute a public health issue in the future (Schillaci et al. 2008). Our approach to studying zoonotic transmission will dictate our success in detecting and preventing their

72

•

Lisa Jones-Engel and Gregory Engel

spread among human populations. Above we have outlined the rationale for focusing increased attention on primate-borne zoonotic transmission in Asia and put forth a paradigm, risk analysis, by which infectious risks can be estimated and resources allocated toward prevention. Risk analysis holds the promise of bringing to bear scientific methodology to meet the challenge of future emerging infectious diseases. In its fully conceived form, this ideally entails making comprehensive assessments of zoonotic risk across all phyla and human populations. Clearly, such an undertaking requires a substantial organization, or reorganization of the current approach to zoonotic disease prevention, and a commensurate commitment of resources. In particular, more attention needs to be paid to the human-animal interface, where cross-species transmission occurs. Seen in this light, our work on human-primate disease transmission may provide a glimpse into the challenges that will be faced in this ambitious undertaking.

References Acha, P.N. and B. Szyfres. 1980. Zoonoses and communicable diseases common to man and animals. Washington, D.C.: Pan American Health Organization. Adams, H.R., J.M. Sleeman, I. Rwego, and J.C. New. 2001. Self-reported medical history survey of humans as a measure of health risk to the chimpanzees (Pan troglodytes schweinfurthii) of Kibale National park, Uganda. Oryx 35: 308–12. Aggimarangsee, N. 1992. Survey for semi-tame colonies of macaques in Thailand. Natural History Bulletin of Siam Society 40: 103–66. Agoramoorthy, G. and M.J. Hsu. 2005. Use of nonhuman primates in entertainment in Southeast Asia. Journal of Applied Animal Welfare 8: 141–49. Alcendor, D.J., P.A. Barry, E. Pratt-Lowe, and P.A. Luciw. 1993. Analysis of the rhesus cytomegalovirus immediate-early gene promoter. Virology 194: 815–21. Alford, C.A. and W.J. Britt. 1993. Cytomegalovirus. In B. Roizman, R.J. Whitley, and C. Lopez, eds. The Human Herpesviruses. New York: Raven Press, 227–55. Amblard, J., P. Obiang, S. Edzang, C. Prehaud, et al. 1997. Identification of the Ebola virus in Gabon in 1994. Lancet 349: 181–82. Arita, I., Z. Jezek, L. Khodakevich, and K. Ruti. 1985. Human monkeypox: a newly emerged orthopoxvirus zoonosis in the tropical rain forests of Africa. American Journal of Tropical Medicine and Hygiene 34: 781–89. Bailes, E., F. Gao, F. Bibollet-Ruche, V. Courgnaud, et al. 2003. Hybrid origin of SIV in chimpanzees. Science 300: 1713. Bergsagel, D.J., M.J. Finegold, J.S. Butel, W.J. Kupsky, and R.L. Garcea. 1992. DNA sequences similar to those of simian virus 40 in ependymomas and choroid plexus tumors of childhood. New England Journal of Medicine 326: 988–93. Blewett, E.L., G. White, J.T. Saliki, and R. Eberle. 2001. Isolation and characterization of an endogenous cytomegalovirus (BaCMV) from baboons. Archives of Virology 146: 1723–38. Blewett, E.L., J. Lewis, E.L. Gadsby, S.R. Neubauer, and R. Eberle. 2003. Isolation of cytomegalovirus and foamy virus from the drill monkey (Mandrillus leucophaeus) and prevalence of antibodies to these viruses amongst wild-born and captive-bred individuals. Archives of Virology 148: 423–33.

The Risk and Context of Emerging Primate-Borne Zoonoses

•

73

Boneva R.S., W.M. Switzer, T.J. Spira, V.B. Bhullar, V. Shanmugam, M.E. Cong, L. Lam, W. Heneine, T.M. Folks, L.E. Chapman. 2007. Clinical and virological characterization of persistent human infection with simian foamy viruses. AIDS Research and Human Retroviruses 23:1330-7 Brack, M. 1987. Agents Transmissible from Simians to Man. Berlin: Springer-Verlag. Brooks, J.I., R. Erling, R.G. Pilon, J.W. Smith, et al. 2002. Cross-species retroviral transmission from macaques to human beings. Lancet 360: 387–88. Callahan, M.E., W.M. Switzer, A.L. Matthews, B.D. Roberts, et al. 1999. Persistent zoonotic infection of a human with simian foamy virus in the absence of an intact orf-2 accessory gene. Journal of Virology 73: 9619–24. Calattini, S., S.A. Chevalier, R. Duprez, S. Bassot, et al. 2005. Discovery of a new human Tcell lymphotropic virus (HTLV-3) in Central Africa. Retrovirology 9(2): 30. Carbone, M., H.I. Pass, P. Rizzo, M. Marinetti, et al. 1994 Simian virus 40-like DNA sequences in human pleural mesothelioma. Oncogene 9: 1781–90. Centers for Disease Control and Prevention. 1987. B-virus infection in humans—Pensacola, Florida. MMWR Morbidity and Mortality Weekly Report 36: 289–90, 295–96. ———. 1989. B-virus infection in humans—Michigan. MMWR Morbidity and Mortality Weekly Report 38: 453–54. ———. 1998. Fatal Cercopithecine herpesvirus 1 (B Virus) infection following a mucocutaneous exposure and interim recommendations for worker protection. MMWR Morbidity and Mortality Weekly Report 47: 1073–76. Cicala, C., F. Pompetti, and M. Carbone. 1993. SV40 induces mesotheliomas in hamsters. American Journal of Pathology 142: 1524–33. Daszak, P., G.M. Tabor, A.M. Kilpatrick, J. Epstein, and R. Plowright. 2004. Conservation medicine and a new agenda for emerging diseases. Annals of the New York Academy of Sciences 1026: 1–11. Decision Tree Writing Group. 2006. Clinical response decision tree for the mountain gorilla (Gorilla beringeii) as a model for great apes. American Journal of Primatology 68: 909–27. Engel, G.A., L. Jones-Engel, M.A. Schillaci, K.G. Suaryana, A. Putra, et al. 2002. Human exposure to herpesvirus B-seropositive macaques, Bali, Indonesia. Emerging Infectious Diseases 8: 789–95. Engel, G., L. Hungerford, L. Jones-Engel, D. Travis, et al. 2006. Risk assessment: a model for predicting cross-species transmission of SFV from macaques (M. fascicularis) to humans at a monkey temple in Bali, Indonesia. American Journal of Primatology 68: 934–48. Eizuru, Y., K. Tsuchiya, R. Mori, and Y. Minamishima. 1989. Immunological and molecular comparisons of simian cytomegaloviruses isolated from African green monkey (Ceropithicus aethiops) and Japanese macaque (Macaca fuscata). Archives of Virology 107: 65–75. Falcone, V., J. Leupold, J. Clotten, E. Urbanyi, et al. 1999. Sites of simian foamy virus persistence in naturally infected African green monkeys: latent provirus is ubiquitous, whereas viral replication is restricted to the oral mucosa. Virology 257: 7–14. Fiennes, R. 1967. Zoonoses of Primates: The Epidemiology and Ecology of Simian Diseases in Relation to Man. Ithaca: Cornell University Press. Fuentes, A. and S. Gamerl. 2005. Disproportionate participation by age/sex classes in aggressive interactions between long-tailed macaques (Macaca fascicularis) and human tourists at Padangtegal Monkey Forest in Bali, Indonesia. American Journal of Primatology 66: 197–204.

74

•

Lisa Jones-Engel and Gregory Engel

Fuentes, A., M. Southern, and K.G. Suaryana. 2005. Monkey forests and human landscapes: is extensive sympatry sustainable for Homo sapiens and Macaca fascicularis on Bali? In J. Patterson and J. Wallis, eds. Commensalism and Conflict: The Human-Primate Interface. Norman, OK: ASP Press, 168–95. Fuentes, A. 2006. Human culture and monkey behavior: assessing the contexts of potential pathogen transmission between macaques and humans. American Journal of Primatology 68: 880–96. Georges, A.J., E.M. Leroy, A.A. Renaut, C.T. Benissan, et al. 1999. Ebola haemorrhagic fever. Journal of Infectious Diseases 179(Suppl 1): S65–75. Gessain, A. and R. Mahieux. 2000. Epidemiology, origin and genetic diversity of HTLV-1 retrovirus and STLV-1 simian affiliated retrovirus. Bulletin de la Societe de Pathologie Exotique 93: 163–71. Gillespie, T.R., C.A. Chapman, and E.C. Greiner. 2005. Effects of logging on gastrointestinal parasite infection risk in African primates. Journal of Applied Ecology 42: 699–707. Grant, R., C. Vinh, R. Watanabe, R. McLain, et al. 2005a. A novel type-D simian retrovirus (SRV) in newly imported M. nemestrina similar to endogenous langur monkey retrovirus; evidence of cross-genus transmission. American Journal of Primatology 66: 77–78. Grant, R., R. Watanabe, L. Kuller, B. Poland, et al. 2005b. Comparison of wild-type SRV in macaques from Singapore and Indonesia. Journal of Medical Primatology 34: 332–33. Hahn, B.H., G.M. Shaw, K.M. De Cock, and P.M. Sharp. 2000. AIDS as a zoonosis: scientific and public health implications. Science 287: 607–14 Hankenson, F.C., N.A. Johnston, B.J. Weiglerand R.F. Di Giacomo. 2003. Zoonoses of occupational health importance in contemporary laboratory animal research. Comparative Medicine 53: 579–601. Heneine, W., W.M. Switzer, P. Sandstrom, J. Brown, S. Vedapuri, et al. 1998. Identification of a human population infected with simian foamy viruses. Nature Medicine 4: 403–7. Hobson, K. 2003. Pet problems: from monkeypox to HIV, animals bring us new ills. US News and World Report Jun 23(134): 40–41. Holmes, G.P., J.K. Hilliard, K.C. Klontz, A.H. Rupert, et al. 1990. B-virus (Herpesvirus simiae) infection in humans: epidemiological investigations of a cluster. Annals of Internal Medicine 112: 833–39. Holmes, G.P., L.E. Chapman, J.A. Stewart, S.E. Straus, et al. 1995. Guidelines for the prevention and treatment of B-virus infections in exposed persons. Clinical Infectious Diseases 20: 421–39. Hueston, W.D. and K.D. Walker. 1993. Macroepidmiologic contributions to quantitative risk assessment. Review of Science and Technology 12(4): 1197–1201. Huff, J.L. and P.A. Barry. 2003. B-virus (Cercopithecine herpesvirus 1) infection in humans and macaques: potential for zoonotic disease. Emerging Infectious Diseases 9: 246–50. Hummeler, K., W.L. Davidson, W. Henle, A.C. LaBoccetta, and H.G. Ruch. 1959. Encephalomyelitis due to infection with Herpesvirus simiae (Herpesvirus B): report of two fatal laboratory cases. New England Journal of Medicine 261: 64–68. Jones-Engel, L., G.A. Engel, M.A. Schillaci, R. Babo, and J. Froehlich. 2001. Detection of antibodies to selected human pathogens among wild and pet macaques (Macaca tonkeana) in Sulawesi, Indonesia. American Journal of Primatology 54: 171–78. Jones-Engel, L., G.A. Engel, M.A. Schillaci, J. Froehlich, et al. 2004. Prevalence of enteric parasites in pet macaques in Sulawesi, Indonesia. American Journal of Primatology 62: 71–82. Jones-Engel, L., G.A. Engel, and M.A. Schillaci. 2005a. An ethnoprimatological assessment of

The Risk and Context of Emerging Primate-Borne Zoonoses

•

75

disease transmission among humans and wild and pet macaques on the Indonesian island of Sulawesi. In Commensalism and Conflict: The Primate-Human Interface, eds. J. Patterson and J. Wallis. Norman, OK: ASP Press, 196–221. Jones-Engel, L., G. Engel, M. Schillaci, A. Rompis, et al. 2005b. Primate to human retroviral transmission in Asia. Emerging Infectious Diseases 7: 1028–35. Jones-Engel, L., G.A. Engel, J. Heidrich, M. Chalise, et al. 2006a. Temple monkeys and health implications of commensalism, Kathamandu, Nepal. Emerging Infectious Diseases 12: 900–906. Jones-Engel, L., G.A. Engel, M.A. Schillaci, B. Lee, H. Heidrich, et al. 2006b. Considering human-primate transmission of measles virus through the prism of risk analysis. American Journal of Primatology 68: 868–79. Jones-Engel, L., et al. 2008. Diverse contexts of zoonotic transmission of simian foamy viruses in Asia. Emerging Infectious Diseases 14:1200–1208. Jones-Engel L., K. Steinkraus, S. Murray, G.A. Engel, R. Grant, N. Aggimarangsee, B.Y-P. Lee, C. May, M.A. Schillaci, T.Chutthapati, C. Somgird, L. Vojtech, J. Zhao, M. Linial. 2007. Sensitive assays for simian foamy virus reveal a high prevalence of infection in commensal, free-ranging, Asian monkeys. Journal of Virology 81:7330–7. Jongwutiwes, S., C. Putaporntip, T. Iwasaki, T. Sata, and H. Kanbara. 2004. Naturally acquired Plasmodium knowlesi malaria in human, Thailand. Emerging Infectious Diseases 10: 2211–13. Karesh, W.B., R.A. Cook, E.L. Bennett, et al. 2005. Wildlife trade and global disease emergence. Emerging Infectious Diseases 11: 1000–1002. Kaur, A., N. Kassis, C.L. Hale, M. Simon, et al. 2003. Direct relationship between suppression of virus-specific immunity and emergence of cytomegalovirus disease in simian AIDS. Journal of Virology 77: 5749–58. Kellar, J.A. 1993. The application of risk analysis to international trade in animals and animal products. Review of Science and Technology 12: 1023–44. King, N.W., R.D. Hunt, and N.L. Letvin. 1983. Histopathologic changes in macaques with an acquired immunodeficiency syndrome (AIDS). American Journal of Pathology 113: 382–88. Krainer, M., T. Schenk, C.C. Zielinski, and C. Muller C. 1995. Failure to confirm presence of SV40 sequences in human tumours. European Journal of Cancer 31A: 1893. Le Guenno, B., P. Formenty, M. Wyers, P. Gounon, et al. 1995. Isolation and partial characterization of a new strain of Ebola virus. Lancet 345: 1271–74. Leighton, F.A. 2002. Health risk assessment of the translocation of wild animals. Review of Science and Technology 21: 187–95. Lerche, N.W., K.G. Osborn, P.A. Marx, S. Prahalda, et al. 1986. Inapparent carriers of simian AIDS type D retrovirus and disease transmission in saliva. Journal of the National Cancer Institute 77: 489–96. Lerche, N.W., W.M. Switzer, J.L. Yee, V. Shanmugam, et al. 2001. Evidence of infection with simian type D retrovirus in persons occupationally exposed to nonhuman primates. Journal of Virology 75: 1783–89. Leroy, E.M., P. Rouquet, P. Formenty, S. Souquière, et al. 2004. Multiple Ebola virus transmission events and rapid decline of Central African wildlife. Science 303: 387–90. Lonsdorf, E.V., D. Travis. A.E. Pusey, and J. Goodall. 2006. Using retrospective health data from the Gombe Chimpanzee Study to inform future monitoring efforts. American Journal of Primatology 68: 855–67.

76

•

Lisa Jones-Engel and Gregory Engel

Mack, T.M. and J. Noble. 1970. Natural transmission of smallpox from man to performing monkeys. An ecological curiosity. Lancet 11: 752–54. Marx, P.A., M.L. Bryant, K.G. Osborn, D.H. Maul, et al. 1985. Isolation of a new serotype of SAIDS type-D retrovirus from Celebes black macaques (Macaca nigra) with immune deficiency and retroperitoneal fibromatosis. Journal of Virology 56: 571–78. Maul, D.H., C.H. Miller, P.A. Marx, M.L. Bleviss, et al. 1985. Immune defects in simian acquired immunodeficiency syndrome. Veterinary Immunology and Immunopathology 8: 201–14. Metcalf, H.E., J.H. Blackwell, and J.A. Acree. 1996. Application of risk assessment to international trade in animals and animal products. Annals of the New York Academy of Sciences 791: 280–95. Meiering, C.D. and M.L. Linial. 2001. Historical perspective of foamy virus epidemiology and perspective. Clinical Microbiology Reviews 14: 165–76. MGVP 2002 Employee Health Group. 2004. Risk of disease transmission between conservation personnel and the mountain gorillas: results from an employee health program in Rwanda. EcoHealth 1: 351–61. Minamishima, Y., B.J. Graham, and M. Benyesh-Melnick. 1971. Neutralizing antibodies to cytomegaloviruses in normal simian and human sera. Infection and Immunity 4: 368–73. Morse, S.S. 1995. Factors in the emergence of infectious disease. Emerging Infectious Diseases 1: 7–15. Murray, S.M., L.J. Picker, M.K. Axthelm, and M.L. Linial. 2006. Expanded tissue targets for foamy virus replication with simian immunodeficiency virus-induced immunosuppression. Journal of Virology 80: 663–70. Nerrienet, E., M.L. Santiago, Y. Foupouapouognigni, E. Bailes, et al. 2005. Simian immunodeficiency virus infection in wild-caught chimpanzees from cameroon. Journal of Virology 79: 1312–19. Neumann-Haefelin, D., U. Fleps, R. Renne, and M. Schweizer. 1993. Foamy viruses. Intervirology 35: 196–207. Nigida, S.M., L.A. Falk, L.G. Wolfe, and F. Deinhardt. 1979. Isolation of a cytomegalovirus from salivary glands of white-lipped marmosets (Saguinus fuscicollis). Laboratory Animal Science 29: 53–60. Ohtaki, S., H. Kodama, R. Hondo, and T. Kurata. 1986. Activation of cytomegalovirus infection in immunosuppressed cynomolgous monkeys inoculated with varicella-zoster virus. Acta Pathologica Japonica 36: 1553–63. Ostrowski, S.R., M.J. Leslie, T. Parrott, S. Abelt, and P.E. Piercy. 1998. B-virus from pet monkeys: an emerging threat in the United States? Emerging Infectious Diseases 4: 117–21. Peeters, M., V. Courgnaud, B. Abela, P. Auzel, et al. 2002. Risk to human health from a plethora of simian immunodeficiency viruses in primate bushmeat. Emerging Infectious Diseases 8: 451–57. Reed, K.D., J.W. Melski, M.B. Graham, R.L. Regnery, et al. 2004. The detection of monkeypox in humans in the Western Hemisphere. New England Journal of Medicine 22(350): 342–50. Rouquet, P., J-M. Froment, M. Bermejo, A. Kilbourne, et al. 2005. Wild animal mortality monitoring and human Ebola outbreaks, Gabon and Republic of Congo, 2001–2003. Emerging Infectious Diseases 11: 283–90. Santiago, M.L., F. Range, B.F. Keele, Y. Li, et al. 2005. Simian immunodeficiency virus infection in free-ranging sooty mangabeys (Cercocebus atys atys) from the Tai Forest, Cote

The Risk and Context of Emerging Primate-Borne Zoonoses

•

77

d’Ivoire: implications for the origin of epidemic human immunodeficiency virus type 2. Journal of Virology 79: 12515–27. Schillaci, M.A., L. Jones-Engel, G. Engel, Y. Paramastri, et al. 2005. Prevalence of enzootic simian viruses among urban performance monkeys in Indonesia. Tropical Medicine and International Health 10: 1305–14. Schillaci, M.A., L. Jones-Engel, G. Engel, and R. Kyes. 2006. Exposure to human respiratory viruses among urban performing monkeys in Indonesia. American Journal of Tropical Medicine and Hygiene 75: 716–19 Schillaci M.A., L. Jones-Engel L, G.A. Engel, A. Fuentes . 2008. Characterizing the threat to the blood supply associated with nonoccupational exposure to emerging simian retroviruses. Transfusion 48:398-401. Shah, K.V. and J.A. Morrison. 1969. Comparison of three rhesus groups for antibody patterns to some viruses: Absence of active simian virus 40 transmission in the free-ranging rhesus of Cayo Santiago. American Journal of Epidemiology 89: 308–15. Shah, K.V. and N. Nathanson. 1976. Human exposure to SV40: review and comment. American Journal of Epidemiology 103: 1–12. Singh, B., L. Kim Sung, A. Matusop, A. Radhakrishnan, et al. 2004. A large focus of naturally acquired Plasmodium knowlesi infections in human beings. Lancet 363: 1017–24. Slattery, J.P., G. Franchini, and A. Gessain. 1999. Genomic evolution, patterns of global dissemination, and interspecies transmission of human and simian T-cell leukemia/lymphotropic viruses. Genome Research 9: 525–40. Switzer, W.M., V. Bhullar, V. Shanmugam, M. Cong, et al. 2004. Frequent simian foamy virus infection in persons occupationally exposed to nonhuman primates. Journal of Virology 78: 2780–89. Travis, D.A., L. Hungerford, G.A. Engel, and L. Jones-Engel. 2006. Disease risk analysis: a tool for primate conservation planning and decision making. American Journal of Primatology 68: 855–67. Tsai, C.C., W.E. Giddens Jr., W.R. Morton, S.L. Rosenkranz, et al. 1985. Retroperitoneal fibromatosis and acquired immunodeficiency syndrome in macaques: epidemiologic studies. Laboratory Animal Science 35: 460–64. Vandamme, A.M., M. Salemi, and J. Desmyter. 1998 The simian origins of the pathogenic human T-cell lymphotropic virus type I. Trends in Microbiology 6: 477–83. Vogel, P., B.J. Weigler, H. Kerr, A. Hendrickx, and P.A. Barry. 1994. Seroepidemiologic studies of cytomegalovirus infection in a breeding population of rhesus macaques. Laboratory Animal Science 44: 25–30. Walsh, P.D., R. Biek, and L.A. Real. 2005. Wave-like spread of Ebola Zaire. PLoS Biology 3: e371. Wheatley, B. 1998. The Not So Sacred Monkeys of Bali. Prospect Heights, IL: Waveland Press. Wolfe, N.D., W. Heneine, J.K. Carr, A.D. Garcia, et al. 2005. Emergence of unique primate T-lymphotropic viruses among central African bushmeat hunters. Proceedings of the National Academy of Sciences 31(102): 7994–99. Wolfe, N.D., W.M. Switzer, J.K. Carr, V.B. Bhullar, et al. 2004. Naturally acquired simian retrovirus infections in central African hunters. Lancet 363: 932–37. Yamashita, M., E. Ido, T. Miura, and M. Hayami. 1996. Molecular epidemiology of HTLV-I in the world. J Acquired Immune Deficiency Syndromes and Human Retrovirology 13 Suppl 1: S124–31.

•3• Viral Panic, Vulnerability, and the Next Pandemic Ann Herring

Introduction The problem of infectious disease in human societies, past and present, is an important site for anthropological theorizing because it sits at the juncture between the microcosmos, evolution, and human behavior. It forms a natural bridge between the nature/culture divide. In this essay, I discuss the intersection between the social and biological worlds through a consideration of the prospect of an avian influenza pandemic in the twenty-first century and its connections, real and constructed, to the 1918 influenza pandemic. More specifically, I explore a narrative line that is embedded in the discourse on avian influenza. During the course of any epidemic, social responses surface in parallel to the challenge of the disease itself as the epidemic takes shape, becomes visible, and then is acknowledged by the people and societies vulnerable to it. Explanations emerge as a means of regaining control and asserting rationality over the crisis. Structures of blame inevitably arise through the process of explanation, and as managing the epidemic becomes a vehicle for social control. As disease and death subside, moral lessons are drawn (Rosenberg 1992). Narratives have a powerful influence on public concern about health crises and may influence health policy. For this reason, it is important to identify and critique the narratives and moral lessons that run through scholarly and media discussions of epidemics, here exemplified by avian influenza. In the case of avian influenza, failure to identify the connections between “bird flu” and the social, economic, and political contexts that influence who is actually vulnerable to it, creates panic about an inevitable global pandemic that threatens everyone. It also masks who is likely to be at the greatest risk of acquiring and dying from avian influenza. This is dangerous from a

Viral Panic, Vulnerability, and the Next Pandemic

•

79

public health policy perspective. In this essay I propose that the concept of syndemic offers a useful way to consider both the terrain and the microbe together, and to develop a more textured analysis of who may be vulnerable to “the next pandemic.” Viral Panic By the third quarter of the twentieth century, interest in infectious disease had waned—at least in a western medical context—and epidemiologic transition theory had relegated pestilence and famine to the past. Degenerative and human-induced diseases, such as cancer and cardiovascular diseases, predominated (Omran 1971). Frederick Cartwright’s leap of faith in 1983 nicely captures the conviction of the period: “It is my belief that, unless control breaks down through widespread famine or atomic warfare, both of which are possibilities, our world has seen the last of the great killing pandemics” (1983: 279). He was not alone. In 1967, William H. Stewart, surgeon-general of the United States, declared that “victory over infectious disease was imminent” (Armelagos 1998: 24). Then HIV/AIDS emerged to shake the foundations of epidemiological thought. The worldwide pandemic demonstrated that infectious disease was not a vestige of the past but an inevitable aspect of living in the organic world (Lederberg 1988), even the affluent Western world (Morse 1991:387). In his book History of AIDS: Emergence and Origin of a Modern Pandemic, Mirko Grmek captures the complete reversal in thinking that accompanied the emergence of HIV/AIDS: “Influenza was the last of the classic pestilences; AIDS, both unpredicted and unpredictable within the framework of the old nosology, is the first of the postmodern plagues” (Grmek 1990: ix). As the security of the age of degenerative and human-made diseases (Omran 1971) has given way to the age of emerging, re-emerging, and antimicrobial resistant diseases (Barrett et al. 1998), anxiety has grown as a breathtaking array of emerging and re-emerging diseases has been recognized, along with the many factors that contribute to their new visibility (Waltner-Toews 1995: 46). We live in an era obsessed with killer germs, says Nancy Tomes, in an era of “viral panic” (2000: 194). A “post-AIDS, post-Cold War crisis of confidence” has emerged as the old twentieth century belief in the biomedical conquest of infection has faded in the face of insurmountable evidence to the contrary (Tomes 2000: 192). There is a new sense of vulnerability and uncertainty with respect to infectious disease, rekindling fears of mortality on the scale of historic plagues and spurring research into the origins and circumstances that allowed epidemics to erupt and flourish in the past. Since it came into view in 1981, HIV/AIDS has “stimulated more interest in history than any other disease of modern times” (Fox and Fee 1988: 1). Vulnerability to a Pandemic A body of opinion now considers emerging infectious diseases and epidemics as inevitable (Klempner and Shapiro 2004: 2334), natural features of human life in

80

•

Ann Herring

a dynamic ecosystem (Lederberg 1993), connected to human-induced changes in that ecosystem (Last 1999). Among the emerging diseases currently generating viral panic—and apocalyptic terror that it represents the seeds of “the next pandemic”—is avian influenza (H5N1). As we wait for the next pandemic, discussions of viral evolution understandably have assumed enormous significance. There are three major forms of influenza (A, B and C), but only influenza A gives rise to pandemics. Influenza A, the 8-stranded RNA virus associated with human pandemics, has the capacity to evolve rapidly through genetic recombination with influenza strains from animal species (Palese 1993: 226), a process through which it can evolve suddenly and dramatically through genetic shift. These new combinations of genes, in turn, produce variation in the two antigens, hemagglutinen (H) and neurmaminidase (N) that sprout from its surface coat. When this process of hybridization and genetic shift occurs, a new strain of influenza emerges. Ultimately, influenza is a zoonotic disease of avian origin; all known influenza A subtypes originated from the aquatic bird reservoir (Webster 1998). It spreads efficiently via droplet nuclei and has a short incubation period, which enhances its ability to spread rapidly from person to person. It “is probably one of the oldest emerging viruses” and may have been responsible for epidemics in ancient Greece and Rome (Webster 1993: 37). The antiquity of influenza pandemics, their reservoir in aquatic birds, and the emergence of a new avian virus H5N1 leads to questions about how far away we are from the next pandemic. There are three steps in the process: 1) transmission of a new influenza viral subtype to humans; 2) viral replication that produces disease in humans; and 3) efficient human-to-human transmission of the virus. Since 1997, the first two conditions have been met on several occasions. As for the final condition, efficient human-to-human transmission, “The question ... is when such changes will happen” (Monto 2005: 324). “It could happen tonight, next year, or even ten years from now” (Osterholm 2005: 36). The last of the classical pestilences is the impending scourge of the twenty-first century. When the H5N1 strain of avian influenza infected and killed six people in Hong Kong in 1997, the World Health Organization ordered the slaughter of all chickens to prevent the third step, efficient human-to-human transmission, and a worldwide pandemic of bird flu. The virus was not eradicated and avian influenza, endemic in poultry in many parts of Asia, continues to evolve. In 2003, the Z strain of H5N1 emerged. Pathogenic to a wider range of species compared to other strains, the new strain is also resistant to first-line antiviral drugs, such as amantadine and rimantadine (Monto 2005: 323). The Z strain has widened its geographical range. In 2004 it had spread to nine countries in East and Southeast Asia (Li et al. 2004) and was identified in the Middle East, Africa, and Europe in 2006 (WHO 2006a). It is expected to infect poultry operations in North and South America (Butler and Ruttimann 2006), though this had not transpired at the time of writing (WHO 2008a). As more poultry are infected,

Viral Panic, Vulnerability, and the Next Pandemic

•

81

and as increasing numbers of people are exposed to H5N1, “all the prerequisites for the start of a pandemic have been met save one—namely, genetic changes in this virus that would allow it to achieve efficient human-to-human transmission” (Stöhr 2005: 4). Person-to-person spread of avian influenza nevertheless has been documented. PCR analysis of a cluster of seven relatives who contracted H5N1 in a remote village in Sumatra, Indonesia, indicates that a father contracted the virus after prolonged, close contact with his ailing 10-year-old son whom he nursed in hospital (Rosenthal 2006). Studies of human cases of avian influenza show, however, that the virus does not spread easily between people. It tends to colonize the lower lung and favours cell receptors in the deepest branches of the respiratory tract. Its preference for deeply buried tissue has limited the ability of H5N1 to spread from person to person by coughs and sneezes (Shinya et al. 2006). This seems to have inhibited achievement of the final step along the road toward a pandemic: efficient human-to-human transmission. Anchoring Avian Influenza to the 1918 Influenza Pandemic Fears that a killer bird flu is on the horizon—along with the massive damage that may accompany it—are anchored in the 1918 influenza pandemic and H1N1, the influenza A virus associated with it. Anchoring is a mechanism whereby the understanding of a new disease is linked and configured in terms of past epidemics (Joffe 1999). This is a process of representation through which a crisis is made understandable and less threatening by connecting it to familiar historical events, metaphors, or symbols. Anchoring a potential H5N1 outbreak to the 1918 influenza pandemic serves to enhance the climate of viral panic. This also occurred when media representations of SARS linked it to the 1918 influenza pandemic and the Black Death (Washer 2004). In much the same way, anchoring the vCJD/BSE to HIV/AIDS in Britain in the late 1990s increased fear (Washer 2006). Discussions of avian influenza’s potential to produce an unimaginable death toll draw parallels to the 1918 outbreak in which some fifty to one hundred million people may have perished worldwide (Johnson and Mueller 2005). The rapid spread of the disease, sudden onset of symptoms among otherwise healthy people, and excess mortality among young adults in the prime of life are frequently reported. The symptoms and medical histories of people who died from H5N1 and from H1N1 in 1918 are described as “disturbingly similar” (Garrett 2005: 14) and H5N1 seems to have an affinity for previously healthy young adults and children. There are other, less alarming and destructive pandemics that could be anchored to avian influenza, notably 1957 (“Asian influenza pandemic,” H2N2), 1968 (“Hong Kong pandemic,” H2N2), and 1977 (“Russian flu” or “Russian threat,” H1N1). They are not invoked in discussions of avian influenza or other frightening new diseases, such as SARS. This is because the 1918 pandemic is constructed as “the catastrophe against which all modern pandemics are measured” (Pandemics and Pandemic

82

•

Ann Herring

Threats Since 1900: 1). It is “the mother of all pandemics” (Taubenberger and Morens 2006) and the gold standard for emerging and re-emerging disease. The 1918 pandemic is anchored, in turn, to ancient plagues. Its devastating death toll, for example, is said to have outranked the Black Death and the Plague of Justinian (Walters 1978: 856). At the time of the 1918 pandemic, when no one knew what was causing healthy people to sicken and die with astonishing speed, and from frightening symptoms, some suggested that the Spanish Flu actually was the Black Death in new guise (MacDougall 1985: 2090–91). In this way, the discourse about avian influenza is connected, through the 1918 pandemic, to medieval plague—the classic image of pestilence and plague. The connection between the H1N1 1918 virus and H5N1 avian influenza tightened in October 2005 with the publication of the genome for H1N1, an internationally newsworthy event (see Appendix 1 of this chapter for more details). Initial phylogenetic analysis had suggested that the 1918 variant of H1N1 was closely related to a classical swine flu strain (Reid et al. 1999). A later, more comprehensive analysis resulted in a different conclusion: the strain’s genome was primarily avian (Taubenberger et al. 2005). This heightened worries about the risks to global health from avian influenza. This anxiety was magnified in January 2006 when H5N1 virus samples taken from people in Turkey were discovered to carry mutations believed to have the potential to facilitate person-to-person spread. Later, this conclusion was judged “premature” and “overinterpreted” in light of the genetic complexity of the influenza virus and the fact that virulence and transmissibility are multigenic traits (Basu 2006: 258). But it is evident from the hasty conclusions that scientific researchers are not immune to the influence of viral panic. Even differences between the two viruses provoke anxiety. A high case rate and low mortality rate are well known features of the 1918 outbreak; the vast majority of people who contracted influenza recovered from it. There was considerable variability in the death toll from influenza (see below), but influenza mortality averaged about 3 percent, exceeding the less than 0.1 percent mortality typical for other influenza epidemics (Dull and Dowdle 1980). This is much lower still than the case-fatality rates for the 1997 Hong Kong outbreak of avian influenza in which 18 percent of affected children and 57 percent of adults older than 17 years of age died (Snacken et al. 1999). WHO mortality rates for the 362 reported, laboratory-confirmed human cases of H5N1 average 63 percent (WHO 2008b), contributing to the fear that avian influenza is “far more dangerous” than the 1918 variant (Garrett 2005: 3). Who is Vulnerable to Avian Influenza? According to classic epidemiological theory, virtually everyone is vulnerable to H5N1 avian influenza. This is because there are no antigens from previous exposures that would confer immunity to individuals, and herd immunity to communities. Still, human cases of H5N1 are not found in all age groups. Analysis of the forty-four cases

Viral Panic, Vulnerability, and the Next Pandemic

•

83

of H5N1 documented in 2004, for instance, shows that avian influenza was concentrated in previously healthy children and young adults for whom the case fatality rate was 73 percent (Stöhr 2005). This nexus of illness and death among the young looks suspiciously like the behavior of a disease that has afflicted the population before and to which older adults may have already acquired immunity through exposure in childhood. Such an interpretation is consistent with the observation that H5 viruses have been present in human populations since the late 1950s (Wade 2006). In addition, immunity to the N1 antigen has been insufficiently studied (Kilbourne 2006: 13). To have a better sense of who is actually vulnerable to contracting and dying from avian influenza, we need to know more about the seroprevalence of H5N1 in communities that have been affected by it. Furthermore, there has been insufficient study of the social and economic context of vulnerability to infection. The higher risks of contracting avian influenza, especially among females in the 10–29 age category, may be linked to their roles in poultry farming, such as culling and de-feathering birds, or in food preparation (WHO 2006b: 256). Human cases of avian influenza tend to cluster among relatively impoverished people, mostly rural farm families, in countries with developing economies in Southeast Asia (WHO 2007). Important social and economic factors, such as subsistence farming and poverty—that contribute to human vulnerability to all infectious diseases—are receiving little attention in the face of H5N1 viral panic in the West (Lockerbie and Herring in press). To whom is avian influenza actually “emerging” (Farmer 1999)? Vulnerability to Stigma (Shame and Blame) Since the 1930s, all serious outbreaks of influenza have developed in Southeast Asia (Scholtissek 1992). The focus of blame for avian influenza, therefore, has centered on Asian countries, the geographical epicenter identified for most new variants of influenza (Scholtissek 1994) and, so far, the region hardest hit by H5N1 in poultry and humans (WHO 2007). It has been suggested that aquaculture, a common form of agriculture in this region, favors cross-species exchange of influenza genes. Aquaculture brings ducks, pigs, and humans together in close contact. Specific receptors in the pig’s throat allow both bird and mammalian influenza viruses to enter pig cells, intermingle, swap genes, and generate new variants of influenza virus (Ito et al. 1998). Swine therefore can act as “mixing vessels” for influenza strains, resulting in novel trans-specific strains (Scholtissek 1992, 1994; Hollenbeck 2006). Implicated as the origin of new influenza viruses, Asian agricultural and health practices have consequently received extensive attention and censure. In discussing communities afflicted with avian influenza, images are offered up of filth and backwardness (lack of modernity), subsistence farmers living in close proximity to animals, densely packed open markets, and poverty (Figure 3.1).

84

•

Ann Herring

Figure 3.1. “Live-poultry markets like this one in Hanoi speed the spread of the virus from farm to farm when vendors take leftover birds back home, along with any flu viruses they’ve picked up” (Appenzeller 2005:13).

Such images have a rapid and global stigmatizing impact (Lee et al. 2005: 2044). The message is that small-scale farming, aquaculture, and open-air markets common in Asia are dangerous to global health. By implication, Asian subsistence farmers and market vendors are not good citizens of the world and are threats to world health. Yet it is the large-scale, international poultry industry that creates conditions that favor the emergence of new avian influenzas, not the small-scale poultry farmers typically depicted in media and scientific accounts (GRAIN 2006). The virus spreads slowly among small village chicken flocks and has difficulty persisting under such lowdensity conditions; in contrast, it spreads and amplifies quickly in densely packed factory farms. Integrated trade networks offer efficient routes for the spread of infection; in most cases, trade has been the agent of viral diffusion (Butler 2006). The international trade in day-old chicks, eggs, live birds, meat, and secondary products, such as chicken manure, feathers, and animal feed, create the circumstances in which avian influenza can spread globally. In Laos, for example, 90 percent of chicken production comes from small-farm and backyard operations, yet the only outbreaks of H5N1 on these farms have come from those next to the country’s small number of factory farms (GRAIN 2006: 9). Countries with avian influenza lose international markets for agricultural products and risk global censure, such as China faced in the wake of SARS (Washer 2004:

Viral Panic, Vulnerability, and the Next Pandemic

•

85

2565). Yet destroying infected poultry flocks literally wipes out the livelihoods and food security of small-scale farmers. By 2006, over 150 million birds had been culled in Asia, and avian influenza is estimated to have cost Asian economies in excess of $15 billion dollars (Anand 2005–6: 18). And while the focus of attention remains on a Southeast Asian epicenter for “the coming plague,” this was certainly not the source for the 1918 pandemic (Herring and Padiak 2008). The current debate about the 1918 pandemic locates its probable epicenter either in the USA (Barry 2004a, b; Burnet and Clark 1942; Crosby 1989; Jordan 1927) or in Western Europe (Oxford et al. 2002, 2005). A Syndemic Approach to Vulnerability To recapitulate, current panic about the next pandemic focuses on the avian influenza virus, H5N1, whose origins are Asian. The mutability of influenza strains, their shifting antigenic coats, ability to infect human and animal species, to evolve and spread rapidly across boundaries, are elements of the “mutation-contagion package” of fear (Ungar 1998). Anchored to the 1918 influenza pandemic, H5N1 avian influenza contributes to the foreboding that a global cataclysm of unmatched dimensions lies just beyond the horizon. In invoking the 1918 pandemic as the model for the coming pandemic, it is evident that some features of that outbreak have been stressed, such as the deaths of fifty to a hundred million people, while others, such as the extensive variation in death tolls, have received less attention. In presenting the 1918 outbreak in this way, what is highlighted and what is obscured? The wider terrain and the social context of the period also warrant careful scrutiny (Farmer 1999). To this end, the concept of syndemic provides a useful framework for exploring vulnerabilities to pandemics, past and present. A syndemic is a set of interactive and mutually enhancing epidemics involving disease interactions at the biological level that develop and are sustained in a community or population because of harmful social conditions and injurious social connections (Singer and Clair 2003: 429). The utility of the concept is wellillustrated by the example of a whooping cough epidemic in 1927 in the Canadian north (see Appendix 2 to this chapter). Let us consider the first facet of the syndemics concept: mutually enhancing epidemics involving disease interactions. Studies of the effects of the reconstructed H1N1 virus in Macaca fascicularis suggest that the 1918 virus provoked a severe respiratory infection and aberrant expression of the immune response that may help to explain its unusual virulence (Kobasa et al. 2007). That said, scholars have known since the 1918 outbreak that the majority of influenza sufferers recovered within about a week. About 20 percent developed severe secondary infections that gave rise to fatal pneumonia, sometimes within twenty-four hours. The deadly complication of influenza pneumonia killed 40 to 50 percent of people with secondary infections (Burnet and Clark 1942: 88).

86

•

Ann Herring

Clearly, it is not sufficient to focus solely on the H1N1 virus; secondary infections played a significant role in the virulence of the 1918 pandemic (Crosby 1989; Kilbourne 2006). Tubercular infection, for example, may have contributed to its lethality. In the USA, tuberculosis mortality peaked along with influenza in 1918 (Figures 3.2 and 3.3). Excess mortality associated with the 1918 pandemic, it is argued, reflects interaction between the pathogens associated with two co-occurring epidemics: influenza and tuberculosis (Noymer and Garenne 2000, 2003; Noymer 2006). In other words, having tuberculosis increased the chances of dying from influenza. Analysis of Union Army veteran data reveals a statistically significant association between having tuberculosis and dying from influenza in 1918, as well as during interpandemic years, with tubercular individuals being four times more likely to die from influenza than those free of the disease (Noymer 2006). This selective mortality effect had long-term consequences for national patterns of mortality in the USA, resulting in a dramatic reduction in tuberculosis mortality in the aftermath of the pandemic (Noymer and Garenne 2000, 2003; Noymer 2006). In contrast, deaths from tuberculosis decreased in England and Wales during 1918–19 (Johnson 2003: 137). From these examples, it is evident that the expression of the 1918 pandemic differed, depending on local infectious disease ecologies and histories.

Figure 3.2. Influenza death rates under age 45 (without infants), 1911–1945, United States (Herring et al. 2006, drawn from data in Grove and Hetzel 1968; Linder and Grove 1947).

Viral Panic, Vulnerability, and the Next Pandemic

•

87

Figure 3.3. Tuberculosis death rates under age 45, 1900–1960, United States (Herring et al. 2006, drawn from data in Grove and Hetzel 1968; Linder and Grove 1947).

Let us now turn to the second facet of the syndemics concept: the maintenance of interacting epidemics because of harmful social conditions. The point is especially interesting in the context of the 1918 outbreak. This is because emerging diseases are usually understood to be “democratic” in the sense that everyone is theoretically vulnerable because no one has antigens that confer resistance to the new pathogen. In other words, there are not supposed to be health inequalities in the face of a newly emerging disease. The 1918 influenza pandemic, however, was anything but democratic. It took a disproportionate toll among young adults, pregnant women, tubercular individuals, immigrant and economically disadvantaged neighborhoods, and marginalized communities that lacked access to health care (Johnson 2003; Jones 2005; Lux 1997; Mamelund 2006; Noymer 2006; Noymer and Garenne 2000, 2003; Taubenberger and Morens 2006). Some communities escaped infection altogether; others in the same region were devastated by it (Herring 1994; Herring and Sattenspiel 2003; Herring and Sattenspiel 2007). Recent recalculations of mortality on national and continental scales reveal how variable the death toll from the 1918 pandemic actually was (Johnson and Mueller 2002). African nations, for instance, show a range extending from 10.7 per

88

•

Ann Herring

1,000 (Egypt) to an extraordinarily high or 445 per 1,000 (Cameroon), with most national estimates hovering between 20 and 50 per 1,000. Rates were lower in the Americas, ranging from 1.2 per 1,000 (Argentina) to 39.2 per 1,000 (Guatemala); the estimated rates for Canada and the US are about 6 per 1,000. In Europe, influenza mortality rates were even lower, ranging from 2.4 per 1,000 (Russia) to 12.7 per 1,000 (Hungary). The constellation of biosocial conditions that contributed to this diversity has barely been explored and warrants close scrutiny, as the implications are important for future pandemics. Implications and Conclusions A syndemic approach—which considers biological synergies and their connections to harmful social circumstances—is a useful way to begin a discussion of inequalities in the experience of the 1918 pandemic, both locally and globally (Singer and Clair 2003). To develop local profiles of vulnerability, careful analysis of disease interactions and their distribution within and between socioeconomic groups needs to be conducted using historical mortality series for 1918 and beyond. The long-term impact of the pandemic on morbidity and mortality has scarcely been assessed beyond the suppression of life expectancy at birth that resulted from the deaths of so many young people. There are barely any national histories or systematic analyses of its connection to social conditions during World War I (Phillips 2004: 130–31). Until such studies are undertaken, the incorrect, stereotypical view of the H1N1 strain of influenza as a universal and relentless killer will continue to be communicated to the public. This is not just a historical problem; it has important implications for public health policy. The “Spanish Flu” is the model against which catastrophic pandemics are compared; it is the “mother of all pandemics” (Taubenberger and Morens 2006). Fears about avian influenza have been linked to it in the scientific literature and media reports and it is a key element of popular narratives about “the next pandemic.” Anchored to the 1918 pandemic, and in the absence of analysis of the social, economic, and political circumstances that determine the virus’s distribution within and between societies, H5N1 will continue to generate viral panic about an inevitable global pandemic that threatens everyone. Viral panic will be fuelled as long as the complexities of the virus’s interactions with other pathogens, and their links to underlying social inequalities, remain superfluous and unexplored in comparison to the allure of the microbe itself. Failure to explore who is likely to be at greatest risk of acquiring and dying from avian influenza is dangerous from a public health policy perspective. Appendix 1 Was the 1918 Pandemic Caused by a Bird Flu Virus? The development of sensitive PCR techniques, coupled with the successful search for tissue samples from individuals who died from influenza during 1918, allowed the

Viral Panic, Vulnerability, and the Next Pandemic

•

89

genome of the virus to be studied from a molecular perspective. A multidisciplinary team headed by Jeffrey Taubenberger of the Armed Forces Institute of Pathology (AFIP) located preserved tissue samples from autopsied individuals who died during the 1918 epidemic: a 21-year-old soldier from Fort Jackson, South Carolina, and a 30-year-old man stationed at Camp Upton, New York. Both men had died of influenza on September 26, 1918 (Taubenberger et al. 1997). The AFIP team later obtained biopsied lung tissue taken from the frozen body of an Inuit woman buried at the Teller Mission on the Seward Peninsula in Alaska. Using the three sets of tissue samples, the team eventually sequenced the entire gene for hemagglutinin (H1), the surface antigen that allows influenza viruses to infect cells. The initial results seemed to support the idea that the closest relative of the 1918 sequences is the oldest classical swine flu strain, characterized as influenza A/Sw/Iowa/30 (Reid et al. 1999). In the autumn of 2005, the team announced that it had completely decoded the 1918 influenza genome (H1N1). The new results contradicted their earlier interpretation. Rather than the product of reassortment with swine influenza, the new molecular research indicated that the 1918 virus had an almost entirely avian genome (Taubenberger et al. 2005). The results suggested that H1N1 most likely jumped from birds to humans shortly before the pandemic (Taubenberger and Morens 2006), a time frame that may have been as long as several years. Publication of the findings electrified scientists and the popular press alike, fanning viral panic. Yet the plausibility of the avian-origins hypothesis of H1N1 in 1918 has been challenged on a number of fronts, before and after publication of the genome results. Hollenbeck (2005: 89), for instance, stresses the role of pigs as intermediate hosts necessary to convert avian strains to human strains, the lack of evidence that H5N1 has adapted to humans, and the rarity of avian influenza prior to the 1997 H5N1 outbreak in Hong Kong. Gibbs and Gibbs (2006) contend that errors were made in interpreting the virus’s phylogenetic relationships, arguing that the 1918 virus is closer to mammalian than to avian viruses. Instead of leaping to humans shortly before the pandemic, they counter that it may have evolved in pigs or people for an unknown period of time prior to the pandemic. Antonovics and colleagues (2006) also disagree with the proposed avian derivation for the 1918 outbreak. Chastising the team for inflaming the public, they say: “This alarming implication, which is based on misinterpretation of the phylogenetic data, is completely unjustified and could seriously distort the public perception of disease risk, with grave economic and social consequences” (p. E9). Taubenberger and colleagues (2006: E10) responded that they “never maintained that the virus entered the human population in 1918…[and that] phylogenetic analysis on its own cannot definitively resolve this issue.” Evidently, there is still much to be learned about the origins of the 1918 influenza virus.

90

•

Ann Herring

Appendix 2 Applying the Syndemic Approach: Whooping Cough at York Factory. A seemingly obscure epidemic of whooping cough in 1927 at York Factory, Manitoba, Canada, illustrates the utility of a syndemic approach (Singer and Clair 2003). The severity of this particular epidemic cannot be understood as a singular event in isolation from concurrent and preceding epidemics, nor without placing these interacting epidemics within the context of deleterious social conditions characteristic of the place at that time (Herring and Young 2005). York Factory is located on a flat, marshy peninsula on the western shore of Hudson Bay near the mouth of the Hayes River. Established in 1714 as a fur trade post, it was the main port of entry for European trade goods to western Canada and quickly became the Hudson’s Bay Company’s (HBC’s) single most important trading post on the bay (Beardy and Coutts 1996). Cree and Assiniboine middlemen were the lynchpins of the business, acquiring furs from a far-flung network of groups in the interior plains and woodlands (Ray 1974:72). Some settled semi-permanently in the immediate vicinity of the post at York Factory. The Home Guard Cree, as they came to be called, trapped, hunted and fished for the company and were the backbone of its success. Over the centuries of its operation, York Factory boomed to prosperity along with the fur trade and then declined during the course of the nineteenth century. Game was depleted in the Northwest, and declining fur harvests prompted the Hudson’s Bay Company to close many of its trading posts. Places like York Factory were basically “trapped out” by the mid nineteenth century and their economies were failing. The disease ecology changed over the course of its history, in concert with the westward expansion of the American frontier, growth of urban disease pools, and improved transportation efficiency, which allowed diseases with short periods of infectivity to spread more easily into the Canadian north (Hackett 1991, 2002). Tuberculosis had become a major health problem, both as a specific cause of death and as an underlying condition that reduced resistance to other infectious diseases (Stone 1925: 79). The soaring tuberculosis problem resulted in a tuberculosis death rate among Indians in the western provinces of Canada that was ten to twenty times higher than that for non-Aboriginal people (Stewart 1936: 675). It is against this backdrop of declining economic and health conditions that a virulent epidemic of whooping cough struck York Factory in the autumn of 1927. Whooping cough is the common name for pertussis, which means “violent cough.” Its name is derived from the diagnostic “whoop” cough: a high-pitched intake of air followed by rapid, consecutive coughs. This almost unmistakable symptom allows pertussis to be diagnosed with relative accuracy in nonmedical settings. It is a strictly human infection, primarily affecting children under the age of 6, and most often caused by the bacterium Bordatella pertussis (Cherry 1999). About 50 percent of cases occur in children under the age of 2, and most deaths occur among infants

Viral Panic, Vulnerability, and the Next Pandemic

•

91

(under the age of 1). A highly contagious infection easily spread within households, whooping cough is acquired by droplet infection through close contact with sufferers, often adults. The 1927 whooping cough epidemic at York Factory was exceptionally severe. Approximately 40 percent of all deaths recorded in the parish registers that year were attributed to it (ACCA 1864–1929). About 35 five percent of the whooping cough deaths occurred among infants—where most deaths are expected to occur— but it is unusual to see older children succumbing to the disease in such large numbers. The mortality rate for children under 6 years of age was an astonishing 237 per 1,000. When the deaths of two teenage girls are taken into account, whooping cough mortality under the age of 15 reached 157 per 1,000. Why was the epidemic so lethal? A syndemic perspective requires close consideration of other epidemics afflicting the community, whether they could have been acting synergistically with whooping cough, and whether co-occurring epidemics can be tied to deleterious social conditions that allowed each to flourish and capacitate the other. A closer look at 1927 shows that it was a terrible year at York Factory. In addition to the autumn whooping cough epidemic, the community had been devastated by an influenza epidemic the previous February. In fact, about 32 percent of the recorded deaths in 1927 were attributed to this deadly outbreak. Entries in the York Factory Post Journal indicate how overwhelming the epidemic was: “We are having difficulty to get men and dogs. Most of the men are sick with Flu” (HCBA 1794–1939, Feb. 21, 1927). “The flu epidemic … [is] very vicious … everyone laid up” (HBCA 1794–1939, Feb. 26, 1927). Influenza took its greatest toll among adults; 87 percent of the deaths occurred in people between the ages of 21 and 65, the age group most productive in fur trade activities. With the whole community laid up during a crucial time in the annual fur harvest, the outbreak not only debilitated the people but undermined the local economy that year “Disgusted with [the fur] trade. This has been the poorest spring trade for many years” (HBCA 1794–1939, June 13, 1927). To recapitulate, in 1927 a severe influenza epidemic killed mostly adults and led to a poor fur harvest, followed by a fall whooping cough epidemic that killed mostly children. All of this occurred against a backdrop of endemic tuberculosis. Were the epidemics intertwined? There is every possibility that they were. All three are respiratory diseases that affect the lungs. As an underlying condition, tuberculosis opens up already compromised immune systems and debilitated lungs to other respiratory infections (Noymer 2006), which would have made tubercular members of the community more vulnerable to influenza and whooping cough. Active tuberculosis, moreover, can exacerbate influenza infection, making it worse. Influenza, in turn, enhances bacterial lung disease, impairs normal recover mechanisms, and impairs the immune system (Couch 1981). The virus has a lethal synergy with pneumococcus bacteria when infection with influenza precedes pneumococcal infection (McCullers and Rehg 2002). Severe influenza pneumonia in humans, more-

92

•

Ann Herring

over, is frequently caused by combined viral-bacterial infections (Scheiblauer et al. 1992). Influenza infection can also provoke latent tuberculosis and chronic nephritis to erupt into active cases (Couch 1981). Consequently, when the whooping cough epidemic broke out in September, its effects would have been increased by preexisting tubercular disease and by the devastating influenza epidemic from which the community was still recovering. Whooping cough affects the respiratory tract, destroys cells in the respiratory lining, and makes it necessary to cough to remove mucous from the lungs. This may have provoked existing lung disease in the form of tuberculosis. Adults already suffering with tuberculosis, in turn, may have been more likely to experience whooping cough. At the very least, adults infected with Bordatella pertussis would have infected susceptible children with whom they were in contact. In other words, we are most likely seeing endemic and epidemic diseases interacting synergistically, thereby magnifying the effects of each and increasing the community’s disease burden. But synergies become syndemics when they are underlain by harmful social conditions and injurious social connections. What deleterious social conditions existed in 1927? York Factory was a dying community. It had lost its strategic importance in the international trade network as the northern sea route from Europe declined and as trade with the US increased, prompting a shift in trade toward the new steamship and railway routes in the south. The surrounding region had never been rich in game and small mammals. After over two hundred years of harvesting, York Factory’s fur-bearing mammal resource base was severely depleted. To make matters worse, competition from non-Aboriginal trappers and non-HBC outfits was on the rise. Fluctuating fur prices, periodic supply shortages, and over hunting produced environmental degradation. Medical parties in the 1930s and 1940s identified worrying levels of malnutrition in many parts of the Canadian north, including York Factory (Herring et al. 2003; Herring and Sattenspiel 2007). By the 1920s, the lack of a sustainable economy and the difficulties in living off the land accelerated out-migration to more prosperous places in the south. Erstwhile center of the North American fur trade, York Factory was being abandoned in the early twentieth century as residents migrated to more prosperous locations further south (Beardy and Coutts 1996). A syndemic perspective on the 1927 whooping cough outbreak makes it possible to see how intertwined epidemics of respiratory disease—tuberculosis, influenza, and whooping cough—were the biological expression and emblem of declining health conditions and growing impoverishment at York Factory in the early twentieth century (Herring and Young 2005).

References ACCA (Anglican Church of Canada Archives, General Synod Office, Toronto, Canada). 1864–1929. York Factory Burials. Diocese of Keewatin Records, Ms. 217, Reel #2.

Viral Panic, Vulnerability, and the Next Pandemic

•

93

Anand, N.N. 2005–6. The economic impact of a flu pandemic. Lab Business, Winter 2005– 2006: 18–21. Antonovics, J., M.E. Hood, and C.H. Baker. 2006. Was the 1918 flu avian in origin? Nature 440: E9. Appenzeller, T. 2005. Tracking the next killer flu. National Geographic October: 2–31. Armelagos, G. J. 1998. The viral superhighway. The Sciences 38: 24–30. Barker, T. 2005. The avian flu and the end of the world. The Manitoban Online 93 no. 10, October 26. Barrett, R., C.W. Kuzawa, T. McDade, and G.J. Armelagos. 1998. Emerging and re-emerging infectious diseases: the third epidemiologic transition. Annual Reviews of Anthropology 27: 247–71. Barry, J.M. 2004a. The Great Influenza: the Epic Story of the Deadliest Plague in History. New York: Viking. ———. 2004b. The site of origin of the 1918 influenza pandemic and its public health implications. Journal of Translational Medicine 2(1): 3. Basu, P. 2006. Panic over bird flu premature, experts say. Nature Medicine 12(3): 258. Beardy, E. and R. Coutts. 1996. Voices from Hudson Bay: Cree Stories from York Factory. Montreal: McGill-Queen’s University Press. Burnet, F.M. and E. Clark. 1942. Influenza. London: Macmillan. Butler, D. 2006. Doubts hang over source of bird flu spread. Nature 439: 772. Butler, D. and J. Ruttiman. 2006. Avian flu and the New World. Nature 441(11): 137–39. Cartwright, F.F. 1983. Pandemics past and future. In Disease in Ancient Man, ed. Gerald D. Hart. Toronto: Clarke Irwin. Cherry, J.D. 1999. Pertussis in the preantibiotic and prevaccine era, with emphasis on adult pertussis. Clinical Infectious Diseases 28 Suppl 2: S107–11. Cohn, S.K., Jr. 2002. The Black Death Transformed: Disease and Culture in Early Renaissance Europe. London: Arnold. Couch, R.B. 1981. The effects of influenza on host defences. Journal of Infectious Diseases 144(3): 284–91. Crosby, A.W. 1989. America’s Forgotten Pandemic: The Influenza of 1918. Cambridge: Cambridge University Press. Dull, H.B. and W.R. Dowdle. 1980. Influenza. In Maxcy-Rosenau Preventive Medicine and Public Health, ed. J. Last. New York: Appleton-Century-Crofts. Farmer, P. 1999. Infections and Inequalities: The Modern Plagues. Berkeley: University of California Press. Fox, Daniel M. and Elizabeth Fee. 1988. AIDS: The Burdens of History. Berkeley: University of California Press. Garrett, L. 2005. The next pandemic? Foreign Affairs 84(4): 3–22. Gibbs, M.J. and A.J. Gibbs. 2006. Was the 1918 pandemic caused by a bird flu? Nature 440: E8. GRAIN. 2006. Fowl Play: The Poultry Industry’s Central Role in the Bird Flu Crisis. GRAIN Briefing, February 2006. [Online] Available: http://www.grain.org/briefings [15 March 2006]. Grmek, M.D. 1990. History of AIDS: Emergence and Origin of a Modern Pandemic. Princeton: Princeton University Press. Grove, R.D. and A.M. Hetzel. 1968. Vital Statistics Rates in the United States 1940–1960. Washington, D.C.: U.S. Government Printing Office.

94

•

Ann Herring

Hackett, F.P. 1991. The 1819–20 Measles Epidemic: Its Origin, Diffusion and Mortality Effects upon the Indians of the Petit Nord. MA thesis, University of Manitoba, Winnipeg. ———. 2002. A Very Remarkable Sickness: Epidemics in the Petit Nord, 1670 to 1846. Winnipeg: University of Manitoba Press. HBCA (Hudson’s Bay Company Archives, Winnipeg, Canada). 1794–1939. York Factory Post Journal, B.239/a/186–96. Herring, D.A. 1994. ‘There were young people and old people and babies dying every week’: the 1918–1919 influenza pandemic at Norway House. Ethnohistory 41: 73–105. Herring, D.A., S. Abonyi, and R.D. Hoppa. 2003. Malnutrition among northern peoples of Canada in the 1940s: an ecological and economic disaster. In Human Biologists in the Archives: Demography, Health, Nutrition and Genetics in Historical Populations, eds. D.A. Herring and A.C. Swedlund. Cambridge: Cambridge University Press. Herring, D.A. and R.D. Hoppa. 1999. Endemic tuberculosis among nineteenth century Cree in the central Canadian Subarctic. Perspectives in Human Biology 4(1): 189–99. Herring, D.A. and J. Padiak. 2008. The geographical epicentre of the 1918 influenza pandemic. In Proceedings of the Eighth Annual Conference of the British Association for Biological Anthropology and Osteoarchaeology, eds. M. Brickley and M. Smith. BAR International Series 1743. Oxford: Archaeopress. Herring, D.A. and L. Sattenspiel. 2003. Death in winter: Spanish flu in the Canadian subarctic. In The Spanish Influenza Pandemic of 1918–19: New Perspectives, eds. H. Phillips and D. Killingray. London and New York: Routledge. Herring, D.A. and L. Sattenspiel. 2007. Social contexts, syndemics, and infectious disease in northern Aboriginal populations. American Journal of Human Biology 19(2): 190–202. Herring, D.A. and A.C. Swedlund, eds. 2003. Human Biologists in the Archives: Demography, Health, Nutrition and Genetics in Historical Populations. Cambridge: Cambridge University Press. Herring, D.A. and C. Young. 2005. A syndemic perspective on respiratory disease epidemics in the central Canadian subarctic. American Journal of Human Biology 17(2): 243. Hollenbeck, J.E. 2005. An avian connection as a catalyst to the 1918–1919 influenza pandemic. International Journal of Medical Science 2(2): 87–90. Ito, T., J.N.S.S. Couceiro, S. Kelm, L. G. Baum, et al. 1998. Molecular basis for the generation in pigs of Influenza A viruses with pandemic potential. Journal of Virology (September): 7367–73. Joffe, H. 1999. Risk and The Other. Cambridge: Cambridge University Press. Jordan, E.O. 1927. Epidemic Influenza: A Survey. Chicago: American Medical Association. Johnson, N.P.A.S. and J. Mueller. 2002. Updating the accounts: global mortality of the 1918– 1920 Spanish influenza pandemic. Bulletin of the History of Medicine 76: 105–15. Jones, E.W. 2005. Co-operation in all human endeavour: Quarantine and immigrant disease vectors in the 1918–1919 influenza pandemic in Winnipeg. Canadian Bulletin of Medical History (CBMH/BCHM) 22(1): 57–82. Kaplan, M.M. and R.G. Webster. 1977. The epidemiology of influenza. Scientific American 237(6): 88–106. Kilbourne, E.D. 2006. Influenza pandemics of the 20th century. Emerging Infectious Diseases 12 (1): 9–14. Klempner, M.S. and D.S. Shapiro. 2005. Crossing the species barrier—one small step to man, one giant leap to mankind. New England Journal of Medicine 350: 1171–72.

Viral Panic, Vulnerability, and the Next Pandemic

•

95

Kobasa, D., S.M. Jones, K. Shinya, J.C. Kash, et al. 2007. Aberrant innate immune response in lethal infection of macaques with the 1918 influenza virus. Nature 445(18): 319–23. Last, J. 1999. Fouling and cleansing our nest: human-induced ecological determinants of infectious disease. Perspectives in Human Biology 4(1): 145–57. Lederberg, J. 1988. Pandemic as a natural evolutionary phenomenon. Social Research 55(3): 344–59. Lee, S., L.Y.Y. Chan, A.M.Y. Chau, K.P.S. Kwok, and A. Kleinman. 2005. The experience of SARS-related stigma at Amoy Gardens. Social Science & Medicine 61: 2038–46. Li, K.S., Y. Guan, J. Wang, et al. 2004. Genesis of a highly pathogenic and potentially pandemic H5N1 influenza virus in eastern Asia. Nature 430: 209–13. Linder, F.E. and R.D. Grove. 1947. Vital Statistics Rates in the United States 1900–1940. Washington, D.C.: U.S. Government Printing Office. Lockerbie, S. and D.A. Herring. In press. Global panic, local repercussions: economic and nutritional effects of bird flu in Vietnam. In Anthropology and Public Health: Bridging Differences in Culture and Society, 2nd ed., eds. R. Hahn and M. Inhorn. Oxford: Oxford University Press. Lux, M. 1997. The Bitter Flats: The 1918 influenza epidemic in Saskatchewan. Saskatchewan History 49(1): 3–13. MacDougall, H. 1985. The fatal flu. Horizon Canada 8(8): 2089–95. Mamelund, S-V. 2006. A socially neutral disease? Individual social class, household wealth and mortality from Spanish influenza in two socially contrasting parishes in Kristiania 1918–19. Social Science & Medicine 62: 923–40. McCullers, J.A. and J.E. Rehg. 2002. Lethal synergism between influenza virus and Streptococcus pneumoniae: characterization of a mouse model and the role of platelet-activating factor receptor. Journal of Infectious Diseases 186: 341–50. Monto, A.S. 2005. The threat of an avian influenza pandemic. New England Journal of Medicine 352(4): 323–25. Morse, S.S. 1993. Examining the origins of emerging viruses. In Emerging Viruses, ed. S.J. Morse. Oxford: Oxford University Press. Noymer, A. 2006. Testing the influenza tuberculosis selective mortality hypothesis with Union Army data. [Online] Available: http//:www.paa2006.princeton.edu/abstractViewer.aspx ?submission Id=61257. [20 May, 2006]. Noymer, A. and M. Garenne. 2000. The 1918 influenza epidemic’s effects on sex differentials in mortality in the United States. Population and Development Review 26(3): 565–81. ———. 2003. Long-term effects of the 1918 ‘Spanish’ influenza epidemic on sex differentials of mortality in the USA. In The Spanish Influenza Pandemic of 1918–19: New Perspectives, eds. H. Phillips and D. Killingray. London and New York: Routledge. Omran, A.R. 1971. The epidemiologic transition: a theory of the epidemiology of population change. Millbank Memorial Fund Quarterly 42(4): 509–37. Osterholm, M.T. 2005. Preparing for the next pandemic. Foreign Affairs 84(4): 24–37. Oxford, J.S., A. Sefton, R. Jackson, W. Innes, et al. 2002. World War I may have allowed the emergence of Spanish influenza. Lancet—Infectious Disease 2(2): 111–14. Oxford, J.S., R. Lambkin, A. Sefton, R. Daniels, et al. 2005. A hypothesis: the conjunction of soldiers, gas, pigs, ducks, geese and horses in Northern France during the Great War provided the conditions for the emergence of the Spanish influenza pandemic of 1918–1919. Vaccine 23: 940–45.

96

•

Ann Herring

Palese, P. 1993. Evolution of influenza and RNA viruses. In Emerging Viruses, ed. S.J. Morse. Oxford: Oxford University Press. Phillips, H. 2004. The re-appearing shadow of 1918: trends in the historiography of the 1918– 19 influenza pandemic. Canadian Bulletin of Medical History (CBMH/BCMH) 21(1): 121–34. Ray, A.J. 1974. Indians in the Fur Trade: Their Role as Hunters, Trappers and Middlemen in the Lands Southwest of Hudson Bay, 1660–1870. Toronto: University of Toronto Press. Reid, A.H., T.G. Fanning, J.V. Hultin, and J.K. Taubenberger. 1999. Origin and evolution of the 1918 Spanish influenza virus hemagglutinin gene. Proceedings of the National Academy of Science 96: 1651–56. Rosenberg, C. 1992. Explaining Epidemics. Cambridge: Cambridge University Press. Rosenthal, E. 2006. Human-to-human infection of bird flu virus is confirmed. New York Times, June 24: A8. Scheiblauer, H., M. Reinacher, M. Tashiro, and R. Rott. 1992. Interactions between bacteria and influenza A virus in the development of influenza pneumonia. Journal of Infectious Diseases 166: 783–91. Scholtissek, C. 1992. Cultivating a killer virus. Natural History 1: 3–6. ———. 1994. Source for influenza pandemics. European Journal of Epidemiology 4(10): 455–58. Shinya, K., M. Ebina, S. Yamada, M. Ono, et al. 2006. Avian flu: influenza virus receptors in the human airway. Nature 440: 435–36. Singer, M. and S. Clair. 2003. Syndemics and public health: reconceptualizing disease in biosocial context. Medical Anthropology Quarterly 17(4): 423–41. Snacken, R., A.P. Kendal, L.R. Haaheim, and J.M. Wood. 1999. The next influenza pandemic: lessons from Hong Kong 1997. Emerging Infectious Diseases 5(2) [Online]. Available: http://www.cdc.gov/pub/EID/vol4no3/ascii/webs.txt [7 March 1999]. Stewart, D.A. 1936. The Red Man and the White Plague. Canadian Medical Association Journal 35: 674–76. Stöhr, K. 2005. Avian influenza and pandemics—research needs and opportunities. New England Journal of Medicine 352(4): 405–7. Stone, E.L. 1926. Health and disease at the Norway House Indian Agency. Reprinted in Native Studies Review 5(1), 1989: 237–56. Taubenberger, J.K. and D.M. Morens. 2006. 1918 influenza: the mother of all pandemics. Emerging Infectious Diseases 12(1): 15–22. Taubenberger, J.K., A.H. Reid, A.E. Krafft, K.E. Bijwaard, and T.G. Fanning. 1997. Initial genetic characterization of the 1918 Spanish influenza virus. Science 275(5307): 1793–96. Taubenberger, J.K., A.H. Reid, R.M. Lourens, R. Wang, et al. 2005. Characterization of the 1918 influenza virus polymerase genes. Nature 437(7060): 889–93. ———. 2006. Taubenberger et al. Reply. Nature 440: E9–10. Tomes, N. 2000. The making of a germ panic, then and now. American Journal of Public Health 90(2): 191–98. Ungar, S. 1998. Hot crises and media reassurance: a comparison of emerging diseases and Ebola Zaire. The British Journal of Sociology 49(1): 36–56. U.S. Department of Health and Human Services. Pandemics and pandemic threats since 1900. [Online]. Available: http:/www.PandemicFlu.gov [23 May 2006]. Wade, N. 2006. Studies suggest avian flu pandemic isn’t imminent. New York Times: 23 March.

Viral Panic, Vulnerability, and the Next Pandemic

•

97

Walters, J.H. 1978. Influenza 1918: the contemporary perspective. New York Academy of Medicine 54(9): 855–64. Waltner-Toews, D. 1995. Changing patterns of communicable disease—who’s turning the kaleidoscope? Perspectives in Biology & Medicine 39: 43–55. Washer, P. 2004. Representations of SARS in the British newspapers. Social Science & Medicine 59: 2561–71. ———. 2006 Representations of mad cow disease. Social Science & Medicine 62: 456–66. Webster, R.G. 1993. Influenza. In Emerging Viruses, ed. S.J. Morse. Oxford: Oxford University Press. ———. 1998. Influenza: an emerging disease. Emerging Infectious Diseases 4(3): 436–41. World Health Organization (WHO). 2006a. Cumulative number of confirmed human cases of avian influenza A/(H5N1) reported to WHO. [Online]. Available: http://www.who .int/csr/disease/avian_influenza/country/cases_table_2006_02_02/en/index.html [1 February 2006]. ———. 2006b. Epidemiology of WHO-confirmed human cases of avian influenza A (H5N1) infection. Weekly Epidemiological Record 26, 81: 249–60. ———. 2008a. Areas reporting confirmed occurrence of H5N1 avian influenza in poultry and wild birds since 2003. [Online]. Available: http://gamapserver.who.int/mapLibrary/ Files/Maps/Global_SubNat_H5N1inAnimalConfirmedCUMULATIVE_20080303 .png [7 March 2008] ———. 2008b. Cumulative number of confirmed human cases of avian influenza A/(H5N1) Reported to WHO. [Online]. Available: http://www.who.int/csr/disease/avian_influenza/ country/cases_table_2008_02_20/en/index.html [28 February 2008].

PART II

• • • Generational and Developmental Change

• • • Thinking about Health through Time and across Generations Darna L. Dufour

Introduction The three chapters in this section focus on issues of health and risk through time and across generations. Ellison and Jasienska provide a theoretically oriented chapter that uses issues related to reproductive health as examples. Núñez-de-la-Mora and Bentley present a case study of migration and changes in risk factors for breast cancer in two generations of migrants. Sellen presents a review of a select number of cases of child growth in sub-Saharan Africa and an analysis of child growth as a measure of health. All three chapters provide important insights on how long-term health outcomes are shaped by life histories and responses to environmental challenges. Adaptation, Health, and the Temporal Domain Ellison and Jasienska challenge us to think about two very familiar concepts, adaptation and health, and the relation between them. Their intent is well taken because “health” is not a concept we have spent much time discussing, and yet is one that frames so much of what we do. The authors begin with adaptation. Adaptation is a big concept that refers to the seemingly simple idea of the “fit” between organisms and their environment. Who would argue that organisms don’t fit into their environment? However intuitively appealing and central the concept, adaptation has always been difficult to define operationally and to measure empirically. As the authors note, the term adaptation is

102

•

Darna L. Dufour

used in two different ways. It refers to both the process of achieving a fit between organisms and their environment, and to the end product—that is, an adapted organism—or a trait that contributes to the state of adaptation of that organism. It is the former, more dynamic sense of process in which Ellison and Jasienska are interested. Specifically, they suggest that the way to understand the process of adaptation is to focus on the responses (behavioral, physiological, and genetic) of organisms to their environments over time. These responses are examples of phenotypic plasticity, that is, the different phenotypes an organism can exhibit under different environmental conditions. As they note, the entire range of these plastic responses is referred to as the norm of reaction, and it is this norm of reaction that is the target of natural selection. West-Eberhard (1989, 2003) provides excellent treatments of phenotypic plasticity within evolutionary biology. How is adaptation related to health? Ellison and Jasienska propose a conceptualization of health that contrasts with the common understanding of health in terms of dichotomous states (health versus illness) at any given point of the lifespan. They argue that health is best viewed as one measure of the success of the organism’s responses to environmental challenges. The plastic responses that we would call adaptive—that is, those that enhance the “fit” between organism and environment— would result in a better state of health. Because adaptation is a process that occurs through time, health as a measure of success needs to be evaluated over the long term, even over the course of the lifespan. Ellison and Jasienska admit that it would be a daunting task to evaluate the lifelong history of responsiveness to environmental challenges, especially given the inherent problems of evaluating risks, benefits, and trade-offs using the same metric. They do, however, provide a framework that could serve as a checklist of things to consider in developing the kind of integrated measure of health they propose. To illustrate their ideas, Ellison and Jasienska focus on questions of reproductive health. One example they provide is that of amenorrhea, and specifically the amenorrhea associated with athletic training. In medicine, amenorrhea is considered an abnormal condition, and one that should be corrected. They argue, however, that if we assume that ovarian function is an aspect of the phenotype that is responsive to the environment, then the amenorrhea associated with athletic training can be considered a plastic response in ovarian function associated with high energy expenditure, and not as pathology. However, as they also note, to consider the amenorrhea an adaptive response, and hence a “healthy response,” one would have to be able to show that it resulted in overall benefit to fitness. This could be done by demonstrating that the amenorrhea served as a mechanism to avoid wasted reproductive effort and thus contributed to long-term fitness; it was not simply a response to the constraint of limited energy availability. This is difficult to show. So, as the example of amenorrhea reveals, it is easier to demonstrate plastic responses than to determine their adaptive value (Rose and Lauder 1996; Schlichting and Pigliucci 1998), and hence whether or not they could be considered “healthy.”

Thinking about Health through Time and across Generations

•

103

The concept of accommodation is useful here because it provides another way of evaluating plastic responses. Accommodation refers those plastic responses to environmental stress that are not wholly successful because, while enhancing survival, they result in a significant loss of function (Frisancho 1996). Hence, accommodations are not considered “healthy” per se, for function is compromised; they are workable but not wholly beneficial kinds of responses. The term accommodation is used within the theoretical framework of human adaptability that is concerned with understanding how humans adapt to environmental stresses. This is an area of research where the outcome variable is usually adequacy of function or indicators like growth and work capacity, not reproductive fitness. However, function and other proximal indicators like growth are assumed to be related to long-term fitness (Huss-Ashmore 2000). Changes in Risk Factors for Breast Cancer in Migrant Women Núñez-de la Mora and Bentley report on a case study of risk factors for breast cancer in migrants from Bangladesh to the United Kingdom. They note that South Asians in general have a lower incidence of breast cancer than Europeans, but that the incidence among South Asian migrants living in the UK is moving toward that of the host population and away from the low rates on the Indian subcontinent. They focus on factors known to increase lifetime exposure to ovarian steroids (female reproductive hormones) because such exposure is associated with increased risk of developing breast cancer. These factors are considered risk factors, that is, characteristics of individuals that increase the chance of developing a particular disease, in this case breast cancer. Risk factors are disease specific, but not necessarily causal. For example, aging is a risk factor for breast cancer, but chronological age per se does not cause breast cancer. The migration study that Núñez-de la Mora and Bentley review is one with an unusually strong design. The migrants originated from a known village in Bangladesh and all live in the same London community, but they migrated at different stages of development (some as children, and some as adults) and at different periods of time (some a generation earlier than others). Hence, the study was able to consider changes in risk factors associated with age of migration (development effects) as well as changes between generations. They found that migration was associated with an increase in certain risk factors for breast cancer: a reduction in the proportion of infants breast-fed, a reduction in the duration of breast-feeding and an increase in obesity. The first two are risk factors because they increase lifetime exposure to ovarian steroids. Obesity is a risk factor because it results in higher levels of circulating estrogen, one kind of steroid. The researchers also found two additional risk factors in second generation migrants: an increase in mother’s age at birth of her first child, and a shift away from the traditional Bangladeshi diet toward one common in the UK. The latter is considered a risk factor because the traditional Bangladeshi diet is assumed to be relatively protective against breast cancer, and that of the UK is not.

104

•

Darna L. Dufour

In sum, migrants had more risk factors for breast cancer than sedentees (women who remained in Bangladesh), and the number of risk factors was greater in the second generation of migrants. These finding are in general agreement with other studies that have demonstrated changes in risk factor profiles with migration. Family Structure and Child Growth in sub-Saharan Africa The chapter by Daniel Sellen reviews studies from sub-Saharan Africa indicating that the type of marriage the mother is in (monogamous versus polygamous) influences the health of her children. In three of the four studies he reviews, child growth was poorer (or mortality higher) in polygamous versus monogamous families (but in one case, only for male children). In the fourth study, mortality was also higher in polygamous than monogamous families, and, interestingly, the growth of living children was poorest when the mother was the second co-wife in a polygamous union. In several of the cases, the effect of the mother’s marital status was more pronounced under conditions of greater poverty or seasonal deficit. Sellen is not able to explain the processes that would account for these observations because the necessary research has not been done, but the pattern he identifies suggests that the mother’s martial status is indeed a factor in child health. In examining these case studies, Sellen uses two measures of health, namely, child mortality and child growth. To use growth as a measure of health Sellen relies on well-established associations between growth and both morbidity and mortality: children who are small for their age exhibit higher rates of illness and death than do children of normal size. As used here, “small for age” and “size” refer to children’s height-for-age and/or weight-for-age in comparison to international reference values. For example, a child will be classified as small for age, or growth retarded, if his or her height-for-age (or weight-for-age) is less than or equal to a z-score of –2 (two standard deviations below the median for moderate growth retardation) (WHO 1986). Growth retardation is another example of what is referred to as accommodation. Sellen emphasizes that growth is an indicator of a child’s past responses to the environment. This is because growth is a cumulative process of adding tissue, and accumulation occurs faster in high-quality environments (environments more optimal from the point of view of nutrition, disease, and psychosocial interactions) and slower in low-quality ones. Hence, the growth achieved, at any one point in time, is an indicator, albeit a nonspecific one, of the overall quality of the environments the child has experienced and been responding to. Taking this idea one step further, Sellen refers to poor growth as a marker of the adversity of past environments. Discussion My brief summaries here emphasize themes running through the three chapters. One of these themes is the idea of viewing health over the lifespan rather than as a con-

Thinking about Health through Time and across Generations

•

105

dition of the organism at any given moment in time. Ellison and Jasienska conceptualize health as a measure of the success of responses to environmental challenges, and since these responses continue throughout the life history of the organism, the authors stress the necessity to think about them occurring along a temporal continuum. They argue for the need to develop an understanding of health that integrates the history of responses to environmental challenges. Núñez-de la Mora and Bentley describe selected features of the behavior and biology of migrants in the present that affect lifetime levels of reproductive hormones and hence risk of developing breast cancer in the future. Sellen focuses on child growth as a measure of health, on the impact of one generation (parental) upon the health of the next generation, and on the value of assessing cultural systems (marriage) and their consequences on child development. He conceptualizes growth as an integrated measure of health, one that sums responses to the environment over the course of the lifespan. He notes that growth is a good example of the kind of integrated measures of health that Ellison and Jasienska argue for. Risk is the second major theme. In general, risk refers to the chance of injury, damage or loss. In relation to health, risk refers to the probability of, or vulnerability to, a specific disease or illness (Walker et al. 2003). Risk is about expectations, but it is not predictive (Krieger 2001). That is, not all women with risk factors for breast cancer will actually end up with breast cancer, and not all children with poor growth will fall ill or die within the next year. What biological anthropology brings to the table in the discussion of risk is the clear understanding that current health risks are contingent on past responses to environmental challenges, as well as the concept of trade-offs over a lifetime. This idea is most clearly expressed in the chapter by Ellison and Jasienska. They note, for example, that high levels of female reproductive hormones over the course of a lifetime are protective against osteoporosis but, at the same time, are a risk factor for breast cancer. The three chapters share a common approach, which is the idea of looking at responses of humans to environmental variables. This is, of course, an approach grounded in evolutionary theory and the goal of understanding how humans “adapt” to their environments. But exactly what people are responding to in the local environment is not always clear. Consider the migrant study. Núñez-de la Mora and Bentley describe changes in the reproductive behavior of women and their body composition (obesity) and, based on the rich ethnographic information they gathered, a suite of associated behaviors, such as increases in the length of schooling. If we consider all of these as responses to the environment, and I think we can, the interesting question then becomes, what are the relevant features in the new environment that the migrants are responding to? Are they responding to a position of social and economic disadvantage, local economic opportunities, the status associated with higher education, the availability of infant formula, the abundance of low-cost food, the availability of public transportation, or the good taste of fish and chips? All of the above? None of the above? Our understanding of the relevant features of the new UK

106

•

Darna L. Dufour

“environment” is fuzzy. And with good reason—the environment has highly complex and multifaceted sociocultural dimensions. Nonetheless, a logical next step in pursuing research on migrants could be to build on the ethnographic insight gained in the first study to better define the environmental factors that are the most relevant. Sellen’s paper on family structure and child growth also describes responses, namely, the growth of children, and specifically the poorer growth of some children in polygamous households. Exactly what factors in the environment to which the poor growth of children is a response is unclear. Again, a logical next step in the research would be to better define those factors. In general, we think of poor growth as a response to environments in which food resources are inadequate and/or disease prevalent, or both, since nutrition and disease act synergistically (Waterlow 1994). Food availability and disease are certainly important, but they are rather general features of the environment. We would need a more detailed understanding of both in this specific case. And, if we think about the environment of the young child, the mother predominates, and hence the care the child receives from its mother is probably the most biologically relevant feature of the child’s environment. Focusing on the mother brings us back to Sellen’s point that further investigation is needed to understand the maternal effects on child growth and survival and how these are affected by relatively subtle distinctions in social status. I am using here the word environment in the broad sense of all the surroundings and conditions of the group or population, including physical environment and social and cultural context. How we conceptualize the environment of the study group is important. Not all features of the environment are relevant or significant. Determining which features are most relevant and significant for any particular question for any given human group demands a biocultural approach and the kind of detailed fieldwork that anthropologists can do so well (Dufour 2006). To better define the features of the environment associated with specific health outcomes is an important contribution that anthropology can make to research on health. There are important differences between the chapters, as well as similarities. One of the most interesting differences is how the authors approach the concept of health. Ellison and Jasienska are explicitly concerned with how we conceptualize health, and argue that we need to systematically think through what we mean by it. They question the usefulness of the common dichotomy of health versus disease. They argue that, like adaptation, health can be assessed only over the long term, when all the costs and benefits have been added up. Núñez-de la Mora and Bentley are not concerned with defining health. They use the term in the conventional way. Their focus is on risk factors for a specific type of ill health, breast cancer, and they are more concerned with risk of future disease than with issues of health per se. They are less explicit than Ellison and Jasienska about the issue of trade-offs, such as, for example, the fact that although the same behaviors may increase risk of breast cancer they also reduce risk of osteoporosis. Sellen also focuses on one particular kind of health, that assessed by child growth, using standard thresholds to distinguish normal from poor growth. He

Thinking about Health through Time and across Generations

•

107

considers health at only one point in the life span, childhood. So while he agrees with Ellison and Jasienska in conceptualizing health as an integrative function of responses to past environments, he takes the practical stance that growth can be dichotomized into good versus poor, something Ellison and Jasienska argue against doing. Taken together, the three chapters in this section are a stimulating combination of theory and empirical research. They broaden our understanding of the temporal dimensions of health, deepen our understanding of the social factors that underpin the biology we measure, and suggest directions for future research. Little more could could be asked by us as readers. References Dufour, D.L. 2006. The 23rd annual Raymond Pearl Memorial Lecture. Biocultural Approaches in Human Biology. American Journal of Human Biology 18(1): 1–9. Frisancho, A.R.1996. Human Adaptation and Accommodation. Ann Arbor: University of Michigan Press. West-Eberhard, M.J. 1989. Phenotypic plasticity and the origins of diversity. Annual Review of Ecology and Systematics 20: 249–78. ———. 2003. Developmental Plasticity and Evolution. New York: Oxford University Press Huss-Ashmore, R. 2000. Theory in human biology: evolution, ecology, adaptability, and variation. In Human Biology: An Evolutionary and Biocultural Perspective, eds. S. Stinson, B. Bogin, R. Huss-Ashmore, and D. O’Rouke. New York: Wiley-Liss. Krieger, N. 2001. Theories for social epidemiology in the 21st century: an ecosocial perspective. International Journal of Epidemiology 30: 668–77. Rose, M.R. and G.V. Lauder. 1996. Post-spandrel adaptationism. In Adaptation, eds. M.R. Rose and G.V. Lauder. New York: Academic Press. Schlichting, C.D. and M. Pigliucci. 1998. Phenotypic Evolution: A Reaction Norm Perspective. Sunderland, MA: Sinauer Associates. Walker, E.A, C.K. Mertz, M.R. Kalten, and J. Flynn. 2003. Risk perception for developing diabetes. Diabetes Care 26: 2543–48. Waterlow, J.C. 1994. Summary of causes and mechanisms of linear growth retardation. European Journal of Clinical Nutrition 48 (supplement 1): S211–12. World Health Organization (WHO). 1986. Use and interpretation of anthropometric indicators of nutritional status. Bulletin of the World Health Organization 64: 929–41.

•4• Adaptation, Health, and the Temporal Domain of Human Reproductive Physiology Peter T. Ellison and Grazyna Jasienska

Introduction Two of the most familiar concepts in human biology are also two of the most problematic: health and adaptation. Health is such a familiar concept that it often goes undefined, even in the context of research, or the definition is at best implicit. Often health is treated as a dichotomous variable, the complement of disease. At other times health is treated as if it were a continuous, or at least a graded variable, allowing for states of “better” or “worse” health. The concept of adaptation has been difficult ever since Darwin’s day (Browne 2002). Theoretical definitions that avoid tautology are hard to come up with, and empirical guidelines for recognizing adaptation and distinguishing it from constraint, neutrality, or even pathology are also difficult to formulate. An important area of overlap between the domains of evolutionary anthropology, human biology, and public health concerns the relationship between adaptation and health. This relationship is often paradoxical or ambiguous. But we cannot expect to fully understand the relationship between these two core concepts if the concepts themselves are not firmly in our grasp. This chapter is devoted to a consideration of the concepts of health and adaptation and the relationship between them, particularly as they are applied to human reproduction. Our suggestion is that we approach a better understanding of both concepts and their interrelationship when we view both as features of human life history with important temporal dimensions. This is perhaps a familiar suggestion in

Human Reproductive Physiology

•

109

itself. But by adopting it we can steer clear of some of the most common obstacles to understanding. Health What do we mean when we say that a person is “healthy?” What do we strive for when we strive to promote human “health?” Rather than being defined by a set of characteristics, health is most often understood as the absence of certain characteristics, that is, the absence of disease, trauma, or pathology. This dichotomous worldview, wherein organisms are viewed as being either in a state of health or unhealth, underlies much of clinical medicine, from the importance of diagnosis (the discrimination of pathology from health) and the reliance on diagnostic criteria (empirical tests or test values that make such discrimination), to an understanding of the physician’s mission as one of treating the sick, injured, or pathological individual in an attempt to restore health. Even preventative medicine, as expressed in its name, is directed at warding off pathology in order to preserve health. In this dichotomous worldview, then, health is a state that an organism may enjoy, a state defined by the absence of pathology, a state that clinical medicine seeks to preserve when it exists and to restore when it is lost. From this perspective, physiological principles such as “homeostasis” take on an almost archetypal significance, reflecting the organism’s own internal mechanisms for preserving correct functioning and coping with destabilizing influences from the outside. At times, however, this conception of health leads to paradox or ambiguity. A good diagnostic test, for example, resolves a given population into a bimodal distribution, one mode representing those with, the other those without the disease. The sharper the bimodality and the less the overlap in the two component distributions, the better the test. Sometimes, however, this approach fails. Luteal phase deficiency in women probably represents such a condition. Luteal phase deficiency refers to a condition in which a woman ovulates but in which the corpus luteum created by ovulation fails to produce enough of the hormone progesterone to support a successful pregnancy. In some women this is a recurring condition and a contributing factor in infertility (Daly 1991). Many clinicians have sought a diagnostic test that would identify this condition with low ambiguity, a test that would resolve the population of menstruating women into two subpopulations, those with and without luteal phase deficiency. All efforts have failed, however (McNeely and Soules 1988; Castelbaum et al. 1994). Rather than some index of luteal progesterone production revealing a bimodal distribution of “healthy” and “diseased” women, all indices reveal a single, unimodal distribution with lower values more likely to be associated with the clinical condition but no clear cut-off value that represents a dramatic change in that probability. No clear boundary exists between health and pathology in this case. Similar situations are surprisingly common, particularly in the diagnosis of chronic diseases such as hypertension or type 2 diabetes.

110

•

Peter T. Ellison and Grazyna Jasienska

To avoid difficulties of this kind we may be tempted to adopt a different concept of health, one that is more quantitative rather than qualitative, one that allows for speaking of the relative health of individuals. When we adopt this perspective, we shift away from thinking in terms of dichotomous diagnostic criteria to quantitative measures of relative risk. Higher blood pressure is associated, for example, with a greater risk of stroke, slow glucose clearance from the blood with greater risk of diabetes, and lower luteal progesterone levels with greater risk of luteal phase deficiency. This perspective is characteristic of epidemiology and public health, where a statistical approach to the world is more familiar. The mission of practitioners from this perspective can also be expressed quantitatively as seeking to improve health and reduce the burden of disease. In part this quantitative, statistical concept of health comes from adopting populations as the unit of analysis rather than individuals, so that “greater health” may simply mean “a greater proportion of individuals in the dichotomous state of health.” When we apply the quantitative concept of health to individuals, however, we encounter new difficulties. How do we measure relative health? Is it measured by probability of sickness or death? By functional capacity (e.g., strength, flexibility, endurance, etc.)? Is a trained athlete healthier than a sedentary person? Is a person with a family history of heart disease less healthy than a person without such a history? Most importantly, by adopting a relative scale of health we do not change the notion of health as a single entity, something that intervention seeks to maximize, both for individuals and for populations. This notion breaks down in the face of ineluctable trade-offs between alternative disease states or risks. If the same practices that lower the risk of osteoporosis also raise the risk of breast cancer, are they healthy or not (see Appendix for this chapter)? Adaptation For evolutionary biologists no concept is more familiar than adaptation. Yet Ernst Mayer could write in 1983 that, “The difficulty of the concept of adaptation is best documented by the incessant efforts to analyze it, describe it, and define it.” Darwin used the term adaptation, as his predecessors had, to refer to the remarkable “fit” between the characteristics of an organism and the exigencies of its environment. As such, adaptation was widely recognized, even by that name, before Darwin, even by those whose writings provided the starkest foil to the Origin of Species, such as William Paley in Natural Theology (1803). Darwin’s great contribution was to argue for natural selection as the process responsible for producing adaptation in the organic world. But as Darwin’s thesis gained acceptance, the definition of adaptation began to change into “the result of natural selection,” introducing the tautology that critics have delighted in pointing out since the nineteenth century. Modern evolutionary biologists tend to adopt this sort of definition, as exemplified by Reeve and Sherman

Human Reproductive Physiology

•

111

(1993): “An adaptation is a phenotypic trait that results in the highest fitness among a specified set of variants in a given environment.” But if adaptation shares the same definition as natural selection (i.e., differential fitness or reproductive success among alternative traits), then it is tautologous to credit natural selection with the generation of adaptation. Even if one adopts this standard, tautologous definition of adaptation, difficulties arise when environments are variable. Traits that arose through natural selection during formative human prehistory may no longer be under positive selection today or may even be under negative selection. That is, traits may have been produced by natural selection and yet not currently be adaptations in the original sense of the word. When the timescale of environmental change is slow this may not pose a great conceptual challenge. We can be comfortable with the notion that climate change may place an organism under new circumstances that its physiology may not “fit” particularly well and yet still view that physiology as a product of natural selection. But when the timescale of change is rapid, the problem can be more acute. Over how many generations should we count descendants (or genes) to determine relative fitness? Does a trait’s claim to being an adaptation change as its relative fitness waxes and wanes across succeeding generations? What about traits that have higher mean fitness than alternative traits but also higher fitness variance, resulting in a greater probability of extinction (Ellison 1994)? Are they adaptive? Aside from the theoretical difficulties, there are significant practical difficulties in operationalizing the concept of adaptation. How do we recognize an adaptation? Must we count offspring and grand-offspring? Not only is this often impractical, it runs afoul of the tautologous logic referred to above. That is, if we operationalize adaptation by measuring relative reproductive success we cannot use it to test whether relative reproductive success produces adaptation. The alternative most often suggested is to recognize elegance of functional design as the hallmark of adaptation. As George C. Williams wrote in Adaptation and Natural Selection (1966), “to prove adaptation one must demonstrate a functional design”; and later, “I will rely on informal arguments as to whether a presumed function is served with sufficient precision, economy, efficiency, etc. to rule out pure chance as an adequate explanation.” This shifts the problem to identification of functional design, something that is also often difficult in practical terms. Gould and Lewontin (1979) famously note the difficulty that can be encountered in distinguishing function from constraint. At times this ambiguity can even extend to encompass outright pathology. If, for example, human ovarian function declines when a woman is under moderate energetic stress, is this an adaptation (e.g., a mechanism designed to avoid wasted reproductive effort), a constraint (e.g., a corollary of down-regulated metabolism in general that is of no particular fitness benefit in itself ), or a pathology (e.g., the failure of physiological homeostasis under stress, ultimately reducing fitness and possibly predisposing to other poor health outcomes)?

112

•

Peter T. Ellison and Grazyna Jasienska

Fitness, Not Health, Is the Legacy of Natural Selection The interaction between health and adaptation is an additional source of conceptual difficulty, at least for some. The recent development of Darwinian (or evolutionary) medicine (Nesse and Williams 1996; Trevethan et al. 1999; Stearns 1999) represents an attempt to explicitly address the failure of clinical approaches to human health to incorporate an understanding of adaptation. Perhaps the core concept in Darwinian medicine is the realization that natural selection optimizes fitness, not health. When we forget this principle, we are often surprised by the way the human organism responds to interventions designed to promote health. Two examples can be cited that demonstrate this seeming paradox with particular reference to human reproduction. In the 1970s an effort was launched to improve infant and child mortality in the Gambia by increasing birth weight and infant weight gain. To achieve this, a program of maternal nutritional supplementation with culturally acceptable, locally produced foods was introduced during both pregnancy and lactation (Prentice et al. 1980). By increasing maternal energy intake the investigators hoped to increase the energy status of both gestating fetuses and nursing infants. They were surprised when the impact of this program on birth weight, milk production, and infant weight gain turned out to be negligible (Prentice et al. 1980, 1983a, 1983b, 1983c, 1987). A significant impact was observed, however, on the time to resumption of menstruation and the interval to the next pregnancy in the supplemented mothers, which became significantly shorter (Lunn 1980, 1981, 1984). More recently, another effort to reduce child mortality and improve child growth and nutritional status was launched in Ethiopia (Gibson and Mace 2006). In this case, the intervention involved the drilling of local wells to reduce the time women spent carrying water and to improve the quality of water available to children. Although infant mortality did decline as a result of this effort, so did birth intervals. More surprisingly, the average nutritional status of children actually declined, perhaps because of increased competition for food within larger families, so that the average health of children after the intervention has to be judged as worse than before. What these examples demonstrate is the fact that humans, like other organisms, have been shaped by natural selection to allocate energy and other resources in ways that optimize reproductive success, a result that may not be the same as optimizing health. The ecological situation in both the Gambia and Ethiopia originally acted to constrain the allocation of energy to reproduction as well to compromise the health and growth of children. When public health interventions relieved the energetic constraints, women’s physiology allocated the newly available energy in ways that increased fecundity, not in ways that improved child health. The lesson from these studies should be kept in mind when considering the relationship between health and fitness. When pathological conditions are life-threatening, it may well be that the most important route to improving fitness for an organism is through improving its own survivorship. But in less dire circumstances trade-offs may be in play between growth, survivorship, and reproduction and their joint im-

Human Reproductive Physiology

•

113

pacts on fitness. The result may be organismal responses to intervention that seem “unhealthy” or otherwise counterintuitive. Many of those trade-offs involve a temporal dimension: current survival versus future survival, current reproduction versus future reproduction, current reproduction versus future survival, current survival versus future growth, and so on (Kozlowski and Wiegert 1986; Perrin and Sibly 1993). In order to remain mindful of these trade-offs and the complexity they introduce, we need to adopt explicitly diachronic perspectives on both adaptation and health. The Temporal Domain of Adaptation and Health Many, if not most, phenotypic traits of organisms are not static but change with time. An organism’s state of health also changes with time in a way that integrates past experience. Diachronic perspectives on adaptation and health focus our attention on the responses that an organism makes to environmental challenges rather than on fixed traits or states. They also lead to a more integrative picture of the entire life of the organism. Environments can change in ways that vary greatly on a temporal scale, from rapid, short term fluctuations, to slow, progressive changes, to punctuated events of a certain frequency and duration (Figure 4.1). The temporal domain for organismal responses to environmental challenges also varies, extending from seconds to generations. At the very short end (seconds) are the rapidly implemented, readily reversible responses governed by the central nervous system. Both conscious and unconscious behavioral adjustments are appropriate responses to environmental challenges that are themselves rapidly transient. At the other extreme, environmental challenges that persist over generations can provoke genetic change in populations through natural selection. In between these extremes, however, is a broad domain governed by physiological changes ranging from rapid and reversible changes in the dynamic state of an organism to slow, often cumulative, changes in the structure and function of an organism mediated by development or even transgenerational processes. In the dynamic perspective, adaptation is not a property of a given phenotype, but more like a function relating phenotype to environment, a function that governs

Figure 4.1. The temporal continuum of adaptive response. Temporal domains of environmental challenge (above) are aligned with temporal domains of organismal responses (below).

114

•

Peter T. Ellison and Grazyna Jasienska

the way the organism responds to environmental challenges that may occur all along the temporal continuum. It is this “response function,” sometimes referred to as a “norm of reaction,” that is the target of natural selection, adapting an organism not to a static environment but to a shifting environment with certain temporal characteristics. Characteristic anatomical features of an organism that are a common focus in standard discussions of adaptation (opposable thumbs, cursorial limbs, etc.) simply represent long-term, genetic responses to chronic features of an organism’s environment. But adaptation does not stop at this level. Beneath the relatively calm and fixed anatomical surface lies a turbulent sea of physiological adaptive response (e.g., changes in vasoconstriction and dilation in response to muscular activity or thermal conditions, changes in hemoglobin levels in response to atmospheric oxygen tension, changes in leukocyte populations in response to viral or bacterial challenge). The pattern of organismal response to environmental challenges itself may change as the organism ages, not as a result of senescent decay in physiological capacity but rather as a result of the differing trade-offs that characterize different stages of development. The trade-offs between survival and growth that a child faces differ from the tradeoffs between survival and reproduction that an adult faces, and may thus result in different responses to disease under nutritional constraint. A dynamic view of adaptation leads naturally to a more dynamic view of health. Health can be considered one measure of the success of an organism’s responses to environmental challenges, recognizing that “success” must be measured over the long term, not just at the current moment. It may be a “healthy” response to constrained energy availability to down-regulate metabolism even at the cost of impaired immune function or compromised physical capacity. Selective hypertrophy of certain muscle groups may be a “healthy” response to frequent demands that approach or exceed the previous capacity of those muscles to respond even if the energy reserves that buffer against malnutrition must be sacrificed. It may be “healthy” to increase fecundity during prime reproductive ages in response to favorable energetic conditions even if the risk of breast cancer later in life is simultaneously increased. Alternatively, it may be “unhealthy” not to respond to chronic energy abundance by reducing energy storage, or to overrespond to the presence of transient allergens. In order to determine the healthiness of an organism and its responses to environmental challenges we need somehow to evaluate the consequence of the entire trajectory of responsiveness that it traces through its lifetime, comparing the integrated value of whatever index of health we decide on (e.g., functionality in daily life, longevity) to the outcome of alternative trajectories. An added challenge lies in the problem of weighting health at different points in time by calculating an integrated survivorship/quality of life value. How much value should be ascribed to higher fecundity or greater strength in young adulthood as opposed to higher mortality risk or greater joint pain in later life? Health economists face a similar problem in trying to establish criteria for the allocation of economic resources in health care. A widely used approach to solving the

Human Reproductive Physiology

•

115

problem is the QALY model of “quality-adjusted life years” that combines life expectancy with indices of life quality, usually based on things like competency in daily life tasks (Cubban 1991; Dolan 2005). Optimizing the allocation of economic resources, it is argued, can be determined in terms of optimizing “quality-adjusted life years” for a given population. Of course for economists, and perhaps for individuals as well, the optimization only applies to the future and is focused on assessing the quality of life at the end of life. To properly conceptualize health in diachronic terms we would need to acknowledge and evaluate trade-offs that have already occurred as well as to consider something like “quality of life” across the life span. This makes the task more difficult, both because past trade-offs are hard to evaluate and because the standards for life quality change with age. Creating an integrated measure of health, however defined, across the lifespan is a significant challenge that has yet to be successfully met. Indeed, we would argue that it is not even well recognized as a challenge by health researchers. In contrast, evolutionary biologists have long recognized the importance of lifespan measures of fitness and adaptation (Charlesworth 1994). Life history theory, an important subfield of theoretical biology, is entirely concerned with this concept, its operationalization, and the testing of its predictions (Stearns 1992; Roff 2001). This fact simply underscores the degree to which trade-offs and the temporal dimension of biological responsiveness has been incorporated into evolutionary thinking as opposed to health research. The temporal domain of health and the trade-offs it implies have featured prominently, however, in recent discussions of early origins of adult disease. Certainly, the recognition of developmental effects on health is not new. Many congenital diseases with origins in fetal life are well known, such as cretinism resulting from maternal iodine deficiency. But beginning with the work of Barker and his colleagues (1989), new attention has been paid to correlations between neonatal characteristics and diseases appearing much later in life, diseases that have been customarily considered chronic or degenerative diseases of aging and not the consequences of physiological responses, or the lack of them, to challenges during fetal life. Gluckman and Hanson (2005), among others, have noted that fetal origins of adult disease may represent trade-offs along the temporal domain of health corresponding to the temporal domain of adaptive response. The fetus may be responding to intrauterine conditions in ways that determine subsequent trajectories of physiological responsiveness. These responses may have been subject to natural selection during human evolution in ways that optimized the reproductive success of individuals in formative environments. The fact that these fetal responses lead to disease in adult life may either be a consequence of recent rapid environmental change, so that signals that used to be reliable cues to the conditions expected in adulthood are no longer reliable, or a consequence of trade-offs between health earlier and later in life. Finch and Crimmins (2004; Crimmins and Finch 2006) have proposed another dynamic view of the correlation between conditions early in life and health outcomes late in life. Studying longitudinal demographic data, they note that age-specific mor-

116

•

Peter T. Ellison and Grazyna Jasienska

tality rates are more highly correlated within cohorts than within periods; that is, mortality rates that are experienced by a group of 70 year olds in a given country are more highly correlated with the mortality rates that their own cohort faced as infants than with the mortality rate among contemporary infants. To account for their observations, Finch and Crimmins advance a hypothesis of the cumulative cost of inflammation, essentially a “wear-and-tear” hypothesis of adult disease, or a hypothesis of “cumulative pathology.” As such, the Crimmins and Finch hypothesis represents a first approximation to a concept of “integrated, lifespan health.” However, this view represents health correlations across the temporal domain as essentially positive: responses that promote health at one age tend to promote health at later ages, and vice versa. The Gluckman and Hanson view (see the chapter by Godfrey and Hanson in this volume), on the other hand, can incorporate negative correlations across the temporal domain: responses that promote health at one age may undermine health at later ages, and vice versa. Both views, however, stress an integrated concept of health based on an integrated concept of physiological responses to environmental challenges (see the Appendix at the end of this chapter). The Temporal Domain of Human Reproductive Function Reproductive ecology is the study of human reproductive physiology as a product of evolution by natural selection (Ellison 1994). In this field, human reproductive function is assumed to be responsive to ecological conditions along the temporal continuum of adaptive response. Variable ovarian function in women has been documented in a wide range of human populations but with certain common characteristics (Ellison et al. 1993). In particular, conditions that limit metabolic energy availability such as food shortages or heavy workloads are associated with suppressed production of ovarian steroids (Ellison et al. 1989; Panter-Brick et al. 1993; Jasienska and Ellison 1998, 2004; Jasienska et al. 2006) (Figures 4.2 and 4.3). Suppression of ovarian function is, in turn, associated with suppression of fecundity, or the likelihood of pregnancy (Lipson and Ellison 1996). Such short-term modulation of ovarian function in response to energetic conditions is hypothesized to represent adaptive modulation of reproductive effort, directing resources toward or away from reproduction in ways that optimize lifetime fitness (Ellison 2001, 2003; Jasienska 2003; Jasienska and Ellison 2004). Because the necessary investment of time and energy in reproduction by women is so much greater than that of men, female reproductive physiology appears to be more sensitive to shorter-term fluctuations in energy availability than does male physiology (Ellison 2001). But as the temporal domain of environmental challenge lengthens, male reproductive physiology also responds (Bribiescas 2006). Down-regulation of reproductive function in males does not appear to involve downregulation of sperm production, which can be maintained with a trivial investment, as much as down-regulation of somatic investment in muscle mass and perhaps behavioral investment in male-male competition (Bribiescas 2001, 2006; Ellison 2001,

Human Reproductive Physiology

•

117

Figure 4.2. Differences in salivary progesterone profiles observed among Lese women in the Ituri Forest of the Congo. Women currently in negative energy balance (losing body weight) due to insufficient energy intake have significantly lower progesterone profiles than women currently in positive energy balance (gaining weight).

2003). Curtailing this investment only makes sense in response to more sustained environmental challenge. The secular trend toward earlier reproductive maturation can be understood as an adaptive response to more sustained or more frequently repeated environmental challenges. The secular trend in menarcheal age in girls is most often attributed to changes in nutrition and energy availability (Frisch 1972; Wyshak and Frisch 1982; Eveleth and Tanner 1991). However, it has also been noted that even tighter correlations exist between menarcheal age and rates of infant mortality a decade or more earlier (Ellison 1981). In either case, it is hypothesized that a more limited investment in childhood growth, either as a result of lower energy intake or higher energy investment in survival, results in later attainment of the appropriate size for the initiation of reproduction (Ellison 1982, 2001). A more controversial hypothesis has also been advanced (Ellison 1990; Lipson 2001), suggesting that the timing of reproductive maturation in women is linked to set-points for subsequent adult reproductive function. A standard principle of life history theory holds that the rate of energy investment in growth should be correlated with the energy investment in reproduction (Charnov and Berrigan 1993). Thus,

118

•

Peter T. Ellison and Grazyna Jasienska

Figure 4.3. Differences in salivary estradiol among Polish women with differing levels of energy expenditure in habitual activity (Jasienska et al. 2006).

slow growth in childhood and late maturation in adolescence should indicate an environment in which lower rates of reproductive investment are optimal. For women, reductions in the rate of reproductive investment are more readily—and likely more successfully—achieved by slowing the pace of reproduction (i.e., lengthening interbirth intervals) than by reducing the investment per birth. Reducing levels of ovarian function to lower, but not eliminate, fecundity can provide a mechanism for lengthening the time to conception, the period of low energetic investment in reproduction that separates consecutive high-investment periods of gestation and lactation. An alternative view, advanced by Vitzthum (2001), holds that female fecundity should not be adjusted to long-term or chronic energy availability, but only to short-term fluctuations about the mean conditions for a given environment. Essentially, this view insists that the temporal domain for reproductive response is limited to horizons of months to years, horizons shorter than the average period of reproductive investment per birth. Empirical resolution of these competing hypotheses will be difficult, given both the longitudinal nature of the data required and the difficulty in empirically determining female fecundity. However, data on hormonal indices of ovarian function, which may also be indices of fecundity (Li et al. 1989; Lipson and Ellison 1996), do appear to show correlations between tempo of adolescent maturation and level of reproductive function in adulthood (Ellison 1996; Lipson 2001).

Human Reproductive Physiology

•

119

Similar patterns of variation in male reproductive function have also been reported, later maturing populations with shorter adult stature having lower levels of testosterone, particularly in young adulthood (Ellison et al. 2002). Analogous arguments to those advanced for females can be made for males if testosterone is interpreted as an index of male somatic, and possibly behavioral, investment in reproductive effort (Bribiescas 2001, 2006; Ellison 2003). Slow growth again signals an environment with limited capacity to support high rates of reproductive investment. Male investment can be effectively reduced by limiting the growth of muscle mass and by curtailing expensive involvement in male-male competition. Lowering testosterone levels may help to achieve such adjustments. Migration studies provide a valuable opportunity to study the interaction of change in environmental conditions and the response of reproductive physiology. The chapter by Núñez-de-la-Mora and Bentley in this volume explicitly addresses such studies. Migration can represent an unusual challenge in evolutionary terms, an abrupt change from one sustained set of environmental conditions to another, potentially a change of considerable magnitude as well. If we think of natural selection as shaping organismal responses to environmental challenges along the temporal continuum, we must acknowledge that abrupt changes between dramatically different chronic environments may rarely have been encountered during formative human evolution. While the responses observed under such situations can provide valuable information about sensitive periods in development, we should be careful about interpreting those responses as adaptive. Recently, research in the fetal origins of adult disease has extended to include effects on adult reproductive function. Perhaps because of the epidemiological orientation of much of this research, most attention to date has focused on correlations between small size at birth and compromised reproductive function in adulthood. De Bruin et al. (1998) noted a correlation between fetal growth retardation and ovarian development in girls, while Ibañez et al. (1998) found that small size at birth was associated with hyperinsulinemia, accelerated pubertal development, and hyperandrogenism in girls. Francois et al. (1997) reported a correlation between small size at birth and low adult male fertility, while Cicognani et al. (2002) report correlations between small size at birth and reduced testicular volume, lower testosterone levels, and higher luteinizing hormone levels in men. We have recently reported a correlation between small size at birth and low estradiol levels in Polish women (Jasienska et al. 2006). It is more difficult to determine whether these associations represent adaptive organismal responses, pathological consequences of disrupted developmental processes, or merely constrained development (Ellison 2005). We have recently argued that an empirical case can be made for small size at birth resulting in an adaptive shift in set-points of ovarian response to energetic stress such that women who are smaller at birth show greater sensitivity to energetic stress in adulthood than women who were larger at birth (Jasienska et al. 2006). An adjustment in the sensitivity of physiologi-

120

•

Peter T. Ellison and Grazyna Jasienska

cal responsiveness is more likely to represent adaptation, we argue, than pathology or constraint. The more difficult problem, however, is to reconcile the temporal domain of the environmental challenge that might result in retarded fetal growth and the temporal domain of responses that seem to have chronic effects on adult function decades later. In theory, shorter-term challenges elicit shorter-term responses and longer-term challenges longer-term responses. Physiological responses that persist over the life course of an individual ordinarily result from challenges over a similar temporal domain. Fetal growth retardation, however, would seem to be a consequence of challenge over a much shorter domain, measured in months. Slow fetal growth might be an adaptive response to limited maternal capacity to meet fetal requirements for energy or other essential nutrients. But persistent effects of a challenge restricted to gestation appears to be a mismatch of temporal domain and thus more likely to represent pathology or constraint than adaptation. Kuzawa (2005; Ellison 2005) has suggested that a closer match of temporal domain might be recognized if the intrauterine environment is not a result of acute conditions faced by the mother, but itself at least partially a congenital legacy of the mother’s own fetal development. If this is the case, then rather than a response to conditions that last only a few months, adjustments in adult physiology that are correlated with fetal growth may be responses to conditions that span generations. These responses would fall on the temporal continuum in the range between more traditional developmental responses and genetic change, the last step in physiological response before full genetic assimilation. Resolving the adaptive status of the adult consequences of fetal growth retardation is one of the more interesting challenges of contemporary human biology. The Temporal Domain of Reproductive Health It is difficult, both conceptually and empirically, to analyze physiological responses that are simultaneously occurring across different temporal domains. Doing so, however, is a key to a fuller understanding of physiological adaptation. It is also a key to a fuller understanding of human health. We should not fall into the trap of equating health and adaptation. Whatever it is we mean by health, it is rarely distilled into a measure of reproductive success. But organismal health is intimately linked to functional responses to environmental challenges. Because those responses are characterized by temporal domain, health must be as well. Immune system responses may occur to cope with transient exposure to foreign organisms or allergens, for example, but they also condition future responses to recurrent exposure. More persistent shifts in leukocyte population composition and T cell subpopulations can result from more persistent exposure to different pathogen/allergen populations (Ellison and Barrett 2004). Reproductive health can also be usefully considered with reference to temporal domain. To the extent that reproductive physiology is responsive to environmental

Human Reproductive Physiology

•

121

challenges across the temporal domain, reproductive capacity may be affected as well. Recognizing these responses as adaptive may influence our understanding of their relationship to health, however. Amenorrhea associated with athletic training in women is still widely considered pathological and treated either with hormonal contraceptives to induce regular menstruation or with advice to reduce the level of training (Chen and Brzyski 1999; Cannavo and Trimarchi 2001). Amenorrhea after childbirth was, in the early twentieth century in the United States, considered to be a consequence of the stress of pregnancy on the uterus, and extended post-partum amenorrhea, in excess of the “normal” two to three months, was considered potentially pathological (Ellison 1995). Both of these responses are now understood as adaptive shifts in ovarian function and fecundity associated with high energy expenditure, either in athletic activity or in milk production (Ellison 2001). The more persistent shifts in set-points for adult gonadal function in both males and females that are associated with more persistent environmental challenges have been discussed above. It is also important, however, to recognize under the term “reproductive health” outcomes that result from responses in reproductive physiology other than those that strictly affect reproductive capacity, such as osteoporosis and cancer. Bone mineralization is strongly affected by estradiol levels in women. Risks of osteoporosis late in life are particularly sensitive to the cumulative effects of estrogen exposure over the life course, including the period of increasing bone mineral density that peaks in early adulthood, and the rate of mineral depletion that occurs during adult life and with increasing speed after menopause. Because the risk of osteoporosis is an integrated function of lifetime estrogen exposure it is difficult to compensate through intervention over a short period for effects cumulated over a long period. Understanding osteoporosis risk as a consequence of physiological responses with a broad temporal domain thus can have relevance for the design of effective treatment and intervention. Similarly, many cancers, including notably breast cancer in women and prostate cancer in men, have risks that are related to both acute and cumulative exposure to gonadal steroids (Ellison 1999; Jasienska et al 2000; Jasienska and Thune 2001; Pike et al 1993). Factors that influence steroid exposure have impacts on the risk of these cancers that vary with temporal domain. Breast cancer risk has also been linked to fetal growth, in this case low birth weight being associated with reduced cancer risk (Michels et al. 1996). Similarly, large size at birth has been associated with an increased risk of metastatic prostate cancer in men (Nilsen et al. 2005). The direction of these correlations is consistent with the direction of the observed effect of fetal growth on adult gonadal function. One aspect of the temporal domain of reproductive health of potential importance remains largely unexplored: the consequence for somatic senescence of the levels of gonadal steroid that characterize adult life. Existing evidence from a number of populations suggests that environments that support vigorous childhood growth and early reproductive maturation are also associated with higher gonadal steroid levels in adulthood among both men and women (Ellison et al. 2002). These pat-

122

•

Peter T. Ellison and Grazyna Jasienska

terns of response and their interpretation have been discussed above. But at older age, post menopause in women and beyond age 45 in men, these population differences largely disappear (Table 4.1). The magnitude of the change in steroid exposure with aging thus varies widely between populations. The magnitude of the change in testosterone levels in men may have direct consequences for body composition and the maintenance of lean body mass. The magnitude of the change in estrogen exposure in women may have direct consequences for menopausal symptoms and the maintenance of bone mineral density. The magnitude of age-related changes in steroid exposure could even have cognitive effects on memory and mood as well as on the risk of degenerative conditions such as Alzheimer’s disease. We become aware of possibilities of this kind only when we consider reproductive health against a broader temporal domain, considering not just acute and cumulative steroid exposure but also amplitude and rate of change across time. Conclusion Temporal domain is a crucial dimension of human biology. It is important, both to our understanding of adaptation and to our understanding of health, to conceive of the phenotype of an organism not as a set of fixed traits, but as a set of responses to environmental challenges, challenges and responses that can be characterized by temporal domain. Natural selection optimizes these responses, ordinarily resulting in a correspondence between the temporal domain of the challenge and the temporal domain of the response. Multiple challenges and responses can occur simultaneously with different temporal domains, making both the physiological task of the organisms and the analytical task of the scientist quite daunting. Nevertheless, many conceptual paradoxes are avoided and many insights gained by making an effort to

Table 4.1. Mean morning salivary testosterone levels in men from four different populations subdivided by age. The populations differ in their average values over all ages and at young ages, but not in the oldest age category (Ellison et al. 2002). All Ages Population

Testosterone SE (pmol/L)

N

15–30 Years Testosterone SE (pmol/L)

30–45 Years N

Testosterone SE (pmol/L)

45–60 Years N

Testosterone SE (pmol/L)

N

USA

259

10 106

335

20 24

288

17

31

238

14

26

Congo

268

12

33

286

15 17

250

19

11

247

34

5

Nepal

240

13

37

251

18 22

224

19

14

225

0

1

Paraguay

192

12

45

197

25 15

187

14

21

192

36

8

p value for one-way ANOVA

0.0002

0.0001

0.0004

0.5143

Human Reproductive Physiology

•

123

adopt this more dynamic view of health and adaptation. Ultimately, we must develop a concept of health that is integrated both across different states and outcomes and across the lifespan. Appendix Building an Integrated Concept of Health Step One: Integrating across Different States of Health and Disease Building an integrated concept of health is more than a theoretical exercise. It can allow for more powerful analysis of medical and public health interventions and their outcomes, helping patients, physicians, and policy makers in the decisions they face. The central problem is finding a common “currency” into which all aspects of health and departures from health can be converted, so that risks, benefits, and trade-offs can be objectively evaluated. When choices must be made between alternatives that are commensurate (measured the same way, or “in the same currency”) there is at least an obvious way to come to an objective decision (i.e., devoid of subjective feelings or emotions). Often we do this by focusing on mortality risk as a common currency. Suppose that removing a gangrenous limb may improve chances of survival from 50 percent to 90 percent. The choice seems clear. But the problem becomes more complicated when the costs and benefits are not so readily commensurate. For many women faced with breast cancer, the choice between alternative treatments presents just such a difficulty. Radical mastectomy might have a greater impact on survival probability in a given case than lumpectomy, but at the cost of loss of the breast and a longer recovery. Drug treatment, chemotherapy, and radiation each may have different effects on survival probability, but they also carry different costs in terms of disability and discomfort. When the differences in survival probability between alternatives are not large, the problem of converting other costs and benefits into a common “currency” in order to compare bottom lines is difficult. More difficult are comparisons in which mortality risk doesn’t figure at all, or not to a significant degree. Suppose that use of a drug—a corticosteroid or betaadrenergic compound, say—can reduce the incidence and severity of allergies but contributes to mild hyperglycemia and hyperactivity. How do we weigh these consequences in evaluating the impact of the drug on health? Physicians are faced with these problems all the time, and pharmaceutical companies lobby them relentlessly on the proper “weightings” to assign to “main effects” and “side effects” of their products. We all face similar decisions, not only in discussions with our doctors, but even in our daily lives. We make them, too, without always being conscious of the “currency” on which our choices are made. How, for example, do we weight the health benefits of exercise or restrained eating against the costs of effort, time, and pleasure? The situation becomes even more complicated when we acknowledge that the conversions involved in rendering different states and outcomes commensurate are

124

• Peter T. Ellison and Grazyna Jasienska

necessarily subjective and can vary dramatically between individuals and groups. The cost of disability or disfigurement may vary depending on a person’s gender, age, or culture, for example, as well as on individual psychology. Cultural differences are particularly important to recognize for those who provide health care in multicultural communities. They also make the establishment of international health policies difficult and at times contentious. As yet there is no clear, agreed upon “common currency” for health, much less an agreed upon “conversion table” for rendering different states and outcomes into that currency. The concept of adaptation has the advantage over the concept of health in this case, since most biologists agree that the various costs and benefits that organisms face can be rendered commensurate by converting them into their respective impacts on inclusive fitness, which has a precise definition. Step Two: Integrating Across the Lifespan How can we derive a concept of health that is integrated across the lifespan of the individual? First, of course, we have to solve the problem presented in Step One and find a way to measure different aspects of health in commensurate terms. That is, we have to find common “units” for combining risks of disease a and disease b into a single quantitative measure of health, h. But suppose we have succeeded in that. Then isn’t an integrated measure of health simply the sum (or, in continuous terms, the integral) of all the time-specific health measures for an individual over his or her life-time, ∑hi = h1 + h2 + h3 . … where the subscripts stand for different “moments” in the individual’s life? Unfortunately, we are faced with another weighting problem. Where in Step One we had to find a way to make different disease states commensurate, now we have to find a way to make health at different points in the life span commensurate. Suppose, using the example of the Ethiopian water project described in the text (Gibson and Mace 2006) that a reduction in infant mortality of 10 percent is linked to an increase in childhood mortality of 10 percent. Do the two exactly balance each other, or does one “count” for more in our integrated measure of health? What if the link is not between infant and childhood mortality but, as implied by the work of Barker et al. (1989) and Gluckman and Hanson (2005) in the text, between infant and middle or old age mortality? If mortality risks at two different ages do not weigh equally, how much of one does it take to equal the other? By focusing simply on mortality risk we may be led to think in terms of lifeexpectancy, arguing that a reduction in mortality in infancy has a larger impact on life expectancy than a similar reduction at old age. Then we might weight health at each age by life expectancy at that age in calculating our integrated measure. This is essentially the approach taken by those health economists who argue for the use of QALYs in health policy analysis (discussed in the text). Essentially, they try to combine mortality risks and non-mortality outcomes into a common currency (“quality”

Human Reproductive Physiology

•

125

of life) and then weight that calculation by life expectancy (“quantity” of life). This approach remains controversial, however, even in the domain of health economics. Once again, the concept of adaptation is more developed in this respect. Life history theory (Stearns 1992; Charlesworth 1994; Roff 2001) is a branch of theoretical evolutionary biology based upon the analysis of trade-offs in fitness (the common currency) at different life stages. Measures such as “reproductive value” (essentially, the expected number of future offspring discounted by the time until they are born) are used to weight impacts on mortality, for example. The problem is much simpler in the context of adaptation, however, since anything that doesn’t affect either the probability of survival or the probability of reproduction can be ignored. References Barker, D.J.P., B.D. Winter, et al. 1989. Weight in infancy and death from ischaemic heart disease. Lancet 2(8663): 381–83. Bribiescas, R.G. 2001. Reproductive ecology and life history of the human male. Yearbook of Physical Anthropology 44: 148–76. Bribiescas, R.G. 2006. Men: The Evolution and Life History of the Human Male. Cambridge, MA: Harvard University Press. Browne, E.J. 2002. Charles Darwin: The Power of Place. New York: Alfred A. Knopf. Cannavo, S. and F. Trimarchi. 2001. Exercise-related female reproductive dysfunction. Journal of Endocrinological Investigation 24: 823–32. Castelbaum, A.J., J. Wheeler, C.B. Coutifaris, L.J. Mastroianni, and B.A. Lessey. 1994. Timing of the endometrial biopsy may be critical for the accurate diagnosis of luteal phase deficiency. Fertility and Sterility 61: 443–47. Charlesworth, B. 1994. Evolution in Age-Structured Populations. 2d edition. Cambridge: Cambridge University Press. Charnov, E.L. and D. Berrigan. 1993. Why do female primates have such long lifespans and so few babies? Or, life in the slow lane. Evolutionary Anthropology 1: 191–94. Chen, E.C. and R.G. Brzyski. 1999. Exercise and reproductive dysfunction. Fertility and Sterility 71: 1–6. Cicognani, A., R. Alessandroni, A. Pasini, P. Rirazzolii, et al. 2002. Low birth weight for gestational age and subsequent male gonadal function. Journal of Pediatrics 141: 376–80. Crimmins, E. and C.E. Finch. 2006. Infection, inflammation, height, and longevity. Proceedings of the National Academy of Sciences 103: 498–503. Cubbon, J. 1991. The principle of QALY maximisation as the basis for allocating health care resources. Journal of Medical Ethics 17:181–84. Daly, D.C. 1991. Current treatment strategies for luteal phase deficiency. Clinical Obstetrics and Gynecology 34: 222–31. de Bruin, J.P., M. Dorland, H.W. Bruinse, W. Spliot, et al. 1998. Fetal growth retardation as a cause of impaired ovarian development. Early Human Development 51: 39–46. Dolan, P., R. Shaw, A. Tsuchiya, and A. Williams. 2005. QALY maximisation and people’s preferences: a methodological review of the literature. Health Economics 14:197–208. Ellison, P.T. 1981. Morbidity, morality, and menarche. Human Biology 53(4): 635–43.

126

•

Peter T. Ellison and Grazyna Jasienska

———. 1982. Skeletal growth, fatness and menarcheal age: a comparison of two hypotheses. Human Biology 54(2): 269–81. ———. 1990. Human ovarian function and reproductive ecology: New hypotheses. American Anthropologist 92: 933–52. ———. 1994. Extinction and descent. Human Nature 5: 155–65. ———. 1994b. Advances in human reproductive ecology. Annual Review of Anthropology 23:255–75. ———. 1995. Breastfeeding, fertility, and maternal condition. In Breastfeeding: Biocultural Perspectives, eds. K.A. Dettwyler and P. Stuart-Macadam. Hawthorne, NY: Aldine de Gruyter. ———. 1996. Developmental influences on adult ovarian function. American Journal of Human Biology 8: 725–34. ———. 1999. Reproductive ecology and reproductive cancers. In Hormones and Human Health, eds. C. Panter-Brick and C. Worthman. Cambridge: Cambridge University Press, 184–200. ———. 2001. On Fertile Ground. Cambridge, MA: Harvard University Press. ———. 2003. Energetics and reproductive effort. American Journal of Human Biology 15(3): 342–51. ———. 2005. Evolutionary perspectives on the fetal origins hypothesis. American Journal of Human Biology 17(1): 113–18. Ellison, P.T. and E.S. Barrett. 2004. Life history perspectives on human disease. In The Changing Face of Disease: Implications for Society, eds. J. Peters, C.G.N. Mascie-Taylor, and S. McGarvey. Boca Raton, LA: CRC Press, 23–39. Ellison, P.T., R.G. Bribiescas, G.R. Bentley, B.C. Campbell, et al. 2002. Population variation in age-related decline in male salivary testosterone. Human Reproduction 17(12): 3251–53. Ellison, P.T., C. Panter-Brick, S.F. Lipson, and M.T. O’Rourke. 1993. The ecological context of human ovarian function. Human Reproduction 8(12): 2248–58. Ellison, P.T., N.R. Peacock, and C. Lager. 1989. Ecology and ovarian function among Lese women of the Ituri Forest, Zaire. American Journal of Physical Anthropology 78(4): 519–26. Eveleth, P.B. and J.M. Tanner. 1991. Worldwide Variation in Human Growth. Cambridge: Cambridge University Press. Finch, C.E. and E. Crimmins. 2004. Inflammatory exposure and historical changes in human life-spans. Science 305: 1736–39. Francois, I., F. de Zegher, C. Speissens, T. D’Hooghe, and D. Vender Schueren. 1997. Low birth weight and subsequent male subfertility. Pediatric Research 42: 899–901. Frisch, R.E. 1972. Weight at menarche: similarity for well-nourished and undernourished girls at different ages, and evidence for historical constancy. Pediatrics 50: 445–50. Gibson, M.A. and R. Mace. 2006. An energy-saving development initiative increases birth rate and childhood malnutrition in rural Ethiopia. PLoS Medicine 3: e87. Gluckman, P. and M. Hanson. 2005. The Fetal Matrix. Cambridge: Cambridge University Press. Gould, S.J. and R.C. Lewontin. 1979. The spandrels of San Marco and the Panglossian paradigm: a critique of the adaptationist programme. Proceedings of the Royal Society of London Series B-Biological Sciences 205: 581–98.

Human Reproductive Physiology

•

127

Ibanez, L., N. Potau, I. Francois, and F. de Zegher F. 1998. Precocious pubarche, hyperinsulinism, and ovarian hyperandrogenism in girls: relation to reduced fetal growth. Journal of Clinical Endocrinology and Metabolism 83: 3558–62. Jasienska, G. 2003. Energy metabolism and the evolution of reproductive suppression in the human female. Acta Biotheoretica 51: 1–18. Jasienska, G. and P.T. Ellison. 1998. Physical work causes suppression of ovarian function in women. Proclamations of the Biological Sciences of London 265: 1847–51. ———. 2004. Energetic factors and seasonal changes in ovarian function in women from rural Poland. American Journal of Human Biology 16: 563–80. Jasienska, G., I. Thune, and P.T. Ellison. 2006. Fatness at birth predicts adult susceptibility to ovarian suppression: an empirical test of the Predictive Adaptive Response hypothesis. Proceedings of the National Academy of Sciences 103:12759–12762. Jasienska, G. and I. Thune. 2001. Lifestyle, hormones, and risk of breast cancer. British Medical Journal 322: 586–87. Jasienska, G., I. Thune, and P.T. Ellison. 2000. Energetic factors, ovarian steroids and the risk of breast cancer. European Journal of Cancer Prevention 9: 231–39. Jasienska, G., A. Ziomkiewicz, S.F. Lipson, I. Thune, and P.T. Ellison. 2006. High ponderal index at birth predicts high estradiol levels in adult women. American Journal of Human Biology 18: 133–40. Jasienska, G., A. Ziomkiewicz, I. Thune, S.F. Lipson, and P.T. Ellison. 2006. Habitual activity and estradiol levels in women of reproductive age. European Journal of Cancer Prevention 15:439–445. Kozlowski, J. and R.G. Wiegert. 1986. Optimal allocation of energy to growth and reproduction. Theoretical Population Biology 29: 16–37. Kuzawa, C.W. 2005. Fetal origins of developmental plasticity: are fetal cues reliable predictors of future nutritional environments? American Journal of Human Biology 17: 5–21. Li, T.C., E.A. Lenton, P. Dockery, A.W. Rogers, and I.D. Cooke. 1989. The relation between daily salivary progesterone profile and endometrial development in the luteal phase of fertile and infertile women. British Journal of Obstetrics and Gynaecology 96: 445–53. Lipson, S.F. 2001. Metabolism, maturation, and ovarian function. In Reproductive Ecology and Human Evolution, ed. P.T. Ellison. New York: Aldine de Gruyter, 235–48. Lipson, S.F. and P.T. Ellison. 1996. Comparison of salivary steroid profiles in naturally occurring conception and non-conception cycles. Human Reproduction 11(10): 2090–6. Lunn, P., M. Watkinson, A.M. Prentice, P. Morrell, et al. 1981. Maternal nutrition and lactational amenorrhoea. Lancet 1(8235): 1428–29. Lunn, P.G., S. Austin, A.M. Prentice, and R.G. Whitehead. 1984. The effect of improved nutrition on plasma prolactin concentrations and postpartum infertility in lactating Gambian women. American Journal of Clinical Nutrition 39: 227–35. Lunn, P.G., A.M. Prentice, S. Austin, and R.G. Whitehead. 1980. Influence of maternal diet on plasma-prolactin levels during lactation. Lancet 1(8169): 623–25. Mayr, E. 1983. How to carry out the adaptationist program? American Naturalist 121: 324–34. McNeely, M.J. and M.R. Soules. 1988. The diagnosis of luteal phase deficiency: a critical review. Fertility and Sterility 50: 1–15. Michels, K.B., D. Trichopoulos, J.M. Robins, B.A. Rosner, et al. 1996. Birthweight as a risk factor for breast cancer. Lancet 348: 1542–46.

128

•

Peter T. Ellison and Grazyna Jasienska

Nesse, R.M. and G.C. Williams. 1996. Why We Get Sick: The New Science of Darwinian Medicine. New York: Vintage. Nilsen, T.I., P.R. Romundstad, R. Troisi, and L.J. Vatten. 2005. Birth size and subsequent risk for prostate cancer: a prospective population-based study in Norway. International Journal of Cancer 113: 1002–4. Paley, W. 1803 (reprinted 2005). Natural Theology. Landisville, PA: Coachwhip Publications. Panter-Brick, C., D.S. Lotstein, and P.T. Ellison. 1993. Seasonality of reproductive function and weight loss in rural Nepali women. Human Reproduction 8: 684–90. Perrin, N. and R.M. Sibly. 1993. Dynamic-models of energy allocation and investment. Annual Review of Ecology and Systematics 24: 379–410. Pike, M.C., D.V. Spicer, L. Dahmoush, and M.F. Press. 1993. Estrogens, progestogens, normal breast cell-proliferation, and breast-cancer risk. Epidemiologic Reviews 15: 17–35. Prentice, A.M., T.J. Cole, F.A. Foord, W.H. Lamb, and R.G. Whitehead. 1987. Increased birthweight after prenatal dietary supplementation of rural African women. American Journal of Clinical Nutrition 46: 912–25. Prentice, A.M., P.G. Lunn, M. Watkinson, and R.G. Whitehead. 1983a. Dietary supplementation of lactating Gambian women. II. Effect on maternal health, nutritional status and biochemistry. Human Nutrition-Clinical Nutrition 37:65–74. Prentice, A.M., S.B. Roberts, A. Prentice, A.A. Paul, et al. 1983b. Dietary supplementation of lactating Gambian women. I. Effect on breast-milk volume and quality. Human NutritionClinical Nutrition 37: 53–64. Prentice, A.M., R.G. Whitehead, S.B. Roberts, A.A. Paul, et al. 1980. Dietary supplementation of Gambian nursing mothers and lactational performance. Lancet 2(8200): 886–88. Prentice, A.M., R.G. Whitehead, M. Watkinson, W.G. Lamb, and T.J. Cole. 1983c. Prenatal dietary supplementation of African women and birth-weight. Lancet 1(8323): 489–92. Reeve, H.K. and P.W. Sherman. 1993. Adaptation and the goals of evolutionary research. Quarterly Review of Biology 68: 1–32. Roff, D.A. 2001. Life History Evolution. Sunderland, MA: Sinauer Associates. Stearns, S.C. 1992. The Evolution of Life Histories. New York: Oxford University Press. ———. 1999. Evolution in Health and Disease. New York: Oxford University Press. Trevethan, W.R., E.O. Smith, and J.J. McKenna. 1999. Evolutionary Medicine. New York: Oxford University. Vitzthum, V.J. 2001. Why not so great is still good enough. In Reproductive Ecology and Human Evolution, ed. P.T. Ellison. New York: Aldine de Gruyter, 179–202. Williams, G.C. 1966. Adaptation and Natural Selection. Princeton: Princeton University Press. Wyshak, G. and R.E. Frisch. 1982. Evidence for a secular trend in age of menarche. New England Journal of Medicine 306: 1033–35.

•5• Changes in Risk Factors for Breast Cancer in Migrant Women An Intergenerational Comparison Among Bangladeshis in the United Kingdom Alejandra Núñez-de la Mora and Gillian R. Bentley

Migrant Studies: Uses, Advantages, and Limitations Since the pioneer work by Boas a century ago (1912), migrant studies have been used as natural experimental models to assess the impact of diverse environments (biological and social) on human plasticity (Lasker 1995, 1969; Lasker and Mascie-Taylor 1988) (see the Box below). Such studies have compared migrants to sedentees (nonmigrants) with the aim of understanding how phenotypic, developmental, demographic, and behavioral patterns change after migration, as well as identifying factors in the new environment responsible for those changes. Migration studies have aided research on the effects of urbanization, modernization, and westernization, in particular in the context of the nutritional and epidemiological transitions (Dufour and Piperata 2004; Ostby et al. 1989; Salmond, Prior, and Wessen 1989). Such studies have been invaluable in shedding light on the effects of biosocial/biocultural practices on disease risk as well as in describing intergenerational trends in health outcomes, such as type 2 diabetes mellitus (Gerber 1984; Misra and Vikram 2004; Serrano-Rios, Goday, and Martínez-Larrad 1999), obesity (Ramirez and Mueller 1980), hypertension (Agyemang, Bhopal, and Bruijn-

130

•

Alejandra Núñez-de la Mora and Gillian R. Bentley

zeels 2004; James et al. 1985), cancer (Liao et al. 2003; Modan 1980; Nelson 2006; Shimizu et al. 1991; Smith et al. 2003; Ziegler et al. 1993), cardiovascular disorders (McKeigue et al. 1988), and asthma (Rottem, Szyper-Kravitz, and Shoenfeld 2005). One of the main strengths of migrant study designs is that they enable the discrimination between genetic and developmental components of the phenotypic adaptations to a new environment. They provide an alternative to longitudinal studies to assess developmental effects, and in practical terms, conducted as cross-sectional designs, they can prove considerably less expensive and time-intensive, easier to implement, and potentially more cost effective than their long-term counterparts. However, migrant designs are not without limitations. One major problem lies in determining which, from a complex array of environmental variables that differ between the localities of origin and destination, is responsible for effects found in the biological comparison of migrants and sedentees. Another limitation refers to the need to assess and control for variation due to selective migration, where migrants can differ from sedentees in a number of biological traits owing to the fact that migrants are not necessarily a random sample of the population represented by the

• • • Characteristics of Migrant Studies Migrant studies • are natural experiments that allow us to investigate the impact of changes in environmental conditions on the phenotype; • aid in identifying how much of the phenotypic variation in a population is related to developmental (plastic) adaptability and how much is due to genetic adaptation; • are used to investigate how different biological and social environments affect phenotypic, developmental, demographic, and behavioral patterns; and • provide a cost-effective alternative to longitudinal studies, especially in the case of long-lived human populations where the collection of data over sequential generations would be expensive and logistically and methodologically very complicated.

• • •

Changes in Risk Factors for Breast Cancer in Migrant Women

•

131

sedentees. Careful experimental design can help ameliorate such limitations (Lasker 1954). In this chapter, quantitative and qualitative data collected during a migrant study are used to explore social and behavioral transformations associated with migration and their potential impact on health. Specifically, findings are used to illustrate how epidemiological risk factors for breast cancer are changing across generations in this population. Following earlier studies (Eaton et al. 1994; Ellison 1999; Greaves 2000; Strassman 1999), this analysis will focus on examining some reproductive and nonreproductive (energetic and dietary) variables that have previously been identified as established risk factors for breast cancer through their association with increased acute, chronic, and cumulative exposure to endogenous ovarian steroids (Bernstein 2002; Henderson et al. 1996; Kelsey, Gammon, and John 1993) (Table 5.1). This analysis is not intended to be a basis for advancing epidemiological predictions regarding risk factors profiles for Bangladeshis in the United Kingdom, since the sample was small and did not represent the population at large. Nor is it intended as an exhaustive review of the factors that affect breast cancer epidemiology. Rather, the aim is to use the detailed information on individual reproductive and lifestyle histories that were collected for women of different generations to determine possible trends in risk factors in this migrant community. The advantages of our study design are that individuals can be unambiguously identified either as first- or second-genera-

Table 5.1. Endocrine-Related Risk Factors for Breast Cancer

1. Acute levels of circulating reproductive ovarian steroids during each menstrual cycle.* 2. Reproductive variables that impact the lifetime cumulative exposure of breast tissue to reproductive steroids: a) Age at menarche* b) Age at menopause* c) Age at first full-term pregnancy d) Parity e) Lactational practices 3. Nonreproductive variables that affect the levels of circulating ovarian steroids and/or menstrual regularity: f ) Obesity and weight gain g) Diet h) Levels of physical activity * Risk factors not analysed in this study.

132

•

Alejandra Núñez-de la Mora and Gillian R. Bentley

tion, and the effect of age at migration and length of residency in the UK can be accounted for. This is in contrast to larger epidemiological surveys where migrant status is often confounded with ethnicity (Falk et al. 2002; Hirani and Primatesta 2001). Breast Cancer Risk among South Asian Migrants in the United Kingdom In the clinical literature, it is commonly perceived that women of South Asian origin (Indians, Pakistanis, and Bangladeshis) are at low risk for breast cancer. This perception is supported by available cancer mortality and incidence data. For example, standardized mortality rates for South Asians in the UK are approximately half those of the general population (Acheson 1998; Balarajan and Raleigh 1993; Barker and Barker 1990; Bhopal 2002; Wild and McKeigue 1997). Estimated age-standardized rates are also considerably lower than those of the native English population (46 per 100,000 vs. 73 per 100,000, respectively) (Winter et al. 1999). Interestingly, South Asians show lower incidence and higher survival rates from breast cancer than women in the general population. These differences are unrelated to differences in age at diagnosis, socioeconomic deprivation, or disease stage at presentation (Farooq and Coleman 2005; Silva et al. 2003). Current evidence, however, indicates that trends are changing for the worse in migrant groups. Consistent with patterns observed in other migrant populations (Shimizu et al. 1991; Ziegler et al. 1993), breast cancer incidence in South Asian groups in the UK is moving in the direction of the host population and away from the low rates prevalent in the Indian subcontinent. For instance, age-standardized breast cancer rates for English South Asians (1990–92) are almost double those reported in the 1983–87 Bombay registry (Smith et al. 2003; Winter et al. 1999). Furthermore, breast cancer rates among South Asians have increased over the last ten years while having decreased among the rest of the population. In Leicester, a city with a large South Asian presence, the incidence ratios between 1990–99, adjusted for age and deprivation tertile, were 1.37 in South Asian versus 0.81 in non-Asian women (Smith et al. 2003). There is also evidence of an increased risk in young age groups including women born in the UK, and those who migrated to the UK during their childhood (Smith et al. 2003). A recent survey of England and Wales found that women of South Asian origin aged 20 to 29 years had higher breast cancer incidence rates than their non– South Asian counterparts (10.1 compared to 6.7 per 100,000) (Winter et al. 1999). Although there are well-established communities of South Asian origin in other parts of the world (mainly in the USA, Canada, and Australia), epidemiological data on breast cancer trends for these migrant groups is scarce. The limited information available, however, is in agreement with the patterns observed for their counterparts in the UK—namely, higher incidence rates compared to those in the Indian subcontinent (Kamath et al. 1999), and lower incidence rates than in the white host population but increasing steadily over time (Jain, Mills, and Parikh-Patel 2005; Parikh-Patel, Mills, and Jain 2006). Studies on breast cancer risk factors among these

Changes in Risk Factors for Breast Cancer in Migrant Women

•

133

South Asian communities are lacking (Kamath et al. 1999), hence the focus of this chapter on the South Asian community in the UK. Reproductive ecologists have argued that dramatic increases in cancer incidence observed among migrant populations, such as South Asians in the UK and other populations in transition (Shimizu et al. 1991; Thomas and Karagas 1987; Trichopolous et al. 1984; Ziegler et al. 1993), parallel the positive secular trends in growth and age at maturation characteristic of economic development and industrial modernization (Henderson and Bernstein 1991). They suggest that the numerous nutritional, health, and lifestyle changes that follow these economic transformations have shifted developmental patterns toward enhanced gonadal steroid production and increased lifetime exposure to these steroids, both of which are associated with higher risk for reproductive cancers (Ellison 1999). Along with the changes in developmental variables brought by improved living standards and general health of populations in transition, there are also important changes in behaviour in response to, and as a consequence of such socioeconomic changes. Such behavioral adjustments potentially influence cancer risk by affecting lifetime exposure to ovarian steroids. For example, a late age at first reproduction, low parity, and low incidence and duration of lactation in affluent groups of developed countries are associated with a higher lifetime risk for breast cancer (Henderson and Bernstein 1991; Kelsey, Gammon, and John 1993). Population-based demographic and health surveys in the UK show that South Asians have reproductive and nutritional patterns recognized as protective factors against breast cancer—namely, high parity, early first birth (ONS 2002; OPCS 1993), a high prevalence of breast-feeding (Thomas 1997), and fiber-rich traditional diets (Kassam-Khamis, Judd, and Thomas 2000; Kassam-Khamis et al. 1999; Silva et al. 2002, 2004). In most cases, however, these findings reflect characteristics observed in South Asians as a group. The extent to which these patterns can be regarded as universal for all subgroups within the “South Asian” category and among different generations of such groups needs to be explored further. The inclusion of different sub-ethnic groups under one heading has been partly the result of the characteristics and limitations of the datasets. Specifically, these relate to the criteria used for identifying and classifying individuals with respect to their ethnic background and migrant status. For instance, until the introduction of the ethnicity question in the 1991 UK census (Sillitoe and White 1992), all individuals of Indian, Bangladeshi, or Pakistani origin were indistinctly classified under the category of “South Asian” based on either their country of birth (Gill et al. 2005) or, when such information was not available, their family name (Nanchahal et al. 2001). Similarly, country of origin was the only indicator of an individual’s migrant status, which, by definition, failed to identify all UK-born second- and subsequent generation groups of South Asian ancestry and thereby biased data sets toward overseasborn individuals (Gill et al. 2005). The origin and reasons behind the use of such classification criteria are beyond the scope of this chapter. However, it is crucial to recognize that, in some data sets in

134

•

Alejandra Núñez-de la Mora and Gillian R. Bentley

the UK, a number of subgroups of rather different cultural, genetic, environmental, and socioeconomic backgrounds and migration status lie grouped under the common term “South Asian.” A similar problem occurs in the USA, where the term “Southeast Asian” is used to designate subgroups of Indian, Bangladeshi, and Pakistani origin. This hidden variation needs to be acknowledged when estimating disease risk and epidemiological trends from these sources, as the relative significance of specific risk factors for a number of health conditions, including breast cancer, are likely to vary within and between ethnic categories (Bhopal 2002; Bhopal et al. 1999). The recent adoption by census and health agencies of a new standard for ethnic group categories that is set on cultural characteristics, as well as the inclusion of a range of variables such as language, religion, and length of residency (Aspinall 2000; Gerrish 2000), has overcome some of the earlier shortcomings and is generating detailed data sets that, it is hoped, will instill an increased analytical rigor in ethnicspecific epidemiological and health policy–related analyses. Currently, the number of studies aimed at addressing differences in risk factors and disease incidence among precisely defined ethnic groups is still scant (Bhopal 2002; Bhopal et al. 1999; Hayes et al. 2002; Kassam-Khamis, Judd, and Thomas 2000; McKeigue 1992; Silva et al. 2002), but some epidemiological data on breast cancer are already available (McCormack et al. 2004). In a population-based casecontrol study of first-generation South Asian migrants, McCormack and collaborators (2004) gathered information on country of origin, religious and linguistic background, and breast cancer risk factors for Indian, Bangladeshi, and Pakistani women. Researchers found significant variation in breast cancer risk between South Asian ethnic subgroups, which was not fully explained by reproductive differences but was partly accounted for by diet and body size. Findings of significant variation in dietary habits and nutritional intake among South Asian subgroups in the UK complement and indirectly support these findings (Kassam-Khamis, Judd, and Thomas 2000; Sevak et al. 2004; Wharton, Eaton, and Wharton 1984). Specific dietary regimes such as long-term vegetarianism often associated with religious beliefs and characteristic of specific subgroups, may be associated with a reduction in breast cancer risk (Silva et al. 2002, 2004). It is clear from these studies, that “South Asians” should not be considered as a homogeneous group with respect to breast cancer risk. Furthermore, this variation is largely related to “lifestyle factors” such as diet and reproductive patterns that characterize and set apart each ethnic subgroup. Acknowledging the existence of a strong cultural component to disease risk is of crucial importance for both understanding disease etiology in a given subpopulation and for designing culturally sensitive public health policy measures. Some cultural aspects such as average age at first birth, breast-feeding patterns, and dietary intake are possible to include in large-scale prospective epidemiological studies. However, the interactions between risk factors, social variables affecting risk factors, and changes of risk over time are best appreciated in small-scale, detailed,

Changes in Risk Factors for Breast Cancer in Migrant Women

•

135

ethnographic research. It is here that biological anthropologists employing a biocultural approach can make a unique, practical, and valuable contribution to studying risk factor determinants of particular diseases. In this chapter, a case study conducted among Bangladeshi migrant women in London is presented as an example of this type of research. The Bangladeshi Migrant Study The original aim of our research was to determine how different environmental conditions experienced during childhood and adolescent development affect levels of adult reproductive function (Núñez-de la Mora et al. 2007). To this end, a cross-sectional migrant study was initiated to compare hormonal profiles of healthy women of reproductive age and similar genetic background who, by moving from a country of poor living standards such as Bangladesh to one of significantly higher standards such as the UK, had been exposed to contrasting environmental conditions during different phases of their life cycle. The migrant study groups were first-generation Bangladeshi women who migrated to the UK as adults (n=62), first-generation Bangladeshi women who moved to the UK as children (n=51), and second-generation women of Bangladeshi descent born in the UK (n=34). A group of nonmigrant Bangladeshi women (n=52) and one of white British women living in the same London neighborhoods as the Bangladeshi women (n=50) were used as references in the hormonal comparisons. Except for the occasional short visit to Bangladesh, all first- and second-generation women had lived uninterruptedly in the UK since migration or since birth, respectively (Table 5.2). Detailed quantitative information on sociodemographic variables, reproductive and migration histories, lifestyle, health, and diet was collected through closed-ended Table 5.2. Groups in the Bangladeshi Migrant Study

Group

Description

Adult migrants

First-generation Bangladeshi women who migrated to the UK as adults (post-menarche) First-generation Bangladeshi women who migrated to the UK as children (pre-menarche) Women born in the UK of Bangladeshi parents Nonmigrant Bangladeshi women resident in Sylhet, Bangladesh White British women born and living in the same London neighborhoods as their Bangladeshi counterparts

Children migrants Second-generation migrants Sedentee women White women

136

•

Alejandra Núñez-de la Mora and Gillian R. Bentley

questionnaires administered on a one-to-one basis. These data were used to make intergroup comparisons of the social, cultural, and biological changes sequential to the migration experience and to substantiate the interpretation of the hormonal results. Bangladeshi migrant women were contacted through local schools, community centers, mosques, and sport centers in the boroughs of Tower Hamlets and Camden in London, where the majority of the Bangladeshi population is concentrated (ONS 2002). As part of the recruitment process in London, a series of workshops on reproductive health were offered at community centers in these neighborhoods. This provided a space to meet and talk to women of different ages, who had been in the UK different lengths of time, and to women at various levels of exposure to the host culture. Interviews were conducted at home, since many women felt more confident and comfortable in their residences. This allowed for a closer, more intimate and personal relationship with each of the participants. Invitations to share meals and tea and to stay at family reunions or even attend weddings were common. In the midst of family life, there were plenty of opportunities to talk about issues that concern women, both young and old, in Bangladesh and London. This gave great insight into the intergenerational differences in perceptions about many aspects of life, such as those involving religion, ethnic and cultural identity, marriage, the roles of women, and aspirations regarding family and professional life. All this information over the course of three years was steadily weaved into our knowledge from first-hand experience of the community and members’ interactions with their host culture and with relations back in Bangladesh. This proved to be the richest source of information on how lifestyles are changing and how socioeconomic variables affect everyday life of women and their community—information that would not have been evident from the quantitative data alone. Such insights form the basis of many of the arguments in the discussion of the quantitative findings obtained through the questionnaires. Sociodemographic Characteristics of Bangladeshis in the United Kingdom The Bangladeshi communities in the UK typically exhibit large family sizes with a very high proportion of young people, low socioeconomic status, high dependence on local authority housing, low levels of education, and a high proportion of unskilled employment (Eade, Vamplew, and Peach 1996). However, our quantitative results revealed differences in socioeconomic indicators between overseas and UKborn migrant groups that reflect their different degrees of acculturation and social integration. Specifically, among the British-born groups there are signs of a shift toward nuclear, smaller and less crowded households, longer house tenure, and social and economic mobility. In relation to issues concerning women, there is evidence for higher educational attainment and employment, and along with this wider participation in economic activities, larger financial responsibilities in the family. Similarly, there is evidence of changes in marital patterns and social interactions away from traditional norms. It would appear that among the new generations, the roles and

Changes in Risk Factors for Breast Cancer in Migrant Women

•

137

situation of women in the Bangladeshi communities are changing in a more radical way than those of their male counterparts. Specifically, young women in this community have acquired new responsibilities outside the home in addition to those expected from them as traditional family carers (Núñez-de la Mora 2005). Some of the consequences of these changes for issues related to lifestyle, family dynamics, and reproductive decision making in the context of women’s health risk factors are discussed below. Breast Cancer Risk Factors Studied among Young Bangladeshi Women in London Reproductive Risk Factors 1. Age at first birth. There is a linear increase in breast cancer risk with increasing

age at first birth that is independent of other known factors (Kelsey, Gammon, and John 1993). It is thought that the protective effect of early age at first full pregnancy is mediated in two ways: 1) by driving final differentiation of the breast tissue and reducing the risk of further mutations; and 2) by permanently increasing the levels of sex hormone binding globulin (SHBG), thus reducing the amount of circulating free estrogens readily available to receptors in the breast tissue (Bernstein 2002). A Kaplan-Meier survival function analysis found significant differences in estimated age at first birth between migrant groups, with significantly older ages (average 4.8 years) at first birth for UK-born generations. Similarly, the period elapsed between menarche and first reproduction was significantly different between groups, with longer waiting times (average 4.5 years) for UK-born women. Additionally, data from the sociodemographic questionnaire revealed that marriage is being delayed among younger generations compared to older first-generation women, presumably as a result of more women entering formal employment and having longer educational careers. Although no data are available on contraceptive use, these findings suggest that intergenerational differences in reproduction decision making may affect breast cancer risk. With a decreasing age at menarche (data not shown) and a delay in the start of the reproductive career among younger generations of British-Bangladeshis, the lapse of uninterrupted menstrual cycles (in the absence of oral contraceptive use) could potentially put this group at comparatively higher risk. 2. Parity. Pregnancy has a dual effect on breast cancer, involving a short-term in-

crease in risk followed by a long-term protective effect (Bruzzi et al. 1988). Pregnancy is associated with high levels of estrogens, progesterone, and prolactin. These high hormone levels induce breast cell differentiation as well as cell proliferation, and this could explain the biphasic effect of pregnancy on breast cancer: pregnancy may be protective by reducing the pool of susceptible stem cells through differentiation, or conversely, it may promote breast cancer by inducing proliferation of cells that have

138

•

Alejandra Núñez-de la Mora and Gillian R. Bentley

already suffered malignant transformation (Bernstein 2002). Nevertheless, there is ample evidence for a protective effect of high parity on breast cancer risk independent of age at first birth (Albrektsen et al. 1994; Kelsey, Gammon, and John 1993; Yuan et al. 1988). All women in this study were of reproductive age (18–39 years old), but many had not begun childbearing. Small sample sizes and unfinished reproductive spans prevent too many generalizations, but available data suggest that the reproductive patterns among all three migrant groups in this study are, regardless of the effect of time since migration or age at migration, in line with high fertility rates previously reported in population-based surveys of Bangladeshis in the UK (Summerfield and Babb 2003). For example, at the time of data collection, the proportion of women with three or more children was approximately 50 percent in all three migrant groups (57 percent, 46 percent, and 50 percent for adult migrants, child migrants, and second-generation women, respectively) compared to only 17 percent for white women. Our data on socioeconomic variables indicate that although young BritishBangladeshi generations are in many respects moving away from the traditional customs of older generations, high fertility rates appear to remain. Future work, with second-generation women who have completed their reproductive life spans, will ascertain whether there are any significant differences in total fertility rates compared to the previous generation. This would determine whether younger Bangladeshi generations will continue to benefit from the protective effect of high fertility characteristic of their older counterparts. 3. Lactation. There is convincing evidence that lactation reduces breast cancer risk

among women by suppressing ovarian function and reducing the lifetime cumulative exposure to ovarian hormones through periods of lactational amenorrhoea (Enger et al. 1997; Newcomb et al. 1994; Yuan et al. 1988). Our results show that breast-feeding incidence, calculated as the proportion of study participant’s offspring who were ever breast-fed, is highest for the Bangladeshborn offspring of adult migrants (100 percent). (Note that some participant women already had children at the time of migration.) No significant differences in incidence were found between babies born in the UK to either group of first-generation migrants (range 76–83 percent). Adult migrant women who gave birth in both countries were less likely to breast-feed their England-born than Bangladesh-born babies (76 percent vs. 100 percent, respectively). For all migrant groups, neither number of years of education nor mother’s age among primiparous women affected breast-feeding incidence. Duration of breast-feeding was measured as the length of time (in months) for which breast-feeding continued. There were no significant differences in breast-feeding duration among UK-born offspring of women of all three migrant groups (range 7.8–4.8). However, adult migrants who gave birth in Bangladesh breast-fed their

Changes in Risk Factors for Breast Cancer in Migrant Women

•

139

offspring on average twice as long as Bangladeshi women who gave birth in the UK. For example, within-subject comparisons show that Bangladesh-born children were breast-fed on average ten months longer than their UK-born siblings. Although average differences in breast-feeding duration between migrant groups living in the UK were not significant, data on breast-feeding prevalence at different ages reveals contrasting patterns between groups. Breast-feeding prevalence was taken as the proportion of all babies who were wholly or partially breast-fed at specific ages. In our sample, the proportion of women who breast-fed for over fifteen months among first-generation migrants in London was 23 percent and 22 percent for adult and child migrants, respectively, and decreased to only 6 percent among secondgeneration women (contrast with 41 percent for sedentee women in Bangladesh and 8 percent for the white women group) (Núñez-de la Mora et al. 2005). These results, although limited owing to small sample sizes, argue for an important modification in breast-feeding patterns among Bangladeshi migrants in the UK. The variables implicated in such changes are likely to be many and varied in nature and probably relate to changes in family structure and lifestyle that have resulted from adapting to a different socioeconomic system and culture. As mentioned earlier, such changes have been particularly radical for females. Many first-, and most secondgeneration women have acquired extra family roles in addition to traditional ones. Time and energy have to be allocated between activities outside the home and domestic ones. With an increasing number of women in employment, many of them providing a large proportion of the family income, constraints typically associated with a decreasing incidence and shorter duration of breast-feeding have emerged for the Bangladeshi community (Núñez-de la Mora 2005). The disappearance of the extended family also has meant that housework and childcare are no longer shared with other female members but, instead, become the responsibility of a single woman. Given the fact that most British-Bangladeshi families are at the lower end of the economic spectrum, paid childcare or extended unpaid maternity leaves that might contribute to an increase in breast-feeding duration in this community are not realistic options. Another factor associated with the economic disadvantage prevalent in most Bangladeshi households is overcrowding. The lack of privacy associated with high household densities was often mentioned by women as an important deterrent to breast-feeding. Exposure to the host culture through everyday life and the media may also contribute to a modification of feeding choices, since there is a striking difference in breast-feeding behaviour of adult migrant women depending on the country where they gave birth. The same women breast-feeding in different environments show radically different patterns more consistent with the prevailing ones in the country of the offspring’s birth. Overall, these findings suggest that changes related to the experience of migration have had an impact on breast-feeding behavior, most notably in the second-generation group. The most apparent change is not in the overall incidence but rather in the reduction in average duration and prevalence of breast-feeding among British-

140

•

Alejandra Núñez-de la Mora and Gillian R. Bentley

born Bangladeshi women. In terms of breast cancer risk, the protective effect of long durations of breast-feeding and the suppression of ovarian function are likely to be undermined among young generations of women. Nonreproductive Risk Factors 1. Obesity and weight gain. Two aspects of the body mass index—obesity and weight

gain as an adult—are associated with a higher breast cancer risk among postmenopausal women (Hunter and Willet 1996). The increased risk in heavy postmenopausal women can be attributed to higher levels of circulating estrogen in these women, since the main source of endogenous estrogen after menopause is the conversion of the androgen precursor androstenedione to estrone in adipose tissue. Obesity is also related to reduced levels of SHBG (steroid-hormone binding globulin) and, therefore, to a higher tissue availability of free estrogens (Bernstein 2002). The International Agency for Research on Cancer estimates that 25 percent of breast cancer cases worldwide are due to overweight/obesity and a sedentary lifestyle. The preponderance of epidemiologic studies indicates that women who are overweight or obese have a 50–250 percent greater risk for postmenopausal breast cancer (McTiernan 2003). We found a high prevalence of overweight condition (25 ≤ BMI < 30) and obesity (BMI ≥ 30) among the Bangladeshi migrants who participated in our study. For instance, overweight rates were 36, 20, and 31 percent for adult migrants, child migrants, and second-generation women, respectively. Overall rates for obesity were 24, 27 and 23 percent for child migrants, adult migrants, and second-generation women, respectively. These figures are similar to those reported for South Asian women in the UK, but considerably higher than those reported for females of the general UK population in equivalent age groups (likely as a result of small sample size) (Erens, Primatesta, and Prior 2001). Obesity was found to be fairly prevalent among both the younger and the oldest age categories in the migrant groups. In the age group closest to menopause (36+), over two-thirds of migrant women had BMI > 25. Among migrant groups, there was a strong effect of parity on BMI; in all groups the highest proportion of parous women was at least one BMI category higher than their non-parous counterparts. However, we found a relatively high proportion of child migrants (25 percent) and secondgeneration women (45 percent) who were already overweight before childbearing. This may be taken as evidence of detrimental changes in lifestyle occurring among Bangladeshis growing up in the UK. A more Western lifestyle may be partly responsible for these changes where modification of dietary patterns and a more sedentary lifestyle result. It is also plausible that the high prevalence of obesity in this group is an illustration of the “thrifty phenotype” phenomenon (see Godfrey and Hanson, Chapter 7) common among groups in economic transition, whether migrant or not (Adair and Prentice 2004; Yudkin 1996). The high prevalence of diabetes and cardiovascular diseases reported for Bangladeshis living in the UK (Erens, Primatesta, and Prior 2001; McKeigue et al. 1988) may support this hypothesis.

Changes in Risk Factors for Breast Cancer in Migrant Women

•

141

If the weight trends for premenopausal women reported here were to continue into later ages, a large proportion of first- and second-generation Bangladeshi women may likely enter menopause with a high BMI. This could translate into potentially higher risk for postmenopausal breast cancer for women in all three migrant groups. 2. Diet. Epidemiological studies have produced rather inconsistent and inconclusive results concerning the role of dietary intake and the risk of breast cancer (Hunter and Willet 1996; Silva et al. 2004; Tavani et al. 2006; Willett 2001). However, data from the literature on dietary quality and steroid metabolism point to a suppressive effect of low-fat/high-fiber diets on steroid hormone levels (Dorgan et al. 2003; Goldin et al. 1994; Rose, Lubin, and Connolly 1997), thus contributing to a reduced breast cancer risk. Additionally, it has been suggested that breast cancer risk may be affected not so much through a high dietary fat intake per se but rather through the effects on body weight and composition. Thus, healthy dietary habits over the life course can be considered as protective by leading to a stable and desirable body weight. The traditional Bangladeshi diet consists of rice as a staple and pulses. In Sylhet, the region in Bangladesh from which the majority of women in this study originate, many varieties of fish, and to a lesser extent, lamb are also important dietary components. The consumption of cooked vegetables is relatively high, while consumption of raw leafy vegetables and fruits is rare. Dairy products are not prominent in the diet except for yogurt and sweetmeats on special occasions. Overall, the traditional Bangladeshi diet could be regarded as relatively healthy on the basis of its high fiber, non-starch polysaccharides (NSP), and omega fatty acid content (Kassam-Khamis, Judd, and Thomas 2000; Silva et al. 2002; Zannath and Edholm 2004). Despite the challenges of adjusting to a radically new lifestyle after migration, results from the diet questionnaire in our study demonstrate that food habits among first-generation Bangladeshi migrants in London are similar to those prevalent in Bangladesh. However, the second generation already shows signs of a more westernized diet as well as different eating habits (Núñez-de la Mora et al. 2004). The data suggest that, with more women attending higher education and joining the workforce, family routines and time budgets are changing. This reflects changes in the structure and character of family meals. The need for convenient, time-saving alternatives to the time-consuming traditional multicourse meals is reflected in the higher consumption of canned and frozen foods. There is evidence of Western foods that are calorically dense and nutritionally poor in quality being steadily introduced into the diet, often under the influence of young, more acculturated children who are attracted to fast foods and ready meals. Similarly, there is a trend across generations (especially among younger women) to eat outside the home. Student and employees reported eating lunch regularly at fast-food places that serve halal dishes. Moreover, among many child migrants and second-generation women there is an increased consumption of soft drinks instead of water during mealtimes. A higher consumption of processed foods, sweets, and chocolates in these groups also contributes to a higher

142

•

Alejandra Núñez-de la Mora and Gillian R. Bentley

intake of carbohydrates and salt. These dietary trends may partly account for the increased prevalence of obesity among young migrants discussed previously. In regard to the traditional diet, a lower intake of pulses, fresh fruits, and vegetables was found among children and second-generation migrants, as was a significant reduction in fish intake, particularly among the latter. Instead, second-generation women are eating red meat (lamb) much more frequently than their first-generation counterparts. This may be as a result of their improved economic situation. Despite these unfavorable changes, there are some in a healthier direction. For example, younger generations of child migrants and second-generation women are eating more brown and whole-grain bread, preparing less fried food, and drinking less full-fat milk than adult migrant groups. In summary, our results show that the traditional Bangladeshi diet, which is high in complex carbohydrates and unsaturated fatty acids from pulses, vegetables, and fish, is shifting toward one rich in energy-dense foods among younger generations. These changes in quality not only undermine the protective characteristics of the diet itself but could also directly contribute to an increased risk of breast cancer through their damaging effects on body weight. 3. Levels of physical activity. There is evidence for a reduced risk of breast cancer associated with lifetime physical activity in pre- and postmenopausal women (Friedenreich and Rohan 1995; McTiernan 2000; McTiernan et al. 1998). For instance, it has been estimated that women who engage in three to four hours per week of moderate to vigorous levels of exercise have a 30–40 percent lower risk for breast cancer than sedentary women (McTiernan 2003). Physical activity is considered a factor for breast cancer because of its potential effect on: 1) delaying age at menarche (Bernstein 2002; Kelsey, Gammon, and John 1993); 2) ovarian function, by suppressing ovulation and reducing circulating steroid levels (Ellison and Lager 1986; Jasienska and Ellison 1998; Jasienska, Thune, and Ellison 2000); and 3) lowering BMI and preventing weight gain (McTiernan 2003). For most Bangladeshi women, household work and walking represent the only type of physical activity undertaken. Analogous to recent data (Erens, Primatesta, and Prior 2001), walking as a physical activity is decreasing in the young generations of our study group. For example, while 87 percent of adult migrants reported walking daily for more than twenty minutes, only 59 percent of second-generation women did so. This reduction in walking appears to be accompanied by an increase in driving among second-generation women. In terms of household work, Bangladeshi women tend to assume all responsibility for chores. Unless second-generation households are able to afford household help, levels of this activity among different generations are unlikely to change. Given the poor tradition of physical exercise and sports activities among Muslim women, it is unlikely that Bangladeshi migrants would be able to take advantage of the protective effects of a lifetime of intense exercise. However, based on observations

Changes in Risk Factors for Breast Cancer in Migrant Women

•

143

and informal conversations during recruitment, it appears that younger women will be more likely to get involved in sport and exercise activities as they become more exposed to Western perceptions of fitness and body image. In London neighborhoods with a high Bangladeshi population, there are community and sport centers that cater to Muslim women and offer female-only swimming and exercise classes. These discrete changes may, in the long-term, increase physical activity levels among British-Bangladeshi women and contribute to lowering high BMIs, which in turn may translate in reduced cancer risk. Conclusion The Bangladeshi migrant community in the UK is experiencing social, cultural, and biological transformations that are relevant for public health. The intergenerational changes in the reproductive and lifestyle factors related to breast cancer risk described here, point to increased risk for malignancy in younger migrant generations. Our work, although focusing on a small sample, indicates that the Bangladeshis in the UK should no longer be seen as a low-risk group for breast cancer. Some developmental risk factors—younger age at menarche, older age at menopause, increased height, and enhanced ovarian function—are also associated with improved health and positive energetic conditions experienced by migrants growing and maturing in the UK (Núñez-de la Mora et al. 2007). For women born in such affluent conditions such “built-in” factors carry a detrimental risk for breast cancer (Okasha et al. 2003) and are unlikely to be reversed in the migrant population. These facts underline the need for public health programs aimed at promoting awareness of reducing breast cancer risk by focusing on those risk factors that are modifiable. Theoretically, all behavioral risk factors should be modifiable; however, in practice, those related to reproductive patterns are linked to and constrained by complex socioeconomic and cultural processes that are generally beyond the control of the individual. Some of the reproductive patterns that would confer a protective effect for breast cancer risk (e.g., high parity and early age at first birth) are at odds with demographic control measures and population trends as well as with young migrant generations’ experience of life in the UK. In fact, as recorded in this study, these reproductive variables tend to move in precisely the opposite direction as women gain access to educational and job opportunities. Thus, diet and physical activity are the most likely candidates for influencing breast cancer risk (McTiernan 2003). These behaviors are affected by a complex array of external factors, but, in contrast with reproductive behaviors, they are already of high priority in the public health agenda. However, as the findings of this and previous research demonstrate, the effort to influence diet and physical activity requires diligent development of programs and policies that are culturally and linguistically appropriate for the target community. This ensures not only the efficacy but also the sustainability of these interventions.

144

•

Alejandra Núñez-de la Mora and Gillian R. Bentley

This case study demonstrates how, through a holistic approach, bio-anthropology can contribute detailed and specific information and insight into the particularities of a given community. Such analytic exercises offer an opportunity to identify the differences in risk behaviors among ethnic groups derived from their specific cultural and socioeconomic characteristics. Moreover, through studies of this kind, which contrast various generations within the same migrant group, it is possible to document changes in risk factor patterns and analyze how these relate to processes of acculturation. A better understanding of this heterogeneity will prevent inaccurate group generalizations regarding perceived lower levels of breast cancer risk among South Asian women and will help public health disease-prevention initiatives to aim at clearer targets. Our results invite us to reflect on how an environment that may be considered “adverse” in terms of the availability, quality, and/or adequacy of resources, such as that in Bangladesh, may in fact be conducive to behavioral and biological characteristics that are protective against a particular condition, in this case, breast cancer. Conversely, an environment that, under the same criteria, would be commonly regarded as favorable, such as that in the UK, may promote biological and lifestyle changes that can increase risk for breast cancer. The idea that an affluent environment can prove damaging for some aspects of health is somehow counterintuitive. However, it constitutes one of the underlying arguments of the explanation for the increased prevalence of chronic and degenerative conditions observed in the westernized world and in populations in transition. The field of “diseases of affluence” will thus require adjustments to the way we think about health risks and demand a more precise understanding of the relationship between changing environments and the corresponding health outcomes. Appendix Breast Cancer Risk Among South Asians: Heterogeneity, Trends, and Prevention • Women of South Asian origin (Indians, Pakistanis, and Bangladeshis) living in Western countries (e.g., UK, USA, Canada) are currently at lower risk for breast cancer than the general host population. However, breast cancer incidence in South Asian groups in these countries is moving in the direction of the host populations and away from the low rates prevalent in their countries of origin in the Indian subcontinent. • The heterogeneity of South Asian populations has rarely been acknowledged in the context of breast cancer risk, but available evidence suggests significant variation in risk for this malignancy among these ethnic subgroups (Indians, Pakistanis, and Bangladeshis). • This study documents how some elements of the nutritional and reproductive patterns typically recognized as characteristic of South Asian populations, and

Changes in Risk Factors for Breast Cancer in Migrant Women

•

145

as protective factors against cancer (such as early age at first birth, prolonged breast-feeding, and high parity), are in fact changing across generations as a result of social, cultural, and biological transformations. • Among Bangladeshi migrant women in London, trends in age at first birth, lactational and dietary practices, and prevalence of obesity, point on the whole to increased risk in younger migrant generations. • Although, theoretically, most breast cancer risk factors are malleable, in practice, efforts must concentrate in diet and physical activity as the most realistic candidates for influencing breast cancer risk. More comparative studies are needed to further assess differences in prevalence of breast cancer risk factors among distinct ethnic groups and across generations. Small-scale, biocultural research offers a valuable opportunity for this.

References Acheson, D. 1998. Independent Inquiry into Inequalities in Health. London: Department of Health. Adair, L.S. and A.M. Prentice. 2004. A critical evaluation of the fetal origins hypothesis and its implications for developing countries. Journal of Nutrition 134(1): 191–93. Agyemang, C., R. Bhopal, and M. Bruijnzeels. 2004. Do variations in blood pressures of South Asian, African and Chinese descent children reflect those of the adult populations in the UK? A review of cross-sectional data. Journal of Human Hypertension 18(4): 229–37. Albrektsen, G., et al. 1994. Breast cancer incidence before age 55 in relation to parity and age at first and last births: a prospective study of one million norwegian women. Epidemiology 5(6 ): 604–11. Aspinall, P.J. 2000. The new 2001 census question set on cultural characteristics: is it useful for the monitoring of the health status of people from ethnic groups in Britain? Ethnicity and Health 5(1): 33–40. Balarajan, R. and V.S. Raleigh. 1993. Health of the Nation—Ethnicity and Health: A Guide for the N.H.S. London: Department of Health. Barker, R.M. and M.R. Barker. 1990. Incidence of cancer in Bradford Asians. Journal of Epidemiology and Community Health 44(2): 125–29. Bernstein, L. 2002. Epidemiology of endocrine-related risk factors for breast cancer. Journal of Mammary Gland Biology and Neoplasia 7(1): 3–15. Bhopal, R. 2002. Heterogeneity among Indians, Pakistanis and Bangladeshis is key to racial inequalities. British Medical Journal 325(7369): 903. Bhopal, R., et al. 1999. Heterogeneity of coronary heart disease risk factors in Indian, Pakistani, Bangladeshi, and European origin populations: cross sectional study. British Medical Journal 319(7204): 215–20. Boas, F. 1912. Changes in bodily form of descendants of immigrants. American Anthropologist 14: 530–62. Bruzzi, P., et al. 1988. Short term increase in risk of breast cancer after full term pregnancy. British Medical Journal 297: 1096–98.

146

•

Alejandra Núñez-de la Mora and Gillian R. Bentley

Dorgan, J.F., et al. 2003. Diet and sex hormones in girls: findings of a randomized controlled clinical trial. Journal of the National Cancer Institute 95(2): 132–42. Dufour, D.L. and B.A. Piperata. 2004. Rural-to-urban migration in Latin America: an update and thoughts on the model. American Journal of Human Biology 16(4): 395–404. Eade, J., T. Vamplew, and C. Peach. 1996. The Bangladeshis: the encapsulated community. In Ethnicity in the 1991 Census: The Ethnic Minority Populations of Great Britain, ed. C. Peach. London: Office for National Statistics. HMSO. Eaton, S.B., et al. 1994. Women’s reproductive cancers in evolutionary context. Quarterly Review of Biology 69(3): 353–67. Ellison, P.T. 1999. Reproductive ecology and reproductive cancers. In Hormones, Health and Behaviour: A Socio-Ecological Perspective, eds. C. Panter-Brick and C. Worthman. Cambridge: Cambridge University Press. Ellison, P.T. and C. Lager. 1986. Moderate recreational running is associated with lowered salivary progesterone profiles in women. American Journal of Obstetrics and Gynaecology 154(5): 1000–03. Enger, S.M., et al. 1997. Breast feeding history, pregnancy experience and risk of breast cancer. British Journal of Cancer 76(1): 118–23. Erens, B., P. Primatesta, and G. Prior, eds. 2001. Health Survey for England: The Health of Ethnic Minority Groups 99. London: HMSO. Falk, R.T., et al. 2002. Does place of birth influence endogenous hormone levels in AsianAmerican women? British Journal of Cancer 87(1): 54–60. Farooq, S. and M.P. Coleman. 2005. Breast cancer survival in South Asian women in England and Wales. Journal of Epidemiology and Community Health 59(5): 402–6. Friedenreich, C.M. and T. Rohan. 1995. A review of physical activity and breast cancer. Epidemiology 6(3): 311–17. Gerber, L.M. 1984. Diabetes mortality among Chinese migrants to New York City. Human Biology 56:449–58. Gerrish, K. 2000. Researching ethnic diversity in the British NHS: methodological and practical concerns. Journal of Advanced Nursing 31(4): 918–25. Gill, P.S., et al. 2005. Limitations and potential of country of birth as proxy for ethnic group. British Medical Journal 330(7484): 196. Goldin, B.R., et al. 1994. The effect of dietary-fat and fiber on serum estrogen concentrations in premenopausal women under controlled dietary conditions. Cancer 74(3): 1125– 31. Greaves, M. 2000. Cancer: The Evolutionary Legacy. Oxford: Oxford University Press. Hayes, L., et al. 2002. Patterns of physical activity and relationship with risk markers for cardiovascular disease and diabetes in Indian, Pakistani, Bangladeshi and European adults in a UK population. Journal of Public Health Medicine 24(3): 170–78. Henderson, B.A. and L. Bernstein. 1991. The international variation in breast cancer rates: an epidemiological assessment. Breast Cancer Research and Treatment 18(1): 11–17. Henderson, B.E., et al. 1996. Breast cancer. Cancer Epidemiology and Prevention, eds. D. Schottenfeld and J.J.F. Fraumeni. Oxford: Oxford University Press. Hirani, V. and P. Primatesta. 2001. Eating Habits—Health Survey for England: The Health of Minority Ethnic Groups ’99. London: Department of Health. Hunter, D.J. and W.C. Willet. 1996. Nutrition and breast cancer. Cancer Causes and Control 7(1): 56–68.

Changes in Risk Factors for Breast Cancer in Migrant Women

•

147

Jain, R.V., P.K. Mills and A. Parikh-Patel. 2005 Cancer incidence in the South Asian population of California, 1988–2000. Journal of Carcinogens 4: 21. James, G.D., et al. 1985. Differences in carcholamine excretion rates, blood pressure and lifestyles among young Western Samoa men. Human Biology 57: 635–47. Jasienska, G. and P.T. Ellison. 1998. Physical work causes suppression of ovarian function in women. Proceedings of the Royal Society of London Series B Biological Sciences 265(1408): 1847–51. Jasienska, G., I. Thune, and P.T. Ellison. 2000. Energetic factors, ovarian steroids and the risk of breast cancer. European Journal of Cancer Prevention 9(4): 231–39. Kamath, S.K., et al. 1999. Breast cancer risk factors in two distinct ethnic groups: Indian and Pakistani American premenopausal women. Nutrition and Cancer—An International Journal 35(1): 16–26. Kassam-Khamis, T., et al. 1999. Development of an interview-administered food-frequency questionnaire for use amongst women of South Asian Ethnic origin in Britain. Journal of Human Nutrition and Dietetics 12(1): 7–19. Kassam-Khamis, T., P.A. Judd and J.E. Thomas. 2000. Frequency of consumption and nutrient composition of composite dishes commonly consumed in the UK by South Asian Muslims originating from Bangladesh, Pakistan and East Africa (Ismailis). Journal of Human Nutrition and Dietetics 13(3): 185–96. Kelsey, J.L., M.D. Gammon and E.M. John. 1993. Reproductive factors and breast cancer. Epidemiological Review 15(1): 36–47. Lasker, G.W. 1954. The question of physical selection of Mexican migrants to the USA. Human Biology 26: 52–58. ———. 1969. Human biological adaptability. Science 166: 1480–86. ———. 1995. The study of migrants as a strategy for understanding human biological plasticity. In The Study of Migrants as a Strategy for Understanding Human Biological Plasticity, eds. C.G.N. Mascie-Taylor and B. Bogin. Cambridge: Cambridge University Press. Lasker, G.W. and C.G.N. Mascie-Taylor. 1988. The framework of migration studies. In Biological Aspects of Human Migration, eds. C.G.N. Mascie-Taylor and G.W. Lasker. Cambridge: Cambridge University Press. Liao, C.K., et al. 2003. Endometrial cancer in Asian migrants to the US and their descendants. Cancer Causes & Control 14(4): 357–60. McCormack, V.A., et al. 2004. Heterogeneity of breast cancer risk within the South Asian female population in England: a population-based case-control study of first-generation migrants. British Journal of Cancer 90(1): 160–66. McKeigue, P.M. 1992. Coronary heart-disease in Indians, Pakistanis, and Bangladeshis— etiology and possibilities for prevention. British Heart Journal 67(5): 341–42. McKeigue, P.M., et al. 1988. Diabetes, hyperinsulinaemia, and coronary risk factors in Bangladeshis in East London. British Heart Journal 60(5): 390–96. McTiernan, A. 2000. Associations between energy balance and body mass index and risk of breast carcinoma in women from diverse racial and ethnic backgrounds in the U.S. Cancer 88(5): 1248–55. ———. 2003. Behavioral risk factors in breast cancer: can risk be modified? Oncologist 8(4): 326–34. McTiernan, A., et al. 1998. Physical activity and cancer etiology: associations and mechanisms. Cancer Causes & Control 9(5): 487–509.

148

•

Alejandra Núñez-de la Mora and Gillian R. Bentley

Misra, A. and N.K. Vikram. 2004. Insulin resistance syndrome (metabolic syndrome) and obesity in Asian Indians: evidence and implications. Nutrition 20(5): 482–91. Modan, B. 1980. Role of migrant studies in understanding the etiology of cancer. American Journal of Epidemiology 112(2): 289–95. Nanchahal, K., et al. 2001. Development and validation of a computerized South Asian names and group recognition algorithm (Sangra) for use in British health-related studies. Journal of Public Health Medicine 23(4): 278–85. Nelson, N.J. 2006. Migrant studies aid the search for factors linked to breast cancer risk. Journal of the National Cancer Institute 98(7): 436–38. Newcomb, P., et al. 1994. Lactation and a reduced risk of premenopausal breast cancer. New England Journal of Medicine 330(2): 81–86. Núñez-de la Mora A. 2005. Developmental effects on adult reproductive hormone levels: a migrant study. Ph.D. thesis, University of London. Núñez-de la Mora A, et al. 2004. Fresh or frozen? Changing food habits among Bangladeshi migrant women in London. American Journal of Human Biology 16(2): 216. ———. 2005. Changes in breastfeeding practices among Bangladeshi women in London. American Journal of Physical Anthropology 126(S40): 159. ——— 2007. Childhood conditions determine adult progesterone levels. PLoS Medicine 4(5): e167 Okasha, M., et al. 2003. Exposures in childhood, adolescence and early adulthood and breast cancer risk: a systematic review of the literature. Breast Cancer Research and Treatment 78(2): 223–76. ONS. 2002. Social Focus in Brief: Ethnicity 2002. London: Office for National Statistics. OPCS. 1993. 1991 Census Report for England: Regional Health Authorities. London: Office of Population Censuses and Surveys, HMSO. Ostby, T., et al. 1989. Type 2 (non-insulin dependent) diabetes mellitus, migration and westernisation: the Tokelau Island migrant study. Diabetologia 32(8): 585–90. Parikh-Patel, A., P.K. Mills and R.V. Jain. 2006. Breast cancer survival among South Asian women in California (United States). Cancer Causes & Control 17(3): 267–72. Ramirez, M.E. and W.H. Mueller. 1980. The development of obesity and fat patterning in Tokelau children. Human Biology 52: 675–87. Rose, D.P., M. Lubin and J.M. Connolly. 1997. Effects of diet supplementation with wheat bran on serum estrogen levels in the follicular and luteal phases of the menstrual cycle. Nutrition 13(6): 535–39. Rottem, M., M. Szyper-Kravitz, and Y. Shoenfeld. 2005. Atopy and asthma in migrants. International Archives of Allergy Immunology 136(2): 198–204. Salmond, C.E., I.A. Prior, and A.F. Wessen. 1989. Blood pressure patterns and migration: a 14year cohort study of adult Tokelauns. American Journal of Epidemiology 130(1): 37–52. Serrano-Rios, M., A. Goday, and T. Martínez-Larrad. T. 1999. Migrant populations and the incidence of type 1 diabetes mellitus: an overview of the literature with a focus on the Spanish-heritage countries in Latin America. Diabetes Metabolism Research Review 15(2): 113–32. Sevak, L., et al. 2004. Validation of a food frequency questionnaire to assess macro- and micronutrient intake among South Asians in the United Kingdom. European Journal of Nutrition 43(3): 160–68.

Changes in Risk Factors for Breast Cancer in Migrant Women

•

149

Shimizu, H., et al. 1991. Cancer of the prostate and breast among Japanese and white immigrants in Los Angeles. British Journal of Cancer 63(6): 963–66. Silva, I.D., et al. 2002. Lifelong vegetarianism and risk of breast cancer: a population-based case-control study among South Asian migrant women living in England. International Journal of Cancer 99(2): 238–44. ———. 2003. Survival from breast cancer among South Asian and non-South Asian women resident in south east England. British Journal of Cancer 89(3): 508–12. ———. 2004. Phyto-Oestrogen intake and breast cancer risk in South Asian women in England: findings from a population-based case-control study. Cancer Causes & Control 15(8): 805–18. Sillitoe, K. and P.H. White. 1992. Ethnic-group and the British census—the search for a question. Journal of the Royal Statistical Society Series A-Statistics in Society 155(1): 141–63. Smith, L.K., et al. 2003. Latest trends in cancer incidence among UK South Asians in Leicester. British Journal of Cancer 89(1): 70–73. Strassman, B.I. 1999. Menstrual cycling and breast cancer: an evolutionary perspective. Journal of Women’s Health 8(2): 193–202. Summerfield, C. and P. Babb, eds. 2003. Social Trends No. 33. Vol. 33. London: Office for National Statistics. Tavani, A., et al. 2006. Consumption of sweet foods and breast cancer risk in Italy. Annals of Oncology 17(2): 341–45. Thomas, D.B. and M.R. Karagas. 1987. Cancer in first and second generation Americans. Cancer Research 47(21): 5771–76. Thomas, M. 1997. Infant Feeding in Asian Families: Early Practices and Growth. London: Department of Health. Trichopolous, D., et al. 1984. The effect of westernisation on urine estrogens: frequency of ovulation and breast cancer risk. A study of ethnic Chinese women in the Orient and the USA. Cancer 53(1): 187–92. Wharton, P.A., P.M. Eaton and B.A. Wharton. 1984. Subethnic variation in the diets of Moslem, Sikh and Hindu pregnant women at Sorrento-Maternity-Hospital, Birmingham. British Journal of Nutrition 52(3): 469–76. Wild, S. and P. McKeigue. 1997. Cross sectional analysis of mortality by country of birth in England and Wales, 1970–92. British Medical Journal 314(7082): 705–10. Willett, W.C. 2001. Diet and cancer: one view at the start of the millennium. Cancer Epidemiology, Biomarkers and Prevention 10(1): 3–8. Winter, H., et al. 1999. Cancer incidence in the South Asian population of England (1990– 92). British Journal of Cancer 79(3–4): 645–54. Yuan, J.M., et al. 1988. Risk factors for breast cancer in Chinese women in Shanghai. Cancer Research 48(7): 1949–53. Yudkin, J.S. 1996. Non-insulin-dependent diabetes mellitus (NIDDM) in Asians in the UK. Diabetic Medicine 13(9): S16–18. Zannath, K. and F. Edholm. 2004. Eating for a Healthy Life: A Guide to Healthy Eating and Bangladeshi Eating Patterns. London: Public Health Department, Camden Primary Care Trust. Ziegler, R.G., et al. 1993. Migration patterns and breast cancer risk in Asian-American women. Journal of the National Cancer Institute 85(22): 1819–27.

•6• Family Structure and Child Growth in sub-Saharan Africa Assessing “Hidden Risk” Daniel W. Sellen

Families, Health, and Research Do social relations within families affect the health of individuals? Despite some influential theoretical work on the household production of health over a decade ago (Berman, Kendall, and Bhattacharyya 1994; Harkness and Super 1994), there is little published evidence for links between social processes of family formation (such as marriage practices) and physical health indicators (such as child growth). A paucity of studies showing such links does not mean that they do not exist, however. It simply means that relationships between family structure and health have not been widely reported. This, in turn, may be because such relationships have not been adequately investigated. The general goal of this chapter is to raise questions about how a priori perceptions of risk influence the research agenda in biological anthropology. I suggest that a vicious cycle of ignorance can become inherent in the research process if real health risks remain “undiscovered” or hidden from scientific view because they are erroneously assumed to be unimportant. To illustrate how this can happen, I provide a review of recent investigations of the relationships between indicators of children’s physical health and mothers’ marital status in sub-Saharan African communities where polygyny is common.

Family Sructure and Child Growth in sub-Saharan Africa

•

151

The results of this review suggest that polygyny is a previously “hidden” risk for poor health outcomes among children in communities where it is practiced. A consistent inverse relationship between measures of polygynous marriage and measures of child growth and survival is found across a number of societies. I discuss the implications of this recent finding for the professional practice of biological anthropologists and the refinement of our explanatory models. I also argue for a renewed focus on family structure and social processes of family formation as predictors of health, risk, and adversity in biocultural studies in sub-Saharan Africa and beyond. In order to provide background on the use and interpretation of growth data in these types of analysis, I also provide a review of how biological anthropologists use child growth as a measure of developmental trajectory, current health, and future biological fitness. I outline what growth measures can tell us about the quality and adequacy of the physical and social environments in which individuals live and about the highly complex interactions between biocultural, ecological, and physiological processes that influence growth and related outcomes. Readers will be able to gain an appreciation of how associations between growth measures and the other healthrelated outcomes result both from direct causal links and from noncausal associations. Background to the Research Questions My interest in investigating the possible relations between family structure and health was triggered by ethnographic observations made while living among the nomadic, cattle-herding Datoga community living in the Eyasi basin region in Tanzania (Box 6.1). My initial theoretical perspective was drawn from human behavioral ecology, which focused my attention on ways to evaluate the adaptiveness of parental behavior in response to variation in ecological, economic, and social constraints. This led to a conceptual framework in which the components of parental fitness were measured by the numbers of children born, their survival to specific ages, and their growth performance at any age. Based on a review of the literature available at the time, I hypothesized that parents in relatively wealthy families (i.e., those with larger herds) had greater access to animal-source foods for consumption or sale, and that their children would therefore be more likely to survive beyond 5 years of age and be less likely to show growth deficits during infancy and early childhood and during the hungry dry season. My measurement methods and analytic strategy assumed that relatively poor growth is an indicator of current poor health, a predictor of risk of future death, and an outcome of past adversity. The findings were unexpected. Although women’s weight and fat status was associated with household wealth in livestock, there was only a weak association between size of the household herd and measures of their children’s growth (Sellen 2003). Similarly, although women’s weight and fat status decreased in the dry season (especially among lactating mothers), children’s weight, height, muscle mass, and fatness

152

•

Daniel W. Sellen

• • • A Polygynous Community in sub-Saharan Africa The ethnic Datoga community living in the Eyasi basin of Tanzania provides an example of a contemporary community in which many households are polygynous. The Datoga are culturally aligned with the neighboring Barabaig and live in a physical environment that is arid and highly seasonal, with less than 600mm of rainfall per year occurring in a bimodal pattern (Tanaka 1969). The Datoga subsist on herds of cattle, sheep, and goats Figure 6.1), consume a diet of maize augmented by milk and meat, and are relatively impoverished relative to neighboring East African pastoralists (Klima 1965; Lane 1996; Sellen 1995; Sieff 1999). Most families are polygynous and live in extended households (Sellen 1995). The Datoga are unusual among East African pastoralists in that women have property rights and can own and exchange livestock and other gifts with members of both their own and their husband’s lineage or clan (Borgerhoff Mulder 1991a,b; Klima 1964, 1965; Tomikawa 1978). Households live in temporary timber-framed wattle and daub houses and are seminomadic in order to exploit limited water and forage resources (Borgerhoff Mulder 1991a). Livelihoods are almost entirely derived from livestock holdings (Lane 1996; Sieff 1997; Tomikawa 1978, 1979). Livestock are essential for household production of animal-source foods (milk, meat, and fat) and are sold or (rarely) exchanged for grain available at local markets. A small proportion of grain consumed is cultivated or acquired through exchange or sale of non-livestock products such as leather goods, beer, wild honey, and medicinal plants or concoctions (Sieff 1999). Few families cultivate grain because of the aridity of the Eyasi basin and the need to move livestock large distances to forage and water (Sellen 2003; Sieff 1995; Tomikawa 1970, 1972). Wage labor is extremely rare and very few families receive remittances from relatives working elsewhere. Under-5 mortality rate and total fertility rate is high (20 percent and 6.9 live births/ woman respectively; Borgerhoff Mulder 1992). Although breastfeeding is universal, young child diets are inadequate (Sellen 1998), child growth is poor (Sellen 1999a), incidence of child morbidity is high (Sellen 2001), and indicators of child undernutrition and adult chronic energy deficiency are prevalent (Sellen 1998, 2000a,b). Seasonal changes in physical nutritional status of women and children are small (Sellen 2000b).

• • •

Family Sructure and Child Growth in sub-Saharan Africa

•

153

varied little by season of measurement after appropriate adjustment for age and sex in relation to international reference data in use at the time of the original analysis (U.S. Department of Health Education and Welfare 1977; data presented in Sellen 2000b). This suggested most children were buffered from seasonal cycles. However, indicators of children’s growth performance varied by age and sex even after taking into account the generally poor growth of the average child in the population and standardizing relative to international reference data (Sellen 2000a). Preadolescent girls (who perform many arduous household tasks such as collecting water) and teenage boys (who are engaged in the rugged work of herding livestock) were more likely to be underweight or stunted (less than two standard deviations below reference data). Most interestingly, although I had originally hypothesized to the contrary based on findings of few costs to polygyny in previous studies of other populations (Borgerhoff Mulder 1991c), children’s growth also varied with the marital status of their mothers (Figure 6.1). The weight and fat status of women themselves was not associated with the number of co-wives or their marital rank. However, achieved weight

Figure 6.1. Growth status of young children, 0-3.5 years, sampled in a Datoga community, Tanzania, according to number of co-wives of the mother (n=86 surviving children). Bars represent estimated residual means and standard errors after adjusting for age and sex of child. Numbers above bars indicate sample sizes. Median lengths and weights of a reference population of American children would fall on the y=0 reference line. Analysis of covariance shows main effects of number of co-wives (HAZ: F 4, 76 = 5.625, p= .001 ; WAZ: F 4, 86 = 4.017, p= 0.005) and child’s age at measurement (HAZ: F 1, 76 = 5.231, p= .025; WAZ: F 1, 85 = 4.787, p= .032) and no significant interaction of these effects.

154

•

Daniel W. Sellen

and height of children was strongly associated with the number of co-wives present and the order of the mother’s marriage to the husband (Sellen 1999, 2000b). Resources and Health in Polygynous Households The unanticipated discovery that children’s health and maternal marital status are linked in the Eyasi Datoga leads me to speculate more widely. Perhaps polygyny is a more important determinant of health than many anthropologists and health scientists currently assume. Social scientists certainly agree that polygyny remains a common practice in many human societies. It is intimately connected to complex systems of social exchange of resources that have the potential to influence health outcomes. But few studies suggest polygyny has direct effects on the health of women and children. It is helpful to consider the distribution of polygyny in time and space and some theories about how the practice influences the allocation of resources within households. It has been estimated that, in the past, approximately 83 percent of preindustrial societies practiced some form of polygyny (Murdoch 1965). Today, polygyny is most prevalent in sub-Saharan Africa (Westoff 2003), where, despite recent declines in some countries, it remains one of the most distinguishing characteristics of family structure (Kayongo-Male and Onyango 1984). Large regional variations are observed, however, with countries in West Africa tending to have the highest percentage of polygynous marriage (e.g., 54.7 percent in Burkina Faso, 53.7 percent in Guinea) and countries in East Africa tending to have the lowest percentage of polygynous marriages (e.g., 9.3 percent in Eritrea. 4.0 percent in Madagascar). Within countries, polygyny tends to be less prevalent in urban areas and declines with education (Westoff 2003), and the prevalence and intensity (number of wives per union) of polygyny are inversely related (Timaeus and Reynar 1998). Polygyny represents an interesting example of a practice that structures social relations in ways that have the potential to generate inequalities in health, risk, and adversity among individuals within and between families. Twenty-five years ago, Becker suggested a theoretical economic model to explain the widespread occurrence of polygyny in terms of the greater net benefits to women than other marriage alternatives (Becker 1981). This model was similar to others already developed by ecologists to account for polygynous mating in other animals in terms of “female choice” for individual males that offer more of the resources needed for female reproduction than other males (Orians 1969). These models argued that polygyny is unlikely to be costly to mothers or their offspring because females assort themselves in what economists refer to as an “ideal free distribution” with respect to male resources—that is, females pair with the male offering the best available resources at the time of her choice, and are “indifferent” to whether this is determined in part by his previous matings. Applied to humans these models predict that, at equilibrium, all mothers should have access to the same resources, and so maternal and child health differentials

Family Sructure and Child Growth in sub-Saharan Africa

•

155

should be negligible. Recently such equilibrium models have been criticized for misrepresenting ethnographic realities. Indeed, the frequently observed and significant power differentials between men provide evidence against key assumptions of the equilibrium models. Taking this line of reasoning, Brabin suggested that addition of wives would lead to fewer resources for everyone (Brabin 1984), and that such “resource dilution” would manifest first as lower nutritional status and second as higher mortality among children of polygynously married mothers. Research Questions The potential health effects of social processes such as marriage formation are not well investigated, however. In particular, the social processes involved in polygyny are little studied as potential determinants of variation in health, risk, and adversity among men, women, and children. Numerous studies in rich economies find positive associations between monogamous marriage and mental health, particularly for men (Horwitz, White, and Howell-White 1996; Markey et al. 2005), but few examine underlying mechanisms (Burman and Margolin 1992). In contrast, very few studies have investigated whether marriage practices affect physical health of children in poorer settings. One comparative study has suggested that demographic (fertility, child survival) and anthropometric (height-for-age and stunting) measures of adversity, risk, and health vary little within polygynous families (Desai 1995). This is puzzling, however, because it is known that there are significant associations between women’s marital status and factors important for the empowerment of women, such as education, occupation, and income. In the absence of data to the contrary, it is often assumed that marriage has few or limited effects on physical health. Therefore, it is difficult to justify studies designed to investigate the potential effects of marriage on health. Studies of the relationship between polygyny and health-related outcomes are also difficult to design for several reasons. Many monogamous unions later become polygynous ones, making it difficult to compare polygynists’ first wives with women in monogamous unions. Relations between co-wives in polygynous families are highly complex, making it difficult to compare groups of higher-order wives from first wives. Interviews often yield current status information rather than a full marital history, making it difficult to know how many women may have been in a different type of union in the past. But what might we find if we did conduct such studies? We can ask some specific questions about the relationship between polygyny and child health in sub-Saharan Africa. Do children in polygynous families face greater or lesser adversity and/or risk of poor health outcomes than children in other kinds of family, such as monogamous, single-parent, or foster families? Within polygynous families, do children of some co-wives face different risks than other children within the household? Answers to these questions are potentially of great relevance in many communities in sub-Saharan Africa. Threats to family and child health are many, and poverty

156

•

Daniel W. Sellen

overdetermines social relations and resource allocation. Biological anthropologists have made some recent progress in testing whether polygyny has negative effects on the health of women or children in sub-Saharan Africa by examining measures of child growth and survival. Several studies have yielded data that can be used to evaluate the “resource dilution” hypothesis by testing for decreased nutritional status or survival among children of polygynously married mothers. Studies in sub-Saharan Africa Kanembous and Goranes, Chad Begin and colleagues (Begin, Frongillo, and Delisle 1999; Begin et al. 1997) found that among a sample of ninety-eight children aged 12–71 months living in rural Chad, children living in households with multiple wives had significantly lower height-forage z-scores (HAZ) than children living in monogamous homes. After controlling for a host of potential confounders, the “polygyny effect” was estimated to be –.89 HAZ. This is a large effect for a population with a mean HAZ of –2.34 and one that is probably biological meaningful. Interestingly, when controls for household socioeconomic status were introduced into multivariate models, the differences between monogamous and polygynously married women’s children increased. This suggested that the higher level of socioeconomic status (SES) of polygynous household had a protective effect: within housholds of poor SES, children of polygynously married women would have even lower nutritional status. Datoga, Tanzania Colleagues and I pooled data collected among a nomadic pastoralist community of Datoga in Tanzania to perform a direct test of the hypothesis that polygyny increases child mortality. This revealed that the probability of a child dying before the fifth birthday was significantly higher in poorer households with more co-wives (Sellen, Borgerhoff Mulder, and Sieff 2000). The under-5 mortality rate (U5MR) was highest among children whose mother shared the household with a single co-wife during the child’s life. The mortality rate was significantly higher among such children in the poorest households than among all other children (p=0.025). More than 50 percent of these children died, compared to approximately 25 percent of children born into other circumstances. This study also revealed a step-wise pattern of decreased nutritional status among children of polygynously married women (Sellen 1999). Interestingly, the results showed that the children of second-married wives suffered the greatest growth deficits and that greater child growth deficits occurred in the poorest households (Figure 6.2). These data provided evidence that, while resource dilution may compromise the growth of some children in polygynous families, particularly in poorer families, these negative effects are not evenly distributed among the children of all co-wives. Rather, Figure 6.2 shows that, irrespective of household wealth in livestock, the children of

Family Sructure and Child Growth in sub-Saharan Africa

•

157

Figure 6.2. Association of growth indicators with household wealth and marital status among surviving children, 0-11 year old, sampled in a Datoga community, Tanzania (n=188). (Redrawn from Sellen 1999). Weight-for-age z-score relative to 1977 NCHS reference medians (Department of Health Education and Welfare, 1977) was significantly associated with wealth after adjustment of child’s sex and age by analysis of covariance. Bars represent means and standard errors estimated by ANCOVA models. Marital status of the mother is designated as M1 (first and only wife), P1 (first wife in a polygynous marriage), P2 (second wife in a polygynous marriage), P3 (third wife in a polygynous marriage), P4+ (fourth or later-married wife in a polygynous marriage). Wealth category was based on reported size of household herd (poorer tercile, 54 TLU).

mothers who married as the second wife experienced significantly greater growth deficits than others in both monogamous and polygynous families. The same pattern of results was consistently found for several anthropometric indicators of child growth and among both children younger than 3 years and surviving children 4–10 years of age (data not shown). Gwembe Tonga, Zambia Gillett-Netting and Perry (Gillett-Netting and Perry 2005) examined anthropometric indicators among children in matrilineal Tonga communities of the Gwembe valley in Zambia. They aimed to test whether dilution of resources increases the levels of undernutrition in polygynous households and also whether gender bias influences child growth outcomes. They found a similar step-wise pattern of poorer growth of children in polygynous households among boys, but not girls. Their interpretation is that resource dilution occurs in polygynous households, girls are generally favored over boys and that girls are therefore protected from adverse outcomes, including poor growth, even when resource are constrained.

158

•

Daniel W. Sellen

Sukuma, Tanzania High levels of poverty and undernutrition are observed in the Datoga, among whom 95 percent of households fall below poverty cut-offs and 45 percent of women present with chronic energy deficiency (CED) indicated by a body mass index (BMI) of less than 18.5 (Sellen 2000b). In order to assess whether the Datoga results were attributable to their poverty or mode of subsistence, Hadley conducted a similar smallscale study in a subsistence agropastoralist community of Sukuma, among whom fewer than 5 percent of women present with BMI indicative of CED (Hadley 2005). The results showed that polygynous marriage was a risk factor for poor child growth, that this was particularly true for achieved height-for-age at any age, and that this relationship was maintained even after controlling for child age, sex, and household characteristics. Analysis of longitudinal measures of growth velocity indicated that this association was most pronounced in the wet season when food insecurity is generally at its highest. This suggests that in resource-rich populations, a polygyny effect may be detectable only during particularly stressful periods of the year. Rakai, Uganda A study from Uganda examined whether polygynous marriage was a risk factor for child mortality in a cohort of more than four thousand mother-child dyads followed for different lengths of time in a population with a high prevalence of HIV (Brahmbhatt et al. 2002). Children of HIV+ mothers were at an elevated risk of dying during the study period. This relationship was modified by mother’s marital status: children of polygynous mothers had the greater odds of dying. The authors hypothesized that this pattern could be due to a reallocation of resources to favor children of healthy mothers. Thus, this study also provides evidence that children of polygynous marriages experience differentially poorer health outcomes than do children of monogamous unions. The Use and Interpretation of Growth Measures Having summarized these studies showing mortality and growth differences in children, it is important to outline the potential of human growth measures for appraising the quality and adequacy of the environment at individual and population levels. Clinicians and epidemiologists have worked for over half a century to develop useful anthropometric measures of human growth (Semba 2001). Over the years many different measures and indices have been constructed and used to define and understand biological processes such as “growth faltering” and “linear growth retardation” (Gibson 1990; Gomez et al. 1956; Seoane and Latham 1971; Waterlow et al. 1977; WHO Expert Committee 1995; WHO working group 1986). Many biological anthropologists have been directly involved in designing protocols for efficient collection of accurate growth data under all sorts of field conditions and in selecting appropriate growth standards against which to assess growth performance across a range of environments (Cameron 2002; Frisancho 1988, 1990; Haas and Habicht

Family Sructure and Child Growth in sub-Saharan Africa

•

159

1990; Martorell 1985, 1989; Ulijaszek 1994). Tried-and-tested protocols for measuring and evaluating growth are now widely used in applied research and public health interventions (de Onis 2001). Growth as a Measure of Past Adversity The size of children is a direct measure of past growth (Latham 1984). Growth failure and resultant small size in relation to healthy referents may not be the direct cause of the other negative outcomes (such as death and illness), but they are extremely effective markers of the environments in which those outcomes have their causal origin (Beaton 1992; WHO Expert Committee 1995). Therefore, achieved size of young children at any given age is frequently used as a proxy measure of the conditions surrounding early growth and development, that is, of the quality of childcare, nutrition, and exposure to pathogens (Box 6.2). Measurable deficits in various healthrelated outcomes are often shown to have a common origin in the adverse biological and social conditions in which early development takes place. Use of growth measures can be frustrating to research scientists because small achieved size often does not indicate the precise etiology of the growth failure (Allen et al. 1992; Pollitt and Gorman 1994). Associations between growth measures and the other health-related outcomes result both from direct causal links and from noncausal associations (Pollitt 2001). The classic example is the difficulty in teasing apart the potentially causal effects of inadequate diet and infectious disease. It is very hard to determine whether a small child has become so through undernutrition or through the effects of illness, since an infectious environment is almost always one in which child diets are of poor quality. It is also very difficult to tease apart prenatal and postnatal factors. Growth as a Measure of Current Health In populations where illness rates are relatively high and a high proportion of children and adults are relatively small for age and sex, it is likely that some of these individuals are likely to be smaller than their genetic potential (Martorell 1989). On investigation, many of the smaller children will show high incidence of illness, particularly from gastrointestinal and acute lower respiratory diseases (Martorell and Ho 1984). Growth-retarded children are more likely to be sick at any given time and to live in materially poor households (Nandy et al 2005). Even if not currently showing symptoms of infectious disease, stunted people of all ages tend to perform relatively poorly in tests of cognitive function (Black 2003; Bryan et al. 2004; Hall et al. 2001; Meyers and Chawla 2000) and physical work capacity (Diaz et al. 1991; Martorell 1995; Spurr 1987). Growth as a Measure of Future Risk Assessment of child growth is of great utility when the goal is to identify individuals at risk of functional deficit, morbidity, and mortality (Haas and Habicht 1990; Keller

160

•

Daniel W. Sellen

• • • Growth as an Indicator of Health, Risk, and Adversity Most biological anthropologists work with a conceptual framework in which poor growth (also referred to as growth retardation or growth faltering) is modeled both as an outcome of adversity in the social or physical environment of individuals and as a risk factor for poor cognitive and functional development, morbidity, and mortality. Similarly, growth according to genetic potential (that is, an absence of growth retardation) is modeled as an indicator of health. A number of feedbacks and synergisms greatly complicate relationships between child growth and three other outcomes of major interest in human biology (development, morbidity, and mortality; Figure 6.3). These outcomes do not have completely distinct causes and are often found to share a common etiology. Thus, the underlying causes of growth retardation are difficult to infer for any particular child. Achieved growth is not a simple, direct, or integrated measure of food intake in relation to requirements. Put simply, poor child growth results from undernutrition, which can be due to infection, lack of access to nutrients, or an interaction of the two (Schroeder 2001). Disease and reduced food intake have direct negative impacts on growth. Subclinical infections and reduced nutrient bioavailability and food malabsorption also play a large role in shaping growth (status and variability) and other health-related outcomes (Lunn 2000; Lunn, Northrop-Clewes, and Downes 1991; Panter-Brick et al. 2001). To summarize, growth retardation due to undernutrition is: • an outcome of adversity • a measure of current poor health • a predictor of future poor health • an indicator of risk

• • •

Figure 6.3. Relationships between environment, growth and other health outcomes

Family Sructure and Child Growth in sub-Saharan Africa

• 161

162

•

Daniel W. Sellen

and Fillmore 1983 Latham 1984; de Onis 2001). Powerful relationships exist between relatively poor growth performance and future functional deficits, morbidity, and mortality. The “growth-morbidity-function-mortality” relationship was first observed empirically in the middle of the last century and became a well-established scientific paradigm in the 1980s. Particularly strong associations exist between growth performance in the first five years of life and various indices of health and nutritional status at subsequent ages. Child growth performance has been used for over a decade as a nonspecific indicator of risk that can be used to predict other poor outcomes with great accuracy (Allen 1994; Grantham-McGregor 1993). Growth-retarded children are more likely to die from infections and other challenges (Rice et al. 2000). Current scientific consensus continues to support the concept that growth faltering may serve well as an indicator of increased risk of death. The relationship is also widely recognized in folk beliefs around the world, and probably was understood in principle since antiquity. Growth and Survival in Different Environments The precise relationship between growth and survival depends on the specific biocultural setting. Until recently, it was thought that the shape of the relationship was assumed to vary across populations, age groups, and with the type of growth measures. However, it is now understood that the use of many different definitions of undernutrition can obscure a fundamental multiplicative effect of increasing undernutrition on mortality that is very similar across populations. The amount of excess mortality caused by undernutrition in a population depends on the prevalence of different categories of undernutrition (e.g., “mild,” “moderate,” or “severe”). Furthermore, the amount of increased excess caused by given levels of undernutrition in a population depends on the “baseline” (i.e., not nutritionally related) mortality rate. In other words, the increased risk associated with undernutrition is disproportionately higher where mortality rates are higher (Pelletier 1994; Appendix 6.1). In sum, child growth performance is a very strong predictor of which individuals will live or die in human populations. Although poor growth does not directly cause child death, it is a predictor of death because of the common etiology of poor growth and a good proportion of child mortality. Poor growth predicts poor cognitive development and morbidity, and is often a factor in the direct causal pathways leading to these negative health outcomes. Hidden Risks Taken together, the results of the community-based studies reviewed here indicate that in small-scale sub-Saharan African societies with a subsistence economy and polygynous marriage system, maternal marital status is a robust predictor of children’s nutritional status. The known associations between poor growth, past adversity, present health, and future risk suggest that the relatively poor growth of some children in

Family Sructure and Child Growth in sub-Saharan Africa

•

163

polygynous marriages is not trivial. Indeed, children of some wives in polygynous unions present with significantly poorer anthropometric scores than do other children and are at higher risk of death. This pattern seems to hold across subsistence types (pastoralism, agropastoralism, and agriculture), inheritance systems (matrilineal and patrilineal), and community nutritional status (very poor to moderate). The advantages seem to accrue mainly to children of first-married wives. At least three non–mutually exclusive mechanisms might account for these complex patterns. First, it is likely that children of first wives may be protected from the negative effects of resource dilution because they are able to access more food, care, or other resources owing to their mother’s seniority in the marriage, or because of their closer proximity to their natal home, from where kin may provide livestock and help with work activities. Second, it is also possible that the older children of first wives may be past the critical age at which most growth deficits occur (birth to 3 years) before additional wives join the family. Third, only relatively wealthier men may be able to acquire multiple wives through payment of bride wealth, resulting in children of higher-order wives being buffered from the negative effects of resource dilution. The study among the Datoga revealed an underlying positive association between per capita wealth and the number of co-wives in households. If per capita wealth is linked to food intake, diet quality, and appropriate care giving, then it is likely that hidden risk of poor growth among children of later-married wives is reduced in larger polygynous families. A relationship between wealth and polygyny may arise because multiple wives increase household wealth, either because their labor contributions to animal husbandry increase livestock fertility and productivity or because gift-giving from their natal families brings more livestock into households. Further investigations are needed before we will be able to tease apart the direct effects of nutrition and infection on children’s growth from the indirect effects of household wealth and mother’s marital status. Meanwhile, the evidence strongly suggests that family wealth and organization influence the proximate determinants of child growth in complex ways mediated by social interactions and care-giver behaviors. In sum, the relationships between the social position of mothers and indicators of growth in children in small-scale polygynous societies are complex and could be explained by a variety of mechanisms that remain largely untested. Further investigation is needed to throw light on the question of whether maternal effects on child growth, development, and survival occur in response to relatively subtle distinctions in social status. But already the results in hand raise important questions about what, if any, policy implications might follow from the elucidation of such “hidden risk” apparently linked to cultural practices. Implications for Applied or Engaged Anthropologists Mother’s marital status emerges as a previously unrecognized source of variation in child growth and survival. The fact that associations between polygyny and child

164

•

Daniel W. Sellen

health have remained undiscovered until recently serves as a salutary example of how a priori perceptions of risk and adversity can influence the research process. Indeed, these remarkably consistent results demonstrate a pressing need to include information about maternal marital status routinely in health surveys and research studies in sub-Saharan Africa. The unmasking of “hidden risks” to children associated with the complex arrangements within polygynous unions is but one example of the ways in which “risk perception” by professionals can influence the “research agenda.” Other examples include the focus on homeless street children at the expense of other groups of children living in rural and/or urban poverty, with expectation of worse health outcomes attributed a priori to the homeless because of their particular family circumstances (Panter-Brick 2002, 2004). The results regarding the apparent negative effects of polygyny on child health raise difficult questions about whether the findings are sufficient to support an argument for health interventions that target those children most at risk. If so, how might they best be designed? When anthropologists detect social inequalities through biological measurements, are they are ethically bound to suggest solutions that will diminish health differentials? In this case, three major challenges must be addressed. First, the data do not show that the practice of polygyny is uniformly negative. Rather, they show that the process of formation of polygynous unions generates new axes of variation in health, risk, and adversity. Clearly, as much as polygyny is associated with poorer growth among children of later-married wives, it also results in better growth among children of firstmarried wives. Second, these gradients can be properly evaluated only against models of what would happen to children if their mothers had entered different unions; this is difficult to do. Third, much more information is required to understand the dynamics that lead to the growth differentials; without this information, it is not possible to make sensible suggestions for intervention. Since no specific “best practice” for addressing any disparities associated with family formation processes in communities with polygyny is at hand, it follows that there is at present no ethical dilemma. Relevance to Central Concepts in Biological Anthropology This review of the data currently available for assessing the association between polygyny (a set of social relations emerging from specific sets of cultural practices) and child growth (an indicator of health risk and adversity) highlights several conceptual themes that recur throughout this book. First, a focus on growth indicators as measurable outcomes provides a specific example of the need for biological anthropologists to be explicit about what their methods can and cannot achieve. Anthropometric indicators of child growth and nutritional status are simple, tried-and-tested tools for investigating the nature and direction of associations between health, risk, and adversity in contemporary popula-

Family Sructure and Child Growth in sub-Saharan Africa

•

165

tions. Nevertheless, when collected along with other more or less conventional indicators, such as measures of social relations (as in the studies discussed in this chapter) or of individual physiological status (as in many other types of study in biological anthropology), these superficially rather dry indicators can be deployed in imaginative analyses to test new and interesting hypotheses. Indeed, the use and interpretation of anthropometry is highly complex and, depending on the conceptual model being tested in a particular research design, measures of achieved growth status or growth variability can be used alternatively as predictors of current health, future risk, and past adversity. Second, developing a strong understanding of context is clearly important if future investigations are to tease out the mechanisms through which the patterns observed emerge. All studies reviewed use a design that takes a “snapshot”: of children’s growth performance or survival in relation to the structure of the family or the marital situation of mothers at the time of observation. There appear to be no studies that compare the growth of successive children born to mothers as the marital union of which they are a part develops. Longitudinal studies of growth variability that capture the household developmental cycle (Fratkin and Smith 1995; Kertzer 1986) are needed if we are to replace our “static” description of the relationship between child growth and polygyny with a “dynamic” understanding of the underlying biocultural processes in play. Third, the possibility that the formation of polygynous family structures generates variability in pathways influencing growth and increased diversity in growth outcomes among children suggests the limitations of analyses of health and adversity that assume homogeneity within social groupings such as families. I suggest that in communities with any kind of complex or extended family structure, whether they are linked to polygyny or not, the position of individuals within families should be included as a key variable in analyses of variability in health, risk, and adversity. This argument is analogous to examining gender as a risk factor associated with nutritional health (Griffiths, Matthews, and Hinde 2002; Travers 1996). Fourth, in raising new questions about what to do about new information on hidden risk, emerging evidence for a negative relationship between polygyny and child growth highlights the challenge of risk communication. On present evidence, it is just as meaningful to infer that polygyny tends to benefit some children in families more than other forms of parenting as it is that some children suffer increased adversity. Certainly the pattern seems consistent and widespread, and has been reported for every community in which it has been investigated. Indications are that polygyny is an important predictor of health for some children when compared to others in the family at certain times, and there is a strong case for further investigation to understand how this comes about. But it is not correct to state that the available evidence suggest that polygnous unions are detrimental to all children they produce, nor that polygyny does not offer substantial benefits to some parents and children in comparison to other forms of family formation.

166

•

Daniel W. Sellen

Fifth, my argument that non-perception or perception of risk drives our research agenda applies to many arenas of research. The data reviewed here suggest that a specific case of “social relationships” (mother’s status within polygynous marriages) should be reinterpreted as a marker of being “at risk” for poor health. An interesting counterpoint is provided by recent studies of homeless street children in the developing world. Although many health professionals assume or expect that such children will present with the poorest health outcomes relative to children living at home, particularly for growth and nutritional status, the evidence from anthropological research warns of the fallacy of a priori categorization: it depends on what kind of home the children live in (Panter-Brick 2004). Such examples illustrate the difficulties inherent in risk narratives (see PanterBrick and Fuentes, Introduction; Herring, Chapter 3). Risk narratives, built on assumptions that health and adversity are independent of social contexts or practices, need to be validated with empirical observations. Returning to the focus of this chapter, polygyny—a social “variable” currently absent from risk narratives—is often thought to be of little systematic relevance to maternal and child health and is therefore not commonly included in analyses of health disparities in sub-Saharan Africa. It is likely to be incorporated into new risk narratives if the results of the selected studies reviewed here are replicated more widely. At that point, marriage formation patterns will no longer be a source of “hidden” risks to children. Summary and Conclusions In this chapter I have reviewed examples of research to investigate the relationships between adversity, risk, and health outcomes in several communities where polygynous marriage is common. All of this research has been designed and conducted by anthropologists who have applied a biocultural approach and tested the assumption that polygynous marriage has little relevance for understanding health and adversity in sub-Saharan Africa. A majority of these studies reveal mother’s marital status as a previously unrecognized source of variation in child growth and survival. I have also discussed how biological anthropologists interpret simple indicators of growth performance as integrated measures of child health. Discussion focused on child growth as general tool for investigating health outcomes and to discover unanticipated risks to children. Despite their lack of causal specificity, growth measures are simpler to use and easier to interpret than some more recently developed biochemical indicators of health. Measurement of growth therefore remains useful for studying complex pathways linking social adversity to undernutrition and related poor health outcomes. A focus on child growth and on “hidden risk” together illustrate the potential of a contextual view from biological anthropology in understanding the intergenerational links between risk, adversity, and long-term health in contemporary populations.

Family Sructure and Child Growth in sub-Saharan Africa

•

167

Appendix Poor Growth and Risk of Death Two Examples Gomez and colleagues (Gomez et al. 1956) were the first to construct a classification of anthropometric status that explicitly included estimates of risk. In a hospital-based sample of Guatemalan children, those admitted with “Gomez 3” levels of undernutrition (weight-for-age below 70 percent of Boston standards) suffered a 33 percent mortality rate. Over the next three decades a similar pattern of lower mortality risks among children presenting with normal, mild, and moderate undernutrition but a markedly increased mortality risk for the severely undernourished was commonly reported (Kielman and McCord 1978; Sommer and Loewenstein 1975). Differences in the actual mortality rates suggested that different mechanistic interactions were at play (Chen, Chowdhury, and Huffman 1980, 1981). In a paper designated by UNICEF as the most important of the decade, biological anthropologist David Pelletier and co-workers used meta-analysis to examine the shape of the relationships between various anthropometric indicators of child growth performance and risk of mortality (Pelletier 1994). This showed that: • most of the apparent differences in the growth-mortality relationship across populations were artifacts of widely different non-nutritionally attributable mortality rates and length of follow-up and disparate definitions of undernutrition; • after standardizing the definitions of mild, moderate, and severe undernutrition across a large number of studies, the risk of death increases exponentially with decreasing anthropometric status; • this is true whether anthropometric status is measured using arm circumference, weight-for-age, height-for-age, or weight-for-height (although the standardized rates of increased risk differ between the measures); and • log-standardized plots of mortality rates and prevalence of undernutrition have similar slopes, indicating that the shape of the underlying growth-mortality function is similar across populations. Pelletier and coworkers (Pelletier et al. 1994a,b) went on to estimate what proportion of deaths within various populations are statistically attributable to various forms of undernutrition. They noted that: • individuals suffering from severe forms of undernutrition are at significantly higher per capita risk of death but are also relatively rare in most populations; • as a consequence, most of the observed deaths attributable to undernutrition occur when a child is defined anthropometrically as mildly or moderately malnourished; and

168

•

Daniel W. Sellen

• the proportion of all deaths attributable to the potentiating (synergistic, multiplicative) effects of undernutrition is therefore fairly constant across populations both within and between countries. This proportion excludes all genetic and non-nutritional environmental components of mortality and is referred to as the “population attributable risk” (PAR). Recent estimates of the PAR of undernutrition range up to 60 percent of child deaths. Of course, the absolute levels of child mortality vary widely among populations and across age groups of children within populations due to background factors such as the burden of infectious disease.

References Allen, L.H. 1994. Nutritional influences on linear growth: a general review. European Journal of Clinical Nutrition 48 (Suppl. 1): S75–89. Allen, L.H., J.R. Backstrand, E.J. Stanek, G.H. Pelto, et al. 1992. The interactive effects of dietary quality on the growth and attained size of young Mexican children. American Journal of Clinical Nutrition 56: 353–64. Beaton, G.H. 1992. Nutrition research in human biology: changing perspectives and interpretations. American Journal of Human Biology 4: 159–77. Becker, G.S. 1981. A Treatise on the Family. Cambridge, MA: Harvard University Press. Begin, F., E.A. Frongillo Jr., and H. Delisle. 1999. Caregiver behaviors and resources influence child height-for-age in rural Chad. The Journal of Nutrition 129: 680–86. Begin, F., J.-P. Habicht, E.A. Frongillo, and H. Delisle. 1997. The Deterioration in children’s nutritional status in rural Chad: the effect of mothers’ influence on feeding. American Journal of Public Health 87: 1356–59. Berman, P., C. Kendall, and K. Bhattacharyya. 1994. The household production of health: integrating social science perspectives and micro-level health determinants. Social Science and Medicine 38: 205–15. Black, M.M. 2003. Micronutrient deficiencies and cognitive functioning. The Journal of Nutrition 133: 3927S–31S. Borgerhoff Mulder, M. 1991a. Datoga pastoralists of Tanzania. National Geographic Research and Exploration 72: 166–87. ———. 1991b. Economic differentiation among Datoga pastoralists of Tanzania. National Geographic Research and Exploration 72: 166–87. ———. 1991c. Behavioural ecology of humans: studies of foraging and reproduction. In Behavioural ecology, 3d ed., eds. J.R. Krebs and N.B. Davies. Oxford: Blackwell Scientific Publications, 69–98 Brabin, L. 1984. Polygyny: an indicator of nutritional stress in African agricultural societies. Africa 54: 31–45. Brahmbhatt, H., D. Bishai, F. Wabwire-Mangen, G. Kigozi, et al. 2002. Polygyny, maternal HIV status and child survival: Rakai, Uganda. Social Science & Medicine 55: 585–92. Brown, K.H. and N.Z. Solomons. 1991. Nutritional problems of developing countries. Infectious Disease Clinics of North America 5: 297–317.

Family Sructure and Child Growth in sub-Saharan Africa

•

169

Bryan, J., S. Osendarp, D. Hughes, E. Calvaresi, et al. 2004. Nutrients for cognitive development in school-aged children. Nutrition Reviews 62: 295–306. Burman, B. and G. Margolin. 1992. Analysis of the association between marital relationships and health-problems—an interactional perspective. Psychological Bulletin 112: 39–63. Cameron, N. 2002. Human Growth and Development. Amsterdam: Academic Press. Chen, L.C., A.K.M.A. Chowdhury, and S.A. Huffman. 1979. Seasonal dimensions of energy protein malnutrition in rural Bangladesh: The role of agriculture, dietary practices, and infection. Ecology of Food and Nutrition 8: 175–87. ———. 1980. Anthropometric assessment of energy-protein malnutrition and subsequent risk of mortality among pre-school children. American Journal of Clinical Nutrition 33: 1836–45. ———. 1981. The use of anthropometry for nutritional surveillance in mortality control programs. American Journal of Clinical Nutrition 34: 2596–99. de Onis, M. 2001. Growth and development. In Nutrition and Health in Developing Countries, eds. R.D. Semba and M.W. Bloem. Totowa, NJ: Humana Press, 71–92. Department of Health Education and Welfare. 1977. NCHS Tables of growth, birth–18 years, United States. Vol. 124. Series 11. Washington, D.C.: Government Printing Office. Desai, S. 1995. When are children from large families disadvantaged? Evidence from crossnational analyses. Population Studies 49: 195–210. Diaz, E., G.R. Goldberg, M. Taylor, M. Savage, et al. 1991. Effects of dietary supplementation on work performance in Gambian labourers. American Journal of Clinical Nutrition 53: 803–11. Fratkin, E., and K. Smith. 1995. Women’s changing economic roles with pastoral sedentarization: varying strategies in alternate rendille communities. Human Ecology 23: 433–53. Frisancho, A.R. 1988. Nutritional anthropometry. Journal of the American Dietetic Association 88: 553–55. ———. 1990. Anthropometric Standards for the Assessment of Growth and Nutritional Status. Ann Arbor: University of Michigan Press. Gibson, R.S. 1990. Principles of Nutritional Assessment. New York: Oxford University Press. Gillett-Netting, R. and A. Perry. 2005. Gender and nutritional status at the household level among Gwembe Valley Tonga children, 0–10 years. American Journal of Human Biology 17: 372–75. Gomez, F., R.R. Galvan, S. Frenk, J. C. Munoz, et al. 1956. Mortality in second and third degree malnutrition. Journal of Tropical Pediatrics 2: 77–83. Grantham-McGregor, S.M. 1993. Assessments of the effects of nutrition on mental development and behavior in Jamaican studies. American Journal of Clinical Nutrition Supplement 57: 303S–309S. Griffiths, P., Z. Matthews, and A. Hinde. 2002. Gender, family, and the nutritional status of children in three culturally contrasting states of India. Social Science & Medicine 55: 775–90. Haas, J.D. and J.-P. Habicht. 1990. Growth and growth charts in the assessment of pre-school nutritional status. In Diet and Disease in Traditional and Developing Societies, vol. 30, Symposia of the Society for the Study of Human Biology, eds. G.A. Harrison and J.C. Waterlow. Cambridge: Cambridge University Press, 160–83 Hadley, C. 2005. Is polygyny a risk factor for poor growth performance among tanzanian agropastoralists? American Journal of Physical Anthropology 126: 471–80.

170

•

Daniel W. Sellen

Hall, A., L.N. B. Khanh, T.H. Son, N.Q. Dung, et al. 2001. An association between chronic undernutrition and educational test scores in Vietnamese children. European Journal of Clinical Nutrition 55: 801–4. Harkness, S. and C.M. Super. 1994. The developmental niche: a theoretical framework for analyzing the household production of health. Social Science and Medicine 38: 217–26. Horwitz, A.V., H.R. White, and S. Howell-White. 1996. Becoming married and mental health: A longitudinal study of a cohort of young adults. Journal of Marriage and the Family 58: 895–907. Kayongo-Male, D. and P. Onyango. 1984. The Sociology of the African Family. London: Longman. Keller, W. and C.M. Fillmore. 1983. Prevalence of protein-energy malnutrition. World Health Statistics Quarterly 36: 129–66. Kertzer, D.I. 1986. A life course approach to coresidence. Current Perspectives on Aging and the Life Cycle 2: 1–22. Kielman, A.A. and C. McCord. 1978. Weight-for-age as an index of risk of death in children. Lancet 10: 1247–50. Klima, G. 1964. Jural relations between the sexes among the Barabaig. Africa 34: 9–19. ———. 1965. Building up a herd. In The Barabaig East: African Catttle Herders. Prospects Heights, IL: Waveland Press, 17–33 Lane, C. 1996. Pastures Lost: Barabaig Economy, Resource Tenure, and the Alienation of Their Land in Tanzania. Nairobi: Initiatives Publishers. Latham, M.C. 1984. Strategies for the control of malnutrition and the influence of the nutritional sciences. Food and Nutrition Bulletin 10: 5–32. Lunn, P.G. 2000. The impact of infection and nutrition on gut function and growth in childhood. Proceedings of the Nutrition Society 59: 147–54. Lunn, P.G., C.A. Northrop-Clewes, and R.M. Downes. 1991. Intestinal permeability, mucosal injury, and growth faltering in Gambian infants. Lancet 338: 907–10. Markey, C.N., P.M. Markey, C. Schneider, and S. Brownlee. 2005. Marital status and health beliefs: different relations for men and women. Sex Roles 53: 443–51. Martorell, R. 1985. Child growth retardation: a discussion of its causes and its relationship to health. In Nutritional Adaptation in Man, eds. K. Baxter and C. Waterlow. London: John Libbey, 13–30. ———. 1989. Body size, adaptation and function. Human Organization 48: 15–20. ———. 1995. Results and implications of the INCAP follow-up study. Journal of Nutrition 125: 1127S–38S. Martorell, R. and T.J. Ho. 1984. Malnutrition, morbidity, and mortality. In Child Survival: Strategies for Research, vol. 10, eds. W.H. Mosley and L.C. Chen. Cambridge: Cambridge University Press, 49–68 Meyers, A., and N. Chawla. 2000. Nutrition and the social, emotional, and cognitive development of infants and young children. In Zero to Three. Washington, DC: National Center for Infants, Toddlers and Families, 5–14. Murdoch, G.P. 1965. Ethnographic Atlas. Pittsburgh: University of Pittsburgh Press. Nandy, S., M. Irving, D. Gordon, S.V. Subramanian, and G.D. Smith GD. 2005. Poverty, child undernutrition and morbidity: new evidence from India. Bulletin of the World Health Organization 83: 210–16

Family Sructure and Child Growth in sub-Saharan Africa

•

171

Orians, G. 1969. On the evolution of mating systems in birds and mammals. American Naturalist 103: 589–603. Panter-Brick, C. 2002. Street children, human rights, and public health: a critique and future directions. Annual Review of Anthropology 31: 147–71. ———. 2004. Homelessness, poverty, and risks to health: beyond at risk categorizations of street children. Children’s Geographies 2: 83–94. Panter-Brick, C., P.G. Lunn, R. Baker, and A. Todd. 2001. Elevated acute–phase protein in stunted Nepali children reporting low morbidity: different rural and urban profiles. British Journal of Nutrition 85: 125–31. Pelletier, D. 1994. The relationship between child anthropometry and mortality in developing countries: implications for policy, programs and future research. Journal of Nutrition 124: 2047S–81S. Pelletier, D., E.A. Frongillo, D.G. Schroeder, and J.-P. Habicht. 1994a. A methodology for estimating the contribution of child mortality in developing countries. Journal of Nutrition 124: 2106S–22S. Pelletier, D., J.W. Low, F.C. Johnson, and L.A.H. Msukwa. 1994b. Child anthropometry and mortality in Malawi: testing for effect modification by age and length of follow-up and confounding by socioeconomic factors. Journal of Nutrition 124: 2082S–2105S. Pollitt, E. 2001. Statistical and psychobiological significance in developmental research. American Journal of Clinical Nutrition 74: 281–82. Pollitt, E. and K. Gorman. 1994. Nutritional deficiencies as developmental risk factors. In Threats to Optimal Development: Integrating Biological, Psychological and Social Risk Factors, vol. 27, Minnesota Symposia on Child Psychology, ed. C. Nelson. Prospect Heights, NJ: Lawrence Erlbaum Associates, 121–44. Rice, A.L., L. Sacco, A. Hyder, and R. Black. 2000. Malnutrition as an underlying cause of childhood deaths associated with infectious diseases in developing countries. Bulletin of the World Health Organisation 78: 1207–21 Schroeder, D.G. 2001. Malnutrition. In Nutrition and Health in Developing Countries, eds. R.D. Semba and M.W. Bloem. Totowa, NJ: Humana Press, 393–426. Sellen, D.W. 1995. The socioecology of young child growth among the Datoga pastoralists of northern Tanzania. Ph.D. thesis, University of California. ———. 1998. Infant and young child feeding practices among African pastoralists: The Datoga of Tanzania. Journal of Biosocial Science 30:481-99. ———. 1999. Polygyny and child growth in a traditional pastoral society: the case of the Datoga of Tanzania. Human Nature: An Interdisciplinary Biosocial Journal 10: 329–71. ———. 2000a. Age, sex and anthropometric status of children in an African pastoral community. Annals of Human Biology. 27: 345–65. ———. 2000b. Seasonal ecology and nutritional status of women and children in a Tanzanian pastoral community. American Journal of Human Biology 12: 758–81. ———. 2003. Nutritional consequences of wealth differentials among pastoral Datoga households in northern Tanzania. Human Ecology 31: 529–70. Sellen, D.W., M. Borgerhoff Mulder, and D.F. Sieff. 2000. Fertility, offspring quality and wealth in Datoga pastoralists: testing evolutionary models of intersexual selection. In Adaptation and Human Behavior: An Anthropological Perspective, eds. L. Cronk, N. Chagnon, and W. Irons. Hawthorne, NY: Aldine de Gruyter, 87–110.

172

•

Daniel W. Sellen

Semba, R.D. 2001. Nutrition and development: historical perspective. In Nutrition and Health in Developing Countries, eds.R.D. Semba and M.W. Bloem. Totowa, NJ: Humana Press, 1–30. Seoane, N. and M.C. Latham. 1971. Nutritional anthropometry in the identification of malnutrition in childhood. Journal of Tropical Pediatrics and African Child Health 17: 98–104. Sieff, D.F. 1995. The effects of resource availability on the subsistence strategies of Datoga pastoralists of North West Tanzania. Ph.D. thesis, University of Oxford. ———. 1997. Herding strategies of the Datoga pastoralists of Tanzania: is household labor a limiting factor? Human Ecology 25: 519–44. ———. 1999. The effects of wealth on livestock dynamics among the Datoga pastoralists of Tanzania. Agricultural Systems 59: 1–25. Sommer, A. and M.S. Loewenstein. 1975. Nutritional status and mortality: a prospective validation of the QUAC stick. American Journal of Clinical Nutrition 28: 287–92. Spurr, G.B. 1987. Effects of chronic energy deficiency on stature, work capacity and productivity. In Chronic Energy Deficiency: Consequences and Related Issues, eds. B. Schürch and N.S. Scrimshaw. Geneva: Nestle/ IDECG (International Dietary Energy Consultant Group), 95–134. Tanaka, S. 1969. Natural environments of Mangola. Kyoto University African Studies 3: 1. Timaeus, I.M., and A. Reynar. 1998. Polygynists and their wives in sub-Saharan Africa: an analysis of five demographic and health surveys. Population Studies-a Journal of Demography 52: 145–62. Tomikawa, M. 1970. The distribution and migrations of the Datoga tribe: the sociological distinction of the Datoga society in the Mangola area. Kyoto University African Studies 5: 1–46. ———. 1972. Cattle brands of the Datoga tribe: human relations in the the Datoga pastoral society in East Africa. Kyoto University African Studies 7: 1–35. ———. 1978. Family and daily life: an ethnography of the Datoga pastoralists of Mangola. Senri Ethological Studies 1: 1–36. ———. 1979. The migrations and inter-tribal relations of the pastoral Datoga. Senri Ethological Studies 5: 1–46. Travers, K.D. 1996. The social organization of nutritional inequities. Social Science & Medicine 43: 543–53. Ulijaszek, S. 1994. Between-population variation in pre-adolescent growth. European Journal of Clinical Nutrition 48 (Suppl. 1): S5–S14. Waterlow, J.C., R. Buzina, W. Keller, J.M. Lane, et al. 1977. The presentation and use of height and weight data for comparing the nutritional status of groups of children under the age of 10 years. Bulletin of the World Health Organisation 55: 489–98. Westoff, C. 2003. Trends in Marriage and Early Childbearing in Developing Countries. Vol. 5. DHS Comparative Reports. Calverton, MD: ORC Macro. WHO Expert Committee. 1995. Physical Status: The Use and Interpretation of Anthropometry. Vol. 854. World Health Organization Technical Report. Geneva: WHO. WHO working group. 1986. Use and interpretation of anthropometric indicators of nutritional status. Bulletin of the World Health Organisation 64: 929–41.

PART III

• • • Gene Evolution, Environment, and Health

• • • Explaining Health Inequalities William W. Dressler

The Concept of Health Inequalities The three chapters in this section, by Keith Godfrey and Mark Hanson, Thom McDade, and Lorena Madrigal and her associates, address a set of related topics that converge on a single major issue in the study of adult chronic disease: how do we understand the social and cultural foundations of health inequalities? The term health inequalities is itself a controversial one. In the United States, the term health disparities is used. Why that is the case is obscure, but there is some evidence that Americans, especially some members of the public health establishment, are uncomfortable with the notion that inequality might be a part of the society, hence the more sanitized term. Whichever term is used, it refers to enduring differences in health status among categories of persons, such that a particular category of person is disadvantaged with respect to health. The term, as far as I can tell, comes from the famous Black Report (Townsend and Davidson 1982). The Black Report (so named for Sir Douglas Black, who chaired the commission) was produced by a panel formed to examine the social distribution of morbidity and mortality in the United Kingdom some thirty years after the establishment of the National Health Service (NHS). There were some interesting political shenanigans involved in its publication, but suffice it to say that the report documented widespread inequalities in health in the UK that were unaddressed by the NHS. But why use a severe term like inequality to describe differences in health status? In so far as epidemiology is all about the documentation of how health is distributed, and as some people are bound to be healthier than others, inequality might seem a rather tendentious way of describing such a distribution. The term was employed

176

•

William W. Dressler

precisely because the distribution of ill-health and higher mortality maps on to measures of social inequality, especially social class and region. Persons lower in the social-class hierarchy (as measured by occupational status) consistently had higher rates of morbidity and mortality, as did persons living in regions in Britain in which levels of material deprivation (i.e., poverty) were consistently higher than in other regions. Because these differences corresponded so clearly to well-recognized socioeconomic disparities, they were termed health inequalities. The study of health inequalities has become a major issue in the international public health establishment (Dressler, Oths, and Gravlee 2005). Much of the research on health inequalities is devoted to the ever-more-precise documentation of those disparities. The chapters in this section, on the other hand, represent efforts to explain those disparities. Each offers new insights into the issue. Predictive Adaptive Responses and Adult Chronic Disease It is convenient to start with the chapter by Keith Godfrey and Mark Hanson, since these authors locate the genesis of health inequalities in the womb. They report on a program of research that was set in motion by the descriptive epidemiological studies of Barker (1991). It was Barker who first identified the link between material deprivation early in life and later risk of adult chronic disease. He did so by examining ecological correlations between rates of low birth weight and adult chronic disease (especially cardiovascular disease) across regions in Britain. Later studies confirmed the same associations in other parts of the world. Some argued that these associations were consistent with an overall genetic predisposition to poor health that aggregated by family and was passed from mother to child, but others placed a greater emphasis on the environment of the mother and the fetus. Using a combination of human and animal studies, Godfrey, Hanson, and their associates have shown that an underlying material deprivation creates an adverse environment to which the fetus must respond and the major feature of that environment at the outset is the mother. That is, if there is an adverse environment for the fetus created by poor maternal nutrition (and other factors as well, as I will argue below), setting the stage for poor fetal growth and development and subsequent low birth weight, there are adaptive responses that the fetus itself can undertake. Using the phrase “predictive adaptive response,” or PAR, Godfrey and associates suggest that the fetus reshapes its own development in response to the maternal environment in such a way that it seems to be anticipating the nature of the environment into which it will be born. As they show, this can take the form of differential support for cell growth and differentiation. Much as in the famous notion of the “thrifty genotype,” the fetus develops so as to conserve the storage and utilization of certain critical nutrients for its own survival. Oftentimes, however, this thrifty fetus has, as it were, guessed wrong about the world into which it will be born. Far from a world characterized by the kinds of depri-

Explaining Health Inequalities

•

177

vations it experienced in utero, it has been born into a world of a caloric surfeit where rates of childhood obesity are on the rise. The child meets, in other words, a world for which it is ill-suited physiologically and metabolically. Early childhood obesity and other stresses lead to a path of adult development that includes an enhanced risk of adult obesity, diabetes, hyperlipidemia, and hypertension. In short, the metabolic syndrome develops, and it places individuals at higher risk of coronary heart disease. These conditions in turn can compromise a pregnancy in an adult woman who was herself low birth weight, and the cycle continues. So, the PAR model can help to explain health inequalities on the basis of higher risk of poor maternal nutrition and low birth weight created by socioeconomic inequalities. Health Inequalities Within and Between Populations of African Descent Lorena Madrigal and her associates also address the issue of health inequalities, but this time ethnic or racial health inequalities. The chapter again confronts questions regarding the genesis of those inequalities. Racial and ethnic health inequalities have become an interesting focus in the study of health inequalities, because they allow for a very clear differentiation of causal models to explain disparities. Even though modern molecular biology has all but consigned the biological concept of “race” to the intellectual waste bin, racial-ethnic health inequalities have lurked in the background of discussions because, with enduring differences in hard-nosed outcomes like blood pressure across groups varying in obvious phenotypic characteristics like skin color, even trained public health researchers (not to mention the general public) succumb to the notion that something of a genetic nature “must” underlie these disparities. This has been compounded by the so-called “slavery hypothesis” that suggests that the infamous “Middle Passage” by which slaves were transported to the New World served as a selective mechanism for salt retention, thus providing a racial-genetic foundation for rates of high blood pressure among people of African descent in the Western hemisphere. As we have argued elsewhere, however (Dressler, Oths, and Gravlee 2005), racial-ethnic health inequalities are consistent with several alternative interpretations, such as group differences in health behaviors (e.g., smoking, alcohol use, exercise patterns), and the psychological and social stresses engendered by living in a racially stratified society. Racial-ethnic health disparities represent a special kind of research problem. The stress model to explain health disparities uses as a part of its foundation the notion that darker skin color is considered a mark of social inferiority in racially stratified societies; if an examination of the evidence associating skin color and blood pressure is then restricted to a single society, there is no valid comparison group for assessing how the skin color/blood pressure association behaves under conditions in which darker skin color is not devalued. This problem demands that these associations be examined cross-culturally, which is precisely what Madrigal et al. do (an approach

178

•

William W. Dressler

that was pioneered by Henry and Cassel 1969). Madrigal and her associates carefully collate and compare the distributions of hypertension prevalence rates among persons of African descent across different studies conducted in North America, Europe, the Caribbean, South America, and Africa. Such a comparison provides the opportunity to assess differing assumptions about the nature of the race-ethnicity/blood pressure association. If one assumes a strong racial-genetic component to blood pressure, one would expect to see similar rates of hypertension everywhere. If one assumes that the slavery hypothesis is true, one would expect to see higher rates of hypertension among persons of African descent in all of the Western hemisphere. If, on the other hand, one assume that racial stratification is important, one would anticipate an eastwest gradient, with the lowest prevalence rates in Africa, intermediate rates in the Caribbean, and highest rates in North America (see Cooper et al. 1997). Generally, the results are consistent with the latter assumption; this is bolstered by the heterogeneity of prevalence rates for studies conducted in South America (see below). So, like Godfrey and Hanson, Madrigal and her associates take a set of health inequalities and suggest how a social and cultural context may underlie what appears to be a purely biological process. As they show, the heterogeneity of prevalence rates among populations of African descent is consistent, at least, with the hypothesis that any racial-genetic component to hypertension is profoundly altered by environmental influences, and that these environmental influences include both social structure and the way in which the very concept of “race” is culturally constructed. Health Inequalities and the Socioeconomic Gradient of Health The third chapter in this section, by Thomas McDade, addresses the question of health inequalities much more directly; indeed, McDade frames his chapter in these terms. Part of his aim is to develop a framework for applying measures of socioeconomic status in such a way as to enhance our understanding of both socioeconomic and racial-ethnic health inequalities. McDade notes that, typically, measures of socioeconomic status, whether examined in themselves in relation to health or used as covariates to try to explain racial-ethnic health disparities, are fairly blunt instruments. That is, they provide very little information relevant to an understanding of social processes involved. He then argues that by taking social theory and ethnographic observation more seriously, hypotheses can be generated that employ conventional measures of socioeconomic status but which combine them in new ways. These hypotheses, if confirmed, can provide more insight into the nature of the social production of disease. McDade illustrates this first with his work from Samoa, where he examined immune function among adolescents. The study of the health effects of social change has been a productive mode of inquiry for many years in anthropology and the public health sciences, and much of this work has been carried out on islands like Samoa in the Pacific. Of course, Samoa has been subject to outside influences for generations,

Explaining Health Inequalities

•

179

starting with the first arrival of European colonizers in the seventeenth century and including the early influence of Christian missions. But the pace of change increased substantially with the globalizing forces set in motion by the Second World War. At a scale never before seen, capitalist labor and consumption markets penetrated Samoan society, such that researchers could, within the region making up Samoa, order communities along a continuum: at one end people followed a traditional horticultural way of life, with their social relationships centered on the village and the extended family while at the other end individuals had entered into wage labor, with social relationships dominated by the nuclear family and non-kin relationships. Samoan society was characterized by an incipient social stratification before the arrival of Europeans. This status ranking was organized by ranked kin groups and the distribution of authority in a series of chiefly statuses known as matai. As modernization has proceeded, this traditional system of statuses has been not supplanted but rather supplemented by new forms of prestige, such forms often taking the shape of consumption of Western material goods and information. Of course, in any developing context, households are going to differ substantially in their capacity to engage in the consumption of Western goods and information. Furthermore, they will also differ in their ranking along traditional dimensions of status. This sets the stage for what, in sociological theory, has been referred to as status incongruence, or the inconsistent ranking of a household on two dimensions of status (Dressler 2004). Theory predicts that status incongruence is a stressful experience; ethnographic sensibilities suggest this, as well. The matai are respected as wise and knowing; so if they do not know of things Western, how can they be so? Similarly, people who have come to dominate Western goods and information must see themselves as having achieved something of note yet what if they have not achieved the status of matai? Are they not simply pretenders to higher status? McDade suggests that these effects can reverberate through the household to affect adolescents, even though they themselves may not be directly involved in consumption activities, and that these effects can cascade down to the cellular level, affecting cell-mediated immune status. In fact, McDade found that adolescents who live in either kind of status-incongruent household do have compromised immune function. Thus, instead of suggesting that different dimensions of status add up to one overall status level, McDade suggests that statuses interact to allocate poor health where one might not at first anticipate it. He then goes on to apply the same reasoning to a very different cultural context, that of African-Americans in the United States. McDade formulates a model to test, using standard measures of socioeconomic status, what I referred to above as a racial-stratification hypothesis. One of the most widely-replicated findings in epidemiology is the inverse association between measures of social stratification and health status (Kaplan 1995); furthermore, numerous studies have shown that this association takes the form of a gradient. As social status increases, the risk of poor health declines by whatever measure, and there is no point along the gradient at which there is a sharp break. That is, people do not reach a certain level

180

•

William W. Dressler

of wealth or education and suddenly cease to be ill. The improvements in health status can be seen virtually at any point on the continuum. McDade suggests that, in American society, given its history of social stratification by race, darker skin color is devalued and skin color becomes a kind of “master status” that overrides the beneficial effect of increasing socioeconomic status. Key to the test of this hypothesis is that the author discards between-group, “racial” comparisons in favor of an intra-ethnic comparison, looking only at the differential effect of income on blood pressure among lighter-skinned and darker-skinned African-Americans. He finds that there is a substantial decline of blood pressure with increasing income for lighter-skinned, but not for darker-skinned, African-Americans. These results are thus consistent with the racialstratification hypothesis and they reduce confidence in other, competing hypotheses. Discussion The foregoing has been a brief summary of the work of each of these authors, but with an emphasis on where and how the chapters converge. Fundamentally, all of the chapters address the question of health inequalities, and all of them emphasize the causal factors in the social environment for generating those inequalities. For Godfrey and Hanson, the emphasis is on how predictive adaptive responses on the part of the growing fetus mediate the effects of early-life material deprivation on adult chronic disease. For Madrigal et al., the emphasis is on bringing a diverse set of data to bear on the question of the slavery hypothesis to explain racial and ethnic health disparities, versus hypotheses that take more completely into account the social processes that serve as the foundation for racial and ethnic inequalities. And, finally, McDade picks up the argument where Madrigal et al. leave off, showing how careful attention to social factors within specific cultural contexts can lead to the formulation of novel hypotheses to account for racial-ethnic health inequalities. To be sure, there are substantial differences among these chapters as well, not the least of which are the levels of explanation sought, from the cellular (Godfrey and Hanson) to the sociocultural (Madrigal et al.; McDade). Nevertheless, a common focus on how social factors “get under the skin” (in the often-used phrase) joins these chapters. It is precisely this focus that will need greater explication in future work in this area. Before I proceed, let me emphasize that my comments here are not intended to be of the “this-is-what-I-would-have-written-had-I-written-these-papers” type. These are all fine presentations and do not need me to offer improvement. Rather, what I will offer are some thoughts regarding what I see to be a useful direction in which this work can be extended. Overall, the direction that I see as most important is a better definition and understanding of the social and cultural context that serves as the foundation for the health inequalities described by these authors. This need can be seen perhaps most clearly in the chapters by Godfrey and Hanson and by Madrigal, et al. Again, Godfrey and Hanson were not trying to describe

Explaining Health Inequalities

•

181

that sociocultural context, emphasizing instead their research on the predictive adaptive response as a mediator of disease, but, nevertheless, that context needs explication. In their presentation, they place the greatest emphasis on maternal nutrition, which, to be sure, is of vital importance in the growth and development of the fetus. At the same time, there has been a considerable amount of energy expended on understanding how variation in social factors surrounding a pregnancy can influence fetal growth and development. In the United States, there is a small but interesting literature on psychosocial stressors that influence birth weight, much of which is disappointing empirically; that is, the results have not been strong. Oths, Dunn, and Palmer (2001) argue that the main problems in that literature have been an overreliance on dichotomous outcomes (i.e., low birth weight vs. “normal” birth weight), weak operational definitions of the stressors, and a reliance on a single measurement of stressors during pregnancy. They conducted a study of job stressors among lowincome women in the rural southern United States. This is interesting right from the start because many people ethnocentrically assume that low-income women do not work (an attitude that Oths et al. encountered in their research) when in fact they do, and moreover they perform work that can be quite arduous. Oths et al. developed measures of job stressors by ethnographically adapting existing measures rather than assuming that those measures would be appropriate for this particular cultural context, and they measured stressors at two points in time during the pregnancy. Finally, they used actual birth weight as the outcome variable, rather than reducing the sensitivity of the measure using a clinical definition of low birth weight. They found that job stressors measured in the first trimester (but not the third trimester) of pregnancy had a profound impact on birth weight, and that this effect was much greater for African-American as against European American women. These results suggest that the nature of the sociocultural context is important in shaping the environment to which the growing fetus adapts. The chapter by Madrigal et al. also points toward the need to better explicate the sociocultural environment within which health inequalities are generated. As this chapter makes clear, so-called “racial” phenotypes (i.e., skin color) are culture masquerading as biology. The risk of a person of African ancestry developing high blood pressure is a function of where the person lives, not of his or her African ancestry. As I noted earlier, there have been a number of different models proposed to account for the risk of high blood pressure associated with darker skin color, and the variables examined range from those that emphasize a purely mechanical causation (e.g., caloric hyper-intake coupled with low physical activity) to an emphasis on the ways in which “race” is culturally constructed as a category and how the meanings associated with race can generate stresses in everyday social interaction (Dressler, Oths, and Gravlee 2005).Thus, there is no shortage of models out there to be used to examine these health inequalities from a sociocultural standpoint. What I think the Madrigal et al. chapter most adds to this literature is their finding regarding the heterogeneity of prevalence estimates of hypertension in African-

182

•

William W. Dressler

descent communities in Latin America. The research on racial-ethnic health inequalities has been dominated by studies conducted in the United States and the Caribbean (and, to a certain extent, Britain). This despite the fact that perhaps as much as two-thirds of the African slave trade went to Latin America, especially Brazil (Curtin 1969). What makes the matter interesting is that the societies of Latin America have different histories with respect to the cultural construction of racial and ethnic differences than do English- and French-speaking societies. Communities of African descent have very different histories within and between these societies. As one small example, there are in Brazil today communities that developed from quilombos, or communities of escaped slaves. At times in Brazilian history, slave-owners decided that it was simply too much trouble to try to recapture escaped slaves, and as a result these communities were left to develop in isolation from the larger society. Recent research in one of these communities documents a very low prevalence of high blood pressure and virtually no rise of blood pressure with age (Jardim, Carneiro, Carneiro, and Baiocchi 1992). At the same time, in other parts of Brazil, the same health inequalities by skin color can be observed, although the magnitude of the difference can be moderated by other cultural factors (Dressler, Balieiro, and Santos 1999). My point is that, given this different portrait with respect to culture history and contemporary social structure, the African-descent communities in Latin American offer a good opportunity to better understand how the association of skin color and blood pressure is shaped by a sociocultural matrix. Finally, McDade has, to a large extent, already done what I am suggesting in taking at least one of the next steps in the direction charted by Godfrey and Hanson and by Madrigal et al. Drawing on his ethnographic work and contemporary social theory, McDade has formulated a more complex hypothesis, and a more complex test of that hypothesis, to account for racial-ethnic health inequalities. It is therefore worthwhile to use McDade’s presentation as a means of drawing out some of the challenges presented in furthering this research. The work of all of these authors falls squarely in a biocultural research tradition. As I have argued for a number of years (Dressler 1995, 2005), one of the fundamental theoretical and methodological issues confronting a biocultural research program is taking the concept of culture seriously. Furthermore, I have argued that for a number of reasons, a cognitive theory of culture is the best tool for this task. In a cognitive theory of culture, the notion of “meaning” is central. Culture consists of a set of shared meanings that enable us to effectively interpret others’ behaviors at the same time that it provides a prototype for our own behaviors in any given situation. In McDade’s paper, he shows how social rank in Samoa can mean something different when it is paired with other indicators of status. We know that it means something different because of the implications it has for adolescent immune function. Under one set of conditions (in this case, having a different rank on an alternate dimension of status), having someone of the valued matai status in your household

Explaining Health Inequalities

•

183

is problematic, as measured by adolescent immune function. Where the two statuses are congruent (and it doesn’t matter whether they are consistently high or consistently low), immune function is superior. Similarly, in his analyses of skin color and blood pressure, the meaning of increasing income changes depending on whether you have lighter or darker skin—and we know this because of the differential effects of income on blood pressure under varying conditions of skin color. For persons of lighter skin color, blood pressure declines as would be expected; however, for persons of darker skin color, the meaning of that increase in income is not the same. Clearly, there is also something else going on here. The notion of shared meaning places all of the emphasis on shared mental representations, but this cannot be the end of the matter. Samoan adolescents from households with disparate status rankings are going out and interacting with other people, and it is that interaction, or the implication that these representations have for concrete social interactions, that must be implicated in their biology. Similarly, when comparing lighter-skinned and darker-skinned African-Americans in the United States, the world clearly must work differently for a higher-income, darker-skinned person than it does for a higher-income, lighter-skinned person. As I have argued elsewhere (Dressler 2005), to further a biocultural research approach, we must take very seriously culture as shared meaning, and that demands that we document that shared meaning. At the same time, people don’t just think things, they do things. We must understand how cultural construction and social structure intersect in ways that both challenge and reward people. McDade gives us the best illustration of such a research model in his Samoan work because, on the basis of his ethnographic description, we (as readers) can develop an intuitive sense of why the matai status is important and also of why it is important to be a consumer of Western goods and information. We can therefore understand why an inconsistency in these statuses might be problematic. This becomes somewhat more difficult with the empirical example from the African-American community, because we don’t have an understanding of the contextual meaning of income or of skin color (although for many Americans, our “native’s” knowledge of the society helps us to speculate why this is so). With the work of Madrigal et al., we need to understand seriously the meaning of “race” in different cultural contexts, because it is that meaning that changes the biological dimension of ethnicity. We have argued elsewhere that any serious study of racial or ethnic groups and health requires at the outset a careful analysis of the cultural models of “race” in any given setting, because everything attributed to race is really a function of what we call “ethnorace,” or the racial categories used in a given community and the cultural theory of biological differences associated with phenotypic difference in that community (Dressler, Oths, and Gravlee 2005). Finally, with the work of Godfrey and Hanson, the environment of the growing fetus is a complex one, made up of the nutrients crossing the barrier between mother and infant, but

184

•

William W. Dressler

also made up of the culturally constructed, socially structured environment of the mother and her varied responses to the convergence of shared understandings and structural constraints. These chapters are all superb examples of a vibrant and exciting biocultural approach. Taking the concept of culture seriously will open up new possibilities. References Barker, D.J. 1991. The foetal and infant origins of inequalities in health in Britain. Journal of Public Health Medicine 13: 63–68. Cooper, R.S., C. Rotimi, S. Ataman, D. McGee, et al. 1997. The prevalence of hypertension in seven populations of West African origin. American Journal of Public Health 87: 160–68. Curtin, Philip D. 1969. The African Slave Trade: A Census. Madison, WI: University of Wisconsin Press. Dressler, William W. 1995. Modelling biocultural interactions in anthropological research: an example from research on stress and cardiovascular disease. Yearbook of Physical Anthropology 38: 27–56. ———. 2004. Social or status incongruence. In The Encyclopedia of Health and Behavior, ed. Norman B. Anderson. Thousand Oaks, CA: Sage Publications, 764–67. ———. 2005. What’s cultural about biocultural research? Ethos 33: 20–45. Dressler, William W., Mauro C. Balieiro, and Jose Ernesto Dos Santos. 1999. Culture, skin color, and arterial blood pressure in Brazil. American Journal of Human Biology 11: 49–59. Dressler, William W., Kathryn S. Oths, and Clarence C. Gravlee. 2005. Race and ethnicity in public health research: models to explain health disparities. In Annual Review of Anthropology, vol. 34. Palo Alto, CA: Annual Reviews, 231–52. Henry, James P. and John C. Cassel. 1969. Psychosocial factors in essential hypertension. American Journal of Epidemiology 90: 171–200. Jardim, Paulo Cesar B. Veiga, Omar Carneiro, Sérgio B. Carneiro, and Mari Nasare Baiocchi. 1992. Pressão arterial em communidade negra isolada remanescente de quilombo— norete de Goiá—Kalunga. Arquivos Brasileiros de Caridiologia 58: 289–93. Kaplan, George A. 1995. Where do shared pathways lead? Psychosomatic Medicine 57: 208–12. Oths, Kathry S., Linda L. Dunn, and Nancy S. Palmer. 2001. A prospective study of psychosocial job strain and birth outcome. Epidemiology 12: 744–46. Townsend, Peter and Nick Davidson, eds. 1982. Inequalities in Health: The Black Report. Harmondsworth, England: Penguin Books.

•7• The Developmental Origins of Health and Disease Keith Godfrey and Mark Hanson

Overview The “developmental origins” hypothesis states that adult coronary heart disease and the associated disorders hypertension and type 2 diabetes, originate through developmental plastic responses to an adverse early environment. Clinical and epidemiological research has established that people who were smaller at birth and had poor growth in infancy have increased rates of coronary heart disease, raised blood pressure, and type 2 diabetes, particularly if their restricted fetal and infant growth was followed by increased childhood weight gain. There is also evidence that both premature birth and overt fetal growth restriction, not just size at birth, are associated with adverse long-term effects. The relations between smaller infant size and an increased risk of ill-health and adult disease extend across the normal range of infant size in a graded manner, suggesting that the developmental plastic responses involve a coordinated change in the pattern of growth, rather than simply a reduction in body growth overall. People with impaired early development become vulnerable to adult cardiovascular and metabolic disorders through a variety of processes. First, they have altered structure in key organs, such as the kidney and vasculature. Differences in organ structure may arise from changes in the numbers and types of cells in these organs, or from changes in gene expression within these cells. They may also have alterations in the settings of the systems that control hormone levels and metabolism, leading, for example, to insulin resistance. A third link is that people who were smaller at birth have enhanced stress responses and are more vulnerable to psychosocial stress and other adverse environmental influences in later life.

186

•

Keith Godfrey and Mark Hanson

The observations emphasize the importance of adopting a life-course approach to health and have led to the hypothesis that coronary heart disease, hypertension, and type 2 diabetes originate through developmental plastic responses made by the embryo, fetus, and infant as part of a prediction regarding the subsequent environment to which it anticipates that it will be exposed. Stimuli that induce changes during critical periods in development result in irreversible phenotypic changes; if the environment in childhood and adult life differs from that predicted during fetal life and infancy, the developmentally induced phenotype may be inappropriate and this may increase the risk of adult disease. In those human societies where economic circumstances and nutrition are rapidly improving, a mismatch between the early prediction and the subsequent reality can cause severe health problems. It is now clear that the associations between early development and later outcomes do not simply reflect genetic influences such as polymorphisms. Rather, the findings indicate that interactions between genetic influences and the early life environment determine risk of disease by changing susceptibility to adverse influences in the adult environment. Research to date has linked particular maternal influences with the later health of the offspring, notably transgenerational effects of the mother’s own intrauterine experience, and her body composition, dietary balance, and endocrine status before and during pregnancy. Understanding maternal influences on the offspring’s developmental plastic responses may allow the design of new interventions to optimize early development and thereby improve health throughout life. Studies in a variety of animal species have readily replicated relationships between early life experience and adult metabolic and cardiovascular function: the breadth of this demonstration suggests that a general biological process underpins it. This has led to the proposal that the observations can best be interpreted as part of a broad set of developmental strategies termed “predictive adaptive responses” (see box below). They have been selected for during evolution because they confer later fitness advantage on individuals of a species. Placing these strategies in this perspective has important implications for understanding changing patterns of human disease. Fetal, Infant and Childhood Growth in Relation to Health in Later Life The Developmental Origins of Health and Disease (DOHaD) hypothesis stemmed from ecological observations of secular trends and geographical variations in the incidence of coronary heart disease. In the early twentieth century the incidence of coronary heart disease rose steeply in Western countries, becoming the most common cause of death. The steep rise has generally been followed by a fall over recent decades, a fall that cannot be accounted for by changes in adult lifestyle (Barker 1998). As longevity continues to increase, it is anticipated that the number of people suffering from cardiovascular disease in Western societies will rise again. The incidence of coronary heart disease is now rising in other parts of the world to which Western influences are extending, most notably in South Asia and Latin America. There are

The Developmental Origins of Health and Disease

•

• • • Evolutionary and Theoretical Perspectives: Predictive Adaptive Responses The observations set out at the beginning of this chapter have led to the hypothesis that adult cardiovascular disease and type 2 diabetes originate through developmental plastic responses made by the embryo, fetus, and infant as part of a prediction regarding the subsequent environment to which it anticipates it will be exposed. Such responses do not necessarily provide an immediate selective advantage. Rather, such responses, which we have termed “predictive adaptive responses” (PARs), act primarily to improve fitness at a later stage in development (Gluckman et al. 2005). Critical periods in development result in irreversible phenotypic changes; if the environment in childhood and adult life differs from that predicted during fetal life and infancy, the developmentally induced phenotype is inappropriate and the risk of adult disease is increased. In those human societies where economic circumstances and nutrition are rapidly improving, a mismatch between the early prediction and the subsequent reality can cause severe health problems. PARs are a form of phenotypic plasticity with delayed selective benefits, a phenomenon seen across a range of vertebrate and invertebrate species. We have argued that PARs have been retained in humans because they conferred survival advantage in the environment of our ancestral hominids (ibid). Palaeolithic diets may have been relatively high in lean protein, and they certainly had a lower glycemic index than do contemporary Western diets. In addition, energy expenditure was much higher. In our evolutionary past, it was advantageous to anticipate a poor future environment, and indeed PARs induce by default a phenotype appropriate for such conditions. Many human PARs, however, now appear to be increasingly inappropriate. This inappropriateness has arisen because the prenatal environment can change only slowly over many generations, whereas the postnatal environment has been enhanced relatively rapidly; indeed, for humans migrating to better conditions it can change dramatically between generations. For this reason the incidence of cardiovascular and related metabolic disease is very high in populations that are in rapid economic transition. In China, for example, chronic diseases now account for an estimated 80 percent of deaths and 70 percent of disability-adjusted life-years lost (Wang et al. 2005).

• • •

187

188

•

Keith Godfrey and Mark Hanson

fears that it will become important even in the poorest societies, such as some in subSaharan Africa (Prentice and Moore 2005). An indication that coronary heart disease might originate during fetal development came from geographical studies of death rates among babies born in Britain during the early 1900s (Barker 1998). Perinatal mortality rates were highest in some of the northern industrial towns and the poorer rural areas in the north and west, a pattern that closely resembles today’s large variations in death rates from coronary heart disease (ibid). The usual certified cause of death in newborn babies during the early 1900s was low birth weight, and one possible conclusion suggested by the geographical association was that low rates of growth before birth are linked to the development of coronary heart disease in adult life. Epidemiological Studies of Size at Birth and Coronary Heart Disease Direct evidence that an adverse intrauterine environment might have long-term consequences for the risk of coronary heart disease, came from follow-up studies of men and women in middle and late life whose body measurements at birth had been recorded. Among people born in Hertfordshire, UK, those who had low birth weight suffered increased death rates from coronary heart disease in adult life (Osmond et al. 1993; Barker 1998). Thus, among 15,726 people born during 1911 to 1930, death rates from coronary heart disease fell progressively with increasing birth weight in both men and women (Figure 7.1). A small rise at the highest birth weights in men could relate to the infants of women with gestational diabetes. Subsequent studies suggested that it was particularly people who were small at birth as a result of growth restriction who were at increased risk of diabetes (Barker et al. 1998). Observation of similar findings has led to wide acceptance that small size at birth is associated with coronary heart disease in later life. For example, confirmation of a link between low birth weight and adult coronary heart disease has come from a study of 70,297 nurses in the United States in which there was a twofold fall in the relative risk of nonfatal coronary heart disease across the range of birth weights (RichEdwards et al. 1997). Follow-up studies of populations with more detailed birth measurements suggest that altered birth proportions are more strongly associated with late outcomes than is birth weight per se. Patterns of altered birth proportions and restricted fetal growth linked with later coronary heart disease may be summarized as a smaller head circumference, shorter crown-heel length, and lower neonatal ponderal index (Barker 1998). Infant and Childhood Growth and Coronary Heart Disease Evidence suggesting both additive and interactive effects of poor prenatal and postnatal growth on the risk of subsequent coronary heart disease is now emerging. Fol-

The Developmental Origins of Health and Disease

•

189

Figure 7.1. Coronary heart disease death rates, expressed as standardized mortality ratios (SMR), in 10,141 men and 5,585 women born in Hertfordshire, UK, according to birth weight. (Derived from Osmond et al. 1993)

low up of men born in Hertfordshire, UK, between 1911 and 1930 found that lower weight at age 1 year was strongly associated with higher rates of coronary heart disease (Osmond et al. 1993), and subsequent analyses have suggested additive effects of poor fetal and infant growth (Barker 1998). Confirmation that smaller and thinner infants at age 1 have increased rates of coronary heart disease in adulthood has come from people born in the 1930s and 1940s in Helsinki, Finland (Eriksson et al. 2001). Consistent with the known association between coronary heart disease and short adult stature, men in Helsinki who developed the disease also tended to have poor weight gain and low rates of height growth in infancy (ibid). The findings from Helsinki also point to the possibility that interactions between the pre- and postnatal environments influence of the risk of coronary heart disease. Although infant growth failure was deleterious both in individuals who were small and in those who were large at birth, childhood weight gain had very different effects in small and large neonates. Among boys who were thin at birth with a ponderal index below the median, rapid weight gain and increasing body mass index during childhood was associated with higher rates of adult coronary heart disease; however, in boys who were not thin at birth, rapid childhood weight gain and increasing body mass index was unrelated to the risk of coronary heart disease (ibid). Findings among girls were similar, and again the risk of coronary heart disease was determined more by the tempo of weight gain than by the body size attained (Forsen et al. 1999).

190

•

Keith Godfrey and Mark Hanson

Potential Confounding Influences and Size of Effects Such findings led to the hypothesis that influences linked to pre- and postnatal growth alter the risk of coronary heart disease. Critics of this hypothesis initially suggested that a) the findings reflect confounding by adult lifestyle influences, b) the associations reflect genetic influences on both early growth and later cardiovascular disease, and c) the size of the developmental influences on later disease is small and of little consequence. Some commentators argued that people whose growth was impaired in utero may continue to be exposed to an adverse environment in childhood and adult life, and it is adversity in this later environment that produces the effects attributed to developmental influences. There is, however, strong evidence against this argument. In several of the studies, data on adult lifestyle factors including smoking, employment, diet, alcohol consumption, and exercise were collected; these lifestyle factors are linked with cardiovascular disease, but allowance for them had little effect on the association between birth weight and coronary heart disease (Frankel et al. 1996; Rich-Edwards et al. 1997; Barker et al. 2001). Influences in adult life, however, add to the effects of the intrauterine environment. For example, the prevalence of coronary heart disease is highest in people who had low birth weight but were obese as adults. It has also been argued that the associations between size at birth and later disease could primarily reflect genetic influences (Hattersley and Tooke 1999). However, birth size has only a modest genetic component and primarily reflects the quality of the intrauterine environment (Morton 1955; Brooks et al. 1995). Moreover, recent findings indicate that it is interactions between the early life environment and genetic influences that are likely to be the principal determinants of disease susceptibility. Evidence for such interactions has come from analyses of glucose-insulin metabolism among 70-year-old people in the Helsinki cohort study; the Pro12Pro polymorphism of the PPAR-γ2 nuclear hormone receptor has been associated with insulin resistance, but the Helsinki analyses showed that this effect was not apparent in those who weighed more than 3.5 kg at birth, and was most marked in those who weighed less than 3 kg (Eriksson et al. 2002). Thus inheriting the Pro12Pro polymorphism may confer susceptibility to disease, but that risk does not become realized in an optimal intrauterine environment. Some commentators have argued that the magnitude of developmental effects on adult cardiovascular risk is small, with one review concluding that the difference in systolic pressure associated with a one-kilogram difference in birth weight was around 2.0 mm Hg (Huxley et al. 2000). In clinical practice this would be a small difference, but the Framingham study indicates it may nonetheless correspond to a substantial proportion of total attributable mortality (Kannel et al. 1971). Where clinical disease has been used as the outcome measure, the effect of early environmental influences is clear. Among 22,846 men in the Health Professionals Follow-up Study there were strong relationships between birth size and the risk of developing clinically significant

The Developmental Origins of Health and Disease

•

191

hypertension or diabetes mellitus, but no relationship with systolic blood pressure (Curhan et al. 1996). The findings demonstrate the importance of studying health outcomes rather than surrogate measures of disease. Assessment of the relative importance of early and later life exposures is difficult as there are few well-characterized cohorts with both perinatal data and health outcomes documented well into later life. Estimates based upon the Helsinki cohort suggest that if lower-birth-weight babies could be brought to the mean for the population, and if growth until one year of postnatal life could similarly be optimized, then the risk of later hypertension would be reduced by 25 percent and that of type 2 diabetes by 57 percent (Barker et al. 2002). Such calculations may underestimate the developmental component because the observed relationship between disease risk and birth size does not imply a causal role of being born small but, rather, reflects the effects of adverse intrauterine influences on fetal growth. It is thought that environmental influences acting during early development are the causal trigger, and much experimental evidence indicates that adverse developmental influences can affect outcomes without birth size being affected. Moreover, the interaction between smaller size at birth and high childhood weight gain indicates the danger of singling out a single period of development in determining the risk of ill-health. A further consideration is that the early life and adult environments may not simply have additive effects but may, instead, interact to influence the risk of coronary heart disease. Poverty and low household income have long been linked with coronary heart disease, but data from the Helsinki cohort suggest that this effect occurs only in individuals of below average thinness at birth (Barker et al. 2001). If interactions between the early life and adult environments are confirmed, this will have important implications for our understanding of the evolutionary implications of developmental responses.

Mechanisms Responsible for a Developmental Influence on Cardiovascular Risk In studies exploring the mechanisms underlying the associations between early growth and later coronary heart disease, there are similar trends between birth weight and major risk factors for cardiovascular disease, including hypertension, raised serum cholesterol, raised plasma fibrinogen, and impaired glucose tolerance as well as type 2 diabetes, obesity, and endothelial dysfunction (Hales et al. 1991; Barker et al. 1998; Ravelli et al. 1999; Leeson et al. 1997). Subsequent studies have examined effects on organs and systems particularly important in the etiology of cardiovascular disease, including blood vessels, lipid metabolism, the kidneys, stress responses, glucose and insulin metabolism, and adipose tissue and obesity (see Appendix 1). Much attention has focused on the possible role of impaired renal development in mediating early life effects on later hypertension. If the materno-placental supply of nutrients does not match fetal requirements in late pregnancy, the fetus diverts blood and nutrients to

192

•

Keith Godfrey and Mark Hanson

maintain brain metabolism, thus reducing blood flow to the fetal trunk, limbs, and kidneys. Mackenzie and Brenner (1995) suggested that retarded fetal growth reduces the number of nephrons, necessitating an increase in pressure in the kidney capillaries to maintain filtration and leading to nephron loss and a self-perpetuating cycle of hypertension and progressive nephron injury. The number of nephrons is established during fetal life and has a wide range. Recent findings showed a significantly reduced median number of glomeruli per kidney of 702,379 in hypertensive patients, as compared with 1,429,200 in normotensive controls (Keller et al. 2003). Several of the maternal exposures that impair fetal growth are associated with alterations in the fetal hypothalamic-pituitary-adrenal and sympathoadrenal axes, and these can alter the set-point of the biological response to stress. Animal studies provide much evidence that resetting of homeostatic neuro-endocrine mechanisms controlling blood pressure are also likely to be involved in mediating developmental effects on later blood pressure (Hanson and Gluckman 2006). People who had low birth weight have increased stress sensitivity, as indicated by hypothalamic-pituitaryadrenal axis activity and sympathoadrenal responses (Phillips et al. 2000; Jones et al. 2006). Patients with high blood pressure tend to have features of increased sympathetic nervous system activity, including a high resting pulse rate, a high cardiac output, and a hyperdynamic circulation. Among men and women in Preston, UK, those who had low birth weight had a higher resting pulse rate (Phillips and Barker 1997). This suggests that retarded growth in utero establishes increased sympathetic nervous activity and contributes to raised blood pressure in later life. Developmental Influences on Type 2 Diabetes Insulin has a central role in fetal growth, and disorders of glucose and insulin metabolism are therefore an obvious possible link between early growth and cardiovascular disease. Although obesity and a sedentary lifestyle are important in the development of type 2 diabetes, they seem to lead to the disease only in predisposed individuals (Barker 1998). Family and twin studies have suggested that the predisposition is familial, but the nature of this predisposition is unknown (Barroso 2005). The disease tends to be transmitted through the maternal rather than the paternal side of the family (Karter et al. 1999). A number of studies have confirmed the association between birth weight and impaired glucose tolerance and type 2 diabetes that was first reported in Hertfordshire, UK (Hales et al. 1991; Curhan et al. 1996). Other studies have reported that it is thinner babies who are most at risk of developing impaired glucose tolerance and type 2 diabetes (Phillips et al. 1994; Lithell et al. 1996). Among the North American Pima Indians, diabetes in pregnancy is unusually common and the association between birth weight and type 2 diabetes is U-shaped, with an increased prevalence in both young people who weighed less than 5.5 pounds at birth and those with birth weights over 9.9 pounds (> 4.5 kg) (McCance et al. 1994; Wei et al. 2003).

The Developmental Origins of Health and Disease

•

193

Table 7.1. Mechanisms linking developmental influences with the risk of cardiovascular disease and type 2 diabetes in later life.

• • • • • •

Endothelial dysfunction Altered lipid metabolism Reduced nephron endowment and consequent hypertension Heightened stress responses Insulin deficiency and resistance Obesity and the adiposity rebound

The increased risk of diabetes among those of high birth weight was associated with maternal diabetes in pregnancy. Findings in other groups of children support an intrauterine origin for type 2 diabetes and suggest that the seeds of diabetes in the next generation have already been sown and are apparent in today’s children (Whincup et al. 1997; Hofman et al. 2004). Both deficiency in insulin production and insulin resistance are thought to be important in the pathogenesis of type 2 diabetes. Evidence suggests that both may be determined in fetal life. Infants who are small for gestational age have fewer insulin producing pancreatic β cells, and there is evidence that nutritional and other factors determining fetal and infant growth influence the size and function of the adult β cell complement (Hales and Barker 1992). Whether and when type 2 diabetes supervenes will be determined by the rate of attrition of β cells with ageing, and by the development of insulin resistance, of which obesity is an important determinant. Follow up of men and women who were in utero during the Dutch famine provides evidence that maternal undernutrition can lead to insulin resistance and type 2 diabetes in the offspring (Ravelli et al. 1998). As in people from South India living in Britain, Fall et al. (1998) found a high prevalence of insulin resistance and central adiposity in Mysore, South India. Men and women with type 2 diabetes were not just insulin resistant but had evidence of insulin deficiency, which was particularly associated with maternal adiposity. These findings led to the hypothesis that maternal adiposity and hyperglycemia during pregnancy may underlie the epidemic of type 2 diabetes in urban and migrant Indian populations (Fall et al. 1998) (Figure 7.2). Adults who had a low birth weight have a high prevalence of the “metabolic syndrome” (Barker et al. 1993b), in which impaired glucose tolerance, hypertension, and raised serum triglyceride concentrations occur in the same individual. Among 30-year-old Mexican-Americans and non-Hispanic white people, 25 percent of those in the lowest third of the birth weight distribution and the highest third of current body mass had the syndrome; by contrast, none of those in the highest third of birth

194

•

Keith Godfrey and Mark Hanson

Figure 7.2. A model to explain the epidemic of type 2 diabetes in urban India. (Derived from Fall et al. 1998)

weight and lowest third of current body mass had the metabolic syndrome (Valdez et al. 1994). Clinical studies provide clues into the processes that link thinness at birth with insulin resistance in adult life. Babies born at term who are thin have a low muscle bulk, with relative sparing of adipose tissue deposition. Thinness at birth may be as-

The Developmental Origins of Health and Disease

•

195

sociated with abnormalities in muscle structure and function that originate in midgestation and interfere with insulin’s ability to promote glucose uptake in skeletal muscle. Magnetic resonance spectroscopy studies show that people who were thin at birth have impairment of their muscle metabolism during exercise (Taylor et al. 1995). When the availability of nutrients to the fetus is restricted, concentrations of anabolic hormones, including insulin, fall, while catabolic hormones, including glucocorticoids, rise; a fetus may reduce its metabolic dependence on glucose and increase oxidation of other substrates, including amino acids and lactate (Barker et al. 1993a). This raises the possibility that hormonal changes or a glucose-sparing metabolism could persist into adult life, contributing to the development of the metabolic syndrome. Bjorntorp (1995) has postulated that glucocorticoids, growth hormone, and sex steroids may play major roles in the metabolic syndrome. Obesity and the Adiposity Rebound The studies described earlier found that people who were small at birth and had low infant weight gain are at particular risk of adult coronary heart disease if they become overweight in childhood. After age 2 the body mass index of normal young children falls, reaching a minimum at around 6 years before rising again—a phenomenon known as the adiposity rebound (Rolland-Cachera et al. 1987). Early age at adiposity rebound has been related to obesity in childhood and adult life (Whitaker et al. 1997). Data from Helsinki showed for the first time that early age at adiposity rebound is a strong predictor of type 2 diabetes (Eriksson et al. 2003), an observation later replicated in Delhi, India (Bhargava et al. 2004). This trend is remarkably strong, even without taking adult obesity into account. Overall, the Helsinki observations showed that slow growth in utero, low weight gain in infancy, and early age at adiposity rebound in later childhood are associated with large increases in the incidence of type 2 diabetes. Surprisingly, thinness at age 1 was a strong predictor of an early adiposity rebound in the Helsinki data. Data from the US National Collaborative Perinatal Project have shown that greater weight gain from birth to age four months is also a major risk factor for later obesity (Stettler et al. 2003), and early life influences on the risk of later obesity are an area of active research. Maternal Influences on Development and Health of Offspring The demonstration that normal variations in fetal size and proportions at birth have implications for health throughout life has prompted a reevaluation of maternal influences on growth and development. Much experimental and epidemiological research now suggests that maternal diet, body composition, and other factors can influence the adult physiology of the offspring, and that these long-term effects can

196

•

Keith Godfrey and Mark Hanson

operate without necessarily affecting commonly measured pregnancy outcomes such as birth weight (Hanson and Gluckman 2005). Research to date has linked specific maternal influences with the later health of the offspring, notably: 1) the mother’s own birth weight and transgenerational effects; 2) maternal body composition, including fat and lean mass; 3) maternal dietary intakes, including macro- and micronutrient balance and 4) maternal endocrine status. There is increasing evidence that fetal development can be affected by nutritional variation even within the normal range of Western diets, and the problem is compounded because many women eat unbalanced diets or constrain their weight by dieting (Godfrey 2000). Furthermore, studies of the Dutch famine indicate that the longer-term effects on offspring may depend on the duration and timing of famine exposure and can be independent of birth size (Roseboom et al. 2001). Although nutrition has received the most attention, other early environmental factors such as infection, season of birth, and smoking may also have long-term effects. We have proposed that the mechanisms through which these maternal influences lead to developmental plastic responses include 1) a mismatch between fetal macroand micronutrient demands, largely determined by the growth trajectory set in early pregnancy, and the materno-placental capacity to meet this demand; 2) alterations in the fetal endocrine milieu; and 3) changes in the placental vasculature, which impact on fetal cardiovascular loading (Godfrey 2000). The developmental plastic responses include alterations in placental growth and epigenetic effects on gene promoter regions, together with changes in regional blood flow and body composition in the fetus. There is now evidence that the consequences of developmental plastic responses can be modified during infancy, and that their effects can be amplified by high childhood weight gain and perhaps by low levels of habitual physical activity, increasing vulnerability to adverse lifestyle influences during adulthood (Godfrey 2000). This conceptual framework is illustrated in Figure 7.3. The Early Trajectory of Fetal Growth Size at birth is the product of the fetal growth trajectory, which is set at an early stage in development, and the materno-placental capacity to supply sufficient nutrients to maintain that trajectory. In Western communities, randomized, controlled trials of maternal macronutrient supplementation have had relatively small effects on birth weight (Merialdi et al. 2003). This has led to the view that regulatory mechanisms in the maternal and placental systems act to ensure that human fetal growth and development is little influenced by normal variations in maternal nutrient intake, and that there is a simple relationship between a woman’s body composition and the growth of her fetus. Subsequent experimental studies in animals and observational data in humans challenge these concepts. They suggest that a mother’s dietary intakes and body composition around the time of conception and during pregnancy can exert major effects on the balance between the fetal demand for nutrients (Kwong et

The Developmental Origins of Health and Disease

•

197

Figure 7.3. A conceptual framework for the developmental origins of health and disease.

al. 2000) and the materno-placental capacity to meet that demand (Duggleby and Jackson 2001). A rapid trajectory of growth increases the fetal demand for nutrients. Although fetal demand for nutrients is greatest late in pregnancy, the magnitude of this demand is thought to be primarily determined by genetic, epigenetic, and environmental effects on the trajectory of fetal growth. Experimental studies of pregnant ewes have shown that, although a fast growth trajectory is generally associated with larger fetal size and improved neonatal survival, it renders the fetus more vulnerable to a reduced materno-placental supply of nutrients in late gestation. Thus, maternal undernutrition during the last trimester adversely affects the development of rapidly growing fetuses with high requirements but has little effect on those growing more slowly (Harding et al. 1992). Rapidly growing fetal lambs were found to make a series of adaptations in order to survive, including wasting of muscle and other soft tissues (Barker et al. 1993a). Characterization of maternal influences on the early growth trajectory of the human fetus is now an important priority. The trajectory of fetal growth is thought to

198

•

Keith Godfrey and Mark Hanson

increase with improvements in periconceptional nutrition, and is faster in male fetuses (Pedersen 1980). The consequent greater vulnerability of male fetuses to undernutrition may contribute to the higher death rates from coronary heart disease among men. Periconceptional and Epigenetic Influences Experiments in animals have shown that alterations in maternal diet around the time of conception can change the fetal growth trajectory. In pregnant rats, feeding a low protein (9 vs. 18 percent casein) diet for just four days in the preimplantation period led to reduced stem cell populations being formed during embryonic development and to small size at birth, compensatory postnatal growth, and raised blood pressure in the offspring as adults (Kwong et al. 2000). The sensitivity of the human embryo to its environment is being increasingly recognized with the development of assisted reproductive technology. Babies conceived by IVF tend, for example, to have a lower birth weight compared with those from natural conceptions (Walker et al. 2000; Winston and Hardy 2002) and have an increased prevalence of specific abnormalities (Arnaud and Feil 2005). Such effects in early gestation focus attention on mechanisms that involve epigenetic processes. The most well described concern imprinted genes that are sensitive to environmental conditions and have an underlying role in fetal overgrowth or “large offspring syndrome” in sheep and cattle (Walker et al. 2000), but it is now recognized that the effect extends to many other, non-imprinted genes too (Lillycrop et al. 2005). The DNA methylation processes underlying such effects depend on the supply of methyl groups from amino acids; recent studies suggest that periconceptional maternal diets rich in folate and other nutrients enhancing methyl group availability can permanently change the DNA methylation of the offspring. It is therefore of interest that folate supplementation of the low-protein diet in the pregnant rat prevents elevated blood pressure and vascular dysfunction in the offspring (Torrens et al. 2006). Materno-Placental Nutrient Supply and Transgenerational Effects The supply of nutrients to the human fetus depends on the long and vulnerable series of steps known as the materno-placental supply line; this includes the mother’s body composition and size, her metabolism, and transport of nutrients to and across the placenta. Part of the materno-placental supply line is established during the mother’s own intrauterine life. Strong evidence for major intergenerational effects in humans has come from studies showing that a woman’s birth weight influences the birth weight of her offspring (Emanuel et al. 1992). We have found a strong effect of lowbirth-weight women tending to have thin babies with a low neonatal ponderal index; in contrast, low-birth-weight fathers tended to have shorter babies, and neonatal thinness was unrelated to paternal birth weight (Godfrey et al. 1997). The effect

The Developmental Origins of Health and Disease

•

199

of maternal birth weight on thinness at birth is consistent with the hypothesis that in low-birth-weight mothers the materno-placental supply line is compromised and unable to meet fetal nutrient demand. These and other observations have led to the conclusion that mothers constrain fetal growth (Gluckman et al. 2005) and that the degree of constraint they exert is set to a degree when they are themselves in utero. Potential mechanisms underlying this effect include alterations in the uterine or systemic vasculature (Torrens et al. 2002), changes in maternal metabolism, and in placental efficiency. The study in Preston, UK, showed that young men and women whose mothers had low birth weight have raised blood pressure even after allowing for their mothers’ blood pressure; father’s birth weight was not related to the offspring’s blood pressure (Barker et al. 2000). This suggests that if the growth of a female fetus is constrained by lack of nutrients, there are persisting changes in its physiology and metabolism that lead to reduced fetal nutrition and raised adult blood pressure in the next generation. Little is known about the effects of intergenerational transmission of retarded fetal growth on coronary heart disease, or on other degenerative disorders, in the next generation. Experimental studies in animals have shown that undernutrition can have effects on reproductive performance, which may persist for several generations. Among rats fed a protein-deficient diet over twelve generations there was a progressive fall in fetal growth rates. When restored to a normal diet, it took three generations before growth and development were normalized (Stewart et al. 1980). The existence of strong intergenerational effects on fetal growth is important for public health today. Improvements in women’s diets are likely to benefit more than one generation. Maternal Diet and Body Composition Evidence for a long-term effect of levels of maternal nutrient intake during pregnancy has come from a follow-up study of children whose mothers took part in a randomized, controlled trial of calcium supplementation in pregnancy (Belizan et al. 1997). Supplementation was associated with lowering of the offspring’s blood pressure in childhood. Follow-up studies of people exposed to the Dutch famine of 1944–45 showed that severe maternal caloric restriction at different stages of pregnancy was variously associated with obesity, dyslipidemia, and insulin resistance in the offspring, with preliminary evidence of an increased risk of coronary heart disease among people conceived during the famine (Ravelli et al. 1998; Roseboom et al. 2001). In the Dutch studies, famine exposure per se was not associated with raised blood pressure in the offspring, but there was evidence for an effect of macronutrient balance. Maternal rations with a low protein density were associated with raised blood pressure in the adult offspring (Roseboom et al. 2001). This adds to findings from Aberdeen, UK, that showed that maternal diets with either a low or a high ratio of animal protein to carbohydrate were associated with raised blood pressure in the offspring during adult life (Campbell et al. 1996). Maternal diets with a high protein

200

•

Keith Godfrey and Mark Hanson

density were also associated with insulin deficiency and impaired glucose tolerance in the offspring (Shiell et al. 2000). While adverse effects of a high-protein diet may seem counterintuitive, these findings are consistent with the results of controlled trials of protein supplementation in pregnancy, which show that high protein intakes are associated with reduced birth weight (Rush 1989). The Aberdeen findings have recently been replicated in a follow-up study of men and women in Motherwell, UK, whose mothers were advised to eat a high-meat protein, low carbohydrate diet during pregnancy. Those whose mothers had high intakes of meat and fish in late pregnancy had raised blood pressure and plasma cortisol concentrations, particularly if the mother also had a low intake of green vegetables (Shiell et al. 2001; Herrick et al. 2004). The long-term effects may be consequences of the metabolic stress imposed on the mother by an unbalanced diet in which high intakes of essential amino acids are not accompanied by the micronutrients required to utilize them. Fetal growth and development depend on maternal nutrient stores and the turnover of protein and fat in the mother’s tissues at least as much as on the mother’s diet: for the fetus to depend primarily on the mother’s diet in pregnancy would be too dangerous a strategy (James 1997). Maternal size and body composition account for up to 20 percent of the variability in birth weight (Catalano et al. 1998). High maternal weight and adiposity has been associated with insulin deficiency, type 2 diabetes, and coronary heart disease in the offspring (Forsen et al. 1997; Shiell et al. 2001; Fall et al. 1998). In data from Helsinki, increasing maternal body mass index had little effect on the offspring’s death rates in tall women, but strong effects in short women (Forsen et al. 1997). One interpretation of these findings is that greater maternal body fatness may increase fetal growth and hence the fetal demand for nutrients; short women may not be able to meet this increased demand as a result of a constrained nutrient supply capacity determined during their own intrauterine development. Of considerable importance is consistent evidence showing strong links between low maternal weight and body mass index and insulin resistance in the adult offspring (Ravelli et al. 1998; Shiell et al. 2000; Mi et al. 2000). Although low maternal body mass index does not appear to be linked with raised blood pressure in the offspring, thin maternal skinfold thicknesses and low pregnancy weight gain have been consistently associated with raised offspring blood pressure (Godfrey et al. 1994; Adair et al. 2001). One of the metabolic links between maternal body composition and birth size is protein synthesis. Women with a greater lean body mass have higher rates of protein synthesis in pregnancy (Duggleby and Jackson 2001). Variation in rates of maternal protein synthesis explains around a quarter of the variability in birth length. Perspectives on the Future The experimental and epidemiological evidence indicates that prevention of a substantial proportion of cardiovascular disease and type 2 diabetes, may depend on

The Developmental Origins of Health and Disease

•

201

interventions at a number of stages of development. Strategies that target infants and young children may give the most immediate benefit, but improving the intrauterine environment is an important long-term goal. In keeping with this, a recent consultation concluded that the emphasis of any programs to combat poor nutrition should target pregnant women and children under 2 years of age (World Bank 2006). This focus reflects the view that, although there will be some flexibility in the later pathways that lead to health or ill-health, reversing many of the developmental effects is unlikely to be possible. In mothers we need to improve the macronutrient balance and micronutrient content of the diet before and during pregnancy. We also need to improve women’s body composition before pregnancy, with avoidance of excessive thinness or overweight. In infants we need to protect growth during the first year after birth by good infant feeding practices, avoidance of recurrent infections, and cognitive stimulation. We need to prevent rapid weight gain among children, especially those who were small or thin at birth or at 1 year. Adults who were small at birth are more vulnerable to obesity and psychosocial stress in adult life: we need to understand this more deeply in order to prevent it. The complexities of fetal growth and development are such that currently available data form only a limited basis for changing dietary recommendations to young women. Future work will need to identify the factors that set the trajectory of fetal growth and the influences that limit the materno-placental delivery of nutrients and oxygen to the fetus. We also need to define how the fetus adapts to a limited nutrient supply, how these adaptations change the structure and physiology of the body, and by what molecular mechanisms nutrients and hormones alter gene expression. A strategy of interdependent clinical, animal, and epidemiological research is required to identify specific recommendations for both whole populations and for vulnerable groups such as teenage pregnancies and single parents. Research is also required to identify the barriers to healthy eating among young women, whose diets are important both for their own health and for the health of the next generation. To improve the evidence base, new intervention studies are needed; it is encouraging, in this respect, that investment is now being made in large-scale population-based trials, such as the Mumbai Maternal Nutrition Study in India investigating the possible benefits of providing micronutrient-rich foods before conception. If, as we believe, a woman’s own fetal growth, and her diet and body composition before and during pregnancy, play major roles in programming the future health of her children, then mothers will want to know what they can do to optimize the intrauterine environment that they provide for their babies. At present we cannot provide mothers with such information. Nutrition is not the only issue, and other aspects of a poor intrauterine environment also need to be tackled, notably in societies where gender inequality is a particular source of maternal adversity. In areas of South Asia, for example, attention needs to be given to the age at which many young women become pregnant and to the physical workloads they undertake before and during pregnancy.

202

•

Keith Godfrey and Mark Hanson

Educational and other intervention measures to improve early development will have to involve men as well as women, and will need more than change at the individual level; building health over the life course demands social and economic measures, together with the political will to achieve these. Such an approach may allow us to reduce the prevalence of major chronic diseases and diminish social inequalities in health. Conclusion The Developmental Origins of Health and Disease (DOHaD) is now an established area of intense biomedical research, with implications for a range of fields from health economics to evolutionary biology and public health. The “journey” that established the DOHaD concept can be viewed as encompassing a number of major “stations” at which new insights were gained and new directions taken (see the Appendix of this chapter). Although necessarily oversimplified, this does provide a synthesis of how an interplay between human and experimental research can lead to rapid progress in understanding a new concept. Cardiovascular disease and type 2 diabetes have higher rates among poorer people and cause much disability in both developed and developing communities. The evidence linking the combination of poor maternal nutrition, impaired fetal/infant development, and increased weight gain in early childhood to later cardiovascular disease is now strong. Many women have diets with a suboptimal balance of nutrients, and wide variations between women in the amounts of lean and fat tissue alter the partitioning of nutrients between mothers and babies. However, little is known about how maternal nutrition influences early growth and development, or about interactions with genetic influences and with nutrition and physical activity after birth. Having reached a peak in the late twentieth century, there is evidence that the health and longevity of many populations in developed societies are now declining. For the first time in many generations, we are facing the reality that children born today may live less long than their parents. The situation is very acute in developing societies where cardiovascular disease and type 2 diabetes are now becoming endemic. If we are to improve the health of future generations, we need urgently to identify the factors that now affect fetal, infant, and childhood growth and development. Appendix The Developmental Origins of Health and Disease (DOHaD) concept— a journey from maps to mechanisms to societies 1. Maps as evidence of geographical variation in deaths The station at which the DOHaD journey started featured maps compiled by David Barker and colleagues. These showed large geographical variations in deaths from coronary heart disease in England and Wales. The geographical variations are a major

The Developmental Origins of Health and Disease

•

203

contributor to inequalities in health between different social groups, but Barker et al. demonstrated that adult lifestyle factors such as diet and smoking differed little between areas with high and low heart disease death rates. Using infant mortality as a general marker of an adverse early environment, a remarkably strong geographical association was found between infant mortality during the early part of the twentieth century and mortality from coronary heart disease sixty or so years later. This geographical association led to the hypothesis that low-birth-weight infants who survived might be at increased risk of coronary heart disease in later adulthood. 2. Examination of records of infant size at birth The second station on the journey was a search for the birth records of people now in their 60s and 70s, followed by ascertainment of coronary heart disease and its risk factors in these individuals. There were strong associations between smaller birth size and adult cardiovascular disease. These were graded across the entire range of size at birth in both males and females, suggesting that they were not a consequence of intrauterine pathology but rather represented biological responses to variations in the prenatal environment. 3. Experimental studies in animals The epidemiological observations on humans were followed by new experimental studies in animals designed to examine whether alterations in prenatal environments could have long-term effects on cardiovascular and metabolic parameters. This third station soon produced compelling evidence that remarkably simple exposures, such as variations in the protein content of a pregnant animal’s diet, or giving drugs to mimic increases in stress hormone levels, had indeed permanent effects on the physiology of the offspring. Elucidation of mechanisms followed, with a series of observations summarized elsewhere in this chapter. 4. Maternal influences In these experimental studies, alterations in maternal nutrition induced long-term effects on cardiovascular risk factors in the offspring without necessarily changing offspring size at birth, and an important fourth station on the journey was evidence supporting this in humans. Thus, the balance of nutrients in the mother’s diet and her body composition were linked with cardiovascular and metabolic function in the offspring in adulthood, even though maternal nutrition had produced no more than modest effects on size at birth. 5. Critical periods A fifth station was the characterization of critical periods during which environmental exposures have particular effects. Controlled experimental studies in several animal species showed the importance of the time around conception and of interactions

204

•

Keith Godfrey and Mark Hanson

between the pre- and postnatal environments, and demonstrated that the impact of exposure during critical periods of development could be transmitted through several generations of the offspring. These observations led to appreciation of the profound evolutionary implications of predictive adaptive responses (PARs). These exist because the intrauterine environment is so different from the extrauterine environment. Theoretical models demonstrate the circumstances under which fitness is enhanced if parents transmit information about the environment to their progeny; factors to consider include the fidelity of the transmission of environmental cues, the degree of predictability of environmental conditions, and the costs of incorrect prediction. The responses have the potential to be particularly disadvantageous in settings where humankind is changing the environment faster than ever before. 6. Epigenesis as key mechanism leading to health outcomes There was some resistance to the DOHaD concept within the scientific community, in part because the underlying molecular and cellular mechanisms were not understood. An important sixth station has therefore been the demonstration of nutrition- and stress- induced epigenetic processes as key mechanisms through which the environment has long-term effects on gene expression. When coupled with increased understanding of gene-environment interactions in other areas, for example in the etiology of cancer, there is growing consensus about the validity of the concept. 7. Opportunities for improving health at the population level While the biological insights arising from research in this area have been considerable, the seventh station—translating these insights into population-wide improvements in health and well-being—is arguably the most important. DOHaD research has given impetus to the recognition that pregnancy and the first two years of life offer tremendous opportunities for improving health in both developed and developing communities. If societies can grasp these opportunities, benefits will be substantial and lasting. Resources now need to be found to enable this seventh station to be reached.

References Adair, L.S., C.W. Kuzawa, and J. Borja, J. 2001. Maternal energy stores and diet composition during pregnancy program adolescent blood pressure. Circulation 104: 1034–39. Arnaud, P. and R. Feil. 2005. Epigenetic deregulation of genomic imprinting in human disorders and following assisted reproduction. Birth Defects Res Part C 75: 81–97. Barker, D.J.P. 1998. Mothers, Babies and Health in Later Life, 2nd ed. Edinburgh: Churchill Livingstone. Barker, D.J.P., J.G. Eriksson, T. Forsen, and C. Osmond, C. 2002 Fetal origins of adult disease: strength of effects and biological basis. Int. J. Epidemiol 31: 1235. Barker, D.J.P., T. Forsen, A. Uutela, C. Osmond, and J.G. Eriksson. 2001. Size at birth and

The Developmental Origins of Health and Disease

•

205

resilience to the effects of poor living conditions in adult life: longitudinal study. BMJ 323, 1273–76. Barker, D.J., P.D. Gluckman, K.M. Godfrey, J.E. Harding, et al. 1993a. Fetal nutrition and cardiovascular disease in adult life. Lancet 341: 938–41. Barker, D.J.P., C.N. Hales, C.H.D. Fall, C. Osmond, et al. 1993b. Type 2 (non-insulindependent) diabetes mellitus, hypertension and hyperlipidaemia (syndrome X): relation to reduced fetal growth. Diabetologia 36: 62–67. Barker, D.J., A.W. Shiell, M.E. Barker, and C.M. Law. 2000. Growth in utero and blood pressure levels in the next generation. J Hypertens 18: 843–46. Barroso, I. 2005. Genetics of type 2 diabetes. Diabet Med 22: 517–35. Belizan, J.M., J. Villar, E. Bergel, A. del Pino, et al. 1997. Long term effect of calcium supplementation during pregnancy on the blood pressure of offspring: follow up of a randomised controlled trial. BMJ 315: 281–85. Bhargava, S.K., H.S. Sachdev, C.H. Fall, C. Osmond, et al. 2004. Relation of serial changes in childhood body-mass index to impaired glucose tolerance in young adulthood. N Engl J Med 350: 865–75. Bjorntorp, P. 1995. Insulin resistance: the consequence of a neuroendocrine disturbance? Int J Obesity 19 (suppl 1): S6–S10. Brooks, A.A., M.R. Johnson, P.J. Steer, M.E. Pawson, and H.I. Abdalla. 1995. Birth weight: nature or nurture? Early Hum Dev 42: 29–35. Campbell, D.M., M.H. Hall, D.J.P. Barker, J. Cross, J., et al. 1996. Diet in pregnancy and the offspring’s blood pressure 40 years later. Br J Obstet Gynaecol 103: 273–80. Catalano, P.M., L.P. Husten, A.J. Thomas, and C.M. Fung. 1998. Effect of maternal metabolism on fetal growth and body composition. Diabetes Care 21: B85–B90. Curhan, G.C., W.C. Willett, E.B. Rimm, et al. 1996. Birth weight and adult hypertension and diabetes mellitus in US men. Am J Hypertens 9, 11A. Duggleby, S.L. and A.A. Jackson. 2001. Relationship of maternal protein turnover and lean body mass during pregnancy and birth length. Clin Sci 101: 65–72. Emanuel, I., H. Filakti, E. Alberman, and S.J.W. Evans. 1992. Intergenerational studies of human birthweight from the 1958 birth cohort. I. Evidence for a multigenerational effect. Br J Obstet Gynaecol 99: 67–74. Eriksson, J.G., T. Forsen, J. Tuomilehto, C. Osmond, and D.J.P. Barker. 2001. Early growth and coronary heart disease in later life: longitudinal study. BMJ 322: 949–53. ———. 2003. Early adiposity rebound in childhood and risk of Type 2 diabetes in adult life. Diabetologia 46: 190–4. Eriksson, J.G., V. Lindi, M. Uusitupa, T.J. Forsen, et al. 2002. The effects of the Pro12Ala polymorphism of the peroxisome proliferator-activated receptor-gamma2 gene on insulin sensitivity and insulin metabolism interact with size at birth. Diabetes 51: 2321–24. Fall, C.H.D., C.E. Stein, K. Kumaran, V. Cox, V., et al. 1998. Size at birth, maternal weight, and type 2 diabetes in South India. Diabet Med 15: 220–27. Forsen, T., J.G. Eriksson, J. Tuomilehto, C. Osmond, and D.J.P. Barker. 1999. Growth in utero and during childhood among women who develop coronary heart disease: longitudinal study. BMJ 319: 1403–7. Forsen, T., J.G. Eriksson, J. Tuomilehto, K. Teramo, et al. 1997. Mother’s weight in pregnancy and coronary heart disease in a cohort of Finnish men: follow up study. BMJ 315: 837–40.

206

•

Keith Godfrey and Mark Hanson

Frankel, S., P. Elwood, P. Sweetnam, J. Yarnell, et al. 1996. Birthweight, body-mass index in middle age, and incident coronary heart disease. Lancet 348: 1478–80. Gluckman, P.D., M.A. Hanson, and H.G. Spencer. 2005 Predictive adaptive responses and human evolution. Trends in Ecology and Evolution 20: 527–33. Godfrey, K.M. 2000. Maternal nutrition and fetal development—implications for fetal programming. In Fetal Origins of Cardiovascular and Lung Disease, ed. D.J.P. Barker. New York: National Institutes of Health, 249–71 Godfrey, K.M., D.J.P. Barker, S. Robinson, and C. Osmond. 1997. Maternal birthweight and diet in pregnancy in relation to the infant’s thinness at birth. Br J Obstet Gynaecol 104: 663–67. Godfrey, K.M., T. Forrester, D.J.P. Barker, et al. 1994. Maternal nutritional status in pregnancy and blood pressure in childhood. Br J Obstet Gynaecol 101: 398–403. Hales, C.N. and D.J.P. Barker. 1992. Type 2 (non-insulin-dependent) diabetes mellitus: the thrifty phenotype hypothesis. Diabetologia 35: 595–601. Hales, C.N., D.J.P. Barker, P.M.S. Clark, L.J. Cox, et al. 1991. Fetal and infant growth and impaired glucose tolerance at age 64. BMJ 303: 1019–22. Hanson, M.A. and P.D. Gluckman, eds. 2006. Developmental Origins of Health and Disease— A Biomedical Perspective. Cambridge: Cambridge University Press. Harding, J.E., L. Liu, P. Evans, M. Oliver, and P. Gluckman. 1992. Intrauterine feeding of the growth-retarded fetus: can we help? Early Hum Dev 29: 193–97. Hattersley, A.T. and J.E. Tooke. 1999. The fetal insulin hypothesis: an alternative explanation of the association of low birthweight with diabetes and vascular disease. Lancet 353: 1789–92. Herrick, K., D.I.W. Phillips, S. Haselden, A.W. Shiell, et al. 2004. Maternal consumption of a high-meat, low-carbohydrate diet in late pregnancy: relation to adult cortisol concentrations in the offspring. J Clin Endocrin Metab 88: 3554–60. Hofman, P.L., F. Regan, W.E. Jackson, C. Jefferies, et al. 2004. Premature birth and later insulin resistance. N Engl J Med 351: 2179–86. Huxley, R.R., A.W. Shiell, and C.M. Law. 2000. The role of size at birth and postnatal catchup growth in determining systolic blood pressure: a systematic review of the literature. J Hypertens 18: 815–31. James, W.P.T. 1997. Long-term fetal programming of body composition and longevity. Nutr Rev 55: S41–S43. Jones, A., K.M. Godfrey, P. Wood, C. Osmond, et al. 2006. Growth and the adrenocortical response to psychological stress. J Clin Endocrinol Metab 2006; 91:1868-91. Kannel, W.B., T. Gordon, and M.J. Schwartz. 1971. Systolic versus diastolic blood pressure and risk of coronary heart disease; the Framingham study. Am J Cardiol 27: 335–46. Karter, A.J., S.E. Rowell, L.M. Ackerson, B.D. Mitchell, et al. 1999. Excess maternal transmission of type 2 diabetes: the Northern California Kaiser Permanente Diabetes Registry. Diabetes Care 22: 938–43. Keller, G., G. Zimmer, G. Mall, E. Ritz, and K. Amann. 2003. Nephron number in patients with primary hypertension. N Engl J Med 348: 101–8. Kwong, W.Y., A. Wild, P. Roberts, A.C. Willis, and T.P. Fleming. 2000. Maternal undernutrition during the pre-implantation period of rat development causes blastocyst abnormalities and programming of postnatal hypertension. Development 127: 4195–4202.

The Developmental Origins of Health and Disease

•

207

Leeson, C.P.M., P.H. Whincup, D.G. Cook, A.E. Donald, et al. 1997. Flow-mediated dilation in 9- to 11-year old children: the influence of intrauterine and childhood factors. Circulation 96: 2233–38. Lillycrop, K.A., E.S. Phillips, A.A. Jackson, M.A. Hanson, and G.C. Burdge. 2005. Dietary protein restriction of pregnant rats induces and folic acid supplementation prevents epigenetic modification of hepatic gene expression in the offspring. J Nutr 135: 1382–86. Lithell, H.O., P.M. McKeigue, L. Berglund, R. Mohsen, et al. 1996. Relation of size at birth to non-insulin dependent diabetes and insulin concentrations in men aged 50–60 years. BMJ 312: 406–10. Mackenzie, H.S. and B.M. Brenner. 1995. Fewer nephrons at birth: a missing link in the etiology of essential hypertension? Am J Kidney Dis 26: 91–98. McCance, D.R., D.J. Pettitt, R.L. Hanson, L.T.H. Jacobsson, et al. 1994. Birth weight and non-insulin dependent diabetes: thrifty genotype, thrifty phenotype, or surviving small baby genotype? BMJ 308: 942–45. Merialdi, M., G. Carroli, J. Villar, E. Abalos, et al. 2003 Nutritional interventions during pregnancy for the prevention or treatment of impaired fetal growth: an overview of randomized controlled trials. J Nutr 133: 1626S–1631. Mi, J., C. Law, K-L. Zhang, C. Osmond, et al. 2000. Effects of infant birthweight and maternal body mass index in pregnancy on components of the insulin resistance syndrome in China. Ann Intern Med 132: 253–60. Morton, N.E. 1955. The inheritance of human birth weight. Ann Hum Genet 20: 123–34. Osmond, C., D.J.P. Barker, P.D. Winter, C.H.D. Fall, and S.J. Simmonds. 1993. Early growth and death from cardiovascular disease in women. BMJ 307: 1519–24. Pedersen, J.F. 1980 Ultrasound evidence of sexual difference in fetal size in first trimester. Br Med J 281: 1253. Phillips, D.I.W. and D.J.P. Barker. 1997. Association between low birthweight and high resting pulse in adult life: is the sympathetic nervous system involved in programming the insulin resistance syndrome? Diabet Med 14: 673–77. Phillips, D.I.W., D.J.P. Barker, C.N. Hales, S. Hirst, and C. Osmond. 1994. Thinness at birth and insulin resistance in adult life. Diabetologia 37: 150–54. Phillips, D.I.W., B.R. Walker, R.M. Reynolds, et al. 2000. Low birth weight predicts elevated plasma cortisol concentrations in adults from 3 populations. Hypertension 35: 1301–6. Prentice, A.M. and S.E. Moore. 2005. Early programming of adult diseases in resource poor countries. Archives of Disease in Childhood 90: 429–32. Ravelli, A.C.J., J.H.P. van der Meulen, R.P.J. Michels, et al. 1998. Glucose tolerance in adults after prenatal exposure to famine. Lancet 351: 173–77. Rich-Edwards, J.W., M.J. Stampfer, J.E. Manson, B. Rosner, et al. 1997. Birth weight and risk of cardiovascular disease in a cohort of women followed up since 1976. BMJ 315: 396–400. Rolland-Cachera, M.F., M. Deheeger, M. Guilloud-Bataille, P. Avons, et al. 1987. Tracking the development of adiposity from one month of age to adulthood. Ann Hum Biol 14: 219–29. Roseboom, T.J., J.J. van der Meulen, A.C. Ravelli, C. Osmond, et al. 2001. Effects of prenatal exposure to the Dutch famine on adult disease in later life: an overview. Mol Cell Endocrinol 185: 93–98.

208

•

Keith Godfrey and Mark Hanson

Rush, D. 1989. Effects of changes in maternal energy and protein intake during pregnancy, with special reference to fetal growth. In Fetal Growth, eds. F. Sharp, R.B. Fraser, and R.D.G. Milner. London: Royal College of Obstetricians and Gynaecologists, 203–33. Shiell, A.W., D.M. Campbell, M.H. Hall, and D.J Barker. 2000. Diet in late pregnancy and glucose-insulin metabolism of the offspring 40 years later. Br J Obstet Gynaecol 107: 890–95. Shiell, A.W., M. Campbell-Brown, S. Haselden, S. Robinson, et al. 2001. A high meat, low carbohydrate diet in pregnancy: relation to adult blood pressure in the offspring. Hypertension 38: 1282–88. Stettler, N., S.K. Kumanyika, S.H. Katz, B.S. Zemel, and V.A. Stallings. 2003. Rapid weight gain during infancy and obesity in young adulthood in a cohort of African Americans. Am J Clin Nutr 77: 1374–78. Stewart, R.J.C., H. Sheppard, R. Preece, and J.C. Waterlow. 1980. The effect of rehabilitation at different stages of development of rats marginally malnourished for ten to twelve generations. Br J Nutr 43: 403–12. Taylor, D.J., C.H. Thompson, G.J. Kemp, P.R.J. Barnes, et al. 1995. A relationship between impaired fetal growth and reduced muscle glycolysis revealed by 31P magnetic resonance spectroscopy. Diabetologia 38: 1205–12. Torrens, C., L. Brawley, F.W. Anthony, C.S. Dance, et al. 2006. Folate supplementation during pregnancy improves offspring cardiovascular dysfunction induced by protein restriction. Hypertension 47: 982–87. Torrens, C., L. Brawley, A.C. Barker, S. Itoh, et al. 2003. Maternal protein restriction in the rat impairs resistance artery but not conduit artery function in pregnant offspring. J Physiol 547: 77–84. Valdez, R., M.A. Athens, G.H. Thompson, B.S. Bradshaw, and M.P. Stern. 1994. Birthweight and adult health outcomes in a biethnic population in the USA. Diabetologia 37: 624–31. Walker, S.K., K.M. Hartwick, and J.S. Robinson. 2000. Long-term effects on offspring of exposure of oocytes and embryos to chemical and physical agents. Hum Reprod Update 6: 564–67. Wang, L., L. Kong, F. Wu, Y. Bai, and R. Burton. 2005. Preventing chronic diseases in China. Lancet 366: 1821–24. Wei, J.N., F.C. Sung, C.Y. Li, C.H. Chang, et al. 2003. Low birth weight and high birth weight infants are both at an increased risk to have type 2 diabetes among schoolchildren in Taiwan. Diabetes Care 26: 343–48. Whincup, P.H., D.G. Cook, F. Adshead, S.J.C. Taylor, et al. 1997. Childhood size is more strongly related than size at birth to glucose and insulin levels in 10-11-year-old children. Diabetologia 40: 319–26. Whitaker, R.C., J.A. Wright, M.S. Pepe, K.D. Seidel, and W.H. Dietz. 1997. Predicting obesity in young adulthood from childhood and parental obesity. N Engl J Med 337: 869–73. Winston, R.M.L. and K. Hardy. 2002. Are we ignoring potential dangers of in vitro fertilization and related treatments? Nature Cell Biol 2: s14–s18. World Bank. 2006. Repositioning Nutrition as Central to Development. Washington, D.C.: World Bank.

•8• Beyond the Gradient An Integrative Anthropological Perspective on Social Stratification, Stress, and Health Thomas McDade The Social Gradient in Health Socioeconomic deprivation is a fundamental form of adversity that leads to poor health by almost any measure. Indeed, socioeconomic status (SES) is among the most powerful predictors of health, yet it also among the least well understood. The terms health disparities (preferred in the US) and health inequalities (preferred in the UK) refer to the uneven population distribution of morbidity and mortality, with a particular emphasis on the disproportionate burden of disease endured by particular racial/ ethnic groups and those at lower levels of SES. African-Americans, for example, suffer from cardiovascular disease, lung cancer, breast cancer, infant mortality, and total mortality at rates that are two to four times higher than those of other racial/ethnic groups in the United States (Keppel, Pearcy, and Wagener 2002). Similarly, a consistent socioeconomic gradient has been documented for cardiovascular disease, diabetes, gastrointestinal disease, arthritis, adverse birth outcomes, accidents, and many (but not all) forms of cancer (Adler and Ostrove 1999). For some in the biomedical research establishment, the social gradient in health is “the major unsolved public health problem of the industrialized world” (Marmot et al. 1997: 901). Despite the fact that race/ethnicity and social class are often conflated (Williams 1999), research has demonstrated that SES alone cannot account for health disparities between ethnic groups and that associations of SES with health are not uniform across such groups (Farmer and Ferraro 2005; Ostrove et al. 2000; Nazroo 1998). It is notable that the social gradient in health has been found in almost every

210

•

Thomas McDade

industrialized population in which it has been studied and is not, therefore, unique to any particular system of health care. In addition, the effects of SES on health are not limited to conditions of poverty; incremental gains in socioeconomic position are associated with positive returns to health across the entire range of SES (Adler and Ostrove 1999). Although race, class, and health are central concerns of many subfields of anthropology, relatively few anthropologists—with some notable exceptions—have contributed directly to understanding health disparities in the United States. Bioarchaeologists have uncovered the historical origins of health inequalities as a consequence of social stratification following agricultural intensification, sedentarism, and population growth (Armelagos, Brown, and Turner 2005; Cohen 1989). Recent work by human biologists has investigated child growth and nutritional status in the American South (Crooks 1999), the effects of environmental pollutants on growth and development in a Native American population (Schell and Tarbell 1998), and racial/ethnic differences in physical activity and obesity (Gordon-Larsen, Adair, and Popkin 2003). Recently, Dressler, Oths, and Gravlee (2005) have made an explicit attempt to construct a biocultural framework for advancing future research on race, stress, and health. Recognizing the importance of SES to health, biomedical research typically uses standardized indicators of race/ethnicity and SES as “control” variables that may confound associations between exposures and outcomes of interest, but this approach cannot capture the complexity of important social processes linking SES and health, particularly across diverse cultural (and subcultural) groups. This chapter reviews evidence for the contribution of psychosocial stress to the social gradient in health, and draws on prior anthropological research to 1) demonstrate the value of minimally invasive biomarker methods for measuring health in community-based settings; and 2) highlight the importance of cultural factors in defining social status and its relevance to stress. Conceptual and methodological tools from cross-cultural research are equally useful in the United States and other Western settings, and they may suggest innovative new directions for research on stratification, stress, and health. Psychosocial Stress and Human Health There are several mechanisms contributing to the social gradient in health, including differential availability of material and social resources, exposure to noxious substances, access to health care, and negative health behaviors. Conversely, poor health may limit opportunities for economic achievement (Adler et al. 1994; Marmot et al. 1997). However, these factors cannot account for all of the health gradient, leading many scholars to consider psychosocial stress as an important mediator of SES effects on disease risk (Adler and Snibbe 2003; Seeman and Crimmins 2001; Siegrist and Marmot 2004).

Beyond the Gradient

•

211

What is stress? In the social/behavioral sciences, the term “stress” has been used to indicate adverse environmental circumstances, perceptions of discomfort or burden, reports of mental, behavioral, emotional and physical problems, or clinical and/ or physiological assessment of biological outcomes. The vast majority of stress research draws on the early work of Hans Selye (1976), whose “General Adaptation Syndrome” provides an implicit framework for research in the behavioral, biological, and biomedical sciences. Although the terms have changed over the years, and different aspects of the stress process have received varying degrees of attention, the conceptual foundation remains largely the same (Elliott and Eisdorfer 1982; McEwen and Stellar 1993). From this perspective, the term stress describes a process that incorporates the following elements: 1) stressor; 2) response; 3) consequences; and 4) moderators. A stressor is an environmental event or situation that disrupts normal functioning and poses an adaptive challenge to the individual. The reaction that follows is a response, representing an attempt by the individual to restore homeostasis, or maintain stability around a new baseline (also known as allostasis) (Sterling and Ayer 1988). Responses, while adaptive for coping with the immediate challenge, may nonetheless have deleterious consequences for an individual’s health, particularly if they are especially severe, frequent, or sustained for a prolonged period of time. Moderators include developmental, genetic, or situational factors that contribute to individual differences in the pathways linking stressors, responses, and/or consequences. Psychosocial stressors initiate the activation of multiple physiological systems, including the sympathetic nervous system (SNS), the sympathetic adrenal medullary system (SAM), and the hypothalamic pituitary adrenocortical (HPA) axis. Epinephrine and norepinephrine (products of SNS and SAM activity), and cortisol (the end product of the HPA axis), are the primary hormonal products of the physiological stress response. It is through these neuroendocrine mechanisms that the adverse effects of stress on health are primarily mediated (Cacioppo et al. 2000; Johnson et al. 1992; McEwen 1998). Stress represents a major means through which our physical and social environments affect our well-being, with demonstrated impact on mental and physical health throughout the life course. Prolonged or severe exposure to psychosocial stressors increases risk for cardiovascular disease, infectious disease, slowed wound healing, diabetes, impaired growth, poor birth outcomes, and reduced reproductive function (Sapolsky 2004). Recent research has employed the concept of “allostatic load” as a summary measure of dysregulation across multiple physiological axes resulting from stress (McEwen 1998; Seeman et al. 2001). Measures of low socioeconomic status have been associated with increased allostatic load in a number of community-based cohorts, and prospective research has associated baseline allostatic load with increased risk for allcause mortality, cardiovascular disease, and declines in cognitive and physical function (Evans and English 2002; Seeman et al. 1997, 2001; Karlamangla et al. 2002).

212

•

Thomas McDade

Models and Methods for Research on Stratification, Stress, and Health: A Samoan Case Study Current research into the contribution of psychosocial stress to health disparities faces two fundamental challenges. First, how do we measure health, particularly when conducting research in diverse cultural and ecological settings? How do we know when someone is “under stress,” and how can we identify individuals at risk for poor health outcomes? Second, how do we derive measures of social status that capture meaningful aspects of everyday experience? In particular, if we expect that stress is a key pathway linking adversity and health, then we have to think carefully about how our models represent these psychosocial processes. On both accounts, conceptual and methodological tools from biocultural anthropology may provide some innovative answers. Methods: Minimally Invasive Tools for Community-Based Health Research Disciplinary traditions are such that the vast majority of stress research in the social/ behavioral sciences features in-depth analysis of psychosocial, economic, and/or cultural processes in diverse populations in conjunction with self-report measures of stress and health. In contrast, biomedical research employs sophisticated assessment of physiology but typically relies on small clinic-based samples and rarely evaluates social contexts beyond standard measures of socioeconomic status or self-reported health behaviors. However, the recent development and application of minimally invasive methods for measuring physiological function in field settings is helping to overcome these obstacles to integrative health research (Ellison 1988; McDade, Williams, and Snodgrass 2007; Worthman and Stallings 1997). There are a number of well-validated and widely used self-report measures of stress and health (e.g., Cohen, Kessler, and Gordon 1997). These measures are administered through interviews or in writing and are easy to apply to large numbers of research participants with relatively minimal cost and effort. These methods are limited, however, by their potential for recall or reporting bias and by their lack of specificity with respect to revealing clinically relevant outcomes and underlying physiological processes (Panter-Brick et al. 2001). In addition, these methods are subject to the problem of “confounded measures,” where self-reports of stress are significantly correlated with self-reports of physical problems in the absence of underlying pathology (Dohrenwend et al. 1984; Watson and Pennebaker 1989). Lastly, the experience of stress is situated and constructed within specific social, cultural, and political-economic contexts (O’Neil 1986; Young 1980), and gaining access to this experience may be particularly difficult when working across linguistic or cultural settings. In other words, when someone reports being “stressed,” how can we be sure what that really means? Biological indicators of health surmount many of these obstacles, and a growing number of social scientists are adapting methods from the biomedical sciences

Beyond the Gradient

•

213

Table 8.1. Measuring Stress: Advantages and Disadvantages of Using Physiological Indicators

Expanding knowledge of stress physiology presents new opportunities for measuring stress at multiple levels, and many studies have employed objective stress biomarkers such as cortisol, catecholamines, blood pressure, and EBV antibodies. Advantages (in contrast to more common self-report methods) 1. Objective data on stress that is not vulnerable to recall, reporting, or observer bias, that is valid across populations despite linguistic or cultural differences, and that provides information on individuals for whom self-report data may not be possible (e.g., infants, children). 2. Access to direct information on the mechanisms linking social and biological phenomena that are causally proximate to important health outcomes, and that can be used to identify individuals most at risk for current or future disease. 3. Opportunities for measuring physiological function in community—rather than clinical or laboratory—settings that encourage the recruitment of more diverse, representative samples, thereby increasing the generalizability of research findings.

Disadvantages 1. Logistical challenges and ethical issues that increase the costs of data collection—in terms of time, money, and participant burden—that require that biomarkers be implemented only in the service of a wellarticulated interdisciplinary research agenda. 2. Epistemological questions regarding the definition of stress: Biomarkers rarely map neatly onto cognitive, self-report measures of perceived stress, leading some to question how they should be interpreted as representing “stress.” 3. Possibility of confounding by factors other than psychosocial stressors, since physiological systems are responsive to a wider range of inputs (e.g., undernutrition, infection, and psychosocial stress, all of which upregulate the HPA axis).

to document the physiological impact of social contexts and processes (National Research Council 2001). A major advantage of these markers is their relative objectivity: Since they are beyond the conscious control of research participants they do not rely on participants’ ability to access or recall relevant health information or their willingness to share this information. This may be particularly advantageous for research with children and for research in diverse populations where linguistic or cultural factors may contribute to variation in the perception, experience, and/or reporting of health (Hahn 1995; Kleinman 1986).

214

•

Thomas McDade

An additional advantage is the fact that biomarkers provide direct information on the physiological pathways that link social contexts, stress, and health. For example, biomarkers of cardiovascular function tend to “track” over time, such that an individual’s risk of developing disease compared to his or her peers remains consistent from childhood into adulthood (Berenson et al. 1995; Li et al. 2004). In other words, a child who has relatively high blood pressure early in life will also likely have high blood pressure later in life, even if his or her blood pressure in childhood does not exceed clinically defined thresholds for hypertension. The measurement of biomarkers can therefore provide insight into the predisease pathways that help identify individuals most at risk for the future development of disease. Of course, requirements for sample collection and processing are major disincentives for implementing biomarkers into population-based health research. Clinical assays of endocrine or immune function typically require significant quantities of serum or plasma, and these samples must be centrifuged, separated, and promptly frozen or assayed to maintain sample integrity. In addition, venipuncture blood sampling, a relatively invasive procedure that requires the skills of a trained medical professional, may be unacceptable in certain cultural contexts and is particularly problematic with children. For these reasons, the vast majority of research on critical aspects of human physiology has been based in clinical or laboratory settings. Biological anthropologists and human population biologists have long-standing interests in documenting the contributions of social, cultural, and ecological contexts to human biological variation (Stinson et al. 2000). To this end, methodological innovation has been central to the discipline as we seek to develop and implement tools that allow us to investigate human physiology and health in diverse populations around the world. These populations are often remote, and in many cases access to basic laboratory facilities, or even electricity, is limited. Anthropometric measures of nutritional status (e.g., height, weight, body composition) have been, and continue to be, powerful tools for documenting the adverse short- and long-term effects of impoverished environments (Bielicki 1986; Bogin 1999). These measures are relatively easy to obtain in community-based research and are sensitive indicators of individual as well as population health (WHO 1995). More recently, anthropologists have turned their attention to investigating the direct physiological consequences of social contexts and to linking these contexts to health outcomes through specific physiological mechanisms (Panter-Brick 1998; Panter-Brick and Worthman 1999). Much of this work has focused on physiological responses to stress. Saliva and urine are relatively easy to collect in field settings and can provide information on HPA and SAM axis activity, as well as measures of reproductive function, that have been linked to stressors across a wide range of cultural and ecological settings (Brown 1981; Ellison et al. 1993; Flinn and England 1995; Hanna, James, and Martz 1986; Pollard, Ungpakorn, and Harrison 1992). Unfortunately, the number of factors that can be measured in saliva or urine is limited since most analytes do not enter these

Beyond the Gradient

•

215

solutions from general circulation in a measurable form. As such, for many research questions, blood samples are still a necessity. The relatively recent application of dried blood spot samples to population-based health research provides a viable option for circumventing limitations associated with collecting blood through venipuncture (McDade et al.2007; Mei et al. 2001; Worthman and Stallings 1997). A sterile, disposable micro-lancet is applied to the participant’s finger to stimulate capillary blood flow, and one to five drops of whole blood (about 50 µL each) are collected on standardized filter paper (Whatman #903) that is designed to diffuse and dry blood at a constant, uniform rate. Samples dry at room temperature for at least four hours and are then stacked and stored until analysis in the laboratory. The paper matrix stabilizes the sample, preserving its contents for an extended period of time (Figure 8.1). The collection of whole blood on filter paper has several advantages over venipuncture that make it ideal for field-based research, namely: 1) collection is relatively painless and noninvasive; 2) samples do not need to be centrifuged, separated, or immediately frozen following collection (blood spots for most analytes are stable at ambient temperatures for up to two weeks and often longer); 3) samples are easily stored and transported; and 4) multiple assays can be performed from a single 50-µL drop of blood. Dried blood spots have been used to assay a growing number of analytes, including markers of infectious disease, stress, immune function, iron deficiency, metabolic activity, and reproductive function. They have been successfully collected in a wide range of cultural and ecological settings, including lowland Bolivia, Papua New Guinea, the Philippines, Kenya, Samoa, Nepal, and the United States.

Figure 8.1. Example of blood spot collection

216

•

Thomas McDade

A primary advantage of dried blood spots is that they make feasible the collection of blood samples from large numbers of people in their homes or other local settings. This is a major boon to community-based research on adversity and health. It is now possible to combine rich measurement of social and cultural contexts with highquality, objective physiological information to gain a better understanding of how social contexts “get under the skin” to shape health. By adapting clinical methods to field-based settings, we can draw larger, more diverse, representative samples that increase the generalizability of research findings and that may help us identify subgroups of individuals, or subsets of environments, that merit special attention. For these reasons, a number of large US-based health surveys (e.g., the National Longitudinal Study of Adolescent Health, the Health and Retirement Study) have recently incorporated the collection of biomarkers to complement their panel of psychosocial and demographic measures. Models: Linking Stratification, Stress, and Health Current health disparities research typically applies measures of income, education, and/or occupation as indicators of an individual’s social position, and enters these measures into statistical models as simple linear predictors of various health outcomes (continuous variables are on occasion categorized to consider nonlinear or threshold effects). While this gradient approach has consistently demonstrated that SES is among the strongest predictors of health, its explanatory power is limited, and it fails to capture the more proximate processes linking stratification, stress, and health. For these reasons there is growing recognition of the need for additional, more meaningful measures and models of social status and health (Dressler and Bindon 2000; Kaplan 1996; Ostrove et al. 2000; Nazroo 1998; Krieger, Williams and Moss 1997). A welcome innovation is the recent development of a subjective SES measure, in which individuals are presented with a drawing of a ladder and asked to mark the rung that corresponds to their position in society (Adler et al. 2000). Subjective SES is moderately correlated with objective SES measures and is an independent predictor of a number of health-related variables (Adler et al. 2000; Ostrove et al. 2000). While the ladder moves beyond income/education/occupation to tap directly into the personal, perceived experience that may mediate associations among SES and health, it continues to work within the classic gradient framework that expects direct, linear associations with health. Culture is a foundational concept for anthropology, and although it has been repeatedly contested within the discipline, few anthropologists would deny that it is a defining attribute of the human species that provides a framework for meeting subsistence and other material needs, structuring social and economic relations, and creating systems of meaning that motivate behaviors and beliefs. Just as we investigate how cultural and ecological factors contribute to human biological variation, we recognize that these factors shape variation in individual psychosocial experience. In particular, the significance of specific social status markers—and their relevance to

Beyond the Gradient

•

217

stress and health—will vary across time and space. It is therefore critical that we develop models of social status and stress that are informed by the cultural and politicaleconomic contexts within which they operate. In this section I discuss my attempt to do just that in a study of adolescent stress in Samoa. The islands of Samoa have served as an important testing ground for evaluating the relationships among culture change, stress, and health (e.g., Baker 1986; Bindon 1997; James et al. 1987; McGarvey and Baker 1979; Pearson, James, and Brown 1993). Historically, Samoans have lived primarily in small seaside villages, surrounded by extended family and engaging in subsistence cultivation of family-owned lands, with some cash-cropping of copra and cocoa (Mead 1928; Shore 1982). Increasingly, Samoans are taking advantage of new opportunities for employment, commerce, and education as Western institutions and lifestyles become an increasingly visible part of the islands. More children are attending Western-style schools for longer periods of time, consumer goods and services are increasingly available and desired, and sons and daughters are frequently traveling to American Samoa, New Zealand, and Hawaii for education or to earn money to send back to the family (Macpherson 1994; Mageo 1988; O’Meara 1990). But despite these intrusions, Samoans have managed to maintain a strong cultural identity that values the fa’aSamoa, or the “Samoan way.” These shifting cultural and economic environments make Samoa an ideal location to investigate the adaptive challenges and opportunities associated with globalization, and a major research effort launched in the 1970s by Baker and colleagues (Baker, Hanna, and Baker 1986) established this as an important area of inquiry in biological anthropology. It was in this tradition that I sought to investigate social status and adolescent stress in the context of cultural transitions in Samoa. A recent suicide epidemic among adolescents and young adults—with rates two to four times higher than in the United States—lends a certain urgency to understanding the underlying causes of psychosocial stress in Samoa, and suggests an uneasy interface between the fa’aSamoa and encroaching Western lifestyles (Bowles 1985; McDade 2002). The study included 352 individuals between the ages of 10 and 20 recruited from nine villages and five neighborhoods, representing the full range of articulation with Western lifestyles in Samoa. Each participant was interviewed to gather demographic and psychosocial information, followed by the collection of standard anthropometric measurements and a finger prick blood spot sample for subsequent biomarker analysis (McDade, Stallings, and Worthman 2000). I measured antibodies against the Epstein-Barr virus (EBV) in blood spot samples as a biomarker of chronic psychosocial stress (McDade et al. 2000). A ubiquitous herpes virus, EBV takes up residence permanently following infection, and cell-mediated immune processes are primarily responsible for maintaining the virus in a latent state (Henle and Henle 1982). Stress-induced immunosuppression allows EBV to reactivate and release viral antigens into circulation, to which a second-tier, humoral antibody response may emerge (Glaser et al. 1991). As a result, levels of antibodies

218

•

Thomas McDade

against EBV antigens provide an indirect measure of an aspect of cell-mediated immune function, such that increased EBV antibody titers indicate lower cell-mediated immunity (Glaser 1987; 1993). Psychoneuroimmunologists have demonstrated the utility of the EBV antibody model in cross-sectional as well as prospective studies. Increased antibody titers have been associated with a wide range of naturalistic stressors, including negative life events, academic stress, strained social relationships, and emotional distress (Esterling et al. 1992, 1993; Glaser et al. 1993; Kiecolt-Glaser et al. 1987, 1987, 1988; Lutgendorf et al. 1994; McDade et al. 2000; McDade 2001, 2002, 2003). Furthermore, in comparison with other measures of immunity, meta-analysis has identified EBV antibodies as among the strongest and most consistent correlates of chronic stress (Herbert and Cohen 1993). Since the duration of time elapsing between a stressor and EBV antibody response is on the order of days or weeks, EBV antibody levels are not subject to shortterm fluctuation, acute context effects, or diurnal variation, and a single sample can thus be used as an immunological measure of chronic stress. This is an advantage over other stress biomarkers, which require multiple samples or are sensitive to the time of day or circumstances under which they are collected. And, as mentioned, using a physiological indicator of chronic stress surmounts many of the shortcomings associated with self-reports. A shortcoming of this method is that results may be confounded by nutritional status or the presence of infection, both of which were controlled for in all analyses with objective measures (McDade et al. 2000). The gradient, part I: Is household SES related to stress?

According to the gradient approach to health disparities, low socioeconomic status should be associated with higher burdens of stress. The challenge in the Samoan case is to derive a meaningful household-level measure of socioeconomic resources. A household-level measure is particularly appropriate in this case since children and adolescents have yet to complete their schooling and establish households independent of their parents. Furthermore, in Samoa, one’s social status is intimately tied to that of the family, and children and adolescents are expected to serve their elders and contribute to the household economy (O’Meara 1990; Shore 1982). Since information on household income was not directly available, each household’s socioeconomic position was estimated by the occupational rank of the father and mother as follows: 1=planter/housewife (no cash income); 2=unskilled wage labor (domestic help, factory work, security, etc.); 3=skilled labor/professional (teacher, government employee, engineer, etc.) (McDade 2001). The relatively low number of individuals with wage-earning jobs (35.9 percent) precluded a finer level of distinction. In addition, remittances from relatives working overseas make important contributions to the Samoan economy (O’Meara 1990), and information was collected on the number of family members regularly sending money home. A summary score of household socioeconomic status was obtained by summing the remittances sent

Beyond the Gradient

•

219

home by wage earners, mother’s occupational rank, and father’s occupational rank (mean SES=4.1, range 2–9). Figure 8.2 presents the association between socioeconomic status and EBV antibody level. A linear effect is evident, in which lower economic status is associated with higher EBV antibody level, indicating reduced cell-mediated immune function and—according to the EBV model—a higher burden of psychosocial stress. This result is consistent with prior work using the gradient approach, but the effect of economic status is small and of marginal statistical significance. The gradient, part II: Is matai absence a source of stress?

An alternative approach is to use a locally defined, historically significant marker of social status specific to the Samoan cultural context. One might expect such a marker to be more meaningful for the everyday lives of Samoan adolescents, and therefore be a stronger predictor of stress. The logic behind this ethnographically informed gradient approach is similar to that of recent work by Dressler and colleagues that uses “cultural consonance” as a tool for assessing the degree to which individuals conform to locally defined cultural models for a “successful” lifestyle (Dressler and Bindon 2000). The matai system of village political organization is a central feature of the fa’aSamoa (O’Meara 1990; Shore 1982). Extended families each elect by consensus a matai to represent them at the council of chiefs (fono) and to exert authority over the family land, property, and labor. The council of chiefs meets frequently to act as

Figure 8.2. Association between household economic status (tertiles, based on parental occupation) and log-transformed EBV antibody level (mean ± SE) in Samoan adolescents (10 to 20 years), controlling for age, sex, region of residence, BMI, and current infection.

220

•

Thomas McDade

the legislative, executive, and judicial bodies of local governance (Shore 1982). While the matai system is currently in flux and chiefs increasingly face emerging sources of power that circumvent the matai system, considerable honor and respect continue to be conferred upon individuals who possess a matai title in all regions of Samoa (O’Meara 1990; Shore 1996). Members of their household also command a certain respect in daily life, and achieve significant social standing due to their affiliation with a matai. This raises the question as to whether matai presence may be an important local marker of social status that is related to adolescent stress. In this study, 47 percent of the adolescents lived in households with a matai titleholder present. Not surprisingly, matai households scored significantly higher on the SES scale than non-matai households (N=157 and 172, respectively; 4.3 vs. 3.8, p160/95 Region

n

Mean

Standard deviation

Median

Africa Caribbean Europe

8 7 8

12.93% 23.7% 31.48%

7.71% 9.7% 19.06%

10.6% 22.9% 28.5%

Fisher’s exact test X2=6.09, df=2, p=0.04

9.1.b. Hypertension defined as=SBP>140 Region

n

Mean

Standard deviation

Median

Africa Caribbean Europe North America

5 7 2 8

18.71% 21.76% 35.35% 37%

7.6% 20.21% 9.26% 10.0%

22.3% 13.0% 35.35% 34.6%

Fisher’s exact test X2=6.35, df=2, p=0.04. Europe excluded from statistical test because of small sample size

9.1.c. Hypertension defined as=BP>160 Region

n

Mean

Standard deviation

Median

Africa Caribbean Europe South America

6 17 2 2

7.22% 24.7% 28.15% 7.8%

1.48% 12.6% 1.2% 0.71%

7.15% 20.6% 28.15% 7.8%

Fisher’s exact test X2=11.05, df=1, p=0.0009. Europe and South America excluded from statistical test because of small sample size.

9.1.d. All studies combined, regardless of hypertension definition used Standard Region n Mean deviation

Median

Africa Caribbean Europe North America South America

13.1% 27.1% 28.9% 33.7% 13.0%

32 43 12 12 7

18.04% 26.09% 31.57% 34.8% 14.1%

Fisher’s exact test X2=24.08, df=4, p=0.0001.

13.39% 12.40% 15.61% 9.1% 8.24%

The Slavery Hypothesis

•

245

Very few studies were available from Afro–South American groups. However, they show a similar picture to what we see in Africa, namely, moderate hypertension rates are found in urbanized areas such as São Paulo and Rio de Janeiro (29.2 and 13 percent in males and females, 14.4 and 20.2 percent in males and females respectively) and low rates are found in rural groups (6.28 percent for males, and 6.9 percent for females). Time Depth Some studies in Caribbean groups from the late 1950s and 1960s report very high hypertension rates. Unfortunately, not all studies used the same variables (urban vs. rural) so the comparison across different papers is not perfect. For example, the hypertension rates in Jamaica reported by Miall et al. in 1962 (33.8 and 36.6 percent in rural and urban females, and 16.8 and 20.6 percent respectively rural and urban males) are much higher than those reported several decades later by Cooper et al. (1997a; 13 percent in males, 20.6 percent in females) and Dressler et al. 1988 (17.8 percent in males, 15.1 percent in females), both of which studies do not divide their subjects by urban/rural residency. Similarly high hypertension rates are reported in 1959 in Grand Bahama and Nassau (27.1 percent in males and 33.8 percent in females respectively) and in 1962 in St. Kitts (40.7 percent in females). Such high hypertension rates in some Caribbean groups in the late 1950s and early 1960s might be due to instrument differences between these early studies and subsequent ones. Alternatively, these early papers might truly indicate that hypertension was high and decreased afterward, perhaps because socioeconomic status improved. In support of this latter explanation is the fact that whereas the females in the 1962 St. Kitts sample had a very high rate of hypertension, the males did not. If the high hypertension rates were a result of the instruments, we would expect to see high prevalence rates in both sexes. Comparisons by Hypertension Definition There are several definitions of hypertension used by these studies. In an effort to control for differences in prevalence rates owing to differences in definitions of hypertension, we divided the studies into three groups according to hypertension thresholds, which included the largest number of studies (BP>160/95, SBP>160 mm Hg, and SBP>140 mm Hg), and tested for regional differences with Kruskal-Wallis tests. Without exception, Africa has the lowest mean prevalence rate, followed by South America, the Caribbean, Europe, and North America. In general, median HPR follows the same pattern observed in mean blood pressure, although for SBP>140 mm HG, median HPR is higher in Africa than in the Caribbean, and higher in Europe than in North America. For each of the three definitions of hypertension, the regions differed statistically. Since observed regional patterns of HPR are consistent regardless of hypertension definition, and initial Kruskal-Wallis tests were all significant, we looked at

246

•

Lorena Madrigal, Mwenza Blell, Ernesto Ruiz, Flory Otárola-Durán

HPRs across all studies (Table 9.1d), without considering study-specific definitions of hypertension. This allowed more statistical power as well as inclusion of South American samples (excluded due to small sample size when analysis was specific to hypertension threshold). Regional Differences Figure 9.5 shows that regions do differ in HPR, regardless of study-specific hypertension definition (X2 = 24.08, df = 4, p< 0.001). HPR in Europe do not differ signifi-

Figure 9.5. Median hypertension prevalence rate by region including both genders, all definitions of hypertension UK US UK Caribbean S. America Africa

Summary of comparisons between regions Caribbean South America

X2= 0.48 df=1,