Including International Studies in the Epidemiology of Diabetes [1st Edition] 9781483215518

Table of contents :
Content:
Advances in Metabolic DisordersPage ii
Front MatterPage iii
Copyright pagePage iv
ContributorsPages xiii-xv
Prologue to International Studies in the Epidemiology of DiabetesPages xvii-xviiiPETER H. BENNETT, MAX MILLER
Comparison of Diabetes Prevalence Rates in Oxford (1946) and Sudbury (1964)Pages 1-11JOHN B. O'SULLIVAN, ROY M. ACHESON
The Incidence of Diabetes Mellitus in Rochester, Minnesota, 1945–1969Pages 13-28P.J. PALUMBO, DARWIN R. LABARTHE
Diabetes in American IndiansPages 29-48KELLY M. WEST
Incidence of Diabetes among the Pima IndiansPages 49-63RICHARD F. HAMMAN, PETER H. BENNETT, MAX MILLER
Blood Sugar and Serum Insulin Levels in Jamaica, West IndiesPages 65-91C. DU V. FLOREY
Variations in Incidence of Diabetes among 10,000 Adult Israeli Males and the Factors Related to Their Development‖Pages 93-110JACK H. MEDALIE, JOSEPH B. HERMAN, URI GOLDBOURT, CHERI M. PAPIER
Epidemiology of Diabetes in South AfricaPages 111-146W.P.U. JACKSON
Prevalence of Diabetes in IndiaPages 147-165O.P. GUPTA, M.H. JOSHI, S.K. DAVE
Vascular Complications in Diabetes in JapanPages 167-200YOSHIO GOTO
Preliminary Studies of the Prevalence and Mortality of Diabetes Mellitus in Japanese in Japan and on the Island of HawaiiPages 201-224RYOSO KAWATE, MICHIHIRO MIYANISHI, MICHIO YAMAKIDO, YUKIO NISHIMOTO
The High Prevalence of Diabetes Mellitus in Nauru, A Central Pacific IslandPages 225-240PAUL ZIMMET, PINCUS TAFT
The Relationships of Diabetes, Blood Lipids, and Uric Acid Levels in PolynesiansPages 241-261I.A.M. PRIOR, R. BEAGLEHOLE, FLORA DAVIDSON, CLARE E. SALMOND
HLA Studies in Diabetes Mellitus: A ReviewPages 263-277JØRN NERUP
EpiloguePages 279-281PETER H. BENNETT, MAX MILLER
The Central Nervous System, Pancreatic Hormones, Feeding, and ObesityPages 283-312STEPHEN C. WOODS, DANIEL PORTE Jr.
The Role of Insulin Resistance in the Pathogenesis of Diabetes MellitusPages 313-331GERALD M. REAVEN, JERROLD M. OLEFSKY
The Pancreas in Idiopathic Diabetes§Pages 333-365BRUNO W. VOLK, KLAUS F. WELLMANN
SomatostatinPages 367-424SUAD EFENDIĆ, TOMAS HÖKFELT, ROLF LUFT
Dental Maturity as a Measure of Somatic Development in ChildrenPages 425-451RUNE FILIPSSON, KERSTIN HALL, JAN LINDSTEN
Subject IndexPages 453-458
Contents of Previous VolumesPages 459-462

Citation preview

ADVANCES IN

Metabolic Disorders Edited by

Rachmiel Levine

Rolf Luft

City of Hope Medical Center Duarte, California

Department of Endocrinology and Metabolism Karolinska Hospital Stockholm, Sweden

Supplementary Volumes 1. Early Diabetes Edited by Rafael A. Camerini-Davalos and Harold S. Cole 2. Vascular and Neurological Changes in Early Diabetes Edited by Rafael A. Camerini-Davalos and Harold S. Cole

ADVANCES IN

Metabolic "Disorders Edited by

Rachmiel Levine

Rolf Luft

VOLUME 9

Including

International Studies in the Epidemiology of Diabetes Edited by

M a x Miller

Peter H. Bennett

Department of Medicine Case Western Reserve University Cleveland, Ohio

National Institute of Arthritis, Metabolism, and Digestive Diseases Phoenix, Arizona

1978

ACADEMIC PRESS

NEW YORK ·

SAN FRANCISCO ·

A Subsidiary of Harcourt Brace Jovanovich, Publishers

LONDON

COPYRIGHT © 1 9 7 8 , BY ACADEMIC PRESS, I N C . ALL RIGHTS RESERVED. NO PART OF THIS PUBLICATION MAY BE REPRODUCED OR TRANSMITTED IN ANY FORM OR BY ANY MEANS, ELECTRONIC OR MECHANICAL, INCLUDING PHOTOCOPY, RECORDING, OR ANY INFORMATION STORAGE AND RETRIEVAL SYSTEM, WITHOUT PERMISSION IN WRITING FROM THE PUBLISHER.

ACADEMIC PRESS, INC.

I l l Fifth Avenue, New York, New York 10003

United Kingdom Edition published by ACADEMIC PRESS, INC. (LONDON) LTD. 24/28 Oval Road, London NW1

LIBRARY OF CONGRESS CATALOG CARD NUMBER:

ISBN 0 - 1 2 - 0 2 7 3 0 9 - 8 PRINTED IN THE UNITED STATES OF AMERICA

64-14568

Contributors Numbers in parentheses indicate the pages on which the authors' contributions begin. R O Y M. ACHESON* Department of Epidemiology and Public Health, Yale School of Medicine, New Haven, Connecticut (1) R. R E A G L E H O L E (241)

Epidemiology

Unit, Wellington Hospital,

University

Wellington, New

Zealand

P E T E R H. R E N N E T T Chief, Epidemiology and Field Studies Branch, National of Arthritis, Metabolism, and Digestive Diseases, Phoenix, Arizona (49, 279)

Institute

S. K. D A V E Postgraduate Department of Medicine and Directorate of Medical tion and Research, New Block, Civil Hospital, Ahmedabad, India (147)

Educa-

FLORA DAVIDSON

Department

of Health, Wellington, New Zealand

(241)

SUAD E F E N D I C Departments of Endocrinology and Histology, Karolinska Hospital and Institute, and the Research Department of the Kabi Group, Stockholm, Sweden (367) RUNE FILIPSSON (425)

Department of Orthodontics, Karolinska

C. D u V. F L O R E Y Department of Community School, London, England (65) U R I GOLDBOURTI" (93)

The Israeli Ischémie

Institute, Stockholm,

Sweden

Medicine, St. Thomas's Hospital

Medical

Heart Disease Study,

YOSHIO GOTOi Third Department of Internal of Medicine, Hirosaki, Japan (167)

Tel-Hashomer,

University

School

O. P. G U P T A Postgraduate Department of Medicine and Directorate of Medical tion and Research, New Block, Civil Hospital, Ahmedabad, India (147)

Educa-

K E R S T I N H A L L Department Stockholm, Sweden (425)

of Endocrinology

Medicine, Hirosaki

Israel

and Metabolism,

Karolinska

Hospital,

* Present address: Department of Community Medicine, Cambridge University School of Clinical Medicine, Cambridge, England. t Present address: Department of Research, Wingate Institute of Physical Education and Sports, Israel. $ Present address: Third Department of Internal Medicine, Tohoku University School of Medicine, Sendai, J a p a n . xiii

Contributors

XIV

RICHARD F. HAMMAN* National Institute of Arthritis, Metabolism, eases, Southwestern Field Studies Section, Phoenix, Arizona (49)

and Digestive Dis-

J O S E P H B. H E R M A N Department tal, Jerusalem, Israel (93)

University

of Internal Medicine C, Hadassah

TOMAS H Ö K F E L T Departments of Endocrinology and Histology, Karolinska and Institute, and the Research Department of the Kabi Group, Stockholm, (367) W. P. U. JACKSON Department of Medicine, University of Cape Town, and Service, Groote Schuur Hospital, Cape Province, South Africa (111)

Hospi-

Hospital Sweden

Endocrine

M. H. J O S H I Postgraduate Department of Medicine and Directorate of Medical tion and Research, New Block, Civil Hospital, Ahmedabad, India (147) RYOSO KAWATE Department Medicine, Hiroshima, Japan

of Internal (201)

Medicine, Hiroshima

D A R W I N R. LABARTHEt Department of Mayo Clinic, Rochester, Minnesota (13) JAN L I N D S T E N den (425)

Medical

University

Statistics

Department of Clinical Genetics, Karolinska

and

Educa-

School of

Epidemiology,

Hospital, Stockholm,

Swe-

R O L F L U F T Departments of Endocrinology and Histology, Karolinska Hospital and Institute, and the Research Department of the Kabi Group, Stockholm, Sweden (367) JACK H. M E D A L I E Department of Family Medicine, Case Western Reserve School of Medicine, Cleveland, Ohio (93) M A X M I L L E R Department of Medicine, Case Western Reserve University, Cleveland, Ohio, and Consultant, National Institute of Arthritis, Metabolism, and Digestive Diseases, Bethesda, Maryland (49, 279) M I C H I H I R O M I Y A N I S H I Department of Internal School of Medicine, Hiroshima, Japan (201) JjORN N E R U P (263)

Steno

Memorial

Hospital,

DK-2820

Medicine,

Hiroshuna

Gentofte, Copenhagen,

YUKIO N I S H I M O T O Department of Internal Medicine, Hiroshima Medicine, Hiroshima, Japan (201)

University

Denmark

University School of

J E R R O L D M. O L E F S K Y Department of Medicine, Stanford University School Medicine and Veterans Administration Hospital, Palo Alto, California (313)

of

* Present address: Preventive Medicine Resident, Department of Epidemiology, Johns Hopkins School of Hygiene and Public Health, Baltimore, Maryland. t Present address: School of Public Health, University of Texas Health Science Center at Houston, Houston, Texas.

xv

Contributors

J O H N B. O ' S U L L I V A N * Department of Epidemiology and Public Health, Yale University School of Medicine, New Haven, Connecticut (1) P. J. P A L U M B O (13)

Department

of Internal

Medicine, Mayo Clinic, Rochester,

Minnesota

C H É R I M. P A P I E R Department of Preventive and Social Medicine, Tel Aviv Sackler School of Medicine, Tel Aviv, Israel (93)

University,

D A N I E L P O R T E , J R . Division of Endocrinology istration Hospital, Seattle, Washington (283)

and Metabolism,

I. A. M. P R I O R (241)

Hospital,

Epidemiology

Unit, Wellington

Veterans

Wellington,

New

Admin-

Zealand

G E R A L D M. REAVEN Department of Medicine, Stanford University School of Medicine and Veterans Administration Hospital, Palo Alto, California (313) C L A R E E. S ALMOND land (241) PiNCUS T A F T (225)

Epidemiology

Unit, Wellington Hospital,

Ewen Downie Metabolic Unit, Alfred Hospital,

Wellington, New Zea-

Melbourne,

Australia

BRUNO W. VoLKt Isaac Albert Research Institute of the Kingsbrook Jewish Medical Center and Department of Pathology, Downstate Medical Center, State University of New York, Brooklyn, New York (333) KLAUS F. WELLMANN:·: Isaac Albert Research Institute of the Kingsbrook Jewish Medical Center and Department of Pathology, Downstate Medical Center, State University of New York, Brooklyn, New York (333) K E L L Y M. W E S T Oklahoma (29)

University

of Oklahoma

Health

S T E P H E N C. W O O D S Division of Endocrinology istration Hospital, Seattle, Washington (283)

Sciences Center, Oklahoma

and Metabolism,

M I C H I O YAMAKIDO Department of Internal Medicine, Hiroshima Medicine, Hiroshima, Japan (201)

Veterans

City,

Admin-

University School of

PAUL Z I M M E T Department of Metabolic Medicine and Epidemiology, rial Hospital, Melbourne, Australia (225)

Southern

Memo-

* Present address: Department of Medicine, Section on Preventive Medicine and Epidemiology, Boston University, Boston, Massachusetts. "·" Present address: Department of Pathology, University of California, Irvine, California. t Present address: Department of Pathology, Beekman Downtown Hospital, New York, New York.

Prologue to International Studies in the Epidemiology of diabetes Epidemiology is the study of the distribution and determinants of disease. A primary use of epidemiology is to determine etiologic factors in disease, a process which usually starts with a descriptive phase and in which the prevalence and disease associations are determined and examined to formulate hypotheses of causation. Subsequent to the definition of prevalence and disease association, incidence studies are performed to provide more powerful methods of examining the meaning of associations, since only this type of study can provide information concerning risk factors. The epidemiology of diabetes is complicated by the involvement of both genetic and environmental determinants, a dual causation which complicates both study design and analysis. Since the field of diabetes epidemiology is relatively new, it is not surprising t h a t most of the data so far published relate to prevalence, in which the frequency of diabetes mellitus at a specific point in time in various populations is described. Incidence studies, on the other hand, provide information concerning the rate of development of new cases of the disease in populations over defined periods of time. Such studies require t h a t the population be examined on at least two occasions with some appreciable intervening time interval, and t h a t consistent methods for recognition of the disease be applied. Since the study design is more complex and more difficult to execute, it is not surprising t h a t incidence studies have been performed only rarely, in spite of the fact t h a t they are much more likely to be productive of new hypotheses and provide the only reasonable means of hypothesis testing in many instances. Since both environmental and genetic determinants are involved in the appearance of diabetes mellitus, some further clarification may be anticipated, now that the discovery of genetic markers relating to diabetes allows more powerful méthodologie approaches than in the past. In this regard, the recent discovery t h a t HLA types are associated with juvenile onset diabetes mellitus opens the door for studies concerning the relationship between the possible environmental determinants and the genetic predisposition to the disease. Despite the fact t h a t studies of the epidemiology of diabetes began in the United States over thirty years ago, the extent of such studies to xv ii

XV111

Prologue

this time has been quite limited, as shown by their lack of prominence at diabetes meetings and within diabetes-related organizations. It is of interest that the Epidemiology Section of the European Association for the Study of Diabetes is the only organizational subgroup whose interest is specifically devoted to the area. In view of the important contributions of epidemiology to our understanding of the etiology of cardiovascular disease, and to such neoplastic diseases as cancer of the lung, breast, scrotum, skin, etc., to cite but a few examples, it is almost certain that more extensive work in the field of diabetes epidemiology will reveal as yet unknown associations and determinants of the disease. Epidemiologie studies of diabetes could, in the final analysis, lead to the prevention or amelioration of the disorder. PETER H. BENNETT M A X MILLER

JOHN B. 0'SULLIVAN*'t and ROY M. ACHESON*$

Comparison of Diabetes Prevalence Rates in Oxford (1946) and Sudbury (1964) I. Introduction II. Materials and Methods A. Populations (Denominators) B. Diagnosis (Numerators) C. Laboratory Methods III. Results A. Published Data Tested B. Current Reanalysis C. Sensitivity Analyses D. Summary Analysis IV. Discussion V. Summary References

1 2 2 2 4 4 4 5 6 7 8 10 10

I. Introduction This paper compares the prevalence rate for diabetes mellitus in the town of Oxford, Massachusetts, in 1946-1947 (Wilkerson and Krall, 1947) with the rate obtained from a nearby study which was designed specifically to facilitate comparison. The second survey was conducted in 1964, some eighteen years later, in the New England town of Sudbury (O'Sullivan et al., 1967). The results bear on current statements * Department of Epidemiology and Public Health, Yale University School of Medicine, New Haven, Connecticut. f Present address: Department of Medicine, Section on Preventive Medicine and Epidemiology, Boston University, Boston, Massachusetts. X Present address: Department of Community Medicine, Cambridge University School of Clinical Medicine, Cambridge, England. 1

2

John B. O'Sullivan and Roy M. Acheson

that diabetes is increasing in frequency (U.S. Department of Health, Education and Welfare, 1975), and consideration is given to the problems of data interpretation that arise when two separate population studies of diabetes are compared. Classical whole-population prevalence studies of the chronic diseases are sufficiently infrequent that such comparisons become necessary despite the obstacles. The comparison of rates for previously known diabetes in the communities suggests a significantly higher frequency in the more recent of the two studies and therefore suggests that there is a secular increase in the prevalence of diabetes mellitus. Further analysis, however, demonstrates that this result does not reflect differences occurring in the eighteen-year interval, and the evidence suggests that the frequency of diabetes mellitus remains unchanged.

II. Materials and Methods A. Populations (Denominators) A total of 70.6% of the inhabitants of the town of Oxford, Massachusetts, were examined in the 1946-1947 study (Wilkerson and Krall, 1947). The 1945 state census was the basis for denominators in the original report. For the present purpose, advantage was taken of the availability of the 1950 federal census and estimates of the population in 1947-1947 made by interpolation between the 1945 and 1950 censuses (O'Sullivan et al., 1966). These population estimates are used as denominators for age-specific rates in this presentation. Data from persons below the age of 15 years in Oxford are excluded, since such data were not collected in the Sudbury study. A special census in the town of Sudbury, Massachusetts, was made by the township in January, 1964, the month this study commenced. Excluding 56 residents who objected for religious reasons and 34 nonresidents who maintained a technical legal residence in the town, there were 5976 persons in the target population. A total of 4626 persons (77.4%) participated. B. Diagnosis (Numerators) The two Massachusetts towns, Oxford and Sudbury, were studied in 1946-1947 and 1964, respectively. Questionnaires combined with postprandial blood and urine glucose tests were employed to define

Diabetes Prevalence in Oxford and Sudbury

3

cases of diabetes in both investigations (Wilkerson and Krall, 1947; O'Sullivan et al, 1967). The diagnostic standards as applied to the data for the original publications required modification for the purposes of this study. To obtain more equitable comparisons of the results from these populations, standards were redefined from existing records both for the acceptance of previously known cases of diabetes and for the definition of previously unknown cases. The previously unknown cases will be labeled "new diabetics," a term more correctly reserved for describing cases in incidence studies, in order to maintain a compatibility of terminology with the original reports from these populations. 1. PREVIOUSLY K N O W N DIABETICS

Subjects classified at the time of the survey as "previously known diabetics" are considered here as diabetic only if they satisfied one of the following reanalysis criteria: (1) were on insulin therapy; (2) had their original diagnosis based on abnormal oral glucose tolerance tests; and (3) were shown to have an elevated postprandial whole blood glucose of at least 200 mg per 100 ml (Folin-Wu) or 180 mg (AutoAnalyzer, Hoffman). This restricted definition of diabetes corresponds with the most certain of several definitions of diabetes previously published and is applied in this study to the original data from both Oxford and Sudbury. Diabetics taking oral hypoglycémies and failing to meet the criteria had their therapy withheld for 48 hours prior to being given a glucose tolerance test (GTT). The necessarily arbitrary nature of any confirmatory requirements will be examined by applying, in a separate analysis, alternative criteria consisting of the same three components but also including glucose tolerance test diagnoses only if the fasting glucose value is at least 140 mg (Folin-Wu) or 120 mg per 100 ml (AutoAnalyzer, Hoffman). 2. N E W DIABETICS

The diagnostic reanalysis criteria for previously unrecognized or "new" cases of diabetes required persons with a blood sugar exceeding the population's 98th percentile level on the first examination to have a value above the 92nd percentile on a second independent examination. The utility of this combination has been demonstrated elsewhere (O'Sullivan and Mahan, 1965a). The specific percentile levels were calculated separately for the two populations, using the 1- to 2-hour postprandial venous blood sugar distributions in Oxford, and for each

4

John B. O'Sullivan and Roy M. Acheson

of several postprandial intervals in Sudbury. This difference in approach was considered necessary because over 90% of the Oxford bloods were drawn within 1 to 2 hours of a meal, a procedure that differed from the more variable time relationship to meals for corresponding data from Sudbury. A check on the influence of criteria on the results was also included for the category of new diabetics. The alternative criteria, applied in a separate analysis, were identical with the standards applied for this purpose to the previously known diabetics. C. Laboratory Methods The Sudbury study depended on the estimation of glucose in venous whole blood by the AutoAnalyzer (Hoffman) method. The Oxford study also analyzed venous whole blood but used the Folin-Wu method as well as the capillary Folin-Malmros method. The differences in methodology were corrected by adding or subtracting constants: 20 mg was added to the AutoAnalyzer (Hoffman) readings to raise them to the Folin-Wu levels (Wilkerson et al, 1962; O'Sullivan and Kantor, 1963), and 15 mg was subtracted from values for capillary blood according to the Folin-Malmros method to adjust to the Folin-Wu venous whole blood method (O'Sullivan and Mahan, 1965b; O'Sullivan, 1969).

III. Results A. Published Data Tested The prevalence rates published from the two surveys of diabetes mellitus are 1.7 and 1.8 per 100 population and are remarkably similar (Wilkerson and Krall, 1947; O'Sullivan et al., 1967). This similarity must be reconsidered, however, in view of the differences in age structure of the two populations and in the multiple steps required for final diagnostic classification. Since persons in the category of previously known or stated diabetics have in the literature been confidently accepted to have undisputed disease, this information is presented separately in Table I. The rates are presented for age groups 15 years and older in both populations including all persons in both populations who stated they had diabetes. Also included in Table I are the results obtained when the age-specific rates for Sudbury are adjusted to the Oxford population (Chiang, 1961). A chi-square test showed the age-

5

Diabetes Prevalence in Oxford and Sudbury TABLE I PREVALENCE OF PREVIOUSLY K N O W N DIABETES IN OXFORD AND SUDBURY

Oxford

Crude prevalence rates for stated diabetes Rate age-adjusted to the Oxford population"

Sudbury

Number of subjects

Number of cases

Rate/ 1000

Number of subjects

Number of cases

Rate/ 1000

3768

40

11"

5976

82

14

3768

66

18'

" Significance of difference between adjusted rates in Sudbury and observed rates in Oxford: p< 0.02. h Age adjustment used age-specific prevalence rates from Sudbury applied to the corresponding population segments in Oxford, and the total number of diabetics thus obtained provided the new overall rate for comparison.

adjusted prevalence rate for stated diabetes in Sudbury to be significantly higher t h a n t h a t for Oxford. B . C u r r e n t Reanalysis Table II presents data from the two populations following application of the identical diagnostic reanalysis criteria to each. The criteria used for both previously known and newly diagnosed diabetes in this reanalysis have been outlined in the section on methods. The results indicate t h a t comparable age-specific prevalence rates in the two towns show some variability, especially among groups of persons aged 65 and over. Table III presents a breakdown of the data in Table II by age and sex; these data follow a consistent pattern, and the numbers at risk are small. In general, however, the age-specific prevalence rates for the two towns are similar and show no significant difference when the ANOVA test for interaction is performed on the logit scale for proportions based on unequal cell frequencies (p = 0.11). The age-adjusted rates for direct comparison shown in Table IV indicate the prevalence of diabetes in Sudbury to be 3 cases per thousand persons higher than t h a t in Oxford. This difference is not statistically significant by chi-square (p > 0.25), a finding t h a t also applies to the subset of previously known diabetics.

6

John B. O'Sullivan and Roy M. Acheson TABLE II AGE-SPECIFIC PREVALENCE RATES FOR DIABETES I N OXFORD (1946) AND SUDBURY (1964) BASED ON SAME REANALYSIS CRITERIA"

Age intervals (years)

Oxford diabetics Number

New

Known

Sudbury diabetics Rate"

Number

New

Known

Rate'

15-24 25-34 35-44 45-54 55-64 65-74 >75

476 558 447 383 270 176 61

0 0 5 9 6 3 4

1 1 1 6 6 5 2

2 2 13 39 44 45 98

762 1239 1452 625 308 170 70

0 2 7 10 9 7 9

0 4 6 8 2 12 2

0 5 9 29 36 112 157

Total

2371

27

22

21

4626

44

34

17

" Test for interaction among age-specific prevalence rates gave p = 0.11 by ANOVA on logit scale with unequal cell frequencies. '' Rate per 1000 population.

C. Sensitivity Analyses The stability of these rates, particularly among older subjects, was explored by applying the diagnostic alternative criteria, outlined above, to the data from both population studies. Table V presents prevalence rates based on these criteria; comparison with Table III shows how sensitive the data are to this procedure, because in some instances there was reversal of the prevalence rates between the sexes in Oxford. TABLE III DIABETES PREVALENCE RATES FOR OXFORD AND SUDBURY BY A G E AND S E X

Age intervals (years)

Oxford diabetics" Number

Sudbury diabetics"

Male

Female

Total

Number

Male

Female

Total

15-24 25-34 35-44 45-54 55-64 65-74 >75

476 558 447 383 270 176 61

0 0 10 33 15 50 100

4 3 17 45 75 42 97

2 2 13 39 44 45 98

762 1239 1452 625 308 170 70

0 6 13 43 29 165 105

0 4 4 13 41 66 176

0 5 9 29 36 112 157

Total

2371

15

26

21

4626

21

13

17

" Rate per 1000 population using the reanalysis criteria.

Diabetes Prevalence in Oxford and Sudbury

7

TABLE IV COMPARABLE AGE-ADJUSTED DIABETES PREVALENCE RATES IN OXFORD AND SUDBURY

Oxford diabetics"

Crude prevalence rates (from Table III) Rate age-adjusted to the Oxford population

Sudbury diabetics"

Number

New

Known

Total

Number

New

2371

11

9

21

4626

10

7

17

2371

14

10

24

Known

Total

" Rate per 1000 population using the reanalysis criteria. Age adjustments as described in Table I.

D. Summary Analysis Table VI summarizes the age- and sex-adjusted comparisons for both the reanalysis and the alternative criteria; it indicates that there is no significant difference in diabetes prevalence rates in the two communities. These analyses were extended further by dichotomizing the populations at age 45. No significant differences emerged when the younger and older groups in both Oxford and Sudbury were compared. Finally, the analytic process was reversed by applying the Oxford ageTABLE V A G E - AND SEX-SPECIFIC PREVALENCE RATES FOR DIABETES USING THE ALTERNATIVE

CRITERIA

Age intervals (years)

Number

Male

Female

Total

Number

Male

Female

Total

15-24 25-34 35-44 45-54 55-64 65-74 >75

476 558 447 383 270 176 61

0 0 10 16 7 37 33

4 3 8 40 68 52 65

2 2 9 29 37 45 49

762 1239 1452 625 308 170 70

0 4 8 19 7 89 0

0 3 0 7 12 44 78

0 3 4 13 10 65 57

Total

2371

9

22

16

4626

10

6

8

Oxford diabetics"

" Rate per 1000 population.

Sudbury diabetics"

8

John B. O'Sullivan and Roy M. Acheson TABLE VI A G E - AND SEX-ADJUSTED DIABETES RATES FOR OXFORD AND SUDBURY USING TWO CRITERIA

Diagnostic criteria Reanalysis Alternative

Town

Number

Diabetics

Oxford Sudbury (adjusted)'' Oxford Sudbury (adjusted)''

2371 2371 2371 2371

49 54 38 25

Probability"

0.50 < P < 0.75 0.10 150

>200

>250

>300

>350

534

237

148

111

94

78

44.9"

16.9

4.7

3.6

2.1

0.0

" Proportion of subjects whose final 2-hour PG was below 150 mg/dl without treatment, after an average of 4.5 years of follow-up.

Richard F. Hamman et al.

52

mation or tests on relatives other than parents were used. In this analysis, "positive family history" means that at least one parent had a 2-hour PG of 200 mg/dl or greater, or was treated with hypoglycémie medications. "Negative family history" means that both parents had 2-hour PGs less than 200 mg/dl at their most recent GTT, and "unknown family history" means that either one or both parents were not tested. 3. OBESITY INDEX

The obesity index is expressed as (weight in kilograms) + (height in meters)2. This index corrects weight for height and is highly correlated with other measures of obesity (Keys et al., 1972). AU subjects were ranked from low to high obesity index, and arbitrarily divided into quartiles for analysis. C. Analysis Subjects were first seen during the years 1965-1973 and examined prospectively until they were lost to follow-up or until the period for this analysis was arbitrarily terminated on December 31, 1974 (withdrawn). Thus, the length of follow-up varies. To make maximum use of these data, standard life table methods (Cutler and Ederer, 1958) were used. Because years of follow-up were fairly homogeneous across different categories of the population (e.g., young subjects with a low obesity index had similar follow-up to older subjects with a high mass index), a binary multiple linear regression model (Feldstein, 1966) was used to adjust simultaneously for multiple variables other than duration of follow-up. This allowed the comparison of rates simultaneously adjusted for different distributions of age, obesity index, family history, and initial glucose level. Differences in adjusted rates were assessed by using analysis of variance (Snedecor and Cochran, 1967).

III. Results and Discussion A. Population There were 1355 eligible subjects3 15 years of age and older with complete data who entered the study between 1965 and 1973. Onethousand and twenty of these individuals (75.3%) were not diabetic on 3

Subjects with incomplete data were excluded from this preliminary analysis.

Incidence of Diabetes among the Pima Indians

53

their first examination and were, therefore, at risk for the development of diabetes. There were also 625 subjects who had a base-line examination, but no follow-up, who, therefore, could not be included in this analysis. They had no significant differences in weight, obesity index, family history, or initial 2-hour PG from those with follow-up examinations. Thus, it seems likely that they were at similar risk for diabetes.

B. Incidence of Diabetes One hundred and thirty-two cases of diabetes developed over the period of the study, or 12.9% of the base-line population. The average duration of follow-up for all subjects was 4.7 years, giving a crude incidence rate of 2.7% per year. Since the number of people at risk of diabetes changed constantly (because of additions and withdrawals), this was an inaccurate estimate of yearly incidence rates, and adjustment of the population remaining at risk was necessary, as is done in the life table method. The life table 8-year cumulative probability (CP) estimate of developing diabetes was 23.7% (95% confidence limits 19.5 to 28.0), giving a probability of 3.0% per year. Table II illustrates how the life table probabilities were calculated. The number of subjects completing 10 years of followup was too small to allow stable rate estimates, and 9 cases of diabetes were excluded from further analysis. The rates shown in Table III for Pima males (25.2%/8 years or 31.5/ 1000/year) and females (23.7%/8 years or 28.3/1000/year) are between five and fifty times as great as the rates reported in other studies (Table IV), although the criterion for the diagnosis of diabetes among the Pima (2-hour PG > 250 mg/dl) was more conservative than that used in any other study. Direct comparison of the Pima rates with the crude annual rates (when available) listed in Table IV must be done with caution, since age, sex, and ethnic composition, as well as the criteria for diagnosis of diabetes, vary. Nevertheless, the Pima have the highest reported incidence of diabetes mellitus in the world. The life table analysis used in this study allowed the calculation of time-based risks, uncorrected for the confounding effects of other variables, and was likely to give good estimates of the proportion of subjects in a given group (e.g., age or sex) that developed diabetes. The binary regression analysis was used to adjust for confounding variables in order to develop estimates of the independent contri-

TABLE II

1020 978 676 392

0-2 3-4 5-6 7-8

8 46 33 36

Q Developed DM in interval

°L* =L-V2W. b P = (Q+L*). r CP„ = 1 - [ ( 1 - Ρ , ) χ ( 1 - Ρ 2 ) χ

L Number at start of interval

Years in study 1003.0 850.0 550.5 267.0 0.0080 0.0541 0.0599 0.1348

P" Probability of DM in interval 0.0080 0.0617 0.1179 0.2368

CP' Cumulative probability

AMONG P I M A INDIAN SUBJECTS AGED 15

x (1 - P,,)], where n = interval from 1 to 5.

34 256 251 250

W Withdrawn

L*" Adjusted number at risk in interval

LIFE TABLE ANALYSIS OF THE DEVELOPMENT OF DIABETES MELLITUS (DM)

0.0056 0.0163 0.0244 0.0425

2SCP Two standard errors ofCP

AND OLDER

Si

a:

£3

Incidence of Diabetes among the Pima

55

Indians

TABLE III EIGHT-YEAR CUMULATIVE PROBABILITY (CP) MELLITUS (%)

Females

Males 2S.E.

CP

Age

CP

15-24 25-54 55+

21.7 30.7 19.2

14.0 10.1 13.1

13.7 29.0 17.7

Total

25.2

7.0

22.6

Relative risk"

1.3

OF DIABETES

BY AGE AT ENTRY AND SEX

Total CP

2S.E.

Relative risk"

8.5 7.5 13.1

16.6 29.3 18.3

7.5 6.0 9.3

1.1 1.8 1.0

5.4

23.7

4.3

2S.E.

1.0

" Relative risks developed from multiple regression adjusted rates. S.E. = standard error of CP.

butions of sex, age, obesity, and family history on the development of diabetes. 1.

S E X AND A G E

Males had a slightly but not significantly (p > 0.1) higher incidence of diabetes among the Pima in all age groups, and overall the relative risk, adjusted for other variables, was small (males/females = 1.3). A male preponderance at ages less than 40 to 50 has been reported by Westlund (1966) and Fitzgerald et al. (1961), but in both studies female rates were higher t h a n male rates after these ages. Palumbo (1975) also reported t h a t males over age 40 had excess incidence in Rochester, Minnesota, but in all studies differences between sexes have been small. The peak incidence of diabetes in the Pima occurred in both sexes between 25 and 54 years of age, and age was a significant independent predictor of risk (p < 0.01). The peak incidence was at an early age, in contrast to other studies reporting age-specific incidence rates. Medalie et al. (1975) found a significant age-associated trend in Israeli males 40 and over, highest in subjects over age 60. Westlund (1966), found first admissions to hospital for diabetes (a crude index of incidence) to peak between 60 and 79 years in both sexes, and Spiegelman and Marks (1946) noted a peak over the age of 65. The highest incidence of fasting hyperglycemia in male veterans was also found a t over 60 years of age (Podulsky and Burney, 1975). It is possible t h a t the earlier peak among the Pima represents a cohort effect where rapid environmental changes have affected those

Follow-up of initial urine screenees, 50-gm oral GTT Mailed questionnaire,

1925-1954

1960-1961, 1966-1967

1916-1950, 1966

Westlund (1966)

Birmingham Diabetes Survey Working Party (1970)

Paffenbarger and Wing (1973)

Age, sex, initial glucose

Weight, blood pressure,

General practice patients, Birmingham, England College students at

Age, sex

Age, sex, weight, initial glucose, family history

Oxford, Massachusetts

One- to twohour postprandial and GTT, followup and retest Hospital discharge of new cases

1946-1963

O'Sullivan and Mahan (1965); O'Sullivan (1969) Oslo, Norway

Age, sex, parity

Birmingham, England, 1951

Record review of new cases in diabetes clinic

1949-1958

Fitzgerald et al. (1961)

Age 1 5 - 8 5 + , sex

United States National Health Survey

Estimation from prevalence and death rates

Related variables

Population

Method

1935-1936

Years of study

Speigelman and Marks (1946)

Author

TABLE IV

Relied on reported cases

Relied on reported prevalence and death certificate data Incomplete definition ofpopulation a t risk; relied on clinic cases at one hospital Cases found at initial survey included among incident cases Ho spital i zation required, date of onset estimated from medical record No calculation of time-based incidence rates

Comments

SUMMARY OF PUBLISHED REPORTS OF THE "INCIDENCE" OF DIABETES MELLITUS"

0.5

4.1

Males 1.1 Females 1.5

5.8

Males 0.8 Females 1.4

Males 2.7 Females 4.5

Estimated incidence per 1000 per year''

g ^ ^ s

a &

*3

ON

1956-1966, 1968

1963-1968

1945-1970

1963-present

1965-1974

Aspevik e£aZ. (1974)

Medalieei a/. (1974, 1975)

Palumbo (1975)

Podulsky and Burney (1975)

Pima Prospective, repeat oral GTT

Random glucose screening, oral GTT; follow-up with 100-gm oral GTT Prospective, repeat oral GTT (1 gm/kg body weight) examination New cases of diabetes at Mayo Clinic Prospective, repeat oral GTT

Household interview

2131 healthy veterans, aged 2 2 + , Boston, Massachusetts 1020 P i m a Indians, Arizona

Rochester, Minnesota

10,000 Israeli male civil servants, aged 40 +

Screenees in mass x-ray program, Bergen, Norway

United States Health Interview Survey

Harvard, Pennsylvania, all males

Age, glucose level, sex, weight, family history

Age, sex, mortality, family history Age, glucose level

Age, weight, peripheral vascular disease, others

Age, sex, geographic region, family income Age, sex, initial glucose

vital capacity, family history, others

Provided accurate incidence rates, based on prospective examinations Only preliminary report available Only preliminary report available

among males with high socioeconomic levels Reported cases only, no medical verification Small numbers of highly selected screenees followed up

Males 31.5 Females 28.3

Males 4.8

1.2

8.0f/

c

Males 2.2 Females 3.7

(l

c

Unsufficient data to calculate. Age and area of birth adjusted.

" All studies used different criteria for diabetes, age distributions were different., and in some studies exact duration of follow-up was estimated. Exercise cautiorL in comparison, b Estimated from nuhlished data if not renorted in this form

1973

National Commission on Diabetes (1976)

based on previous student health data

s*" 8-

8

»*«H

v5-

S

*·*

S5-

3

3.

58

Richard F. Hamman et al.

born after 1920-1930 much more than those born earlier. Thus, the relatively recent introduction of a risk factor(s) for diabetes may have selectively increased the risk of developing diabetes in younger subjects (for example, rapid dietary changes or increasing obesity occurring with western "acculturation" could account for this picture). It is also possible that the low incidence of diabetes in older subjects may be due to the fact that all those susceptible developed the disease at an earlier age. 2. INITIAL TWO-HOUR PLASMA GLUCOSE

The 8-year probability of developing diabetes was strongly related to the level of the initial 2-hour PG (Fig. 1). Eleven percent of subjects in the lowest 2-hour PG category (< 100 mg/dl) became diabetic, reflecting the extremely high incidence of hyperglycemia among the Pima, even in the (low) risk categories. When the base-line PG level was 121 to 140 mg/dl, the adjusted relative risk was 1.80. The relative risk almost doubled with each 20 mg rise in PG level above 140 mg/dl. The independent contribution of glucose above 140 mg/dl, adjusted for other variables, was highly significant (p < 0.001). The effect of glucose levels between 100 and 140 mg/dl was of borderline significance (0.1 >p> 0.05) but was clearly part of an increasing trend. Sixty percent of subjects were predicted to decompensate over 8 years if their initial glucose was greater than 160 mg/dl. However, 40% of such subjects did not develop severe hyperglycemia after one GTT with moderate elevation. Although criteria for diagnosis differ among studies, the more obese, those with a high initial glucose level, and those with a positive family history of diabetes have been found to have a higher incidence of dia80 I

« σ

Ω

büΛ

■

40-

Έ5£ rv zoö k.

GO

0J

Adj. Rel. Riskl 1.00 1 1.761 1.8013.32 [6.2817.21 I

Initial 2hr Plasma Glucose (mg/dl)

FIG. 1. Eight-year cumulative probability (CP) of developing diabetes by initial 2-hour plasma glucose (PG) level. Adjusted relative risk developed from multiple regression adjusted rates.

Incidence of Diabetes among the Pima

Indians

59

betes (O'Sullivan and Mahan, 1965; O'Sullivan, 1969; Birmingham Diabetes Survey Working Party, 1970; Paffenbarger and Wing, 1973; Aspevik et al., 1974; Medalie et al., 1974, 1975). Several of these variables have been studied by O'Sullivan and Mahan (1965; O'Sullivan, 1969) in the 17-year follow-up of the 1946 survey in Oxford, Massachusetts (Wilkerson and Krall, 1947). The results invite direct comparison with those of the Pima because the entire Oxford population was studied, and the data were presented in enough detail to allow calculation of relative risks for different levels of weight, family history, and glucose. Subjects with initial elevated postprandial glucose levels were examined together with an age-stratified random sample of subjects with normal values. Subjects initially identified in the 1946 survey were reexamined only once, 17 years later. Period-prevalence rates, as O'Sullivan noted, were determined. Subjects were considered to have diabetes if they were on insulin or had persistent postprandial hyperglycemia in excess of 200 mg/dl. A whole blood glucose level greater than 170 mg/dl in Oxford was associated with a risk of diabetes 6.8 times t h a t of subjects with glucose levels less t h a n 170 mg/dl. Since this included cases discovered at the initial survey, as well as true incident cases, it probably represents an overestimate of the true relative risk. The most comparable Pima subjects (plasma glucose greater t h a n 180 mg/dl) had 4.1 times the risk of diabetes when compared with subjects who had glucose values below 180 mg/dl at entry. 3. OBESITY INDEX

Figure 2 and Table V indicate that, the more obese a subject is at entry, the higher the future risk of diabetes. The adjusted relative risks were not different in the three lower obesity groups but indicated t h a t the heaviest quartile had almost three times the risk of the lowest

-Q 40 H σ —»

sj

«

y 0j—,^-τ^ 0

2 4 6 Years in Study

FIG. 2. Cumulative probability (CP) of developing diabetes by mass index quartile and years in study. Mass index = kilograms/meters 2 . Quartiles of mass index are: (1) 16.1624.242; (2) 24.243-28.089; (3) 28.090-32.526; (4) 32.527-52.861. Adjusted relative risk developed from multiple regression adjusted rates.

60

Richard F. Hamman et al. TABLE V

EIGHT-YEAR CUMULATIVE PROBABILITY (CP)

OF DEVELOPING DIABETES BY INITIAL

TWO-HOUR PLASMA GLUCOSE AND OBESITY INDEX QUARTILE"

Initial 2-hour plasma glucose

Obesity index quartile

(mg/dl)

1

2

3

4

Total

(A) 130 mg per 100 ml, determined by the Somogyi-Nelson method, were classified as suspects or positive screenees and were followed up. Those suspects with three successive fasting blood glucose level determinations of >

SS*-

122

W. P. U.Jackson

In each decade the total glycosuria and diabetes prevalence were remarkably similar, although there were many diabetics who exhibited no glycosuria and many glycosuries who were normal by GTT. Most glycosuries in the youngest decades were nondiabetic; nevertheless, the frequency of nondiabetic glycosuria rose with age, from 0.9% to 4.1% in subjects over the age of 50. In the second survey both total glycosuria and diabetes prevalence were higher among the more affluent Indians (Table I). These bare figures need a little embellishment. Prevalence of diabetes. As can be judged from the "number screened" column in Table I, the Tongaat Indians were a very young community (cf. Cape Town Indians; see Fig. 2), so t h a t any comparison with other groups surveyed is valid only if age adjustment is performed; differences in method and criteria must also be taken into account. A comparison on these lines can be made with the Birmingham (United Kingdom) Survey (Working Party Reports, 1962,1963), from which an overall prevalence estimate of 6.2% was determined for Britain. From the Birmingham evidence, the comparable Tongaat figure is 6.1%, but on age adjustment it becomes 11.1% as applied to the whole Natal Indian population. If we bear in mind also t h a t the known diabetes rate of 1.8% without age adjustment is approximately double t h a t usually quoted for White communities (Jackson, 1970), it would appear t h a t diabetes is indeed very common among Natal Indians. On the other hand, if we adjust the Tongaat total diabetes figures to compare with the Bedford (United Kingdom) Survey (Butterfield, 1964), we obtain a figure of 13.1%, which is just short of Butterfield's figure of 14%! Comparison with other South African surveys will be made later. O'Sullivan and Williams (1966) reported prevalences ranging from 1.2% to 15.8% when varying criteria were applied to the data obtained from the Sudbury (United States) data. When the same "varying criteria" were applied to the Tongaat data, the prevalence ranged between 4.9% and 8.1%. This much smaller difference between the results of adoption of "lax" and "stringent" criteria of around twofold was also found in the Cape Town data (Marine et al., 1969). Reproducibility of blood glucose values. Some individuals vary greatly in blood glucose responses to repeated GTT (McDonald et al., 1965). In 363 individuals the 2-hour postglucose screening value could be compared with the 2-hour value on GTT, performed later. The correlation is highly significant (r = 0.746, p < 0.001), but a number of subjects nevertheless showed wide individual variations. Body weight (Jackson et al., 1968b). Weights were adjusted for height

123

Diabetes in South Africa

and age by partial regression. No correlation could be found between corrected body weight and screening blood glucose level, nor between weight and blood glucose at any time during GTT in discovered diabetics. The importance of body weight is clearly shown by the increasing prevalence of diabetes with increasing weight, from 1.9% in the lightest group to 17.8% in the heaviest (Table II). Not only is the frequency of diabetes related directly to overweight, it is also related, but inversely, to underweight. 2. CLINICAL FINDINGS

It has been possible to compare three groups of positive screenees from this survey (i.e., known diabetic, discovered diabetic, and pronounced nondiabetic) with an unselected group of 125 clinic-attending Indian diabetics and 279 White diabetics (Jackson et al., 1970a). Important observations include the following: 1. Diabetic retinopathy was discovered by direct ophthalmoscopy in a darkened room after adequate mydriasis in 61% of 125 clinicattending Indian diabetics, 34% of 279 White diabetics, 6% of 89 discovered Indian diabetics, and 0% of 215 nondiabetics (excluding 1 subject who showed exudates only). The duration of diabetes was in general greater among White than among Indian clinic diabetics. We believe that the 6% retinopathy among the survey-discovered hyperglycémies confirms that at least some really are diabetic. (A similar percentage with retinopathy is found among symptomatic maturity-onset diabetics at the time of diagnosis.) 2. Ischémie heart disease (IHD) is more difficult to be sure of (the criteria cited are precisely stated in Jackson et al.9 1970a). Ischémie heart disease was diagnosed in 29.5% of 122 Indian clinic-attending diabetics, 26% of White diabetics, 35% of 52 discovered diabetics, and 26% of 65 nondiabetics. There was no significant difference between the TABLE II DIABETES PREVALENCE AND BODY W E I G H T IN NATAL INDIANS (AGE CORRECTED)

Percent of mean weight

Number

Percent diabetic

125

803 807 267 113

1.9 5.8 9.7 17.8

124

W. P. U. Jackson

sexes. The similarity of these figures might suggest t h a t the rate of IHD is approximately as high among Indian nondiabêties as among White diabetics, and t h a t diabetes in the Indian makes little difference to his likelihood of having IHD. We again noted a tendency for Indians to develop IHD at an earlier age t h a n Whites. Thus, in the 20- to 39year age groups 5% of White diabetics had evidence of IHD, as against 10% Indian nondiabetics and 12.5% Indian diabetics. Dietary note. Monthly per capita consumption of various foodstuffs at Tongaat was closely comparable to t h a t of the Indians living in Durban and included rice 5.6 lb, maize 5.3 lb, flour 3.3 lb, bread 16.8 lb, and sugar (direct and indirect) 6.1 lb. Animal protein consumption was low (beef is forbidden to the Hindu). Some 2% are vegetarians. F a t consumption included butter (and ghee) 0.9 lb, vegetable ghee 0.6 lb, and oils 1 to 2 bottles per month. Certain other aspects of this Tongaat Survey will be considered together with the Cape Town and Mamelodi Survey (below). C. M a m e l o d i

Study (Transvaal B a n t u , G o l d b e r g et ai,

1969)

Mamelodi is an African township near Pretoria in which 2015 subjects over the age of 10 years were selected and screened by the same methods as those used at Tongaat. They can be considered "semiurbanized." Only one person was already known to be diabetic, yet the screening levels were just as high as among the Indians—in fact, the blood glucose distribution curve of the Africans was slightly to the right o f t h a t of the Indians. Since no GTTs were performed, "probable diabetes" was diagnosed if the 2-hour screening value was over 150 mg/dl, which gave us an overall prevalence of 2.9%. Of these 59 "diabetics," 43 had no glycosuria; glycosuria was in fact found in only 1% of these Africans. It is suggested, therefore, t h a t in some African areas hyperglycemia (even up to 800 mg/dl) may be relatively frequent, without symptoms or glycosuria; consequently, diagnosed diabetes remains infrequent.

V. Cape Town Studies We have investigated samples of the five major racial groups in the Cape Town area in an attempt to determine the prevalence of diabetes and glycosuria and to correlate blood glucose and serum insulin levels

Diabetes in South

Africa

125

with such variables as age, sex, weight, race, religion, diet, income, time of day, and fecundity (summarized in Marine et al., 1969; Jackson, 1972). A.

Methods

The methods used have been described in detail elsewhere (Marine et at., 1969; Jackson et al., 1970b). In each survey, areas were chosen t h a t were as nearly as possible representative of the total population of the relevant ethnic group, with regard to age distribution, income range, and religion. Within the chosen areas, randomly selected entire households over the age of 10 (age 15 in some studies) were interviewed and invited to attend for screening. Briefly, screening was performed at a set time (2 hours when practicable) after 50 gm of oral glucose by urine testing (Tes-tape) and capillary blood sampling. For reasons of convenience, screening in most of the studies was done in the late afternoon or evening, at least 4 hours after the preceding meal. All subjects whose screening blood sugar values exceeded an agreed level (159 mg per 100 ml at 1 hour, 119 mg per 100 ml at 2 hours) or who showed glycosuria were invited to undergo full oral glucose tolerance tests, together with matched negative-screening controls. The final diagnosis of newly discovered diabetes was made if two or more GTT values were abnormal (venous plasma, Autoanalyzer, Hoffman method 1 )—i.e., fasting level > 120 mg per 100 ml; maximum level > 185 mg per 100 ml; 2-hour level > 140 mg per 100 ml. In all groups, approximately 90% of positive screenees attended for GTT—among those who did not attend, "presumed diabetes" was diagnosed when the screen 1-hour blood sugar exceeded 250 mg (at 1 hour) or 200 mg (at 2 hours), together with glycosuria. The term "lag (storage) curve" is applied to a GTT in which a peak value over 185 mg per 100 ml occurred at 30 minutes or earlier, with other values normal. "Borderline" is applied to a GTT with one of these three values abnormal (excluding lag curve). Glycosuria is considered "renal" if no blood glucose value was over 159 mg per 100 ml. "Trace" tests are not included. The presence of already "known diabetes" was checked from records or by blood sugar estimations. 1

A Hagedorn-Jensen method was used in early tests (Marine et al., 1969), and the earliest groups were screened by postprandial blood glucose (see below).

126

W. P. U.Jackson

Weight standards were obtained from Documenta Geigy tables, "obesity" being defined as more than 15% above standard weight. B. Age and Sex Distribution of Populations and Screened Samples Age distribution in the Cape Town area, taken from the 1960 census, is shown in Fig. 2. Cape Malays are not recorded separately but are included with Coloured. The African curve is artifically humped at 25 to 44 years by the preponderance of young workers who came from rural territories and live temporarily in Cape Town. Apart from this it is clear that the White curve differs from the non-White curves, with far more even distribution of age groups, and far more people over 45. This White distribution is similar to that in Britain and the United

30

20

io H

0-4

5-14

15-24

25-34 AGE

35-44

45-54

55-65

65*

GROUP

FIG. 2. Age distribution of population of Cape Peninsula, 1960 census. Note comparatively horizontal curve of White age distribution (see text). (Reprinted, with permission, from Diabetes.)

Diabetes in South Africa

111

States, and demonstrates the fallacy of direct comparisons of prevalence among different races, unless either age-corrected or confined to specific age groupings. Our own screened samples were reasonably representative of the population at large. In all races the numbers of males and females were approximately the same in each age group. C. Results 1. INDIANS (Marine et al., 1969) The first 690 selected subjects were screened by postprandial blood and urine testing; the second group of 830 were from a poorer area and were screened 1 hour after glucose loading. Of those selected, 75% were tested (recovery rate). Serum blood glucose levels showed no sex differences at any age. The rise with age occurred in both groups over 35 years of age. Previously known diabetes was present in 4.3% over age 15, in 20.2% over age 55, and in more women than men (Fig. 3). Newly discovered diabetes was found in 5.6% over age 15 in the postglucose group, and in 6.6% in the postprandially screened group. For both groups combined, the prevalence was 12.6% in subjects over age 35 and 20.0% in those over age 55. All but one out the 49 discovered diabetics had significant glycosuria either at screen or GTT or both. Age adjustment to the age distribution found in the White population raised the total diabetes prevalence from 10.2% to 19.1%. 2. MALAYS

Once again we tested a postprandial group of 632, and a postglucose group of 622, with 88% recovery. Screen blood glucose levels were significantly higher in females, tended to rise with age, and were higher than those of the Indians at all ages. Known diabetes was present in 1.4% over age 15, and in 6.9% over age 55; and women outnumbered men (Fig. 3). Diabetes was discovered in 5.2% over age 15, and in 22% over age 55. Asymptomatic diabetes was found in two 17-year-old girls. 3. AFRICANS (Marine et al., 1969) Of 1029 subjects selected, 86% were screened after oral glucose. Screen blood glucose levels were considerably lower than those of

128

W. P. U.Jackson

KEY

AFRICAN (1029)

[:§:J both

I MALAY (1254)

sexes

[ mal·

^Μ

female

j».

I 20X

■^m

10% K N O W N

55 + 10%

D I S C O V E R

20%

30%

40%

E D

FIG. 3. Diabetes prevalence in different racial groups at different ages with sex ratios. These are all Cape Town surveys. "Already known" diabetes is shown to the left, "discovered at survey" to the right. The sex ratios in the small dotted bars were approximately unity. Numbers in parentheses indicate total people in each survey. (Reprinted, with permission, from Postgraduate Medical Journal.)

Indians and Malays at all age groups. Known diabetes was present in 0.9% over age 15, and in 2.5% over age 55. We discovered diabetes in 2.7% over age 15; none of these diabetics was under 35 years of age. 4. WHITES (Jackson et al.,

1969)

A group of 1650 was selected, and 72% of these were screened 2 hours after oral glucose. Mean screen blood glucose values were slightly (not significantly) higher in women at all ages, and rose with age only after 49 years. They were considerably lower than the mean values of the Tongaat Indian and Mamelodi Bantu. Known diabetes prevalence in subjects over age 15 was 0.9%; in those over age 55 it was 5%. Discovered diabetes prevalence in subjects over age 15 was 2.3%; in those over age 55 it was 6.3%. Diabetes, both known and discovered,

129

Diabetes in South Africa

was more common in men t h a n in women (total diabetes prevalence 4.0% for men, 2.5% for women). 5. CAPE COLOURED (Michael et al.,

1971)

Results may be expressed as a flow sheet: Total survey population Known diabetics Total screened Screened positive Presumed diabetic" Received GTT Diagnosed diabetic Total diabetic

1534 968 (63% recovery) 310 284

16 3 56 75

" Blood glucose > 200 + glycosuria.

The recovery of only 63% was low, but this sample did not appear to be biased with regard to family history of diabetes (age and sex bias can be corrected). Screen blood glucose levels showed no significant sex difference, experienced the usual rise with age, and produced a large number of positives (over 159 mg/dl at 1 hour)—almost one-third of the total. Known diabetes in subjects over age 15 was 1.3%, discovered diabetes 7.4%. This survey produced 7 diagnosed diabetics aged 16 or less, all mild in type and only one symptomatic. Four had a diabetic close relative. Only 3 of the 59 discovered diabetics had diabetes-related symptoms on close questioning. Of 41 discovered diabetics who submitted to examination, 11 had evidence of IHD, and 12 were hypertensive (for criteria, see Jackson et al., 1970a). One had gross background retinopathy.

VI. Some Comparisons and Discussion A. Age Distribution There were only minor differences between the age distribution of each of the various surveyed communities and the corresponding general population, but major differences between the various ethnic groups (Fig. 2). Age adjustment (Table III) led to an increase in prevalence rates in all non-White groups.

130

W. P. U.Jackson TABLE III GLYCOSURIA AND DIABETES IN SUBJECTS OVER A G E

15

IN

DIFFERENT SURVEYS"

Glycosuria at screening

Known diabetes

Total diabetes''

7.5 6.9 1.0 7.9 6.5 9.5 13.5

0.9 0.9 0.05 4.3 2.2 1.4 1.3

3.2 (3.6) 3.6 (4.2) 2.9 10.3(19.1) 7.9(14.2) 6.6 8.7(10.7)

White Bantu: Cape Town Mamelodi Indian: Cape Town Tongaat Malay Coloured

" Values are given in percentages. Known plus discovered at survey. In parentheses are prevalences adjusted to age distribution of total White population. b

B. Screen Blood

Glucose

In both sexes in all races the mean blood sugar levels rose with age, although in White people the rise was seen only over the age of 60. In no race was there a significant difference between the first two age groups—i.e., no rise in blood glucose from childhood to young adulthood. There was no significant difference between the sexes in any group. Two-hour postglucose values were much lower among the White subjects t h a n among the Natal Indians at all ages, but the Mamelodi Africans and Natal Indians had very similar mean values. Mean 1-hour figures were very much the same for Indians, Malay, and Coloured people in Cape Town, with the Africans distinctly lower at all ages. Frequency distribution curves of blood glucose levels were unimodal in all groups except possibly the Natal Indians. If we make use of a log scale for numbers of subjects in each blood glucose group, the curve for Natal Indians does seem to have a double hump, possibly indicating a "normal" curve and a "diabetic" curve, which meet at around 175 mg/dl (Fig. 4). The only other double-peaked curves for blood glucose distribution have been recorded from the Pima Indians (Bennett et al., 1971).2 (I must t h a n k Professor John Butterfield for transforming the blood glucose values into the curve shown in Fig. 4.) 2

Other double-peaked curves have been recently described.

131

Diabetes in South Africa

100

150

200

250

300

350

2 hr. venous plasma glucose level

FIG. 4. Blood sugar distribution curve of 2419 Natal Indians. Note the suggestion of a double-peaked curve (see text).

No significant correlation between blood glucose levels and body weight could be found in any racial group or either sex, nor between blood glucose levels and parity in women over 40 years of age. C. Glycosuria (Jackson et al., 1968a) The frequency of glycosuria, both diabetic and nondiabetic, rose with age in all races and was generally more common among males. It was most frequently found among the Coloured people. Glycosuria was rare in children—even renal glycosuria was found only once among almost 1400 subjects under age 15 out of a total of 73 diagnosed cases in the Cape Town surveys. This suggests that, in view of the fact that the blood glucose does not change between childhood and young adulthood, the renal threshold must fall with age in some people. We could find no evidence for any rise in threshold with age. Glycosuria, even after a glucose load under reasonably standardized conditions, was inconsistent in approximately one-third of the 350 subjects who received two loads and showed glycosuria at least once. It is thus inefficient and insensitive as a sole screening test, while postprandial glycosuria screening alone would have missed two-thirds of the Indian diabetics! Nevertheless, 90% of all newly discovered diabet-

132

W. P. U.Jackson

ics showed glycosuria at some time during their testing. We could not confirm the previous assertion t h a t large numbers of Indian diabetics never had glycosuria, although diabetes was more frequent t h a n glycosuria only among Indians (Table III). Glycosuria found on screening more often turned out to be nondiabetic than diabetic: 155 diabetics out of a total of 400 with glycosuria (39%). This proportion varied enormously with age, 4% of young glycosurie men being diabetic, as against 77% of elderly women (cf. Butterfield et al., 1967). The frequency of diabetes diagnosed in the presence of glycosuria varied with the degree of color change, being 29% for 1+ glycosuria, rising to 78% for 4 + glycosuria. Although we teach t h a t Tes-Tape is not a quantitative test, yet the approximately estimated degree of glycosuria has some predictive value (see also Kenny and Chute, 1953). We appreciate t h a t our assessment of glycosuria in these studies is crude and takes no account of the volume of urine voided. D.

Diabetes

Generally speaking, the figures for the prevalence of glycosuria, known diabetes, and discovered diabetes in the different ethnic groups maintain similar ratios with each other, there being approximately twice as much discovered as known diabetes, and rather more glycosuria than total diabetes. The Cape Indians had a particularly high known diabetes r a t e (4.3% over age 15) and the Coloured people a high discovered diabetes rate (7.2%). The glycosuria-at-screening r a t e was lower t h a n the total diabetes rate among Cape Indians, and much the same in the Natal Indians. 1.

AFRICANS

Mamelodi Africans showed very little glycosuria and virtually no known diabetes, yet their blood glucose screening figures were as high and their "diagnosed diabetes" rate almost as high as the corresponding Natal Indian figures. Several of their blood glucose levels were over 600 mg/dl without symptoms being present. It would appear that, at least among some African groups, hyperglycemia may be frequent, but without symptoms and without general awareness of the condition of diabetes. The rather more sophisticated although only partially urbanized Cape Town Africans produced figures all very similar to those of Cape Town Whites. Hence there would appear to be at least as much

Diabetes in South

133

Africa

potential for diabetes among the Bantu Africans as among White people. Unfortunately we have no accurate figures on really rural or tribal Bantu people. 2. W H I T E PEOPLE (Jackson, 1970,

1972)

The prevalence of already known diabetes in Cape Town was similar to t h a t found in White communities in the western world (e.g., Britain, the United States, Canada, Scandinavia). However, the total diabetes prevalence of 3.7% in subjects over age 15 is low compared with other studies based on postglucose blood sugar estimates, which have been around 10% (e.g., Bedford, Birmingham, Berlin, "nationwide United States," and Tecumseh). Other reports of expatriate White populations have also indicated similar comparatively low diabetes frequency (e.g., Australia, Hawaii). 3.

INDIANS

I think we have confirmed t h a t diabetes and hyperglycemia are more common among South African Indians t h a n among Whites. It is not clear that they are more hyperglycémie than the Malays or Coloured people. The finding of glycosuria in an Indian does not necessarily indicate diabetes. Diabetic Indians probably do tend to have frequent and severe vascular disease and rarely become ketoacidotic, even when young. Muslim and Hindu seem equally affected. For some time diabetes has been said to be more common among White women and Indian men t h a n among White men and Indian women. Other recent, properly conducted studies have not supported this claim (Jackson, 1970), and our own studies indicate little or no sex difference. It is generally believed t h a t diabetes is several times as common among Indians in South Africa as in India, but until recently there have been no real population studies from India t h a t could be used for comparison. Certainly Berry et al. (1966), using postprandial glycosuria as a screen and very lax criteria for subsequent diabetes diagnosis, arrived at a total diabetes frequency of only 3% for adults in a northern town. A careful study from Cuttack found diabetes in 11.5% of urban dwellers but in only 0.9% of rural dwellers; around Delhi the comparable figures were 2.4% and 1.5% (Jackson, 1970). Subjects in both studies tended to be thin. It is probable, therefore, t h a t diabetes has really increased among Indians who came to South Africa; nevertheless, its prevalence may be no greater t h a n in the urban Cuttacks.

134

W. P. U.Jackson

4. CAPE COLOURED (Michael et al., 1971) The cause of the frequency of hyperglycemia in this group is not clear. As a whole, the Coloured community is no more adipose than the White (see Table IV), and, despite the large preponderance of obese females (28% versus 7% obese males), the frequency of diabetes in the two sexes is similar. The total calorie, sugar, and fat intakes are known to be lower than the corresponding levels among the White population (Bronte-Stewart et al., 1955). Racially considered, the problem is worse confounded: In Hottentots diabetes is unknown; it is less common among the other progenitors of the Coloured people (European, Malayan and Bantu); yet mix all these, and we find a diabetes prevalence of 7.2% in subjects over the age of 10 years. This apparent increase in diabetes among a mixed racial group is not unique to the Cape Coloured. It seems to have occurred in New Zealand (Prior and Davidson, 1966) and in Hawaii (Sloan, 1963) and, of course, in the United States, where the "black" population have more diabetes than the "white," and far more, evidently, than their progenitors in East Africa. E. Is "Discovered Diabetes" Really Diabetes? The answer to this question is presumably "yes" in those 5 to 10% who admit to clearly diabetic-type symptoms on questioning, in the occasional subject with retinopathy, and in those with gross hyperglycemia (say over 200 mg/dl 2 hours after oral glucose). In all our surveys the fasting blood glucose levels of well over half the "discovered diabetics" were above 120 mg/dl, and a large majority have shown glycosuria. Where clinical assessments were made, ischémie heart disease was more common among the hyperglycémies than among the normoglycemics. In many cases, however, we cannot be sure that hyperglycemia means diabetes, especially since some subjects have normal GTTs on retesting (see below). Our definition of "discovered diabetes" would be acceptable by most criteria, since those we use are on the stringent side. Strictly speaking, however, we should use the term "discovered hyperglycémie," rather than diabetic; will the reader please excuse this license in terminology? F. Body Weight The revalence of obesity in the various racial groups for each sex is shown in Table IV. We have already remarked on the striking relation

Moderate High Moderate High High

White Coloured Bantu Indian Malay

From Walker et al. (1971).

Overall frequency

Ethnic group

TABLE IV

Common Less common Less common Very rare Uncommon

Insulindependent juvenile type

Rare Common Uncommon Common Common

Young insulinindependent type

16 7 7 8 7

Male

23 28 52 20 16

Female

Obesity (%)

Very high Moderate Low Moderate Moderate

Sugar intake (gm/day)"

Very high (110) High (90) Moderate (70) High (80) High (90)

COMMUNITIES

Fat intake

DIABETES AND RELATED FEATURES AMONG C A P E T O W N

Very common Moderate Very rare Very common Moderate

Ischémie heart disease

136

W. P. U. Jackson

between body weight and diabetes in the Natal Indians; on the other hand, the prevalence of diabetes among nonobese Africans of 1.6% rose only to 2.3% among the obese. West and Kalbfleisch (1971) have provided evidence t h a t overweight may be the most important single environmental factor in the emergence of diabetes in different races. Our South African populations do not entirely support this thesis. Thus, the two "most obese" races (White and Bantu) had the lowest frequency of diabetes. Bantu women are fat by tradition and by their husbands' preference; yet we found marginally more diabetes among the men. To our surprise, we found obesity more common in diabetic women under the age of 40 than in women above this age—in the general population in all races, obesity was more common in the over-40s. The figures were: 46% of discovered young female diabetics were obese, as against 23% older diabetics; of the older nondiabetics 23% were obese, as against 13% of the young. We therefore suggest t h a t obesity may be a more important diabetogenic factor in young women t h a n in older women (discussed in more detail elsewhere; see Jackson et al., 1972a).

VIL Further Findings and Analyses A. A Tamilian I n d i a n C o m m u n i t y ( J a c k s o n et al,

197 r 4)

In the course of our population study of Indians in Cape Town, a community of Tamil-speaking Hindus was found to have distinctly higher blood glucose levels and more diabetes than other Indians. The unraveling of the pedigree of this group revealed t h a t all members comprised one large family, with several cross-cousin marriages and even some uncle-niece marriages. Traced back five generations, this whole extended family had arisen from eight married couples. In 1965 there were approximately 400 living members, 311 over 10 years old. Of these 266 were screened and/or received full oral GTT. (These people were not included among the general Indian survey.) In subjects over the age of 10, 6.8% of the community were known diabetics, and a further 10.9% were discovered to be diabetic, including 4 teenagers. In those over the age of 25 years, 37% were diabetic. The sexes were equally affected. There were 93 offspring of diabetic couples; these showed no more

137

Diabetes in South Africa

diabetes than the family as a whole, but seemed to develop diabetes at an earlier age (Jackson et al., 1974). In 1970, retesting of 77 available family members indicated that 12 more were now diabetic, while the mean glucose tolerance of 8 who were previously diagnosed diabetic had significantly deteriorated. This Tamil community was rather poorer than the average Cape Indian family, and had a low food expenditure per unit person. The mean sucrose intake was calculated as 23 kg per unit, compared with 26.3 for all Indians; their fat intake was also the lowest of all groups studied. This family had no more obesity than the general Indian community, the men actually being rather leaner. The prevalence of diabetes in this small community is similar to that found among the Pima Indians (Bennett et al., 1971). The difference between this inbred Tamil group and other local Hindu communities points to the importance of genetic influences, and makes us cautious of accepting the prevalence rates of any small, closed community as indicating anything more than the effects of genetic drift. B. Studies among Bushmen (Joffe et al,

1971)

In 1970 an expedition was made into the Dobe area in the northwestern Botswana (Kalahari) desert, where a group of !Kung Bushmen had congregated. Fifteen apparently healthy adults were chosen for the study. Most were underweight (male mean 46 kg, mean height 158 cm; female mean 38 kg, 145 cm); the mean age was estimated as 40 years. They had consumed fairly large quantities of carbohydrate-rich vegetable food during the week before testing and were not undernourished; their serum albumin levels were normal. Fifty-gram oral GTT produced mean blood glucose values significantly above those of White people, being 95,169, and 121 mg/dl at fasting, 1, and 2 hours, respectively, while the serum insulin levels were significantly below the White values, being 10, 33, and 20 /xU/ml, respectively. They had normal, suppressible growth hormone levels. Jenkins et al. (1974) also found relative carbohydrate intolerance in 6 !Kung Bushmen. This rather unexpected tendency toward glucose intolerance among Bushmen has likewise been found among Central African pygmies, whom the Bushmen resemble also as regards average height and weight, insulinopenia, and normal growth hormone levels.

138

W. P. U.Jackson

C. R e t e s t i n g of a S u r v e y e d P o p u l a t i o n ( J a c k s o n et al, 1 9 7 2 b ) — P r e d i c t i v e Value of G l u c o s e Tolerance Tests The individual variability of blood glucose response to a glucose load and various other observations make it clear that: 1. Basically normal subjects may on occasion show abnormal or even "diagnostic" GTTs. 2. Diabetics in the early stages of their disorder may on occasion show normal GTTs. A single negative result is not conclusive of normality, nor is a single positive response conclusive of abnormality. Individuals who show variability may be assumed to vary in wave form above and below their means. If we examine a population at any one time and then retest the top 10%, a fair proportion of these high values must represent subjects near the peak of their wave. Hence, the later examination will find most of these subjects at a lower part of their wave, and therefore they will have lower glucose values. Consequently, the mean blood glucose levels of this upper 10% should mathematically be lower on retesting, and in fact we have observed this several times. Despite the impossibility of making predictions from single or even repeated mildly abnormal GTTs, there is evidence t h a t such tests may be significant in two connections. First, ischémie heart disease is excessively frequent in subjects with asymptomatic hyperglycemia. Second, follow-up studies have shown t h a t some 12% of subjects with GTT diabetes develop florid or overt diabetes within five years (Working Party Report, 1970) and between 30% and 40% within seventeen years (O'Sullivan, 1969), while subjects with borderline abnormalities are more likely t h a n completely normal subjects to develop diabetes over a period of five years or longer (Working Party Report, 1970; O'Sullivan, 1969; McKiddie et al., 1971; John, 1950; McCullagh et al., 1954). With this in mind we repeated GTTs in 112 subjects one year after they had been similarly tested in the Cape Coloured survey described above. Repeated "borderlines" There were 40 "borderline" subjects (who had shown abnormal results originally on two separate occasions). Two were now clearly diabetic, 35 completely normal. The mean blood glucose values were significantly lower than those found initially. Repeated "discovered diabetics." The GTTs of 3 out of 4 originally

Diabetes in South Africa

139

diagnosed young "diabetics" (under age 20) were now normal; one was still diabetic. The 38 adult "discovered diabetics" now varied considerably according to whether they had dieted or not. Those who had dieted and lost weight (at least 2.3 kg) had mean glucose tolerance curves within the normal range. Nine of the original 15 now gave normal GTTs. The mean blood glucose values of those who had neither attempted to diet nor lost weight were very similar on both tests. Nevertheless, considered individually, 4 of the 14 could be classed as normal on the repeat tests. Repeated rflag storage curves." The mean half-hour level was significantly lower on repeat; only 1 curve of 14 was still of the lag type. Repeats in controls. Tests on 15 subjects who had been originally normal at screening and on GTT were repeated, and their mean blood glucose values were virtually identical. The great improvement ("normalization") of GTTs in "borderline" subjects, in the 3 young "diabetics," and in the 4 diabetics who did not change their habits suggests that many of these people may be inherently normal, but were caught near the apex of their blood glucose fluctuation waves in the initial survey. The greater improvement among diabetics who lost only a very moderate amount of weight bears testimony to the importance of carbohydrate restriction in treating mild glucose intolerance, as had been briefly explained to these subjects after their initial testing. D. Effectiveness of Different Screening Tests (Jackson et ai, 1968b) The most obvious object of diabetes surveys is to detect previously unknown diabetics in a population. "Diabetes" in this regard is defined in terms of hyperglycemia during a glucose tolerance test by arbitrary criteria, with or without glycosuria, symptoms, or raised fasting blood sugar levels. In any diabetes survey it is useful to be able to predict the ability of a screening test to include all diabetics (defined as above), while at the same time excluding nondiabêties. "Sensitivity" is defined as the percentage of all diabetics rated as positive by a given test; "specificity" is the percentage of all nondiabetics rated as negative by the same test.

140

W. P. U.Jackson

In simple terms, the sensitivity indicates the ability of the test to classify as positive those who have diabetes among an apparently well population, while the specificity indicates the ability of the same test to classify as negative those who are not diabetic. Our various surveys utilized different screening tests, including (1) blood glucose levels 1 hour after 50 gm of glucose; (2) blood glucose 2 hours after 50 gm of glucose; (3) blood glucose 1 hour after a heavy meal; (4) glycosuria after glucose; and (5) glycosuria after a meal. The effectiveness of each of these screening techniques can therefore be compared at different levels of blood glucose and different degrees of glycosuria. Glycosuria after a meal as a screening test is so obviously ineffective t h a t it will not be analyzed in detail here. The calculations of sensitivity and specificity in the 2-hour postglucose screening group and of specificity in the other groups are predicated on the premise t h a t all the diabetics in the screened population have been discovered by the methods adopted. This is probably not true of the postprandial and the 1-hour postglucose screening groups, but any minor deficiency makes virtually no difference to the specificity figures. The calculation of sensitivity in these groups has been performed only on those who screened at above 140 mg per 100 ml postglucose and 120 mg per 100 ml postprandial instead of on the whole tested community. In these two groups some subjects (approximately 10%) refused to undergo GTT after the screening procedure; the number of diabetics among these refusals was estimated pro rata. Results. As an example, if the screening levels were set at 160 mg/dl after a meal, only 33% of diabetics would be included, but 98.6% of nondiabetics would be excluded. One hour after glucose, 160 mg/dl would screen in 79% of the diabetics and screen out 93% of the total nondiabetics. A 2-hour postglucose level of 160 mg/dl would include only 58.6% of the diabetics, but exclude 99.7% of the nondiabetics. Glycosuria after glucose was highly specific (i.e., it excluded 99% or more of nondiabetics), but of low sensitivity, rising from 19% for 4 + glucose to 6 1 % for all degrees from 1+ up. Glycosuria after a meal never exceeded 50% sensitivity, thus missing over half the diabetics. It seemed clear t h a t the 1-hour postglucose screen was more efficient than the postprandial test, and the 2-hour postglucose screen was more efficient than the 1-hour test. Thus a screening level of 120 mg/dl in the 2-hour test would include some 94% of all diabetics and exclude 94% of nondiabetics. The other methods cannot approach this efficiency (cf. Mitchell and Strauss, 1964).

Diabetes in South

Africa

141

VIII. Metabolie Considerations A. Varieties of Diabetes In India, diabetes has been known for centuries as the disease of the glutton (Major, 1954), rather t h a n the '"melting down of flesh" as described by Aretaeus (±60 AD; translation, 1856); note also the opening quotation to this chapter. The White race is the odd man out in having such a comparatively high proportion of juvenile-onset, insulin-dependent diabetics. This condition is rare among Indians; in South African Indians it is almost unknown. (Insulin may of course be occasionally needed in Indians of any age to allow optimum control and health, but not for life; ketoacidosis virtually occurs only with severe complicating infection.) Insulin-independent young diabetics, both fat and lean, are fairly common among our Indians (Campbell, 1963); they occur in the other racial groups, but less frequently. In childhood, known diabetes of any type is much rarer among Indians, Africans, and Coloured people t h a n in Whites (Jackson and Huskisson, 1965). In the surveys herein summarized, several young (under 20 but over 12 years) Indian and Coloured hyperglycémies were discovered, but no White or Bantu. Pregnancy. The difference between the White and non-White people becomes more obvious during pregnancy. While virtually all our White pregnant diabetics need insulin, of the last 200 non-White pregnant women attending Groote Schuur Hospital, comprising all four ethnic groups, only 21 have needed insulin at any stage—and we believe in strict control during pregnancy. Campbell (1963) and Notelovitz (1969) have reported much the same among the Natal Indians, and Notelovitz (1970) found t h a t only 14 out of 77 pregnant Bantu diabetics needed insulin. Special types of acquired diabetes. We have already mentioned the fascinating hemosiderosis-diabetes-porphyria syndrome found commonly among Transvaal Africans. In the Cape our particular syndrome is t h a t of chronic calcific pancreatitis, occurring largely but not exclusively among Coloured men and being related to excessive consumption of alcohol (Marks and Bank, 1963). Although chronic pancreatitis is fairly frequent, it probably accounts for only approximately 2% of diabetes among the Coloured people. We found no new cases during our survey of the Coloured community.

142

W. P. U.Jackson

B. Serum Insulin Levels, etc. During our Cape Coloured survey we estimated the serum insulin (IRI) levels after glucose in 300 subjects (Jackson and Keller, 1972), including measurements at 10 and 20 minutes. Using the criteria of McKiddie and Buchanan (1969), we divided our curves into "flat," "low," "medium," "high," and "delayed" responses. Subjects with normal GTTs had very variable insulin response curves, as has been found by many other workers. The same applies to subjects with borderline and diabetic GTTs. Thus, although the mean IRI values and mean curve patterns may differ significantly between one group and another, there is a considerable overlap, which makes it seem to us doubtful whether one can say that such and such an insulin response is "characteristic" of mild diabetes, for example, and particularly doubtful whether one should deduce general pathogenetic principles from such evidence—e.g., that the initial lesion of diabetes is a delay in insulin response to stimulus. Among our own discovered diabetics, only 37% are classed as showing "delayed" responses, while 35% show a so-called normal response. Our borderline subjects showed early and excessive mean insulin responses, while the positive screenees who had normal GTTs had responses intermediate between those of the control subjects and borderlines. From this and other data (Jackson et al., 1972c) we have argued that the earliest lesion associated with glucose intolerance is insulin excess rather than insulin deficiency. There seems to be general agreement that mean serum insulin concentrations following an oral glucose load are significantly lower in the general population of Bantu people than in Whites (Rubinstein et al., 1969; Wapnick et al., 1972; de Bruin and Meyer, 1971). Asmal and Leary (1975) found lower insulin levels among Black maturity-onset diabetics than among Indians. Rubinstein and co-workers found no difference between the insulin levels of nondiabetic Indians and Whites, but we (Keller et al., 1972) and Walker et al., (1972) reported higher levels in Indian students and Indian children, respectively, than in Whites. In Cape Town our investigations suggest that insulin concentrations are lower in Coloured and Malay pregnant women than in pregnant Whites, and that the Coloured community as a whole also have lower insulin concentrations in response to a glucose load (Jackson and Keller, 1972). As mentioned above, Bushmen, like pygmies, show a high insulin response. Growth hormone levels have been reported as high among Africans (Rubinstein et al., 1969), but these findings were not confirmed by Asmal and Leary (1975). Low serum cholesterol levels among Africans

Diabetes in South Africa

143

have been found by very many workers (Brock and Gordon in 1959 already cited seven references). Urbanized Bantu have higher values than rural Bantu. Various reports find that Indians have about the same levels as Whites, with Coloured people intermediate between Whites and Africans. There is less information and no unanimity with regard to triglycéride levels. Can these biochemical findings help to explain the racial differences with regard to diabetes? The most I should like to suggest is that changes in diet probably account for the increase in diabetes and in serum cholesterol levels with urbanization. Coronary artery disease, although occasionally seen nowadays in urban Africans (Seftel et al., 1970), remains extremely uncommon among these people (Sehrire, 1971); the Coloured community, as with serum cholesterol levels, are intermediate between Africans and Whites with regard to the prevalence of IHD.

IX. Some Conclusions Regarding diabetes itself, it seems that we can accept that South African Indians do have a very high prevalence; so also have the Coloured and Malay people. Our small community of Tamil Indians share with the larger Pima Indian group the distinction of being the most diabetic known in the world,3 although ethnically they are entirely different (Tamils being Dravidian, Pimas Mongoloid). The reasons for this excessive prevalence probably differ in the two groups of "Indians"; both, however, belong in a low-income bracket and are considerably inbred. Incidentally, in our surveys in general we found at least as much hyperglycemia in the poorer people as in the more affluent. We found no great difference between the sexes, although among the Coloured, Malay, and Indian there were rather more female diabetics than male. Indians further differ from Whites in the rarity of insulin-dependent diabetes and the frequency of young (but not childhood) insulinindependent diabetes. The other ethnic groups seem somewhere in between, diabetes in childhood being certainly less common than it is among White people. Our Indians also seem to be prone to early and severe vascular disease—both coronary and microangiopathic. The urbanized Bantu, on the other hand, although developing as much diabetes as White people and as much diabetic retinopathy, still show only slight indications of increasing ischémie heart disease. 3

Some Polynesian groups with even more diabetes have recently been described.

144

W. P. U.Jackson

White South Africans have an incidence of myocardial infarction among the highest in the world, a very large mean per capita consumption of sugar (110 lb per head per year, according to the International Sugar Council, 1968), and are considerably obese. Their intake of total and saturated fats is higher than that of the Cape Coloured and Bantu. Yet, comparatively speaking their diabetes prevalence is not great. In their rural state the Bantu eat much carbohydrate (maize is their staple foodstuff); the town dwellers in general eat more animal fat (still little compared with the White population), more protein, more calories, and considerably more sugar and other refined carbohydrate. The Indians are not particularly fat—considerably less obese than the Bantu—and, although they may eat more calories and certainly more sugar than Indians in India itself, yet their consumption is lower than that of White people. It would therefore seem unlikely that these ethnic differences involving prevalence, variety, and complications of diabetes can be entirely explained by environmental factors. Fundamental racial, presumably genetic, differences appear to be ineluctable. ACKNOWLEDGEMENTS

I am particularly grateful to my colleagues throughout South Africa who have toiled with me, and to those others whose work is quoted here. Some are now in other parts of the world—all are mentioned in the text and bibliography. Heartfelt thanks also to my laboratory staff and social workers, and to my regular secretary, Mrs. E. Orkin, who dealt with the references. Our own contributions have been made possible by grants from the Council for Scientific and Industrial Research (now Medical Research Council) to the Endocrine Research Group, from the United States Public Health Service (Grant No. AM 09052) to the University of Cape Town, and assistance from the pharmaceutical houses of Pfizer, Ames, Hoechst, Warner, and Boehringer-Mannheim. I thank Mrs. M. J. McBlain for so stalwartly preparing a rush typing job and Miss L. van Schalkwyk for the map. Figures 2 and 3 have been used previously, in Diabetes and Postgraduate Medical Journal, respectively, and are republished here with permission.

References Aretaeus, the Cappadocian. (1856). "The Extant Works" (F. Adams, ed. and transi.). Asmal, A. C , and Leary, W. P. (1975). S. Afr. Med. J. 49, 810. Bennett, P. H., Miller, M. O., and Burch, T. A. (1971). Lancet 2, 125. Berry, J. N., Chakravarty, R. N., Gupta, H. D., and Malik, K. ( 1966). Indian J. Med. Res. 54, 1025. Bose, J. P. (1907). Br. Med. J. 2, 1051. Brock, J. F., and Gordon, H. (1959). Postgrad. Med. J. 35, 223. Bronte-Stewart, B., Keys, S. A., and Brock, J. F (1955). Lancet 2, 1103. Butterfield, W. J. H. (1964). Proc. R. Soc. Med. 57, 196.

Diabetes in South Africa

145

Butterfield, W. J. H., Keen, H., and Wichelow, M. J. (1967). Br. Med. J. 2, 565. Campbell, G. D. (1960a). Br. Med. J. 2, 537. Campbell, G. D. (1960b). S. Afr. Med. J. 34, 332. Campbell, G. D. (1963). S. Afr. Med. J. 37, 1195. Cosnett, J. E. (1957). S. Afr. Med. J. 31, 1109. Cosnett, J. E. (1959). Br. Med. J. 1, 187. de Bruin, E. J. P., and Meyer, B. J. S. (1971). S. Afr. Med. J. 45, 892. Goldberg, M. D. (1968). M.D. Thesis, Cape Town University. Goldberg, M. D., Marine, N., Ribiero, F., Campbell, G. D., Vinik, A. I., and Jackson, W. P. U. (1969). S. Afr. Med. J. 43, 733. Gordon, H., Vooijs, M., and Keraan, M. M. (1966). S. Afr. Med. J. 40, 1031. Hathorn, M. K., Gillman, T., and Campbell, G. D. (1961). Lancet 1,1314. Jackson, W. P. U. (1970). Acta Diabetol. Lat. 7, 361. Jackson, W. P. U. (1972). Postgrad. Med. J. 48, 391. Jackson, W. P. U., and Huskisson, J. M. (1965). S. Afr. Med. J. 39, 526. Jackson, W. P. U., and Keller, P. (1972). Hormones 3, 361. Jackson, W. P. U., Marine, N., and Vinik, A. I. (1968a). Lancet 1, 933. Jackson, W. P. U., Goldberg, M. D., Marine, N., and Vinik, A. I. (1968b).Lencet 2,1105. Jackson, W. P. U., Vinik, A. I., Joffe, B. I., Sacks, A., and Edelstein, I. (1969). S. Afr. Med. J. 43, 1496. Jackson, W. P. U., Goldberg, M. D., Major, V., and Campbell, G. D. (1970a). S. Afr. Med. J. 44, 279. Jackson, W. P. U., Vinik, A. I., Joffe, B. I., Sacks, A., and Edelstein, L (1970b). S. Afr. Med. J. 44, 1283. Jackson, W. P. U., Campbell, G. D., Goldberg, M. D., and Marine, N. (1972a). Diabetologia 7, 405. Jackson, W. P. U., Kalk, J., Whisson, A., Hardcastle, A., and Toyer, M. G. (1972b). S. Afr. Med. J. 46, 2065. Jackson, W. P. U., van Mieghem, W., and Keller, P. (1972c). Lancet 1, 1040. Jackson, W. P. U., van Mieghem, W., Marine, N., Keller, P., and Edelstein, I. (1974). S. Afr. Med. J. 48, 1839. Jenkins, T., Lehmann, H., and Nurse, G. T. (1974). Br. Med. J. 2, 23. Joffe, B. I., Jackson, W. P. U., Thomas, M. E., Tbyer, M. G., Keller, P., Pimstone, B. L., and Zamit, R. (1971). Br. Med. J. 3, 206. John, H. J. (1950). Am. J. Dig. Dis. 17, 219. Keller, P., Schatz, L., and Jackson, W. P. U. (1972). S. Afr. Med. J. 46, 152. Kenny, A. J., and Chute, A. L. (1953). Diabetes 2, 187. McCullagh, E. W., Fawell, W. N., and Lane, F. J. (1954). J. Am. Med. Assoc. 156, 925. McDonald, G., Fisher, G. F., and Burnham, C. (1965). Diabetes 14, 473. McKiddie, M. T., and Buchanan, K. D. (1969). Q. J. Med. [N. S.] 38, 445. McKiddie, M. T., Scott, R. M., and Cole, R. (1971). Postgrad. Med. J. 47, 605. Major, R. M. (1954). "A History of Medicine," p. 67. Blackwell, Oxford. Marine, N., Vinik, A. L, Edelstein, L, and Jackson, W. P. U. (1969).Diabetes 18,840-857. Marks, I. N., and Bank, S. (1963). S. Afr. Med. J. 37, 1039. Michael, C, Edelstein, L, Whisson, A., MacCullum, M., O'Reilly, L, Hardcastle, A., Toyer, M. G., and Jackson, W P. U. (1971). S. Afr. Med. J. 45, 795. Mistry, S. D. (1965). S. Afr. Med. J. 39, 691. Mitchell, F. L., and Strauss, W. T. (1964). Lancet 1, 1185. O'Sullivan, J. B. (1969). Diabetologia 4, 207. O'Sullivan, J. B., and Williams, R. F. (1966). J. Am. Med. Assoc. 198, 579. Politzer, W. M. (1955). S. Afr. J. Lab. Clin. Med. 1, 270.

146

W. P. U.Jackson

Politzer, W. M., and Schneider, T. (1961). Proc. Nutr. Soc. South. Afr. 2, 42. Politzer, W. M., and Schneider, T. (1962). S. Afr. Med. J. 36, 608. Prior, I. A. M., and Davidson, F. (1966). N.Z. Med. J. 65, 375. Notelovitz, M. (1969). S. Afr. Med. J. 43, 394. Notelovitz, M. (1970). S. Afr. Med. J. 44, 1171. Rubinstein, A. H., Seftel, H. C , Miller, K., Bersohn, I., and Wright, A. D. (1969). Br. Med. J. 1, 748. Schrire, V. (1971). S. Afr. Med. J. 45, 634. Seftel, H. C. (1967). Med. Proc. 13, 3. Seftel, H. C , and Abrams, G. J. (1960). Br. Med. J. 1, 1207. Seftel, H. C, and Schultz, E. (1961). S. Afr. Med. J. 35, 6. Seftel, H. C, Keeley, K. J., Isaacson, C, and Bothwell, T. H. (1961). J. Lab. Clin. Med. 58, 837. Seftel, H. C, Keeley, K. J., and Walker, A. R. P. (1963). S. Afr. Med. J. 37, 1213. Seftel H. C , Kew, M. C , and Bersohn, I. (1970). S. Afr. Med. J. 44, 8. Sloan, N. R. (1963). J. Am. Med. Assoc. 183, 419. Wainwright, J. (1969a). S. Afr. Med. J. 43, 83. Wainwright, J. (1969b). S. Afr. Med. J. 43, 136. Walker, A. R. P. (1966). S. Afr. Med. J. 40, 841. Walker, A. R. P., Mistry, S. D., and Seftel, H. C. (1963). S. Afr. Med. J. 37, 1217. Walker, A. R. P., Holdsworth, C. M., and Walker, E. J. (1971). S. Afr. Med. J. 45, 516. Walker, A. R. P., Bernstein, R. E., and Du Plessis, I. (1972). S. Afr. Med. J. 46, 1916. Wapnick, S., Kanengoni, E., Wicks, A. C. B., and Jones, J. J. (1972). Lancet 2, 300. West, K. M., and Kalbfleisch, J. M. (1971). Diabetes 20, 99. Wood, M. M. (1960). Med. Proc. 6, 140. Working Party Report. (1962). Br. Med. J. 1, 1497. Working Party Report. (1963). Br. Med. J. 2, 655. Working Party Report. (1970). Br. Med. J. 3, 301.

O. P. GUPTA,* M. H. JOSHI,* and S. K. DAVE*

Prevalence of Diabetes in India I. Introduction II. Materials and Methods III. Results and Discussion A. Age and Sex B. Economic Status C. Occupational Status D. Diet E. Heredity F. Parity G. Obesity H. Glycosuria IV. Summary References

147 149 152 153 154 156 156 157 158 158 159 160 163

I. Introduction The study of the prevalence of diabetes in any region gives an indication of how widespread and important the problem is in that part of the world. It is of great help in studying its etiologic factors, planning the organization of medical care, and stimulating research so as to find a cure or possibly a means of preventing the disease. Although we do not know precisely the cause of diabetes, many studies have emphasized that the disease is the sequela of the interaction of various etiologic factors. These factors vary from place to place. Therefore, the study of the prevalence of diabetes in the various geographical regions of a country like India, where so many ethnic, dietary, climatic, economic, and social variations are found, is of great importance. * Postgraduate Department of Medicine and Directorate of Medical Education and Research, New Block, Civil Hospital, Ahmedabad, India. 147

148

0. P. Gupta et al.

Although worldwide interest in the epidemiologic study of diabetes began as early as 1929 (J. V. De Porte, quoted by Butterfield, 1969), proper attention to the subject has been paid only recently. The prevalence of diabetes has been reported from different parts of the world, varying from the lowest of 0.03% of Alaskan Eskimos (Scott and Mouratoff, 1970), 1.03% in Bedford (Butterfield, 1964), 1.67% in India (Gupta et al., 1970), 1.88% in Norway (Jorde, 1969), 2.0% in Japan (Kuzuya and Kosaka, 1970), 2.0% in Africa (Campbell, 1969), 4.5% in Cleveland (Kent and Leonard, 1968), 4.7% in Oxford (United States) (Wilkerson and Krall, 1947), 6.82% in Colombia (Medina and Cortazar, 1971), and 14.7% in Cape Town (Marine and Jackson, 1966), to more than 20% in the Pima Indians of Arizona (Bennett and Miller, 1972). Earlier reports indicated that diabetes is not common in Asia, but since the end of World War II interest in the study of diabetes has increased (Patel, 1969). The earliest report on the prevalence of diabetes in India was published in 1938 by A. Chakravarty (quoted by Tulloch, 1962). He examined the urine in a population aged 5 years and above and reported the prevalence of diabetes as 0.7%. Subsequent reports have shown variations in the incidence of diabetes ranging from 0.7% (Datta et al., 1966) to 12.6% (Tripathy, 1970). As was mentioned in a previous communication (Gupta et al., 1970), these studies were often confined to analysis of hospital records of outpatients and/or inpatients, and diagnosis was based on the presence of glycosuria (Patel, 1969; Ajgaonkar and Sathe, 1960; Ajgaonkar, 1972; Ramadwar, 1965; Pai et al., 1966; Shanker, 1966; Sarojkumari and Padmavati, 1966; Misra et al., 1966) or on the detection of glycosuria in visitors in exhibitions or other populations of various cities (Patel et al., 1966; Vishwanathan, 1966; Satyanarayan Rao et al., 1966). More recently, studies have been carried out after giving a glucose load or by postprandial examination of blood for glucose in selected groups of population (Ganguly et al., 1964; Ahuja et al., 1966; Berry et al., 1966; Patel et al., 1966). Tripathy (1970) conducted a study in a general population after a glucose load and reported a prevalence rate of 12.6%. Gupta et al. (1970) for the first time conducted a house-tohouse survey in a general population over 5 years of age and reported a prevalence rate of 1.67% in the urban area of Ahmedabad in Western India. As the above studies did not reflect the prevalence of diabetes in all geographical regions of India and particularly the prevalence of diabetes in rural populations, the Indian Council of Medical Research decided to conduct a collaborative study of the prevalence of diabetes in

Prevalence of Diabetes in India

149

TRIVANDRUM

FIG. 1. Map of India showing six geographical regions for study of prevalence of diabetes (Indian Council of Medical Research, 1975).

six different geographical regions in India (Fig. I), 1 covering both urban and rural populations in a statistically significant number of cases, adopting a common methodology for all centers by using a printed form, and subjecting the data to IBM computer analysis. The study we conducted at Ahmedabad representing the western geographical region of India is presented here, along with a comparison of the results reported from various centers and other geographical regions. Studies on the prevalence of diabetes carried out by various authors in India so far are listed in Table I.

II. Materials and Methods A house-to-house survey was conducted in the urban part of Ahmedabad and in eleven villages situated up to 30 km from Ahmedabad. Details with reference to age, sex, socioeconomic status, (income, education, occupation), dietetic history, height, weight, history of diabetes in the subject and family, and parity were recorded on a printed form designed for computer analysis. This study covered 25,000 subjects in the urban area and 20,000 in Ahmedabad (Gupta et al.), Calcutta (Chhetri et al.), Cuttack (Tripathy et al.), Delhi (Ahuja et al.), Poona (Mutalik et al.), and Trivandrum (Pai et al.), ail 1975. (Unpublished data of I.C.M.R.)

Urine Blood and urine Blood and urine GTT Urine Urine 10 and above 5 and above 15 and above 20 and above All ages Adults Varanasi Ahmedabad Ahmedabad Hyderabad Jabalpur Madras

1966 1970 1975 1972 1966 1970

9. 10. 11. 12. 13. 14.

Blood Urine Blood and urine Blood Blood 15 and above 5 and above 15 and above All ages 20 and above

Chandigarh Calcutta Calcutta Pondichery Lucknow

1966 1938 1975 1966 1964

4. 5. 6. 7. 8.

Gour Gupta et al. Gupta, 0 . P. (unpublished data) Jaya Rao et al. Misra et al. Moses

Berry et al. Chakravarty (quoted by Tulloch, 1962) Chhetri et al. (unpublished data) Dattae^aZ. Ganguly et al.

15 and above 15 and above

Delhi Bombay

1975 1968

Urine Blood Blood and urine Urine

Test used

2. Ahuja, M. M. (unpublished data) 3. Ajgaonkar and Sathe

All ages

Age (years)

Delhi

Place

1966

Year

1. Ahuja et al.

Author

TABLE I

PREVALENCE RATE OF DIABETES REPORTED BY VARIOUS AUTHORS IN INDIA

9.4 7.6 1.23 2.5" 7.0" 1.53 0.73 1.63 0.70 2.3 1.1' 2.7 1.67 2.16 2.4 1.70 3.8

Prevalence rate (%)

Pateletal.

Pa tel and Dhirwani Sainani et al. Ramadwar, D. K. Sarojkumari and Padmavati Satyanarayan Rao et al. Shanker et al. Tripathy et al. Tripathy, B. B. (unpublished data) Vaishnava et al. Vishwanathan et al.

" Work area. " Offices. '' Rural.

21. 22. 23. 24. 25. 26. 27. 28. 29. 30.

20. 1963 1959 1966 1965 1966 1966 1966 1970 1975 1964 1966

Bombay Bombay Bombay Nagpur New Delhi Hyderabad Hubli Cuttack Cuttack Vellore Madras

Poona Trivandrum Trivandrum Banglore

1975 1975 1966 1973

16. 17. 18. 19.

Mutalik, G. M. (unpublished data) Pai et al. Pai étal. Parmeshwara, V.

Calcutta

1973

15. Mukerjee, A. B.

20 and above All ages 10 and above 15 and above All ages 20-60

All ages All ages All ages 20 and above

15 and above 15 and above 20 and above 5 and above

All ages

Blood Urine Blood Blood Urine Urine GTT Blood Blood Urine Urine Urine Urine Urine Blood Blood Blood Blood and urine

and blood

and urine and urine

0.7 2.84 1.48 1.81 8.7 0.81 1.19 2.36 0.98 2.24 2.4 2.26 4.12 2.20 12.67 1.40 2.56 11.3

sfc'

Î5

^

*> 5**.

S5-

0

^ ^

Ï5

^

"Ό

S^ ^

152

0. P. Gupta et al.

the rural area. Out of these, 3516 subjects in the urban area and 3495 subjects in the rural area, above the age of 15 years, were given a 75-gm oral glucose load on an empty stomach in early morning. After 2 hours, blood and urine samples were collected for sugar estimation. Blood sugar was estimated by the Somogyi-Nelson method, and urine sugar was examined by oxidase strips. Subjects having a blood sugar value above 100 mg% were subjected to a standard oral glucose tolerance test (SOGTT) by giving 40 gm of glucose per square meter of body surface area. Fasting, 1-, 2-, and 3-hour venous blood samples were then taken. The criteria for diagnosis of diabetes were (1) a summed value of four samples exceeding 500 mg or (2) a fasting value greater than 100 mg%, a first-hour value greater than 170 mg%, a second-hour value greater than 120 mg%, and a third-hour value greater than 110 mg%. At least two values in (2) must be abnormal. The results showed that 263 subjects in the urban population and 367 in the rural population were positive. Out of these, 58 urban and 3 rural cases were known diabetics, 17 urban and 23 rural cases did not cooperate, and 188 urban and 333 rural cases were subjected to SOGTT. In the urban population 49 cases were found to have an abnormal GTT, and 42 in the rural population. Three positive cases in the rural area expired during the survey.

III. Results and Discussion As shown in Table II, the prevalence of diabetes in the urban and rural populations was 3.04% and 1.287%, respectively. This difference was statistically significant. However, the prevalence rate of newly detected diabetics was nearly equal in both groups of population—i.e., urban 1.39% and rural 1.2%. As may be seen from Table I, reports of the prevalence of diabetes from different regions of the country vary widely. This variation is also found in published reports of the prevalence rate from different parts of the world. Some of them have been referred to earlier; others from the remaining part of the world are too numerous to be mentioned here. The wide variation in prevalence rates could be due to differences in diagnostic criteria, selection of population samples, methods of conducting surveys, and many other variable factors. In the study of the prevalence of diabetes in six different geographical regions in India, the above variables were standardized (see Table VII). Even so, prevalence rates in the urban populations range from 0.95% in Delhi to 3.04% in Ahmedabad. Similarly, in the rural populations the preva-

153

Prevalence of Diabetes in India TABLE II PREVALENCE RATES IN URBAN AND RURAL POPULATIONS

Total number studied Males Females Total number of diabetics Known cases New cases Prevalence rate Known cases New cases

Urban

Rural

3516 1909 1607 107 58 49 3.04% 1.65% 1.39%

3495 1891 1604 45 3 42 1.287% 0.087% 1.2%

X2 = 24.74, d.f. = 1; P < 0.001 (significant). t = 5.5 > 1.96; P < 0.001 (significant).

lence rates vary from 0.60% in Cuttack to 1.28% in Ahmedabad. From these studies it is apparent that regional variations in the prevalence of diabetes may occur because of genetic or evnironmental differences. A. Age and Sex Table III shows that in both urban and rural populations the prevalence of diabetes increases with advancing age. The maximum prevaTABLE III COMPARISON OF PREVALENCE RATES IN URBAN AND RURAL POPULATIONS ACCORDING TO AGE AND SEX

Age group (years)

Male

16-20 21-30 31-40 41-50 51-60 Above 60

_ 0.58 2.33 4.65 13.56 16.64 3.93

2.0

Total

Urban prevalence rate (%)

Rural prevalence rate (%)

Total

Male

_

_

0.21 1.53 4.60 6.50 15.34

0.38 1.91 4.63 10.33 16.47

_ —

3.04

Female

Urban: * 2 = 11.09, d.f. = 1; P < 0.001 (significant). Rural: χ2 = 0.04; d.f. = 1; P > 0.50 (insignificant).

Female

Total

_

_

0.40 1.32 1.25 5.66

0.93 0.21 0.70 1.34 4.76

0.34 0.31 0.99 1.29 5.32

1.32

1.25

1.29

154

0. P. Gupta et al.

lence rate was from the fifth decade on. Most of the authors cited (Gupta et al, 1970; Patel et al, 1966; Vishwanathan et al., 1966; Datta et al., 1966; Pai et al, 1966; Mukerjee, 1973; Sarojkumari and Padmavati, 1966; Kent and Leonard, 1968; Joslin, 1959; Tulloch, 1962; Sakumoto et al., 1970; Bennett and Miller, 1972) have made similar observations, and it is now almost universally accepted that the majority of diabetics manifest the disease after the age of 40. Similar observations were made by Chhetri et al. (1975), Tripathy et al. (1975), Mutalik et al. (1975), and Pai et al. (1975) from various centers of the present collaborative study (see Table VII). The male-to-female ratio was 2:1 in the urban population and 1:1 in the rural population. In the urban population a preponderance of males has also been reported by Gupta et al. (1970), Ahuja et al. (1966), Vishwanathan et al. (1966), Pai et al. (1966), Misra et al. (1966), Datta (1973), Gour (1966), Ajgaonkar and Sathe (1960), Parmeshwar (1973), Ganguli et al. (1964), Vaishnava et al. (1964), A. G. Shaper, 1959 (quoted by Tulloch, 1962), and Sakumoto et al. (1970). However, some authors in India and most of the authors from Western countries have reported a higher prevalence in females (Sainani et al., 1966; Shanker, 1966; Patel et al, 1966; Joslin, 1959; Jaya Rao et al, 1972; Tulloch, 1962; Bennett and Miller, 1972). Other authors have found an equal incidence in males and females (Mukerjee, 1973; Sasaki et al, 1970; R. G. Fernando, 1965, quoted by Patel 1969; Prior and Davidson, 1969; Kawamoto, 1970; H. J. Germar and A. G. Villanueva, quoted by Patel et al 1969). The studies in the rural areas are limited. Some authors have reported a slight preponderance of males over females (Jaya Rao etal, 1972; Tripathy, 1970; 1975; Chhetri et al, 1975; Sakumoto et al, 1970). It is likely that in the future, with better nutrition and increased longevity, the prevalence rate of diabetes in females, compared with that in males, will correspond to the prevalence of this disease in females reported from Western countries. B. Economic Status Figure 2 shows that in the urban population the prevalence of diabetes increases with higher economic status, a finding that is statistically significant. The prevalence rate ranges from 2.16% in the lowest income group to 18.40% in the highest income group. In the rural population the prevalence of diabetes does not show any statistical significance based on income groups. Various authors have reported a high incidence of diabetes among the rich class of the population. The

155

Prevalence of Diabetes in India — · (URBAN)

20

— · (RURAL)

18 h 16

£

14

U

12

UJ

o

UJ

a.

OO) cviro

Ο0> ^-ιο

0CÖ

INCOME IN RUPEES PER CAPITA PER ANNUM (Rs.10= $ 1 approx.) URBAN ; x 2 = 7.94, d.f. =9;/? 0.10). The mortality of diabetes and the consumption of nutrients among the countries showed a similar relationship (Goto, 1963).

184

Yoshio Goto

Kawate et al. (1974, 1976) studied diabetics of Japanese origin residing in Hawaii and found that the cause of death in 182 diabetics was arteriosclerotic heart disease (41%), cerebral vascular disease (25%), nephropathy (12%), infection (7%), diabetic coma (3%), malignant neoplasm (2%), and the remaining 10% from other causes. It is noted in their report that the percentage of the cardiac deaths was 27% in the series of 1952-1961 and 47% in the series of 1962-1972. They found that the frequency of hypercholesterolemia was significantly higher in Japanese in Hawaii. Kawate et al. (1976) compared the nutritional intake of Japanese men living in Japan and of those living in Hawaii. The average total was 2132 calories in Japan and 2274 calories in Hawaii; the protein intake was 76 gm in Japan and 94 gm in Hawaii; the fat intake was 36 gm in Japan and 85 gm in Hawaii; and the carbohydrate intake was 335 gm in Japan and 260 gm in Hawaii. These results indicate clearly that the high fat intake may be a cause of the high incidence of cardiac death in diabetic Japanese in Hawaii. As shown in Fig. 1, the average fat intake of Japanese people has increased during the last twenty years, but it is still very low compared with that in the United States. It is well known that American people are more obese than Japanese and that the serum lipid level of the Japanese people is lower than that of Americans. This may be one of the reasons for the low incidence of ischémie heart disease among Japanese diabetics.

IV. Vascular Complications in Diabetic Patients: Clinical Observations A. Diabetic Retinopathy 1. AGE AT ONSET AND DURATION OF DIABETES

Many investigators have reported the incidence of diabetic retinopathy in Japan (Saito et al., 1956; Kato et al., 1959; Rudnick and Anderson, 1962; Tokuda et al., 1963; Katsuse, 1966; Fukuda et al., 1969; Kojima et al., 1971; Kuzuya and Kosaka, 1971). However, the incidence might be different depending on the age distribution of the patient population and also on the duration or control of their diabetic state. For this reason, we observed the incidence by grouping age at diabetes onset and duration of the disease; in other words, the incidence was studied by each age decade and also matched against the duration of—i.e., at 3-year, 5-year, 10-year, and 15-year periods. In this study,

185

Vascular Complications in Diabetes in Japan

we used a multi-institutional survey method to avoid an inequality in patient distribution frequently experienced in the specialized outpatient clinic. The details of this study were described in our original report (Goto et al., 1976); Table VI is reproduced from that report. The relationships among the frequency of the retinopathy, the age of diagnosis, and the duration of the disease are diagrammed in Fig. 5. At the time of diagnosis, which is regarded as the age of onset, no case with retinal changes was observed in the under-10-year onset cases, but two cases were observed in the 10-19-year-old group. Both were 18year-old females. It may be said, therefore, that diabetic retinopathy does not appear in cases under 15 years old at the time of diagnosis. The frequency of retinopathy at the time of diagnosis increases with age at onset of diabetes, and retinopathy occurs and developes with the TABLE VI FREQUENCY OF DIABETIC RETINOPATHY OBSERVED BY AGE OF ONSET, SEX AND DURATION OF DIABETES"

Duration of diabetes (years) Age —

Satiety

VLH\

*~^

t Insulin

(Smaller meals)(I Glucose)

/VMH

I

0

TGIucose

VLHN

Λ

Eating

UInsulin)

FIG. 2. Diagram depicting the influence of insulin-sensitive glucoreceptors of the CNS. In the glucoregulatory system, a n increase of glucose uptake triggers reflexes in both the VMH and the VLH simultaneously. The VMH causes a reduced meal size (increased satiety) and the VLH causes increased insulin secretion. A decrease of glucose uptake activates the VMH to increase blood glucose levels and the VLH to increase eating.

The glucoregulatory system is characterized by neurons that are sensitive to glucose uptake. The function of such glucoreceptors is to maintain a somewhat normal level of blood glucose through negative feedback reflexes, as discussed in Section V. Furthermore, the glucose uptake rate by these glucoreceptors varies in proportion to the amount of insulin present. If blood glucose levels become dangerously low, the insulin-sensitive glucoreceptors initiate both feeding and activation of the sympathetic nervous system (cf. Fig. 2). Both responses would lead to increased glucose levels, the first by bringing new fuels into the body and the second by mobilizing available energy stores. If glucose levels become excessively elevated, insulin secretion is facilitated and satiety-inducing neurons may be activated. That these glucoreceptors are sensitive to CSF as well as plasma glucose was demonstrated by experiments in which either 2-DG (Frohman et al., 1973; Müller et al., A CDecreased CSF Increased CSF Λ\ lin Insulin ^ ^ ^^^ ^ - v^ Insulin /VMH

Satiety

VL>T\

tGlucose (I Insulin)

/VMH

I VLH\

t Insulin Eating (Ψ Glucose)

FIG. 3. Diagram depicting the influence of CNS insulin receptors. Unlike the reflexes of the glucoregulatory system, which excite portions of the VMH and the VLH simultaneously (Fig. 2), reflexes of this weight-regulatory system excite only one or the other hypothalamic nucleus as a unit. An increase of insulin in the CSF is hypothesized to be the signal indicating an increase of adiposity that causes the VMH to reduce meal size and to mobilize stored fuels. A decrease of insulin in the CSF signifies a reduction of adiposity and causes the VLH to initiate eating and to secrete insulin to increase the uptake and storage of substrate.

304

Stephen C. Woods and Daniel Porte, Jr.

1973; Miselis and Epstein, 1970) or insulin (Anderson and Hazelwood, 1969; Chen et al, 1975; Chowers et al, 1961, 1966; Woods and Porte, 1975) was added to the CSF and appropriate glucoregulatory reflexes initiated. The weight-regulatory system is more difficult to characterize. This system presumably responds to correlates of body adiposity by an appropriate reflex adjustment of behaviors that alter body fat stores. Ideally, some neurons would be able to monitor the amount of fat in the body and relay the information to the CNS. Such neurons have never been described. On the other hand, the amount of fat in the body correlates well with basal plasma (and presumably CSF) insulin levels. All that is therefore required is an insulin receptor situated so as to monitor CSF insulin levels. As pointed out above, there is substantial evidence that such receptors exist. These receptors presumably influence neurons that have either a genetically determined normal rate of activity or an environmentally determined rate, or both. In particular, the level of other hormones such as growth hormone (Woods et al, 1974b) may be expected to be important. When an animal becomes relatively obese, more insulin is secreted. In the present model, this increase of insulin would be monitored by the CSF insulin receptors, which, in turn would trigger responses leading to smaller meals (increased satiety) and to decreased insulin secretion (for less storage of newly ingested fuels) (cf. Fig. 3). The result would be a decrease in weight to the set point of the animal. When an animal is fasted, it loses weight and its insulin levels decrease. This change of insulin would be monitored by the CSF insulin receptors, and reflexes would be triggered leading to larger meals (decreased satiety) and increased insulin secretion (for storage of the available fuels). Again, CSF insulin would provide the signal necessary for normal maintenance of body weight. There is evidence that the glucoregulatory system is relatively more important if both systems are activated at the same time. This would appear reasonable since it would be maladaptive for the hypothalamus to be concerned with fat stores in the face of an emergency alteration of glucose levels. The administration of insulin to the CSF presumably provides an input to both systems. The insulin would cause glucoreceptors to take in glucose more rapidly and to respond as if glucose levels were suddenly elevated; i.e., they would trigger pancreatic insulin secretion. This, in fact, has been found to occur (Woods and Porte, 1975). On the other hand, the weight-regulatory insulin receptors, also sensitive to CSF insulin, would react as if the body had suddenly become too fat and would therefore act to decrease insulin secretion. Chen et al., (1975) have found that when insulin is administered to the CSF of

The CNS, Pancreatic Hormones, Feeding, and Obesity

305

anesthetized dogs there is a rapid increase of pancreatic insulin secretion as measured at the pancreaticoduodenal vein. However, even though CSF insulin values remained elevated for over three hours, the pancreatic insulin secretion was transient and rapidly declined. Normally, rapid changes of the insulin concentration of the CSF would not occur, thus rendering this approach somewhat unphysiological. Nevertheless, experiments in which CSF insulin has been manipulated have helped to reveal the nature of the glucoregulatory system. The finding of an increase of insulin secretion after the intraventricular administration of insulin would not be predicted by the glucostatic theory, for example. Lesions of the ventral hypothalamus provide an experimental situation in which the system is suddenly changed. The most striking effect of a VMH lesion is that relating to body weight. Both increased eating and hyperinsulinemia also occur, of course. However, as detailed above, the major effect of the lesion is to increase the set point for weight without affecting the ability to regulate that weight. Therefore, we are suggesting that the amount of insulin seen by the CNS, presumably through the level in the CSF, serves as a signal that influences weight-maintaining reflexes. When the signal is increased, weight is reduced through decreased feeding, and vice versa. A lesion of the VMH might act simply to change the sensitivity of the system such that more insulin is required to turn off the weight-increasing responses. The animal therefore responds to such a lesion with an increase in insulin secretion and overeating until appropriate levels of insulin are achieved in the CSF. This is accomplished by an elevated weight which is then rigorously defended. The opposite occurs after a lesion of the VLH. Any procedure which affects insulin secretion chronically should interact with this system. For example, rendering an animal experimentally diabetic by destroying its insulin secretory capacity has two effects. First, there is a decreased insulin signal to the CNS, so the brain responds as if the animal were too thin; it elicits overeating (diabetic hyperphagia). Equally important, the lack of insulin reduces glucose utilization by insulin-sensitive tissues, thus eliciting a chronic state in which an emergency decrease of glucose exists in these tissues and therefore chronic activation of the sympathetic glucoregulatory system. This also causes overeating as well as hyperglycemia. An experimentally diabetic animal therefore overeats but cannot gain weight since the lack of insulin precludes fuel storage. Vagotomy reportedly lowers both body weight and basal insulin levels (Powley and Opsahl, 1976) as well as glucose-stimulated insulin secretion (see Woods and Porte, 1974). The decrease of weight would

306

Stephen C. Woods and Daniel Porte, Jr.

not be predicted by the present theory since lowered insulin levels should initiate overeating. However, the vagus is a mixed nerve and provides a large sensory input to the CNS (Paintal, 1973). It may provide an input to the ventral hypothalamus necessary for normal eating. Ball (1974) provided evidence for this concept in rats with stimulating electrodes implanted in the VLH. Stimulation of the same electrode elicited both feeding and self-stimulation. When the vagus was severed in these rats, the current necessary to maintain selfstimulation was relatively unchanged, whereas that necessary to elicit feeding was greatly increased. Vagotomy also renders an animal relatively insensitive to the feeding stimulus provided by 2-DG (Booth, 1972). These studies suggest that the vagus provides a tonic facilitory effect on feeding centers. Any treatment that interferes with vagal afférents may then modulate the expected feeding responses. In summary, the theory presented in this article predicts that alterations of the substrates and peptide hormones of the CSF will be part of the feedback system for metabolic regulation. Although the concept is undoubtedly too simple, it does serve to integrate a great deal of data and to provide a framework around which research might be designed. In addition, this approach emphasizes the inadequacy of the glucostatic theory for satiety and weight regulation. ACKNOWLEDGMENTS

Preparation of this paper was aided by institutional funds from the Veterans Administration and by grants number AM-05498, AM-12829, and AM-17844 from the National Institutes of Health. We thank Edwin L. Bierman, Jeff Halter, Andy Goldberg, R. Paul Robertson, and Phillip H. Smith for their useful suggestions.

References Ahlskog, J. E. (1974). Brain Res. 82, 211. Ahlskog, J. E., and Hoebel, B. G. (1973). Science 182, 166. Alberti, K. G. M. M., Christensen, N. J., Christensen, S. E., Hansen, Aa. P., Iversen, J., Lundbaek, K., Seyer-Hansen, K., and Orskov, H. (1973). Lancet 2, 1299. Alvarez-Buylla, R., and Alvarez-Buylla, E. R. (1975). J . Comp. Physiol. Psychol. 88,155. Alvarez-Buylla, R., and Carrasco-Zanini, J. (1960). Acta Physiol. Lat.Am. 10, 153. Anand, B. K., and Dua, S. (1955). Indian J. Med. Res. 4 3 , 113. Anderson, D. K., and Hazelwood, R. C. (1969). J. Physiol. (London) 202, 83. Arees, E. A., Veltman, B. I., and Mayer, J. (1969). Exp. Neurol. 25, 410. Bagdade, J. D., Bierman, E. L., and Porte, D., Jr., (1967). J. Clin. Invest. 46, 1549. Baig, H. A., and Ezdozien, J. C. (1965). Lancet 2, 662. Balagura, S. (1971). In "The Hypothalamus" (L. Martini, ed.), pp. 181-193. Academic Press, New York. Balagura, S., and Kanner, M. (1971). Physiol. Behav. 7, 251.

The CNS, Pancreatic Hormones, Feeding, and Obesity

307

Ball, G. G. (1974). Science 184, 484. Bassett, J. M. (1974). Aust. J. Biol. Sei. 27, 157. Bergman, R. N., and Miller, R. E. (1973). Am. J. Physiol. 225, 481. Bernard, C. (1849). C. R. Soc. Biol. Ses Fil. 1, 60. Bernardis, L. L., and Frohman, L. A. (1970). Neuroendocrinology 6, 319. Bernardis, L. L., and Goldman, J. K. (1976). J. Neurosci. Res. 2, 9 1 . Bernstein, I. L., Lotter, E. C , Kulkosky, P. J., Porte, D., Jr., and Woods, S. C. (1975). Proc. Soc. Exp. Biol. Med. 150, 546. Blass, E. M., and Kraly, F. S. (1974). J. Comp. Physiol. Psychol. 86, 676. Bloom, S. R., Edwards, A. V., and Vaughan, N. J. A. (1973). J. Physiol. (London) 233, 457. Bloom, S. R., Edwards, A. V., and Vaughan, N. J. A. (1974a). J. Physiol. (London) 236, 611. Bloom, S. R., Vaughn, N. J. A., and Russell, R. C. G. (1974b). Lancet 2, 546. Booth, D. A. (1968). J. Comp. Physiol. Psychol. 65, 13. Booth, D. A. (1972). Physiol. Behav. 8, 1069. Booth, D. A., and Brookover, T. (1968). Physiol. Behav. 3 , 439. Booth, D. A., and Pitt, M. E. (1968), Physiol. Behav. 3, 447. Bray, G. A. (1976). "The Obese Patient." Saunders, Philadelphia. Bray, G. A., and York, D. A. (1971). Physiol. Rev. 5 1 , 598. Breisch, S. T., Zelman, F. P., and Hoebel, B. G. (1976). Science 192, 382. Brobeck, J. R., Tepperman, J., and Long, C. N. H. (1943). Yale J. Biol. Med. 15, 831. Caffyn, Z. E. Y. (1972a). J. Pathol. 106, 49. Caffyn, Z. E. Y. (1972b). J. Pathol. 106, 57. Cegrell, L. (1968). Acta. Physiol. Scand. Suppl. 314, 17. Chen, M., Woods, S. C., and Porte, D., Jr. (197r5). Diabetes 24, 910. Chlouverakis, C., and Bernardis, L. L. (1972). Diabetologia 8, 179. Chlouverakis, C., and Hojnicki, D. (1974). Metabolism 2 3 , 133. Chowers, L, Lavy, S., and Halpern, L. (1961). Exp. Neurol. 3, 197. Chowers, L, Lavy, S., and Halpern, L. (1966). Exp. Neurol. 14, 383. Coleman, D. L., and Hummel, K. P. (1969). Am. J. Physiol. 217, 1298. Conway, M. J., Goodner, C. J., Werrbach, J. H., and Gale, C. C. (1969). J . Clin. Invest. 48, 1349. Coore, H. G., and Rändle, P. J. (1964). Biochem. J. 93, 66. Correll, J. W. (1963a). Fed. Proc, Fed. Am. Soc. Exp. Biol. 22, 574. Correll, J. W. (1963b). Science 140, 387. Coscina, D. V , and Stancer, J. C. (1977). Science 195, 416. Curry, D. L., and Bennett, L. (1974). Biochem. Biophys. Res. Commun. 60, 1015. Curry, D. L., and Joy, R. M. (1974). Endocrine Res. Commun. 1, 229. Debons, A. F., Krimsky, L, From, A., and Cloutier, R. J. (1969). A m. J. Physiol. 217,1114. Debons, A. F., Krimsky, L, and From, A. (1970). Am. J. Physiol. 219, 938. Delgado, J. M. R., Roberts, W. W., and Miller, N. E. (1954). Am. J. Physiol. 179, 587. Dubois, M. (1975). Proc. Natl. Acad. Sei. U.S.A. 72, 1340. Epstein, A. N. (1960). A m . J. Physiol. 199, 969. Epstein, A. N., Nicolaidis, S., and Miselis, R. (1975).In "Neural Integration of Physiological Mechanisms and Behavior" (G. J. Mogenson and F. R. Calerasu, eds.), pp. 1 4 8 168. Univ. of Toronto Press, Toronto. Fenstermacher, J. D., Patlak, C. S., and Blasberg, R. G. (1974).Fed. Proc,Fed. Am. Soc Exp. Biol. 33, 2070. Fibiger, H. C , Zis, A. P., and McGeer, E. G. (1973). Brain Res. 55, 135. Fischer, U., Hommel, H., Zeigler, M., and Michael, R. (1972). Diabetologia 8, 104.

308

Stephen C. Woods and Daniel Porte, Jr.

Fleming, D. G. (1969). Ann. N. Y. Acad. Sei. 157, 985. Friedman, M. I., and Stricker, E. M. (1976). Psychol. Rev. 83, 409. Frohman, L. A., and Bernardis, L. L. (1971). Am. J. Physiol. 221, 1596. Frohman, L. A., Ezdinli, E. Z., and Javid, R. (1967). Diabetes 16, 443. Frohman, L. A., Müller, E. E., and Cocchi, D. (1973). Horm. Metab. Res. 5, 2 1 . Fryer, J. H., Moore, N. S., Williams, H. H., and Young, C. M. (1955). J. Lab. Clin. Med. 45, 684. Gasnier, A., and Mayer, A. (1939). Ann. Physiol. Physicochim. Biol. Physiol. 38, 353. Gerich, J. E., Langlois, M., Noacco, C., Schneider, V., and Forsham, P. H. (1974). J. Clin. Invest. 53, 1441. Gerich, J. E., Lovinger, R., and Grodsky, G. M. (1975). Endocrinology 96, 749. Gerich, J. E., Charles, M. A., and Grodsky, G. M. (1976). Ann. Rev. Physiol. 38, 353. Girardier, L., Seydoux, J., and Campfield, L. A. (1976). J. Physiol. (Paris) 72, 801. Gold, R. M. (1973). Science 182, 488. Goldfien, A., Gullixson, K. S., and Hargrove, G. (1966). J. Lipid Res. 7, 357. Goldfine, I. D., Abraira, C , Gruenwald, D., and Goldstein, M. S. (1970). Proc. Soc. Exp. Biol. Med. 133, 274. Goldman, J. K., Schnatz, J. D., Bernardis, L. L., and Frohman, L. A. (1972). Metabolism 21, 132. Gonzalez, M. E , and Novin, D. (1974). Physiol. Psychol. 2, 326. Goodner, C. J., and Porte, D., Jr. (1972). In "Handbook of Physiology, Endocrine Pancreas" (D. F. Steiner and N. Freinkel, eds.), Sect. 7, Vol. I, pp. 597-609. Am. Physiol. Soc. Washington, D. C. Goodner, C. J., Tustison, W. A., Davidson, M. B., Chu, P., and Conway, M. J. (1967). Diabetes 16, 576. Goodner, C. J., Koerker, D. J., Werrbach, J. H., Toivola, P. T. K., and Gale, C. C. (1973). Am. J. Physiol. 224, 534. Grodsky, G. M. (1970). Vitam. Horm. (New York) 28, 37. Grossman, M. I. (1974). Gastroenterology 67, 730. Guillemin, R., and Gerich, J. E. (1976). Ann. Rev. Med. 27, 379. Hadden, D. R., and Beif, M. D. (1967). Lancet 2, 589. Hâkanson, R., Liedberg, G., and Lundquist, I. (1971). Experientia 27, 460. Haies, C. N., and Kennedy, G. C. (1964). Biochem. J. 90, 620. Han, P. W., Yu, Y. -K., and Chow, S. L. (1970). Am. J. Physiol. 218, 769. Hansen, A. P., and Lundbaek, K. (1976). Diab. Metab. 2, 203. Hatfield, J. S., Millard, W. J., and Smith, C. J. V. (1974). Pharmacol. Biochem. Behav. 2, 223. Herberg, L. J. (1960). Nature (London) 187, 245. Herberg, L. J., and Coleman, D. L. (1977). Metabolism 26, 59. Hervey, G. R. (1959). J. Physiol. (London) 145, 336. Hetherington, A. W., and Ranson, S. W. (1940). Anat. Rec. 78, 149. Hoebel, B. G. (1971). Annu. Rev. Physiol. 33, 533. Hoebel, B. G. (1975).In "Nebraska Symposium on Motivation: 1974" (J. K. Cole and T. B. Sondenegger, eds.), pp. 49-112. Univ. of Nebraska Press, Lincoln. Hoebel, B. G., and Teitelbaum, P. (1966). J. Comp. Physiol. Psychol. 61, 189. Hommel, H., Fischer, IL, Retzlaff, K., and Knöfler, H. (1972). Diabetologia 8, 111. Hustvedt, B. E., and Lovo, A. (1972). Acta. Physiol. Scand. 84, 29. Inoue, S., and Bray, G. A. (1977). Endocrinology 100, 108. Iversen, J. (1973). Diabetes 22, 381. Janowitz, H. D., Hanson, M. E., and Grossman, M. I. (1949). Am. J. Physiol. 156, 87. Kajinuma, H., Kaneto, A., Kuzuya, T., and Nakao, K. (1968). J. Clin. Invest. 28, 1384.

The CNS, Pancreatic Hormones, Feeding, and Obesity

309

Kaneto, A., Kajinuma, H., Kosaka, H., and Nakao, K. (1968). Endocrinology 85, 651. Kaneto, A., Kosaka, K., and Nakao, K. (1967b). Endocrinology 80, 530. Kaneto, A., Miki, E., and Kosaka, K. (1974). Endocrinology 95, 1005. Keesey, R. E., and Powley, T. L. (1975). Am. Sei. 63, 558. Kennedy, G. C. (1967). Jrc "Handbook of Physiology; Food and Water Intake" (C. F. Code, ed.), Vol. 1, Sect. 6. Am. Physiol. Soc, Washington, D. C. Kennedy, G. C., and Parker, R. A. (1963). Lancet 2, 981. Kern, H. F., and Grübe, D. (1972). Proc. Int. Cong. Endocrinol. 4th, 1972, pp. 224-228. Koerker, D. J., Ruch, W., Chideckel, E., Palmer, J., Goodner, C. J., Ensinck, J., and Gale, C. C. (1974). Science 184, 482. Kulkosky, P. J., Porte, D., Jr., and Woods, S. C. (1975). Experientia 3 1 , 123. Kurotsu, T., Tabayashi, C , and Ban, T. (1953). Med. J. Osaka Univ. 3, 529. Lebovitz, H. E., and Feldman, J. M. (1973).Fed. Proc, Fed. Am. Soc. Exp. Biol. 32,1797. Lefebvre, P. J., and Unger, R. H. (1972). "Glucagon: Molecular Physiology, Clinical and Therapeutic Implications." Pergamon, Oxford. LeMagnen, J. (1971). In "Handbook of Sensory Physiology, Chemical Senses I. Olfaction" (L. M. Beidler, ed.), Vol. 4, pp. 465-482. Springer-Verlag, Berlin and New York. Lichko, A. E. (1959). Pavlov J. Higher Nerv. Activ. 9, 731. Linquette, M., Fourlinnie, J. C , and Lagache, G. (1969). Ann. Endocrinol. 30, 96. Lotter, E. C , and Woods, S. C. (1977). Physiol. Behav. 18, 293. Luft, R., Efendic, S., Hokfelt, T., Johansson, O., and Arimura, A. (1974). Med. Biol. 52, 428. Lundquist, I. (1973). Proc. Scand. Soc. Study Diabetes 9, 18. MacKay, E. M., Callaway, J. W., and Barnes, R. H. (1940). J. Nutr. 20, 59. Maker, H. S., and Lehrer, G. M. (1972). In "Basic Neurochemistry" (R. W. Albers, G. J. Siegel, R. Katzman, and B. W. Agranoff, eds.), pp. 169-189, Little, Brown, Boston, Massachusetts. Malaisse, W. J. (1972). In "Handbook of Physiology, Endocrine Pancreas" (D. F. Steiner and N. Freinkel, eds.), Sec. 7, Vol. I, pp. 237-260. Am. Physiol. Soc, Washington, D. C. Malaisse, W. J., Malaisse-Lagae, F , Wright, P. H., and Ashmore, J . (1967'a). Endocrinology 80, 975. Malaisse, W. J., Malaisse-Lagae, F , and Mayhew, D. (1967b). J. Clin. Invest. 46, 1724. Margolis, R. U., and Altszuler, N. (1967). Nature (London) 215, 1375. Margolis, R. U., and Altszuler, N. (1968). Proc. Soc. Exp. Biol. Med. 127, 1122. Margules, D. L., and Olds, J. (1962). Science 135, 374. Marliss, E. B., Girardier, L., Seydoux, J., Wollheim, C. B., Kanazawa, Y., Orci, L., Renold, A. E., and Porte, D., J r . (1973). J. Clin. Invest. 52, 1246. Marshall, J. F., and Teitelbaum, P. (1973). Brain Res. 55, 229. Marshall, J. F , Richardson, J. S., and Teitelbaum, P. (1974). J. Comp. Physiol. Psychol. 87, 808. Martin, J. M., Konijnendijk, W., and Bouman, P. R. (1974).Diabetes 2 3 , 203. May, K. K., and Beaton, J. R. (1968). Proc. Soc. Exp. Biol. Med. 127, 1201. Mayer, J. (1953). New Engl. J. Med. 249, 13. Mayer, J., and Thomas, D. (1967). Science 156, 328. Meissner, H. P. (1976). J. Physiol. (Paris) 72, 757. Miller, R. E. (1970). Endocrinology 86, 642. Miselis, R., and Epstein, A. N. (1970). Physiologist 13, 262. Miselis, R., and Epstein, A. N. (1971). Am. Zool. 11, 624. Mogenson, G. J. (1976). In "Hunger: Basic Mechanisms and Clinical Implications" (D. Novin, W. Wyrwicka, and G. A. Bray, eds.), pp. 473-485. Raven, New York.

310

Stephen C. Woods and Daniel Porte, Jr.

Müller, E. E., Frohman, L. A., and Cocchi, D. (1973). Am. J. Physiol. 224, 1210. Myers, R. D., and Sharpe, L. G. (1968). Physiol. Behav. 3, 987. Nicolaides, S. (1969). Ann. N.Y. Acad. Sei. 157, 1176. Nicolaidis, S., and Meile, M. J. (1972). J. Physiol. (Paris) 65, 151A. Novin, D., VanderWeele, D. A., and Rezek, M. (1973). Science 181, 858. Olds, J. (1960). In "Electrical Studies on the Unanesthetized Brain" (E. R. Ramsey a n d D. S. O'Doherty, eds.), pp. 127-142. Harper (Hoeber), New York. Olds, J., and Milner, P. (1954). J. Comp. Physiol. Psychol. 47", 419. Olds, M. E., and Olds, J. (1963). J. Comp. Neurol. 120, 259. Oomura, Y. (1976), In "Hunger: Basic Mechanisms and Clinical Implications" (D. Novin ed.), pp. 145-157. Raven, New York. Orci, L., Perrelet, A., Ravazyola, M., Malaisse-Lagae, F., andRenold, A. E. (1973a).Eur. J. Clin. Invest. 3, 443. Orci, L., Unger, R. H., and Renold, A. E. (1973b). Experientia 29, 1015. Orci, L., Baitens, D., Dubois, M. P., and Rufener, C. (1975). Horm. Metab. Res. 7, 400. Owen, O. E., Reichard, G. A., Jr., DiTullio, N. W., and Shuman, C. R. (1970).Diabetes 19, 397. Paintal, A. S. (1973). Physiol. Rev. 53, 159. Panksepp, J. (1975). Psychol. Rev. 82, 170. Panksepp, J., and Nance, D. M. (1972). Physiol. Behav. 9, 447. Parra-Covarrubias, A., Rivera-Rodriguez, I., and Almarez-Ugalde, A. (1971). Diabetes 20, 800. Pfaffman, C , Norgren, R., and Grill, H. J. (1977). In "Food Intake and the Chemical Senses" (Y. Katsuki, ed.). Univ. of Tokyo Press, Tokyo. Polak, J. M., Pearse, A. G. E., Grimelius, L., Bloom, S. R., and Arimura, A. (1975).Lancet 1, 1220. Pollay, M. (1974). Fed. Proc, Fed. Am. Soc. Exp. Biol. 33, 2064. Porte, D., Jr. (1969). Arch, of Intern. Med. 123, 252. Porte, D., Jr., and Bagdade, J. D. (1970). Annu. Rev. Med. 2 1 , 219. Porte, D., Jr., and Robertson, R. P. (1973). Fed. Proc, Fed. Am. Soc. Exp. Biol. 32,1792. Porte, D., Jr., Girardier, L., Seydoux, J., Kanazawa, Y., and Posternak, J. (1973). J. Clin. Invest. 52, 210. Porte, D., Jr., Woods, S. C , Chen, M., Smith, P. H., and Ensinck, J. (1975). Pharm. Biochem. Behav. 3, 127. Porte, D., Jr., Smith, P. H., and Ensinck, J. (1976). Metabolism 25 (Suppl. 1), 1453. Powley, T. L. (1977). Psychol. Rev. 84, 89. Powley, T. L., and Keesey, R. E. (1970). J. Comp. Physiol. Psychol. 70, 25. Powley, T. L., and Opsahl, C. A. (1974). Am. J. Physiol. 226, 25. Powley, T. L., and Opsahl, C. A. (1976). In "Hunger: Basic Mechanisms and Clinical Implications" (D. Novin, W. Wyrwicka, and G. A. Bray, eds.), pp. 313-326. Raven, New York. Preshaw, R. M., Cooke, A. R., and Grossman, M. I. (1966). Gastroenterology 50, 171. Quaade, F., and Juhl, O. (1962). Am. J. Med. Sei. 243, 438. Rafaelson, O. J. (1967). Ada Med. Scand. Suppl. 476, 75. Rezek, M. (1976). Can. J. Physiol. Pharm. 54, 650. Robertson, R. P., and Porte, D., Jr. (1973). Diabetes 22, 1. Routtenberg, A. (1967). Science 157, 838. Russek, M., and Grinstein, S. (1974).In "Neurohumoral Coding of Brain Function" (R. D. Myers and R. R. Drucker-Colin, eds.), pp. 81-97. Plenum, New York. Saito, M., Murakami, E., Nishida, T., Fujisawa, Y , and Suda, M. (1975). J. Biochem. 78, 475.

The CNS, Pancreatic Hormones, Feeding, and Obesity

311

Saito, M., Murakami, E., Nishida, T., Fujisawa, Y., and Suda, M. (1976a). J. Biochem. 80, 563. Saito, M., Murakami, E., and Suda, M. (1976b). Biochim. Biophys. Acta 4 2 1 , 177. Sakata, K., Hayano, S., and Sloviter, H. A. (1963). Am. J. Physiol. 204, 1127. Sailer, C. F., and Strieker, E. M. (1976). Science 192, 385. Salter, J. M., and Best, C. H. (1953). Brit. Med. J. 2, 353. Scremin, O. U. (1965). J. Comp. Neurol. 139, 3 1 . Sem-Jacobsen, C. W. (1968). "Depth-Electrographic Stimulation of the H u m a n Brain and Behavior." Thomas, Springfield, Illinois. Sharma, K. N. (1967). In "Handbook of Physiology, Food and Water Intake" (C. F. Code, ed.), Sec. 6, Vol. I, pp. 225-237. Am. Physiol. Soc, Washington, D.C. Sharma, K. N., Anand, B. K., Dua, S., and Singh, B. (1961). Am. J. Physiol. 2 0 1 , 593. Shimazu, T., and Fukuda, A. (1965). Science 150, 1607. Shimazu, T., and Ogasawara, S. (1975). Am. J. Physiol. 228, 1788. Sims, E. A. H., Goldman, R. F , Gluck, C. M., Horton, E. S., Kelleher, P. C , and Rowe, D. W. (1968). Trans. Ass. Am. Physicians. 81, 153. Smith, G. P. (1976). In "Obesity in Perspective" (G. A. Bray, ed.). U.S. Govt. Printing Office, Washington, D.C. Smith, G. P., and Epstein, A. N. (1969). Am. J. Physiol. 217, 1083. Smith, M. H. (1966). J. Comp. Physiol. Psychol. 6 1 , 11. Smith, O. A., J r . (1961). In "Electrical Stimulation of the Brain" (D. E. Sheer, ed.), pp. 5 4 - 6 1 . Univ. of Texas Press, Austin. Smith, P. H., and Porte, D., J r . (1976). Ann. Rev. Pharm. Tox. 16, 269. Smyth, C. J., Lasichak, A. G., and Levy, S. (1947). J. Clin. Invest. 26, 439. Steffens, A. B. (1970). Physiol. Behav. 5, 147. Steffens, A. B., Mogenson, G. J., and Stevenson, J. A. F. (1972). Am. J. Physiol. 222, 1446. Steiner, D. F , and Freinkel, N., eds. (1972). "Handbook of Physiology, The Endocrine Pancreas," Sec. 7, Vol I. Am. Physiol. Soc, Washington, D.C. Stephan, F., Reville, P., Thierry, P., and Schlienger, J. L. (1972). Diabetologia 8, 196. Stul'nikov, B. V. (1973). Byull. Eksp. Biol. Med. 76, 3. Sudakov, K. V. (1966). Fed. Proc, Fed. Am. Soc. Exp. Biol. 25, T-43. Szabo, O., and Szabo, A. J. (1972). Am. J. Physiol. 223, 1349. Taborsky, G. J., Jr., and Bergman, R. N. (1976). Diabetes 25, 322. Tannenbaum, G. A., Paxinos, G., and Bindra, D. (1974). J. Comp. Physiol. Psychol. 86, 404. Teitelbaum, P., and Epstein, A. N. (1962). Psychol. Rev. 69, 74. Teitelbaum, P., and Wolgin, D. L. (1978). Prog. Brain Res. (In press.) Tonge, D. A., and Oatley, K. (1973). Physiol. Behav. 10, 497. Unger, R. H. (1974). Metabolism 2 3 , 581. Ungerstedt, U. (1970). A eta Physiol. Scand. 80, 35A. Ungerstedt, U. (1971). Acta Physiol. Scand. Suppl. 367, 95. Valenstein, E. S. (1973). "Brain Control." Wiley, New York. Van Itallie, T. B., Beaudoin, R., and Mayer, J. (1953). Am. J. Clin. Nutr. 1, 208. Wagner, J. W., and DeGroot, J. (1963). Am. J. Physiol. 204, 483. Weir, G. C , Knowton, S. D., and Martin, D. B. (1974). Endocrinology 95, 1744. Weiss, B., and Maickel, R. P. (1968). Int. J. Neuropharmacol. 7, 395. Woods, S. C. (1972). Am. J. Physiol. 223, 1424. Woods, S. C , and Kulkosky, P. J. (1976). Psychosom. Med. 38, 201. Woods, S. C , and Porte, D., J r . (1974). Physiol. Rev. 54, 596. Woods, S. C , and Porte, D., J r . (1975). Diabetes 24, 905.

312

Stephen C. Woods and Daniel Porte, Jr.

Woods, S. C , and Porte, D., J r . (1976). In "Hunger: Basic Mechanisms and Clinical Implications," pp. 273-280. (D. Novin ed.). Raven, New York. Woods, S. C , and Shogren, R. E., Jr. (1972). J. Comp. Physiol. Psychol. 8 1 , 220. Woods, S. C , Makous, W., and Hutton, R. A. (1969). J. Comp. Physiol. Psychol. 69, 301. Woods, S. C , Hutton, R. A., and Makous, W. (1970). Proc. Soc. Exp. Biol. Med. 133, 964. Woods, S. C , Alexander, K. R., and Porte, D., Jr. (1972). Endocrinology 90, 227. Woods, S. C , Chen, M., and Porte, D., Jr. (1974a). Diabetes 23, 341. Woods, S. C , Decke, E., and Vasselli, J. R. (1974b). Psychol. Rev. 81, 26. Woods, S. C , Kaestner, E., and Vasselli, J. R. (1975). Psychol. Rev. 82, 165. Woods, S. C , Vasselli, J. R., Kaestner, E., Szakmary, G. A., Milburn, P., and Vitiello, M. V. (1977). J. Comp. Physiol. Psychol. 91, 128. Yin, T. H., and Tsai, C. T. (1973). J. Comp. Physiol. Psychol. 85, 258. York, D. A., and Bray, G. A. (1972). Endocrinology 90, 885. Young, T. K., and Liu, A. C. (1965). Chung-Kuo ShengLi Hsueh Tsa Chih 19, 247.

GERALD M. REAVEN* and JERROLD M. OLEFSKY*

The Role of Insulin Resistance in the Pathogenesis of Diabetes Mellitus I. Introduction II. Measurement of Plasma Insulin Concentrations in Patients with Diabetes Mellitus A. Plasma Insulin Levels in Various Clinical Types of Diabetes Mellitus B. Rate of Appearance of Insulin in Plasma of Patients with Diabetes Mellitus C. Role of Obesity in the Plasma Insulin Response of Patients with Diabetes D. The "Appropriateness" of the Plasma Insulin Response in Diabetes Mellitus III. Insulin Action in Vivo A. Nonketotic Diabetics B. Ketotic Diabetics IV. Insulin Action in Vitro V. Summary References

313 316 316 318 319 320 322 322 325 327 327 329

I. Introduction Evidence that diabetes mellitus might be of pancreatic origin was furnished in 1889 by von Mering and Minkowski (1889-90), who demonstrated that total pancreatectomy in dogs was followed by hyperglycemia, glycosuria, ketosis, and death. This observation was soon followed by the epochal studies of Banting and Best (1922a,b), who indicated that the inevitable sequelae of total pancreatectomy * Department of Medicine, Stanford University School of Medicine and Veterans Administration Hospital, Palo Alto, California. 313

314

Gerald M. Reaven andjerrold M. Olefsky

could be controlled if dogs were treated with parenteral injections of pancreatic extract obtained following ligation of the pancreatic ducts. Although these observations did not prove that human diabetes is caused by a deficiency of pancreatic secretion, their impact has been such as to dominate thinking about the etiology of diabetes, and there has been general agreement, at least until quite recently, that human diabetes is due entirely to insulin deficiency. Given this apparent consensus, it is interesting to realize that within ten years of the discovery of insulin Himsworth (1932) pointed out "that in the diabetic patient insulin appears to vary in efficiency at different times." Based upon this clinical observation, he raised the possibility that there might be cases of diabetes that were not pancreatic in origin. Four years later he suggested that diabetes could be differentiated into "insulin sensitive" and "insulin insensitive" types on the basis of the blood glucose response of a subject to insulin administered immediately following an oral glucose load (Himsworth, 1936). Himsworth continued this line of investigation and in 1949 (Himsworth, 1949) published an extensive summary of the accumulated evidence in support of the notion that insulin insensitivity, and not insulin deficiency, was present in many patients with diabetes. Furthermore, it appeared to Himsworth that this differentiation of patients with diabetes into two groups corresponded to the two clinical forms of diabetes—patients who were insulin sensitive tended to be ketosis-prone, while the middle-aged, nonketotic diabetic tended to be insulin insensitive. At about the same time that Himsworth was proposing that diabetics should be separated into two groups on the basis of their insulin sensitivity, other workers were attempting to develop techniques that would permit direct assessment of plasma insulin levels, and in 1950 Bornstein and Trewhella (1950) published estimates of levels of plasma insulin activity in patients with diabetes. These measurements were made by injecting plasma samples into rats that had been alloxanized, adrenalectomized, and hypophysectomized. Using this approach, they found that untreated patients with diabetes could be divided into two groups—those with and those without plasma insulin activity. These studies were extended, and in 1951 Bornstein and Lawrence (1951) reported that plasma insulin activity was absent in diabetic patients with weight loss and ketosis, but that plasma insulin activity was essentially normal in overweight diabetics without ketosis. A similar difference in the plasma insulin activity of the two clinical forms of diabetes was also reported by Vallance-Owen and associates (1955) when insulin activity was estimated by measuring

Insulin Resistance in the Pathogenesis of Diabetes

315

insulin-stimulated glucose uptake following incubation of plasma from normal and diabetic subjects with rat diaphragm in vitro. Thus the view that human diabetes was not entirely due to insulin deficiency that had evolved from estimates of insulin action in the patient was now supported by direct estimates of circulating insulin levels. However, the various methods for estimating insulin activity were far from ideal; as experience with the different methods increased, the situation became less clear, and unanimity as to the levels of plasma insulin activity in the various forms of human diabetes was not attained (Rändle, 1957). This uncertainty as to levels of plasma insulin in patients with diabetes ended abruptly in 1960 when Yalow and Berson published their study of the immünoassay of endogenous plasma insulin in man (Yalow and Berson, 1960). In this monumental work they described an immunological method for measuring insulin that combined specificity with the degree of sensitivity needed to measure the minute concentrations of insulin present in the circulation. When they used this new method to compare plasma immunoreactive insulin levels in normal subjects to those of patients with nonketotic diabetes, they found that the insulin levels were on the average higher in the diabetic patients. On the basis of these results they concluded "that the tissues of the maturity-onset diabetic do not respond to his insulin as well as the tissues of the nondiabetic subject respond to his insulin." Or, to use Himsworth's terminology, patients with diabetes were "insulin insensitive." In the subsequent 15 years many measurements have been made of plasma immunoreactive insulin response to glucose in normal and diabetic subjects. Although there has been considerable refinement of the experimental approach, the initial observations of Yalow and Berson have been amply confirmed. However, their interpretation of the significance of these findings in the pathogenesis of diabetes mellitus has not been as readily accepted, and the role of resistance to the action of insulin in the etiology of diabetes remains a hotly debated issue. It is the purpose of this review to address this question directly, and we will attempt within it to evaluate critically the thesis that resistance to the action of endogenous insulin is an important factor in the pathogenesis of human diabetes. However, at the onset, we should identify both the scope and the limitations of our effort. Since we are interested in the etiology of human diabetes, we shall focus on the results of studies carried out in humans. In this regard, we shall be primarily concerned with relating the results of measurements of plasma insulin levels, and of insulin's biological action, to the issue being raised. In order to avoid

316

Gerald M. Reaven andjerrold M. Ole/sky

confusion, we will attempt, whenever possible to define precisely the degree of hyperglycemia present in the patients being studied. It is our opinion that considerable confusion has evolved from the absence of such distinction when any aspect of diabetes is under consideration. Thus, we will differentiate between ketotic and nonketotic diabetics, and in the latter category, between patients with and without significant fasting hyperglycemia (fasting plasma glucose >125 mg%). Finally, we would be less than candid if we did not explicitly state our own prejudice at the onset—we do not believe that primary insulin deficiency is the sole cause of what is currently defined as human diabetes mellitus. The acknowledgment of our bias does not ensure objectivity, but it at least serves as a warning.

II. Measurement of Plasma Insulin Concentrations in Patients with Diabetes Mellitus A. Plasma Insulin Levels in Various Clinical Types of Diabetes Mellitus With the advent of a specific immunoassay for measurement of plasma insulin (Yalow and Berson, 1960), considerable information has accumulated in the last 15 years regarding plasma insulin levels in patients with diabetes. As usual, as data accumulate following the introduction of a new experimental approach, their interpretation becomes more complicated. Thus, the original finding of Yalow and Berson of normal to elevated levels of plasma insulin in response to oral glucose in patients with nonketotic diabetes raised the possibility that diabetes might be a disease of insulin insensitivity. However, it soon became apparent that very low insulin concentrations were found in patients with ketotic diabetes and in nonketotic subjects with significant fasting hyperglycemia (Berson and Yalow, 1962; Hales and Rändle, 1963; Ehrlich and Bambers, 1964). This separation of patients with diabetes into two categories—those with and those without a normal immunoreactive insulin response to oral glucose challenge—is consistent with Himsworth's (1936) suggestion that diabetics are insulin insensitive or insulin sensitive, and is also compatible with the results of earlier attempts to estimate insulin activity by bioassay (Bornstein and Trewhella, 1950; Bornstein and Lawrence, 1951; Vallance-Owen et al., 1955). Unfortunately, upon closer inspection it is clear that insulin responses to oral glucose cannot be simply divided

Insulin Resistance in the Pathogenesis of Diabetes

317

into two categories. Thus, as oral glucose tolerance begins to deteriorate, absolute plasma insulin response to oral glucose is equal to or greater than normal (Yalow and Berson, 1960; Berson and Yalow, 1962, 1965; Hales and Rändle, 1963; Buchanan and McKiddie, 1967; Reaven and Miller, 1968; Chiles and Tzagournis, 1970). However, with increasing degrees of glucose intolerance and the appearance of significant fasting hyperglycemia, the plasma insulin response becomes attenuated, and with severe fasting hyperglycemia the insulin response is much less than in normal control subjects (Berson and Yalow, 1962; Hales and Rändle, 1963; Ehrlich and Bambers, 1964; Yalow et al., 1965; Buchanan and McKiddie, 1967; Reaven and Miller, 1968; Chiles and Tzagournis, 1970). Although it is difficult to define any sharp cut-off point, it is clear that substantial numbers of patients currently diagnosed as having diabetes are not insulin deficient. For example, one-third of the patients studied by the University Group Diabetes Program (1970) had fasting glucose levels less than 100 mg% and almost certainly had insulin responses to oral glucose that were comparable to those of normal subjects. At about the same time that it was becoming evident that many diabetics were not insulin deficient, proinsulin was described in an islet cell adenoma (Steiner and Oyer, 1967). It soon became clear that proinsulin was also present in the plasma (Rubenstein et al., 1968; Roth et al., 1968). Since proinsulin could be detected by insulin radioimmunoassay and since it has much less biological activity than insulin (Steiner et al., 1968; Shaw and Chance, 1968), the possibility arose that the increased plasma immunoreactive insulin levels seen in diabetics might actually be due to excessive secretion of proinsulin. This formulation would account for the apparent insulin resistance of patients with diabetes by attributing the combination of hyperglycemia and hyperinsulinemia to the pancreatic secretion of proinsulin. However, several studies have indicated that the high levels of IRI in patients with diabetes cannot be accounted for by proinsulin (Gordon and Roth, 1970; Melani et al., 1970; Duckworth and Kitabchi, 1972). Indeed, in a recent paper it has been suggested that proinsulin accounts for the greatest proportion of immunoreactive insulin in patients with severe hyperglycemia and hypoinsulinemia (Gordon et al., 1974). Thus, it still appears necessary to conclude that many patients with diabetes are hyperglycémie in spite of having plasma insulin levels that are equal to or greater than normal. The persistence of hyperglycemia under these conditions requires us to consider the possibility that insulin resistance exists in these patients.

318

Gerald M. Reaven andjerrold M. Olefsky

B. Rate of Appearance of Insulin in Plasma of Patients with Diabetes Mellitus Until now we have focused on the fact that the total mean integrated plasma insulin response to oral glucose of patients with nonketotic diabetes is frequently equal to or greater than normal. However, the first radioimmunoassay study of plasma insulin (Yalow and Berson, 1960) revealed another difference between normal and diabetic subjects. Thus, the plasma insulin response of normal subjects to a 100 gm oral glucose load was greatest between 30 and 60 minutes, and declined by 120 minutes. In contrast, plasma insulin levels in patients with nonketotic diabetes showed a lesser increase at 30 minutes, but the levels continued to rise throughout the two-hour interval. This difference in pattern of insulin response between normal and diabetic subjects has been frequently observed, and in a recent editorial it has been stated that "the commonest metabolic abnormality which has been revealed in diabetes mellitus or in people with a genetic predisposition to the development of diabetes is delayed insulin release in response to a glucose stimulus" (Plasma Insulin in Diabetes, 1970). Indeed, it has even been suggested that the hyperglycemia in patients with nonketotic diabetes is secondary to this observed delay in the time it takes for plasma insulin levels to peak (Seltzer et al., 1967). In this formulation the delay in reaching the peak insulin response is implicitly equated with decreased early insulin response in diabetic subjects. However, the fact that plasma insulin levels peak at a later time following oral glucose in patients with maturity onset diabetes does not necessarily mean that the insulin response was significantly lower at the earlier time intervals. And, if the insulin response is not lower at the earlier time intervals, the delay in the rate of appearance cannot be used as an explanation for the hyperglycemia. In the initial study of Yalow and Berson (1960), the mean plasma insulin levels were lower at 30 minutes in the patients with diabetes. However, the patients studied were relatively heterogenous in their degree of glucose intolerance: several of the subjects had fasting hyperglycemia and the mean fasting glucose level was 127 mg%. Since it is clear that insulin response decreases with severity of glucose intolerance (Berson and Yalow, 1962; Hales and Rändle, 1963; Ehrlich and Bambers, 1964; Yalow et al, 1965; Buchanan and McKiddie, 1967; Reaven and Miller, 1968), it is possible that the lower mean insulin level at 30 minutes in the diabetic patients is due to the presence in this group of patients with fasting hyperglycemia. Indeed, if plasma glucose and insulin response to oral glucose is compared in normal

319

Insulin Resistance in the Pathogenesis of Diabetes

subjects and patients with chemical diabetes, the absolute plasma insulin level in the diabetic patients is equal to or greater t h a n normal at every time interval during the tolerance test (Berson and Yalow, 1965; Chiles and Tzagournis, 1970; Reaven et al., 1971; Danowski et al., 1973; Tchobroutsky et al., 1973). Under these conditions, the fact t h a t peak insulin levels are reached later during the glucose tolerance test in the diabetic patients can hardly be used as an argument t h a t an absolute decrease in amount of insulin released during the early part of the test is responsible for the hyperglycemia. Furthermore, it has been shown t h a t glucose tolerance does not deteriorate when two oral glucose tests are performed three hours apart (Yalow et al., 1969), in spite of the fact t h a t a delayed insulin response is seen during the second test. Thus, the delay in time of the peak insulin response in patients with chemical diabetes seems most reasonably attributed to the persistent hyperglycemia in such patients, and the persistence of hyperglycemia in the presence of hyperinsulinemia serves as evidence for the presence of insulin resistance in these patients. It should be emphasized that the delay in the peak insulin response t h a t occurs in patients with significant fasting hyperglycemia is also associated with an absolute decrease in the magnitude of the insulin response throughout the tolerance test. Under these conditions, insulin deficiency seems to serve as a reasonable explanation for the hyperglycemia. C. R o l e of O b e s i t y in t h e P l a s m a Insulin R e s p o n s e P a t i e n t s with D i a b e t e s

of

Shortly after the introduction of the insulin radioimmunoassay, a relationship between obesity and hyperinsulinemia was noted (Karam et al., 1963; Bagdade et al., 1967). For example, K a r a m et al. pointed out t h a t an excessive insulin response to oral glucose was frequently seen in obese subjects (Karam et al., 1963). In view of the frequent association between obesity and nonketotic diabetes, a good deal of attention was then focused on the possible role of obesity in determining the insulin response of diabetic patient. Karam and associates extended their studies to include diabetic subjects, and indicated t h a t the plasma insulin response of nonketotic diabetics to oral glucose was less t h a n normal in thin patients and greater t h a n normal in obese diabetics (Karam et al., 1965). On the basis of these results, they concluded t h a t the plasma insulin response was lower in patients with diabetes as compared to normal subjects when matched for weight, and t h a t the

320

Gerald M. Reaven and J err old M. Olefsky

increased levels of plasma insulin that had been reported in diabetics were secondary to the presence of obesity in these patients. Unfortunately, although Karam et al. matched diabetics with controls on the basis of weight, they did not match their diabetic patients on the basis of their degree of hyperglycemia. Since their thin diabetic patients were significantly more hyperglycémie than the obese diabetics, their diminished insulin response could simply reflect the fact that patients with severe diabetes, thin or obese, secrete less insulin. Indeed, most studies indicate that when normal subjects and diabetic patients are matched for weight and degree of glucose intolerance, diabetic patients as a group have insulin levels equal to or greater than normal subjects until significant fasting hyperglycemia supervenes (Berson and Yalow, 1965; Yalow et al., 1965; Buchanan and McKiddie, 1967; Reaven and Miller, 1968; Chiles and Tzagournis, 1970). Furthermore, although obese subjects tend to have a greater insulin response to oral glucose than do nonobese individuals, not all obese subjects exhibit hyperinsulinism in response to oral glucose. There is considerable heterogeneity in the insulin response of obese subjects during an oral glucose tolerance test, and there is a marked overlap between the insulin response of obese and nonobese subjects with similar glucose tolerance to an oral glucose challenge (Berson and Yalow, 1965; Buchanan and McKiddie, 1967; Reaven and Miller, 1968). Finally, although loss of weight may lead to a marked decrease in the insulin response of obese subjects (Simsei al., 1968), a similar degree of reduction can also be seen when the degree of weight loss that is achieved still leaves the patient obese (Yalow et al., 1965; Olefsky et al., 1974). Thus, although obesity is a determinant of the insulin response to oral glucose, its impact should not be overestimated, and it is impossible to attribute the increased insulin response that is seen in many nonketotic diabetics to the coexistence of obesity. D. The "Appropriateness" of the Plasma Insulin Response in Diabetes Mellitus The presence of greater than normal levels of insulin in subjects with nonketotic diabetes initially led Yalow and Berson to point out that insulin deficiency did not exist in these patients and therefore some other factor must account for the hyperglycemia (Yalow and Berson, 1960). Theoretically, one could use these same data to defend the thesis that insulin deficiency does exist in these patients. Thus, it could be argued that the greater degree of hyperglycemia in diabetic subjects

Insulin Resistance in the Pathogenesis of Diabetes

321

should elicit an even greater insulin response than it does, and that the absence of this "appropriate" response indicates that the diabetic state is associated with reduced insulin secretory capacity. The manner of calculating the appropriateness of the insulin response had varied from such simple manipulations as determining the insulin/glucose ratio (Perley and Kipnis, 1966) or the ratio between the incremental changes in plasma glucose and insulin levels (Seltzer et al., 1967), to much more complicated analysis using an analog computer model (Cerasi, 1967). These approaches have in common two features. First, they focus on what the insulin response should be rather than what it actually is. Second, they are based on the assumption that the plasma insulin response to an oral glucose load in normal subjects is a linear function of the coexisting plasma glucose response. Unfortunately, there seems to be little experimental data in support of this assumption, and considerable evidence that indicates that plasma insulin levels of a normal individual increase in response to an increase in oral glucose load without a concomitant rise in the plasma glucose response (Hales and Rändle, 1963; Castro et al., 1970; Peterson and Reaven, 1971; Reaven and Olefsky, 1974). Other attempts to define the appropriateness of the pancreatic response in diabetics have been based upon determining the plasma insulin response under conditions in which the glycémie stimulus was supposedly equalized in normal and diabetic subjects. This condition was attained by producing "normal" and "diabetic" blood glucose profiles by infusing glucose to simulate the mean response to a 100 gm oral glucose load (Perley and Kipnis, 1967). When this was done, the insulin response was lower in the diabetic subjects, and it was concluded that pancreatic insufficiency was present in these patients. However, it must be again emphasized that this argument is based upon the assumption that insulin response is a linear function of coexisting glucose concentration, and that both groups were responding to an equivalent glycémie stimulus. But were both groups responding to the same "glycémie stimulus"? In order to achieve similar glucose levels in the two groups of patients it was necessary to infuse more than twice as much glucose into the normal subjects. It seems theoretically possible that the amount of glucose infused may also affect the intensity of the insulin response, and in these terms the two groups of subjects were not necessarily responding to identical glycémie stimuli. Indeed, recent experimental data strongly support this notion, and demonstrate that insulin responses are more closely correlated to the amount of glucose infused than to glucose concentration and that plasma insulin levels rise throughout a prolonged glucose infusion,

322

Gerald M. Reaven andjerrold M. Olefsky

even when the plasma glucose levels are falling during the latter part of the infusion (Olefsky et al., 1973a). If control of plasma glucose concentration is considered to be a function of insulin secretion alone, it is obvious that hyperglycemia is always due to an insufficient insulin response. However, if this thesis is being questioned, it is not sufficient to indicate that arbitrary transformations of experimental data yield ratios that prove that absolute elevations of plasma insulin are in reality indications of insulin deficiency. The view that the insulin response should be related to the inciting glycémie stimulus appears to be a reasonable one. However, there is another side to the coin, and that relates to the blood glucose response to insulin. In this regard, we are not prepared to accept a lower insulinogenic index as proof that reduced insulin secretion is the cause of impaired glucose tolerance. Nor are we prepared to ignore the fact that similar loads of oral glucose elicit a greater insulin response but are disposed of less quickly in diabetic subjects.

III. Insulin Action in Vivo A. Nonketotic Diabetics Considerable in vivo evidence exists supporting the idea that nonketotic maturity onset diabetic patients are resistant to the action of insulin. Historically, an initial experimental approach was to administer a glucose load to normal and diabetic subjects and to measure the efficiency with which the glucose was removed from plasma (Forbath and Hetenyi, 1966; Butterfield et al., 1967; Whichelow and Butterfield, 1971). These studies either involved direct measurement of glucose uptake via the forearm perfusion technique (Butterfield et al., 1967; Whichelow and Butterfield, 1971) or utilized isotopic tracers to measure glucose turnover (Forbath and Hetenyi, 1966). Studies of this nature have uniformly found that the efficiency of glucose disposal is decreased in nonketotic diabetic subjects with fasting hyperglycemia. However, since plasma insulin levels were not measured, it is difficult to determine whether this decreased glucose uptake is a result of inadequate amounts of circulating insulin or is related to insulin resistance. Since circulating insulin levels in the stimulated state are often found to be subnormal in the nonketotic diabetic patient with fasting hyperglycemia (Berson and Yalow, 1962; Hales and Rändle, 1963; Ehrlich and Bambers, 1964; Yalow et al., 1965; Buchanan and McKid-

Insulin Resistance in the Pathogenesis of Diabetes

323

die, 1967; Reaven and Miller, 1968; Chiles and Tzagournis, 1970), the idea can be proposed that the decreased glucose disposal rates were secondary to inadequate insulin levels. In order to measure insulin responsiveness more directly, a number of studies have been performed in which glucose disposal was measured under conditions in which the insulin stimulus was controlled (Beam etal., 1951; Heller etal., 1958; Kalant etal., 1963; Zierler and Rabinowitz, 1963; Stocks and Martin, 1969; Alford et al., 1971). Basically, these approaches involve the administration of comparable amounts of insulin to normal and diabetic patients followed by the measurement of some aspect of glucose removal. The findings of these latter studies have been consistent with the concept that nonketotic diabetic patients are insulin resistant. For example, Zierler and Rabinowitz (1963), using the forearm perfusion technique, found that nonketotic diabetic patients with fasting hyperglycemia take up less glucose in the basal state than do normals. Furthermore, when they administered equal amounts of insulin to both groups and then measured glucose extraction, they found that forearm muscles of diabetic patients had decreased insulin responsiveness as compared to normal. They also observed an inverse relationship between the severity of diabetes and the ability of forearm tissues to increase glucose uptake in response to insulin. This abnormality could not be corrected by prior insulin administration (up to six weeks), indicating that insulin unresponsiveness was not simply a function of preexisting insulin lack. They concluded that maturity onset diabetic patients are resistant to insulin. Using a different approach, Alford et al. (1971) administered an intravenous glucose load followed by insulin (0.1 units/kg) to normal patients, patients with chemical diabetes, and nonketotic patients with fasting hyperglycemia. By measuring the rate of fall in glucose concentration following the administration of insulin they found that the effectiveness of insulin decreased as diabetes progressed, i.e., patients with fasting hyperglycemia had the poorest response to insulin, the group with chemical diabetes had an intermediate response to insulin, and the normals had the best response to insulin. Stocks and Martin (1969) also measured the glucose concentration following the administration of insulin (0.1 units/kg) in normal and nonketotic diabetic subjects with fasting hyperglycemia and calculated a glucose assimilation index from their data as a measurement of insulin responsiveness. They found that the glucose assimilation index was significantly greater in the normal subjects and concluded that nonketotic diabetic subjects were resistant to the action of exogenous insulin. Kalant et al. (1963) administered insulin to normal and nonketotic

324

Gerald M. Reaven andjerrold M. Olefsky

diabetic patients 1 and measured subsequent plasma glucose concentrations and glucose outflow using an isotopic technique. In the diabetic subjects they found an attenuated hypoglycémie response to the administered insulin and were able to correlate this attenuated plasma glucose response to a diminished rate of glucose outflow in response to insulin. All of the above studies were conducted when the plasma glucose and insulin concentrations were in non-steady-state conditions. If insulin is working inefficiently in these patients, it is difficult to quantitate this phenomenon while the rates of plasma insulin entry and removal and glucose entry and removal are rapidly changing. In order to circumvent this problem we have designed an experimental approach in which we can directly measure the ability of different groups of patients to dispose of comparable glucose loads under the influence of identical insulin stimuli during steady-state conditions (Shen et al., 1970). To achieve this, we simultaneously infuse constant amounts of insulin, glucose, epinephrine, and propranolol for 150 minutes. Steady-state plasma concentrations of insulin and glucose are achieved within 90 minutes after the start of the infusion, and are measured every 10 minutes during the final 60 minutes of the study. This approach is based upon the known ability of epinephrine and propranolol to suppress endogenous insulin secretion, and thus comparable steady-state plasma levels of exogenous insulin are achieved in all subjects [(102 μϋ/πύ; coefficient of variation, 10% (Olefsky et al., 1973b)]. Thus, we are able to measure the ability of similar circulating levels of exogenous insulin to promote the disposal of comparable glucose loads in a variety of subjects. In these studies the height of the steady-state plasma glucose concentration (SSPG) is a direct reflection of a subject's overall efficiency of insulin-mediated glucose disposal. Using this technique we have found that SSPGs are elevated in patients with chemical diabetes as compared to normals (220 ± 20 vs 120 ± 10 mg%) (Ginsberg et al., 1974). Diabetic patients with fasting hyperglycemia have even greater elevations of SSPG levels during these infusions (380 ± 40 mg%) (Ginsberg et al., 1975). Thus, these data provide direct evidence for a graded increase in the level of insulin resistance as the severity of hyperglycemia increases in nonketotic diabetic subjects. However, many patients with nonketotic diabetes are overweight, and the relationship between obesity and insulin resistance has been well described (Rabinowitz and Zierler, 1962). Consequently, one must ask what the relationship is between obesity and 1

In the report by Kalant et al. (1963) the subjects are simply described as "of the maturity-onset type beginning after the age of 40."

Insulin Resistance in the Pathogenesis of Diabetes

325

the insulin resistance in the diabetic patients described in the above studies. In most of the studies (Zierler and Rabinowitz, 1963; Forbath and Hetenyi, 1966; Butterfield et al, 1967; Stocks and Martin, 1969; Shen et al., 1970; Whichelow and Butterfield, 1971; Alford et al., 1971; Olefsky et al., 1973; Ginsberg et al., 1974, 1975) control and diabetic groups were matched for weight, and thus the difference in insulin resistance noted between diabetics and normals must be independent of any effect of obesity. Furthermore, in our own studies (Shen et al., 1970; Olefsky et al., 1973b), we included only subjects who were nonobese, and within the range of percent adiposity of our nonobese subjects no relationship was found between insulin resistance and percentage of adiposity. This latter finding suggests that significant levels of obesity must be reached before adiposity has an appreciable influence on insulin resistance. Thus, while obesity can promote insulin resistance, the insulin resistance described in chemical and nonketotic maturity onset diabetic subjects is clearly independent of this factor.

B. Ketotic Diabetics When applied to ketotic diabetics the term insulin resistance usually denotes a patient who requires large amounts of insulin (>200 U/day) for control (Field, 1962). These patients are unusual and often have high titers of insulin antibodies to explain their insulin resistance (Williams and Porte, 1968). A few patients have been described who do not have insulin antibodies and yet need thousands of units of insulin a day to avoid ketosis (Field, 1962). However, the great majority of ketotic diabetic patients do not require more than 100 U/day for control, and the question must be asked whether these subjects respond normally to insulin. No clear-cut answer emerges from the literature. One reason for this is that ketotic diabetics are hardly a homogenous population, since a subject's insulin sensitivity is at least partially dependent on the degree of diabetic control, with the resulting changes in metabolic milieu (Rändle et al., 1963; Walker et al., 1963). Intrinsic heterogeneity of this population may also exist. For example, Martin and Stocks (1967) studied insulin sensitivity in a group of 43 ketotic diabetic patients by calculating the rate of fall of plasma glucose concentration following intravenous insulin (0.1 U/kg). These workers found that 50% of the subjects were insulin insensitive as compared to controls. Furthermore, they found no relationship between insulin sensitivity and anti-insulin antibody titers, age, degree of obesity, dura-

326

Gerald M. Reaven and ]err old M. Olefsky

tion of diabetes, or insulin requirements. They did find a striking relationship between the insulin insensitivity and microangiopathic and/or atherosclerotic vascular disease (Martin and Stocks, 1968). In a subsequent study (Martin and Pearson, 1971) these same workers found, as Bondy et al. (1949) had previously noted, t h a t in these patients the hypoglycémie response to insulin was closely related to insulin's ability to decrease net hepatic glucose production and was not related to stimulation of extrahepatic glucose uptake. These findings suggest t h a t hepatic insensitivity to insulin is an underlying mechanism in those ketotic diabetic patients who are insulin resistant. On the other hand, Butterfield et al. (1967) and Zampa et al. (1967), studying insulin sensitivity of the perfused forearm, found t h a t ketotic diabetic patients take up less glucose then normal, suggesting peripheral insensitivity to insulin. Thus ketotic diabetic patients appear to be a heterogenous group with some subjects manifesting insulin resistance—at least during some stage of their diabetic syndrome. In those subjects who do demonstrate insulin resistance, data consistent with a relative inability of insulin to inhibit hepatic glucose production (Martin and Pearson, 1971; Bondy et al., 1949) and to promote glucose uptake by extra hepatic tissues have been reported (Zampa et al., 1967). Another much discussed manifestation of the insulin insensitivity of ketotic diabetic patients is the large amounts of insulin t h a t have been thought necessary to correct ketoacidosis. This long-held belief (Bondy, 1969) has been regularly challenged, and a series of recent studies have rather conclusively shown t h a t nearly all diabetic subjects in ketoacidosis—regardless of severity—can be safely managed on relatively "small" doses of insulin (Sonksen et al., 1972; Page et al., 1974; Semple et al., 1974; Kidson et al., 1974). It is difficult, however, to use these studies to prove t h a t ketotic diabetic patients are not insulin resistant. Although the insulin doses used were small relative to the usual therapeutic doses, they still exceeded the average daily pancreatic output and, based on published information in normals, certainly resulted in "elevated" plasma insulin levels. Thus subjects being treated with this regimen are responding to elevated plasma insulin levels. Furthermore, for obvious reasons, control studies were not done in which normal subjects were made hyperglycémie and ketotic and treated with similar insulin regimens. Consequently we do not know how a normal (but hyperglycémie, ketotic) subject would respond to "low dose" insulin therapy. Thus, these types of observations, while useful in formulating a rational therapeutic approach to the ketoacidotic diabetic, do not tell us if they are normally insulin sensitive.

Insulin Resistance in the Pathogenesis of Diabetes

327

IV. Insulin Action in Vitro Information concerning the in vitro glucose metabolism of tissues from diabetic patients and the responsiveness of these tissues to insulin is limited. However, the available studies do suggest that diabetic tissue does not behave normally. For example, Gal ton et al. (1971) studied glucose metabolism in adipose tissue fragments from 30 diabetic subjects (only two of which required insulin therapy). They found that when diabetic and normal subjects were matched for weight, glucose utilization by adipose tissue from diabetic subjects was decreased at all media glucose concentrations. Studies of insulin responsiveness were not performed. However, Kahlenberg and Kalant (1964) studied glucose consumption by adipose tissue from normal and nonketotic diabetic subjects in both the basal and insulin-stimulated states. They found that adipose tissue from diabetic patients demonstrated decreased glucose uptake in both the basal and insulin-stimulated states. This decreased insulin responsiveness was apparent if the data were expressed in either absolute terms or as a percentage of stimulation above basal. In similar studies, Bjorntorp (1966) measured the ability of insulin to promote glucose oxidation in adipose tissue isolated from normal or nonketotic diabetic subjects (fasting blood sugar > 150 mg/100 ml). He too was able to show that adipose tissue from the diabetic subjects demonstrated decreased insulin responsiveness as compared to normal. Finally, Kalant and Schucher (1962) were able to demonstrate decreased insulin responsiveness using leukocytes obtained from nonketotic diabetic subjects. Clearly these studies are consistent with the hypothesis that tissues from nonketotic diabetic patients are resistant to the action of insulin.

V. Summary In the previous sections of this review, we have summarized a considerable amount of evidence that implicates resistance to the action of insulin as an important factor in the pathogenesis of diabetes in man. The situation is perhaps clearest in patients classified as having chemical diabetes. These individuals are not insulin deficient—their plasma insulin response to an oral glucose load is in absolute terms at least equal to that of normal individuals at all time intervals measured. Furthermore, insulin is not as effective in promoting glucose uptake in these patients as in normal subjects. Thus, the hypothesis that

328

Gerald M. Reaven and ]errold M. Olefsky

hyperglycemia in spite of hyperinsulinemia means that the tissues of such diabetics do not respond normally to their insulin is supported by direct estimates of insulin's action in vivo. Indeed, in these patients it is possible to attribute carbohydrate intolerance entirely to insulin resistance, and there seems to be no reason to postulate the concomitant presence of insulin deficiency. Unfortunately, the situation is more cloudy in the case of patients who have nonketotic diabetes, but who also have significant fasting hyperglycemia. These patients appear to be both insulin deficient and insulin resistant. Thus, as we have described earlier, they respond to an oral glucose load with an insulin response that is less than normal and they are resistant to the in vivo and in vitro action of insulin. These individuals appear to have two explanations for their hyperglycemia. Although both abnormalities may exist, and may both be primary lesions, the law of parsimony would be served if one could be shown to be secondary to the other. For example it could be postulated that the basic lesion in patients with significant fasting hyperglycemia is an absolute deficiency of insulin, resulting in a fall in glucose uptake rate. In an attempt to promote glucose transport, counterregulatory hormones are secreted, gluconeogenesis is increased, and significant fasting hyperglycemia ensues. The hyperglycemia stimulates the pancreas, resulting in fasting insulin levels, which are now equal to normal, and the combination of normal insulin levels and fasting hyperglycemia promotes glucose uptake and inhibits ketone body formation. Under these conditions, i.e., those of accelerated gluconeogenesis, the subject is resistant to the action of insulin, but this insulin resistance is secondary to insulin deficiency. If this formulation is correct, all patients with insulin deficiency should become insulin resistant. We tested this hypothesis by directly estimating insulin resistance in patients with insulin-deficient diabetes secondary to chronic pancreatitis (Ginsberg et al., 1975). The mean (±SE) steadystate plasma glucose response in five such subjects [mean (±SE) fasting plasma glucose = 142 ± 12 mg%] was 104 ± 17 mg%, as compared to 296 ± 23 mg% in six patients with idiopathic diabetes and a comparable degree of fasting hyperglycemia (154 ± 12 mg%). Thus, insulin deficiency in patients with moderate fasting hyperglycemia does not necessarily lead to insulin resistance, and it appears from the data that insulin resistance must be considered as primary a defect as insulin deficiency in these subjects. Furthermore, Rabinowitz and Zierler (1962) gave insulin to nonketotic diabetic subjects with fasting hyperglycemia for up to six weeks and could not detect any amelioration of the insulin resistance as measured by the forearm perfusion technique. These results provide further evidence that insulin resis-

Insulin Resistance in the Pathogenesis of Diabetes

329

tance is not simply secondary to insulin deficiency in nonketotic diabetic patients with fasting hyperglycemia. The situation is most obscure in patients with ketotic diabetes. There is no doubt that insulin deficiency exists in these subjects, but there is less clarity as to the presence of insulin resistance. It has generally been assumed that uncontrolled ketotic diabetics were severely insulin resistant, and this belief has served as the cornerstone for the use of large amounts of insulin in the treatment of this situation. However, as described earlier, the recent successful use of lower-dose insulin therapy in the treatment of diabetic ketoacidosis has certainly served to question that assumption. On the other hand, we are aware of no studies in which the insulin resistance of ketotic diabetics has been directly compared to control subjects, and it is possible that such patients could require a great deal less insulin than previously believed and still be insulin resistant as compared to normal subjects. Finally, even if ketotic diabetics were found to be insulin resistant, the more severe degree of insulin deficiency in such subjects could lead to the development of the insulin resistance. Thus, our ability to ascribe the hyperglycemia of patients with diabetes to either insulin resistance or insulin deficiency becomes more difficult as the severity of the hyperglycemia increases, and more information will be required before we can sort out the relative contributions of both lesions in various groups of patients with diabetes. In conclusion, we feel that the evidence summarized in this review makes it clear that insulin resistance exists in diabetics, and that hyperglycemia in these patients can no longer be considered as being solely due to insulin lack. Obviously, the cause of the insulin resistance could theoretically be due to secretion of abnormal forms of insulin, circulating insulin antagonists, or cellular unresponsiveness to the action of insulin. These issues are currently the focus of extensive investigation in many laboratories, but we have considered this subject to be outside the purview of our review. As the results of these studies become available we will be better able to place in proper perspective the relative roles of insulin lack and insulin resistance in the development of the diabetic syndrome. References Alford, F. P., Martin, F. I. R., and Pearson, M. J. (1971). Diabetologia 7, 173. Bagdade, J. D., Bierman, E. L., Porte, D., Jr. (1967). J. Clin. Invest. 46, 1549. Banting, F. G., and Best, C. H. (1922a). J. Lab. Clin. Med. 7, 251. Banting, F. G., and Best, C. H. (1922b). J. Lab. Clin. Med. 7, 464.

330

Gerald M. Reaven andjerrold M. Olefsky

Bearn, A. G., Billing, B. H., and Sherlock, S. (1951). Lancet 2, 698. Berson, S. A., and Yalow, R. S. (1962). In "Immunoassay of Hormones, Ciba Foundation Colloquium on Endocrinology" (G. E. W. Wolstenholme and M. P. Cameron, eds.), Publ. No. 14, p. 182. Berson, S. A., and Yalow, R. S. (1965). Diabetes 14, 549. Bjorntorp, P. (1966). Acta Med. Scand. 179, 229. Bondy, P. K. (1969). In "Duncan's Diseases of Metabolism" (P. K. Bondy, ed.), Chapter 6, p. 253. Bondy, P. K., Bloom, W. L., Whitner, V. S., and Farrar, B. W. (1949). J. Clin. Invest. 28, 1126. Bornstein, J., and Trewhella, P. (1950). Aust. J. Exp. Biol. Med. Sei. 28, 569. Bornstein, J., and Lawrence, D. D. (1951). Brit. Med. J. 2, 1541. Buchanan, K. D., and McKiddie, M. T. (1967). Diabetes 16, 466. Butterfield, W. J. H., Abrams, M. E., St. John, D. J. B., and Whichelow M. J. (1967). Metabolism 16, 19. Castro, A., Scott, J. P., Grettie, D. P., McFarlane, D., and Bailey, R. E. (1970).Diabetes 19, 842. Cerasi, E. (1967). Acta Endocrinol. (Copenhagen) 55, 163. Chiles, R., and Tzagournis, M. (1970). Diabetes 19, 458. Danowski, T. S., Khurana, R., Nolan, S., Stephan, T., Gegick, C , Chae, S., and Vidalan, C. (1973). Diabetes 22, 808. Duckworth, W. C , and Kitabchi, A. E. (1972). Am. J. Med. 53, 418. Ehrlich, R. M., and Bambers, G. (1964). Diabetes 13, 177. Field, J. B. (1962). Annu. Rev. Med. 13, 249. Forbath, N., and Hetenyi, G., Jr. (1966). Diabetes 15, 778. Galton, D. J., Wilson, J. P. D., and Kissebah, A. H. (1971). Eur. J. Clin. Invest. 1, 399. Ginsberg, H., Olefsky, J. M., and Reaven, G. M. (1974). Diabetes 2 3 , 674. Ginsberg, H., Kimmerling, G., Olefsky, J. M., and Reaven, G. M. (1975). J. Clin. Invest. 55, 454. Gorden, P., and Roth, J. (1970). J. Clin. Invest. 48, 2225. Gorden, P., Hendricks, C. M., and Roth, J. (1974). Diabetologia 10, 469. Hales, C. N., and Rändle, P. J. (1963). Lancet 1, 790. Heller, N., Kalant, N., and Hoffman, M. M. (1958). J. Lab. Clin. Med. 52, 394. Himsworth, H. (1932). Lancet 2, 935. Himsworth, H. (1936). Lancet 1, 127. Himsworth, H. (1949). Lancet 1, 465. Kahlenberg, A., and Kalant, N. (1964). Can. J. Biochem. 42, 1632. Kalant, N., and Schucher, R. (1962). Can. J. Biochem. Physiol. 40, 899. Kalant, N., Csorba, T. R., and Heller, N. (1963). Metabolism 12, 1100. Karam, J. H., Grodsky, G. M., and Forsham, P. H. (1963). Diabetes 12, 297. Karam, J. H., Pavlatos, F C , Grodsky, G. M., and Forsham, P. H. (1965). Lancet 1,286. Kidson, W., Casey, J., Kraegen, E., and Lazarus, L. (1974). Brit. Med. J. 2, 691. Martin, F. I. R., and Stocks, A. E. (1967). Aust. Ann. Med. 16, 298. Martin, F. I. R., and Stocks, A. E. (1968). Brit. Med. J. 2, 81. Martin, F. I. R., and Pearson, M. J. (1971). Metabolism 20, 859. Melani, F , Rubenstein, A., and Steiner, D. (1970). J. Clin. Invest. 49, 497. Olefsky, J. M., Bachelder, T., Farquhar, J. W., and Reaven, G. M. (1973a). Metabolism 22, 1277. Olefsky, J. M., Farquhar, J. W., and Reaven, G. M. (1973b). Diabetes 22, 507. Olefsky, J. M., Reaven, G. M., and Farquhar, J. W. (1974). J. Clin. Invest. 53, 64.

Insulin Resistance in the Pathogenesis of Diabetes

331

Page, M. McB., Alberti, K. G. M. M., Greenwood, R., Gumaa, K. A., Hockaday, T. D. R., Lowy, C , Nabarro, J. D. N., Pyke, D. A., Sonksen, P. H., Watskins, P. J., and West, T. E. T. (1974). Brit. Med. J. 2, 687. Perley, M., and Kipnis, D. M. (1966). Diabetes 15, 867. Perley, M., and Kipnis, D. M. (1967). J . Clin. Invest. 46, 1954. Peterson, D. T., and Reaven, G. M. (1971). Diabetes 20, 729. Plasma Insulin in Diabetes (1970). Lancet 1, 1211. Rabinowitz, D., and Zierler, K. L. (1962). J. Clin. Invest. 4 1 , 2173. Rändle, P. J. (1957). In "Ciba Foundation Colloquim on Endocrinology" (G. E. W. Wolstenholme and M. P. Cameron, eds.), Publ. No. 11, p. 115. Rändle, P. J., Garland, P. B., Hales, C. N., and Newsholme, E. A. (1963). Lancet 1, 785. Reaven, G., and Miller, R. (1968). Diabetes 17, 560. Reaven, G. M., and Olefsky, J. M. (1974). J. of Clin. Endocrinol. Metab. 38, 151. Reaven, G. M., Shen, S. W., Silvers, A., and Farquhar, J. W. (1971). Diabetes 20, 416. Roth, J., Gorden, P., and Pastan, I. (1968). Proc. Natl. Acad. Sei. U.S.A. 6 1 , 138. Rubenstein, A., Cho, S., and Steiner, D. (1968). Lancet 1, 1353. Seltzer, H. S., Allen, E. W., Herron, A. L., Jr., and Brennan, M. T. (1967). J . Clin. Invest. 46, 323. Semple, P. R., White, C., and Manderson, W. G. (1974). Brit. Med. J. 2, 694. Shaw, W. N., and Chance, R. E. (1968). Diabetes 17, 737. Shen, S. W., Reaven, G. M., and Farquhar, J. W. (1970). J. Clin. Invest. 49, 1970. Sims, E. A. H., Goldman, R. F., Gluck, C. M., Horton, E. S., Kelleher, P. C , and Rowe, D. W. (1968). Trans. Ass. Am. Physicians 8 1 , 153. Sonksen, P. H., Srivastava, M. C , Tompkins, C , and Nabarro, J. D. N. (1972). Lancet 2, 155. Steiner, D. F , and Oyer, P. (1967). Proc. Natl. Acad. Sei. U.S.A. 57, 473. Steiner, D. F , Hallund, O., Rubenstein, A., Cho, S., and Bayliss, C. (1968). Diabetes 17, 725. Stocks, A. E., and Martin, F. I. R. (1969). Brit. Med. J. 4, 397. Tchobroutsky, G., Kopf, A., Eschwege, E., and Assan, R. (1973). Diabetes 22, 825. University Group Diabetes Program (1970). Diabetes 19, 747. Vallance-Owen, J., Hurlock, B., and Pease, N. W. (1955). Lancet 2, 583. von Mering, J., and Minkowski, O. (1889-1890). Arch. Exp.Pathol.Pharmakol. 26,371. Walker, B. G., Martin, F I. R., and Baird, C. W. (1963). Lancet 2, 964. Whichelow, M. J., and Butterfield, W. J. H. (1971). Quart. J. Med. 40, 261. Williams, R. H., and Porte, D., Jr. (1968). In "Textbook of Endocrinology" (R. H. Williams, ed.), 4th ed., Chapter 9, p. 592. Yalow, R. S., and Berson, S. A. (1960). J. Clin. Invest. 39, 1157. Yalow, R. S., Glick, S. M., Roth, J., and Berson, S. A. (1965). Ann. N.Y. Acad. Sei. 131, 357. Yalow, R. S., Goldsmith, S. H., and Berson, S. A. (1969). Diabetes 18, 402. Zampa, G. A., Bracchette, D., Geminiani, G. D., and Borgetti, E. (1967). Diabetologia 3, 35. Zierler, K., and Rabinowitz, D. (1963). Medicine 42, 385.

BRUNO W. VOLK*t and KLAUS F. WELLMANN* ΐ

The Pancreas in Idiopathic Diabetes§ I. II. III. IV. V. VI. VII. VIII. IX. X. XI. XII. XIII. XIV. XV XVI. XVII.

Introduction Gross Pancreatic Alterations Hyalinization of the Pancreatic Islets Vascular Lesions Pancreatic Fibrosis Fatty Atrophy Vacuolization ("Hydropic Degeneration") of the B Cells Hypertrophy of the Islets Inflammatory Lesions of the Islets Regeneration and Ductal Proliferation Fat Deposits in B Cells Degranulation of B Cells Quantitative Changes of the Islets of Langerhans Estimation of Islet Volume Differential Counts of the Islets of Langerhans Quantitative Estimation of Islet Tissue Pathogenetic Considerations of Idiopathic Diabetes References

334 336 337 341 344 347 349 351 352 353 353 354 355 355 356 358 359 361

* Isaac Albert Research Institute of the Kingsbrook Jewish Medical Center and Department of Pathology, Downstate Medical Center, State University of New York, Brooklyn, New York. f Present address: Department of Pathology, University of California, Irvine, California. $ Present address: Department of Pathology, Beekman Downtown Hospital, New York, New York. § The subject of this article is discussed more extensively in Volk and Wellmann (1977). 333

334

Bruno W. Volk and Klaus F. Wellmann

I. Introduction The etiology of diabetes is heterogeneous. Its manifestations may be produced by a number of factors only some of which are known. The disease may have its roots in extrapancreatic alterations such as hyperthyroidism, cortical or medullary adrenal disease, or hyperplasia or tumors of the anterior pituitary gland. In other cases it may be a sequel of afflictions of the pancreas that destroy or considerably diminish the number of its islets, for instance, hemochromatosis, acute or chronic pancreatitis, or pancreatic neoplasia; or it may follow the surgical removal of this organ. In most diabetic patients, however, the etiology of the disease remains obscure, and it is to this large group of cases that the term "idiopathic diabetes" has been applied. Depending upon the age of onset and on the clinical manifestations, two types of diabetes are generally distinguished: the juvenile, or growth-onset, and the adult, or maturity-onset (Table I). Population surveys indicate that in the United States alone approximately ten million people are afflicted with diabetes (1); yet, in about two-fifths of these cases the disease is unrecognized because it is oligo- or asymptomatic, because medical care is inadequate, or because of individual neglect (Marble, 1971). According to the National Commission on Diabetes, 35,000 deaths per annum are directly attributable to diabetes, but this figure climbs to 300,000 if all the complications of the disease are taken into account. Total mortality TABLE I CLASSIFICATION OF DIABETES MELLITUS I N

I.

MAN

Idiopathic diabetes (primary) (a) Growth-onset (juvenile type) (b) Maturity-onset (adult type) II. Disorders of endocrine glands other than the pancreas (a) Tumors or hyperfunction of the anterior pituitary gland (b) Tumors or hyperplasia of adrenal cortex (c) Pheochromocytoma (d) Hyperthyroidism III. Diabetes due to pancreatic pathology (secondary) (a) Acute or chronic pancreatitis (b) Pancreatic lithiasis (c) Extensive malignant pancreatic tumors (d) Hemochromatosis IV. Surgical removal of pancreas

The Pancreas in Idiopathic Diabetes

335

from diabetes thus ranks third, behind cardiovascular disease and cancer. The Commission also ascertained that the prevalence of diabetes increased by 50% between 1965 and 1973, and it is estimated that the number of deaths will double every 15 years if present trends continue. Juvenile diabetics comprise 5-8% of all diabetic cases in the United States (White, 1932, 1960a; Boyd, 1946; Danowski, 1957). In 750 such patients studied by White (1960b), the disease began before the age of four in 60%, whereas it was diagnosed during the first year of life in only 0.5%. In a survey of published cases, Schwartz et al. (1947) also found that the disease was recognized in 0.5% of the afflicted children before they were one year old, while 0.05% of these infants had the diagnosis established during the first month after their births. The morphologic alterations found in the pancreases of diabetic patients are not diagnostic and lack specificity. Such changes include intra- and interacinar fibrosis, arterio- and arteriolosclerosis, and insular fibrosis and hyalinization (Hoppe-Seyler, 1904; Cecil, 1909; Opie, 1910; Weichselbaum, 1911; Moritz and Oldt, 1937; Neumann, 1929; Lazarus and Volk, 1962, p. 196). Although formerly considered characteristic for diabetes, all of these alterations may also be observed in nondiabetic persons. Also, while often in diabetic individuals the volume of islet tissue as well as the relative number and weight of the B cells are diminished, it has been shown that some diabetic pancreases contain more, rather than less, islet tissue and B cell granulation than do those of nondiabetics (Maclean and Ogilvie, 1955; Gepts, 1957, 1958; Bell, 1953b). However even if there is a B cell reduction, this may not in itself account for the diabetes since in experimental animals diabetes will ensue only if as much as 80-95% of the pancreas has been eliminated. Furthermore, diabetics may have normal amounts of extractable insulin in their pancreases (Wrenshall et al., 1952,1954) and normal concentrations of insulinlike substances in their blood (Bornstein, 1953a,b). Observations such as these have served to direct attention to extrapancreatic factors as possible causes or precipitating agents of diabetes. Implicated were certain hormones (Houssay, 1937; Long and Lukens, 1936; Young, 1937), trauma, emotional stress, obesity, viral infections (Gundersen, 1927; Gamble et al., 1969), or an autoimmune process (Pav et al., 1963; Chetty and Watson, 1965). Additional extrapancreatic conditions thought to play a role in the genesis of the disease include heredity, excessive destruction of insulin, and certain proteins or protein-associated agents ("synalbumin") (Vallance-Owen,

336

Bruno W. Volk and Klaus F. Wellmann

1964) capable of binding or holding insulin (Antoniades et al., 1962) and thus opposing its action, yet none of these hypotheses has been confirmed. It has recently been suggested that the metabolic disturbances of diabetes mellitus may not only be the result of relative or absolute insulin deficiency but may also require the presence of glucagon (linger and Orci, 1975). Since glucagon is a product of the A cells of the islets of Langerhans, at least primarily so, renewed attention is being focused upon the pancreas as the organ of pivotal importance for an understanding of the etiology and pathogenesis of diabetes. In this review, we shall attempt to portray the gross and microscopic alterations occurring in the pancreases of diabetic patients. The historical development of recognition of these changes will be traced, and their place within the framework of the modern concept of the pathophysiology of human diabetes will be discussed.

II. Gross Pancreatic Alterations The gross changes observed in the pancreas in idiopathic diabetes, all of them nonspecific, consist primarily of a reduction in weight, an accentuation of the lobular markings, and an increased consistency. Pancreatic weights as low as 28 gm have been recorded by Weichselbaum (1910) in diabetic individuals more than 20 years old. Herxheimer (1932) noted that the diabetic pancreases were thin and small and were often infiltrated by fat; in 105 of 162 cases, they weighed 40-50 gm, while some values were as low as 20 gm. Other authors (Lazarus and Volk, 1962, p. 196; von Halasz, 1909; Seyfarth, 1920a; Terplan, 1966) have obtained similarly low pancreatic weights in their diabetic patients. Eight of 31 diabetic pancreases examined by Terplan (1966) were described as fatty; their weights ranged from 21 to 110 gm. In Maclean and Ogilvie's series (1955), 52% of the adult diabetic pancreases but only 20% of the pancreases from a nondiabetic control group weighed less than 50 gm. The same authors studied 27 juvenile diabetics and found that the average weight of the pancreas in cases with "acute" diabetes was 51.6 gm, while that of "chronic" diabetics amounted to 38.3 gm (Maclean et al., 1959). The lowest pancreatic weight observed by Kraus (1929) among 46 diabetic persons ranging in age from 14 to 77 years was 19 gm; in many cases of juvenile diabetes the weight of the pancreas was half or less than half of normal. Twenty-two of 166 diabetic pancreases but only two of the age- and sex-matched control organs in a series recorded by Vartiainen (1944)

The Pancreas in Idiopathic Diabetes

337

weighed less than 40 gm. Among juvenile diabetics, Terbrüggen (1948) often found pancreatic weights between 17 and 40 gm. Warren and LeCompte (1952, p. 112) encountered small pancreases frequently weighing 50 gm or less in 122 of 730 diabetic patients; in 481 instances the weight was normal or ranged from 50 to 100 gm, while in 127 cases the pancreas was large and weighed more than 100 gm. In contrast to the observations recorded above, Lazarus and Volk (1962, p. 196) in 1Q0 cases failed to note significant differences in pancreatic weight between diabetics and nondiabetics of similar age and sex. In patients with maturity-onset diabetes, the weight ranged from 40 to 160 gm, with 85% of the organs weighing between 80 and 110 gm and with an average weight of 100 gm; in 100 nondiabetic controls, the average weight amounted to 99 gm and the range extended from 50 to 200 gm. These authors pointed out that both the size and the weight of the pancreas may be misleading, since considerable parenchymal atrophy may be masked by the presence of fatty infiltration or by fibrosis. Pancreases weighing 200 gm may thus contain less parenchymal tissue than others weighing only 50 gm. Much islet tissue may still be present even in organs extensively replaced by carcinoma or diffusely affected by pancreatitis or the sequela of duct obstruction. Nonspecific pancreatic changes were observed as early as 1894 by Hansemann (1894), who found 36 instances of "agranular atrophy" and three examples of fibrous induration in a series of pancreases from diabetic patients. Simmonds (1912) recorded 45 cases of atrophy in a group of 150 diabetic pancreases, while 12 showed "coarse" fibrosis and the remaining ones displayed extensive fat infiltration.

III. Hyalinization of the Pancreatic Islets Although found also in nondiabetic persons, hyalinization of the islets is usually considered the most common and most typical lesion of the diabetic pancreas. It was first recorded in 1901 by both Opie (1900-1901) and Weichselbaum and Stangl (1901). Weichselbaum (1911) described hyalinization in 28%, Seyfarth (1920b) and Allen (1922) in 20%, and Kraus (1929) in 10.5% of diabetic pancreases. Lazarus and Volk (1962, p. 209) noted this change in 25% of patients with maturity-onset diabetes; it was mild in 5%, moderate in 6%, and severe in 14% of these cases. An unusually high incidence of islet hyalinization was encountered by Ehrlich and Ratner (1961), who found it in 45 of 91 (49.5%) of their diabetic patients 50 years and older. Weichselbaum (1911) encountered hyalinization mainly in patients

338

Bruno W. Volk and Klaus F. Wellmann

over 50 years of age and only rarely between 27 and 40 years. Bell (1952) failed to observe the lesion in persons less than 20 years old; less t h a n 10% of those between the ages of 20 and 40, about one-fourth of the patients 40 to 50 years old, and 45.7% of those more t h a n 60 years old displayed some islet hyalinization. Bell also noted t h a t this change is rare in juvenile diabetes. In persons over 60 years of age, Seifert (1959) recorded hyalinization in one out of every two diabetic pancreases vs one out often in nondiabetics; he emphasized t h a t the presence of this lesion is not related to the severity and duration of the diabetes. Islet hyalinization was seen by Warren et al. (1966, p. 60) in only 5.7% of diabetic patients less t h a n 40 years of age, as compared to 34.5% in older persons. These authors studied 481 cases of diabetes that had lasted for at least ten years and found t h a t only one of 81 persons with onset of the diabetes prior to the 20th year of life showed islet hyaline, whereas this change was noted in 100 of 341 patients who were 35 years or older when their diabetes manifested itself. In Legg's (1966) series of 223 pancreases of diabetics, hyalinization was absent in individuals under the age of 40, while it was most frequently seen in persons more than 60 years old. While older authors, including Opie (1900-1901a) and Weichselbaum and Stangl (1901), looked upon hyalinization of the islets as typical for diabetes, it has subsequently been established that the lesion also occurs in metabolically intact individuals. As early as 1904, Ohlmacher (1904) encountered extensive hyalinization in a 27 year old nondiabetic patient, and somewhat later Saltykow (1909) recorded four such cases among 21 nondiabetic pancreases. Similar instances were reported by Cecil (1909) and Milne and Peters (1912). Seven of the cases in Wright's (1927) series were nondiabetic and concerned patients over 50 years of age; in two of them, however, the urine contained traces of sugar while the blood sugar values were not recorded. Hyalinized islets were noted by Ahronheim (1943) in five of 50 nondiabetics over 50 years of age. In the same age group, Arey (1943) observed the lesion in 16.6% of nondiabetic and in 71.7% of diabetic cases. Warren (1938, p. 31) also described islet hyalinization in individuals without diabetes. Although Bell (1959) encountered hyalinized islets in five of 200 nondiabetic subjects examined consecutively at autopsy, all 50 years old or older, he noted t h a t in none of these individuals had blood sugar or glucose tolerance tests been performed. The same author observed hyalinization of the islets in a high proportion of subjects over 80 years old (Bell, 1960); he found t h a t when hyalinized islets are present in elderly males there is only a 10% chance t h a t the patient has clinical evidence of diabetes. In his series of more t h a n 4000 autopsy investiga-

The Pancreas in Idiopathic

Diabetes

339

tions on nondiabetics, mild hyalinization was seen in 10% of persons 60 to 80 years old, and 14-18% of those between the ages of 80 and 100. In a smaller series of 178 consecutive autopsies on nondiabetic persons, Ehrlich and Ratner (1961) obtained an incidence of 3.9% for this lesion. According to Hartroft (1958), the severity of the observed hyalinization increases with age in both diabetic and nondiabetic populations, but at any given age is greater in the former. Histologically, pancreatic hyaline is an amorphous, acellular substance located between the islet cells and the intrainsular capillaries (Figs. 1 and 2). It may be present in small amounts near the capillaries, or it may form large masses replacing much or all of the islet tissue. The hyaline material may contain fusiform cells and may surround distinctly visible capillaries (Fig. 2). Involvement of islets in any given case is frequently uneven; while in some areas of the pancreas many or most islets are extensively affected, in other regions only occasional focal hyaline deposits may be present. Even in organs with numerous

FIG. 1. Normal islet. The B cells are black, the A cells are dark gray (straight arrows), and the D cells are light gray (curved arrows). Aldehyde fuchsin tri chrome stain. X500.

340

Bruno W. Volk and Klaus F. Wellmann

FIG. 2. Pancreas of maturity-onset diabetic showing partial hyalinization of an islet. The thin-walled capillaries are surrounded by small amounts of hyaline material. Phosphotungstic acid hematoxylin stain. X280.

involved islets, many others appear to remain intact. Concomitant fibrosis and atrophy of the exocrine parenchyma are frequently observed where hyalinization of islets is severe. While Opie (1900-1901b) was the first to note that there is a histologie resemblance between the insular hyaline and amyloid, he was unable to confirm with amyloid stains that the two substances might be identical. Mallory (1914, p. 521) also suggested that the hyaline deposits are closely related to amyloid. Gomori (1943) recorded amyloid deposits in the islets of occasional diabetic patients, and Bloom (1937) obtained a positive Congo-red stain in the islets of a cat with spontaneous diabetes. Other authors (Ahronheim, 1943; Arey, 1943; Van Beek, 1939) recorded a positive amyloid reaction with methyl violet and iodine green in human islets and believed the deposits to be amyloid; the material was found to be localized in the immediate vicinity of the capillary walls and gave staining reactions similar to, if not identical with, those obtained in liver and spleen in cases of gen-

The Pancreas in Idiopathic Diabetes

341

eralized amyloidosis. The results achieved with the methyl violet method caused Gellerstedt (1938) to suggest that the hyaline actually is amyloid and that the condition should be termed "insular (para)amyloidosis.,, He surmised that insular hyalinization represents a manifestation of senile amyloidosis similar to that seen in brain and heart. Arey (1943), on the other hand, expressed the belief that insular amyloid is an isolated feature not usually associated with amyloidosis of other tissues. Distinct metachromasia with crystal violet and binding of Congo-red, the latter displaying dichroic birefringence in polarized light, was noted by Ehrlich and Ratner (1961) in the hyalinized islets. These authors concluded that the substance that they thought identical with amyloid was localized between capillary walls and argyrophilic fibers. Schwartz (1965) employed fluorescent dyes such as thioflavine-T and proposed that the pancreatic insular hyaline is a manifestation of senile amyloidosis. In ultrastructural studies recorded by Lacy (1964), the hyaline material appears in the form of masses of small interlacing fibrils very similar to those of amyloid. The substance is not associated with the basement membranes but is deposited as foreign material between them; this differs from what is seen in diabetic small vessel disease within the dermal capillaries. Lacy therefore concluded that islet hyalinization represents a pathologic entity sui generis. In electron microscopic studies of islet cell adenomas, Porta et al. (1962) encountered deposits that had the tinctorial characteristics of amyloid.

IV. Vascular Lesions In 1884, Hoppe-Seyler (1904) observed considerable thickening of the walls of the arteries and especially the arterioles in the pancreas of diabetic persons. Cecil (1909) recorded arteriosclerosis in 80% of 90 pancreases and small vessel sclerosis in half of these cases; he also found that the incidence of vascular disease increases with age as reflected by a 40% occurrence figure for 20-30 year olds as against 77% for persons 30-40 years of age. After the 40th year, all but two of the pancreases exhibited arteriosclerosis. Moritz and Oldt (1937) encountered pancreatic arteriosclerosis in 30% of their diabetic patients aged 31-45 years and in 55% of such cases after the age of 61. In persons with coexisting diabetes and hypertension, arterio- and arteriolosclerosis of the pancreas was seen in 73% of individuals 31-45 years old and in 87% of those who had passed their 61st year of age. Other authors (Weichselbaum, 1911; Herxheimer, 1925; Seyfarth,

342

Bruno W. Volk and Klaus F. Wellmann

1950) subsequently confirmed the presence of considerable pancreatic arteriosclerosis. While Moschcowitz (1951) recorded age-dependent arteriosclerosis in the pancreases of diabetic patients, he failed to detect vascular disease in that organ in six individuals with juvenile diabetes. Considerable sclerosis and calcification was noted in 61 diabetic cases described by Warren et al. (1966, p. 106), who often saw little change in the arterial branches leading to the pancreas even where the splenic artery displayed severe sclerosis and calcification. Distinct arteriolosclerosis was reported by Lazarus and Volk (1962, p. 199) in 66% of pancreases from diabetic patients, but in only 34% of nondiabetic age-matched organs in which the disease was frequently less severe. In addition, there was occlusive arteriosclerosis of varying degree in 52% of the diabetic subjects as compared with less advanced changes in 34% of the nondiabetic group. Histologically, the arteriolar walls of diabetic patients show intimai

FIG. 3. Diabetic pancreas showing marked interacinar fibrosis and parenchymal atrophy. There is no fibrosis of the interlobular septa (curved arrows). Several arteries (straight arrows) show marked arteriosclerosis. Periodic acid schiff tri chrome stain. X95.

The Pancreas in Idiopathk

Diabetes

343

FIG. 4. Diabetic pancreas showing extensive perilobular fibrosis and parenchymal atrophy. The larger and small arteries (arrows) show considerable sclerosis. Aldehyde fuchsin trichrome stain. X125.

thickening and hyalinization (Figs. 3 and 4), which often results in almost total obliteration of the lumina. The extent of the involvement of the arterioles does not necessarily parallel the degree of sclerosis in the larger arteries even though the latter may be severely affected in some of the patients and may be almost or completely occluded. Lesions in the smaller blood vessels are often accompanied by parenchymal atrophy, interacinar fibrosis, and at times, fibrosis of the islets. Arteriosclerosis tends to be severe in pancreases containing many hyalinized islets. These observations appear to be in agreement with those of Moschcowitz (1951), who felt t h a t arteriosclerosis may be of importance both in the development of insular hyalinization and in the causation of diabetes mellitus during advancing age. On the other hand, Warren et al. (1966, p. 102) deemphasized the role of arteriosclerosis in diabetogenesis since they encountered this lesion in its severe form in only 7% of their diabetic patients.

344

Bruno W. Volk and Klaus F. Wellmann

V. Pancreatic Fibrosis Fibrosis is the most commonly encountered change in the diabetic pancreas. As early as 1894, Hansemann (1894) described parenchymal atrophy and replacement by newly formed connective tissue infiltrated focally by round cells; he surmised t h a t diabetes was secondary to a disease of the exocrine pancreas such as pancreatitis. Opie (19001901a) believed t h a t chronic interacinar pancreatitis was a characteristic concomitant of diabetes mellitus and suggested t h a t the newly formed connective tissue eventually engulfs the islets of Langerhans. Fibrosis of varying extent was observed by Sauerbeck (1902) in 62% of 176 pancreases of diabetic patients. Hoppe-Seyler (1904) noted widespread inter- and intraacinar connective tissue proliferation in a series of 18 diabetic pancreases; the islets of Langerhans were frequently involved to the point of their completely becoming obliterated. Distinct interacinar fibrosis occurring predominantly in older persons was observed by Cecil (1909) in 71% of 90 pancreases of diabetics; the pancreatic islets were frequently also affected by advanced sclerosis, especially in areas of maximal exocrine parenchymal fibrosis. Vartiainen (1944) found no gross fibrosis of the pancreas in 165 control patients but encountered it 30 times among 165 diabetics. In a series of 405 pancreases from diabetic persons examined by Warren et al. (1966, p. 102), 189 showed slight, 141 moderate, and 74 severe fibrosis, usually interacinar rather t h a n interlobular in location. There was no definite relationship between the occurrence of insular lesions and interacinar fibrosis except in cases of advanced islet fibrosis or hyalinization. In 50 cases of maturity onset diabetes, Lazarus and Volk (1962, p. 202) described fibrosis of varying degree in 58% as compared to a 42% incidence in a control group of nondiabetic patients; in every instance of insular fibrosis there was also fibrosis of the exocrine parenchyma. In the normal pancreas, thin and delicate fibrous tissue septa separate the individual lobules; at times, some augmentation of this tissue may be observed. In the diabetic pancreas, fibrosis is either interacinar or perilobular. Interacinar fibrosis, the more common type, is characterized by diffuse proliferation of fibrous connective tissue between the acini (Fig. 3). Occasionally, round cell infiltrates may be seen in the connective tissue, suggesting a possible inflammatory etiology (Fig. 5). The fibrosis may be generalized but often is focally distributed. The connective tissue frequently engulfs and may invade the islets, thus causing insular fibrosis, which can be extensive. This process involves the formation of peri-insular bands of fibrous tissue that extend into the islets along the paths of the capillaries. Compartmentalization as

The Pancreas in Idiopathic

Diabetes

345

FIG. 5. Diabetic pancreas showing interacinar fibrosis and infiltrates with lymphocytes and plasma cells. Periodic acid schiff tri chrome stain. X125.

well as compression and loss of insular tissue may follow. Most observers (Weichselbaum, 1911; Kraus, 1929; Seyfarth, 1920b) agree t h a t intrainsular fibrosis primarily affects older diabetics although the incidence varies. In perilobular fibrosis, connective tissue bands surround and separate lobules or groups of lobules (Fig. 4). In these pancreases there may also be an increase of fibrous tissue between the acini t h a t is usually more pronounced in the periphery of the parenchyma. The exocrine parenchyma usually is more or less atrophie, and there may be focal proliferation and occasional dilatation of the ducts. Fibrosis or hyalinization of the islets, of both, may be present, especially in older diabetics. Hultquist and Olding (1975) described pancreatic insular fibrosis in six of ten infants born to diabetic mothers; they were between 11 and 142 days old. The fibrous tissue occupied an islet area of 5-10% in three cases, 10-20% in two, and more t h a n 20% in the remaining organ. The three infants with the highest degree of fibrosis had an

346

Bruno W. Volk and Klaus F. Wellmann

abnormally large birth weight, and at least two of them were of "diabetic appearance." There was also enlargement of B cells, many of which contained hyperchromatic nuclei. Amyloid stains were negative, and there were no inflammatory cell infiltrates either in or around the islets. The authors believe t h a t these alterations occur only in infants of diabetic mothers, since they have not been observed in the offspring of metabolically intact women. The etiology of pancreatic fibrosis in diabetic patients is difficult to elucidate. Although in the perilobular variety the pancreatic ducts are frequently obstructed, the islets usually are not affected; but in some cases diabetes ensues when many islets have been replaced by fibrous tissue. In a number of instances of pancreatic fibrosis a history of previous gallbladder disease could be elicited. In all of these cases, a thick fibrous capsule can be seen to surround many of the islets, a lesion t h a t Otani (1927) believed to be evidence of the pathologic process. Another cause of pancreatic fibrosis may be acute or chronic pancreatitis as suggested by several authors (Opie, 1900-1901a; Schwartz, 1965; Comfort et al., 1946; Maimon et al., 1948; Popper, 1952). Evidence of a possible inflammatory etiology is seen in the occasional presence of large mononuclear cells, lymphocytes, and plasma cells within the connective tissue; these are more frequently encountered in diabetic than in nondiabetic patients (Lazarus and Volk, 1962, p. 202) (Fig. 5). It is unlikely t h a t chronic pancreatitis is responsible for many cases of pancreatic fibrosis, since the latter displays a progressive increase with age and is frequently associated with arteriosclerosis. Therefore, vascular impairment and ischemia inducing focal atrophy and fibrous replacement of the exocrine pancreas may be important pathogenetic factors (Robbins, 1957; Blumenthal and Probstein, 1959; Hranilovich and Baggenstoss, 1953). The vascular hypothesis of pancreatic fibrosis appears to derive support from Cecil's (1909) observation t h a t arteriosclerosis is associated with interacinar fibrosis in 80% of cases. Cecil noted conspicuous alterations in the small arteries, frequently including luminal obliteration, and recorded the frequent coexistence of insular fibrosis with sclerosis of the small pancreatic arteries. Hoppe-Seyler (1904) also believed that arteriosclerosis causes interstitial pancreatic fibrosis, a process resulting in exocrine acinar atrophy partly because of ischemia induced by the narrowing of the arteriolar lumina and partly because of the shrinkage of the connective tissue around the lobules. He suggested t h a t this type of pancreatic atrophy should be distinguished from the atrophy of the organ seen in older individuals. Hoppe-Seyler also noted that pancreatic fibrosis was far more severe in diabetic t h a n

The Pancreas in Idiopathic

Diabetes

347

in nondiabetic individuals and coined the term "pancreatitis interstitialis angiosclerotica," which emphasized the presumed vascular origin of the process. The common association of arteriolosclerosis with interacinar as well as insular fibrosis led Lazarus and Volk (1962, p. 202) to assume t h a t a pathogenetic relationship exists between the vascular and parenchymal lesions. A cause and effect relationship has also been inferred by these authors from the observed association of severe arteriosclerosis, at times accompanied by partial or complete occlusion of large pancreatic vessels, with parenchymal atrophy and fibrosis. They therefore proposed t h a t insular fibrosis results from a deficient blood supply to the pancreas primarily due to arteriolosclerosis, but also at times secondary to sclerosis of the larger blood vessels.

VI. Fatty Atrophy Fat has been observed in the pancreas of diabetic patients by many authors. Herxheimer (1906) noted a combination of fatty changes, fibrosis, and acinar atrophy in many cases of long-standing disease. He recorded t h a t in the same organ there were foci of fatty change and severe fibrosis alternating with advanced atrophy of the gland. Gruber (1929) suggested that the fatty changes were frequently located at the site of arteriolosclerotic alterations; he also noted the preservation of islets within the adipose tissue lobules and emphasized the frequent coexistence of atrophy and fibrosis. These lesions were only seen in older individuals, not in younger persons or children. Dieckhoff (1894) surmised t h a t pancreatic lipomatosis is often associated with pancreatic lithiasis, and he also pointed out t h a t tumors in the pancreas may result in sclerosis, lipomatosis, and atrophy of the distal portion of the organ. In the case of a cystadenoma of the head of the pancreas reported by Priesel (1922), the proximal part of the organ was normal, while the distal portions were replaced by adipose tissue containing only intact islets of Langerhans. Lang (1925) described an osteogenic sarcoma of the femur t h a t had metastasized to the body and tail of the pancreas; the pancreatic tissue between the metastatic nodes had been converted to fat and connective tissue in which small and large islets of Langerhans as well as small foci of exocrine tissue survived. In the authors' material, there was fatty infiltration of varying degree in 70 of 102 diabetic pancreases (71.4%) as compared to 22 of 64 nondiabetic organs (34.4%). In general, the fat was focally distributed

348

Bruno W. Volk and Klaus F. Wellmann

and often associated with atrophy or interacinar fibrosis, or both. Not infrequently, the fat was found within the septa that surrounded the pancreatic lobules. Many groups of islets were completely separated from the remaining parenchyma by zones of adipose tissue (Fig. 6). While a large number of these islets displayed considerable hyalinization, there were some islets involved by this process, although to a lesser degree, also seen in other areas of the same pancreases that were not surrounded by fat tissue. Focal or lobular fatty infiltration of the pancreas was often associated with vascular changes. The observation that atrophy, fibrosis, fatty changes, and hyalinization (Fig. 6) were frequently found in the same areas of fatty infiltration argues in favor of a cause and effect relationship. Lazarus and Volk (1962, p. 204) suggested that ischemia on the basis of vascular sclerosis induces a variegated pattern of response in the organ. Thus, in some instances

FIG. 6. Diabetic pancreas showing lipomatosis and atrophie lobules. The residual islets display hyalinization as well as peri- and intrainsular fibrosis. Several small arteries display marked sclerosis and concomitant narrowing of the lumen (arrows). Periodic acid Schiff trichrome stain. X50.

The Pancreas in Idiopathic Diabetes

349

atrophy and fibrosis will ensue, while in others fatty infiltration may follow, either alone or in combination with the above-mentioned changes. Eventually, in some areas the fatty changes may predominate. Lazarus and Volk were unable to explain why the insular tissue may persist even after the surrounding exocrine tissue has become atrophie or has disappeared altogether. The adipose tissue replacing the pancreatic parenchyma possibly represents an ex vacuo proliferation of fat such as it occurs in an atrophie zone. In what has been called lipomatosis of the pancreas, the entire organ is replaced by fat and, to a lesser degree, fibrous tissue with only a few islets identifiable. Pancreatic lipomatosis is believed to be associated with generalized obesity. Undoubtedly, the distinction between what may be called fatty atrophy and genuine lipomatosis is not always easy. Lazarus and Volk (1962, p. 204) described one diabetic and two nondiabetic pancreases, with extensive or subtotal replacement of the exocrine parenchyma by adipose tissue, that contained well preserved islets. They noted that this type of diffuse fatty infiltration of this organ is usually encountered in severely obese individuals. This change, though similar in type in both diabetic and nondiabetic pancreases, is nevertheless usually not as pronounced in the latter.

VII. Vacuolization ("Hydropic Degeneration") of the B Cells In 1901, Weichselbaum and Stangl (1901) observed a peculiar vacuolization of pancreatic islet cells in patients with diabetic coma. They termed this lesion, which was subsequently described in greater detail by Allen (1922), "hydropic degeneration" and thought that it resulted from liquefaction of B cell granules, eventually leading to atrophy of the B cells. Occasionally associated were lymphocytic infiltrates in the peri- and intrainsular connective tissue as well as a moderate augmentation of the connective tissue in the islets and dilatation of the capillaries. Both Heiberg (1911) and Fischer (1915) also noted lymphocytic infiltrates believed by them to represent a response to islet cell degeneration. Weichselbaum (1910) recorded vacuolization of the islets in 98 of 183 diabetic patients (53%) while others (Nakamura, 1924; Martius, 1915; Conroy, 1922) reported this lesion less frequently (8-40% of cases) and some (Karakascheff, 1904-1905; Thoinot and Delamare, 1907) were unable to find it at all. A few authors (Sauerbeck, 1904) observed islet vacuolization in nondiabetic as well as diabetic individuals. Warren (1938, p, 31) noted vacuolization in only 22 of 484 diabetic

350

Bruno W. Volk and Klaus F. Wellmann

pancreases and warned t h a t postmortem autolysis might be mistaken for this change. Gomori (1941) failed to detect hydropic alterations in any of his diabetic patients, but did encounter them in the B cells of five nondiabetics, one of whom had received large amounts of an intravenous dextrose solution. In phase microscopic investigations, Hartroft (1950) saw B cell vacuolization in eight of 46 diabetic persons (18%). Seifert (1959) recorded an incidence of 43% for diabetic and 5% for nondiabetic organs, while Conroy (1922) reported hydropic changes in only one of 12 diabetic pancreases. In a series of 653 pancreases examined by Warren et al. (1966, p. 73), 36 cases with B cell vacuolization (2%) were uncovered; the low incidence was deemed to be due to insulin treatment reducing functional strain in these cells so t h a t these lesions occur less often. While the relation of islet vacuolization to the duration of the disease is not entirely clear, the lesion has been encountered in fulminant forms of diabetes of relatively short duration. Weichselbaum (1910) noted a close correlation between hydropic degeneration and age, and both Kraus (1923) and N a k a m u r a (1924) observed the lesion more often in older persons and not at all in juvenile diabetics. It is possible that the incidence of islet vacuolization was higher in the preinsulin era. Warren (1938, p. 31) described vacuolization in the pancreatic duct epithelium in 73% of diabetic patients with complicating infections who had not received insulin. However, there was no evidence of hydropic B cell changes in two groups of preinsulin diabetics examined by Warren et al. (1966, p. 73); one comprised individuals dying in coma after a short duration of the disease, while persons in the second group died from arteriosclerosis or intercurrent infections after 25 to 30 years of mild diabetes. On the other hand, the lesions were encountered in two cases of a control series of nondiabetic patients. It is believed t h a t B cell vacuolization is due to glycogen infiltration occurring during diabetes. Toreson (1951) observed it in 11 and 26 diabetic persons (42.3%); glycogen was present in large amounts in two and in small quantities in six of these cases and was always associated with cytoplasmic vacuolization. However, glycogen infiltration is apparently rare in pancreatic human duct epithelium. Glycogen infiltration of pancreatic B cells and "ballooning degeneration" are independent phenomena, as shown in dogs by Lazarus and Volk (1962, p. 102; Volk and Lazarus 1964). In animals with growthhormone-induced diabetes, the hydropic lesions were thought to be reversible and independent degenerative alterations since the nuclei remained intact and well preserved. During the early stage of hydropic changes- the B cell granules are usually replaced by small vacuoles, which then increase in size until

The Pancreas in Idiopathic Diabetes

351

they occupy most or all of the cytoplasm. The nuclei appear unchanged and are surrounded by glycogen. Sometimes seen are basophilic cytoplasmic masses called Körnchen by Weichselbaum (1910) and shown to be RNA by Gepts (1964, 1965). Ultrastructural investigations have established that glycogen first accumulates diffusely in the B cells and then becomes concentrated in focal areas (Williamson and Lacy, 1961; Volk and Lazarus, 1963). Lazarus and Volk (1961) surmised that the glycogen infiltration follows the distention and vacuolization of the endoplasmic reticulum, which are signs of increased secretory activity. While vacuolization appears to be reversible during the earlier stages, the cytoplasm later atrophies and the cells eventually disappear. Vacuolization was also observed by Bastenie (1956, p. 148) in a patient after two weeks of steroid treatment, while Lazarus and Volk (1962, p. 106) encountered it in a woman who became diabetic during steroid therapy for breast cancer; in this patient the B cells, although vacuolated, contained intact nuclei. Vacuolization and glycogen infiltration have also been seen in nondiabetic patients subjected to prolonged glucose administration in the period preceding death (Gomori, 1945).

VIII. Hypertrophy of the Islets True hypertrophy of the pancreatic islets may occur in diabetic as well as nondiabetic individuals. While most of the islets may be considerably enlarged, usually only a few are genuinely hypertrophied (Cecil, 1909). The critical diameter above which the islets are considered hypertrophie has been stated to be 300 μηι (Weichselbaum, 1908) or 400 μτη (Warren et al., 1966, p. 80) (121). Occasionally it may be difficult to distinguish large islets from small adenomata. In most cases, there are no obvious reasons to account for insular hypertrophy. Pathogenetically, the so-called giant islets apparently form by coalescence of several small islets (Eder, 1955). Hypertrophy of the islets of Langerhans is rather common. Two common distinguishable types have been described by Cecil (1909, 1911). In one, the islets increase without a change in the character of the individual cells. In the other, the cells become columnar or "ribbonlike/' with the nuclei located in their centers. These cells are arranged in the form of serpentine formations, an architectural trait entirely different from the normal pattern even though the relationship between the cells and the capillaries is maintained. Cecil encountered 34

352

Bruno W. Volk and Klaus F. Wellmann

examples of hypertrophy and probable regeneration among 100 autopsy cases of diabetic patients, none in 33 cases of chronic pancreatitis, and one in 17 instances of pancreatic carcinoma. Ogilvie (1944,1964) saw enlarged islets mainly in obese persons and attributed them to the effect of a pituitary pancreotropic factor. Maclean et al. (1959) noted t h a t hypertrophy of the islets occurs relatively often in young early diabetic patients. The B cell ribbons were considered to be characteristic for insular hypertrophy by LeCompte (1960); he found t h a t in many diabetic pancreases, especially from juvenile diabetics, the ribbons are composed of small cells with scanty cytoplasm, which apparently are A cells. Large islets were encountered by Hüttl (1936) at autopsy in diabetic persons with pancreatic ligations. In a series of 1376 diabetic pancreases studied by Warren et al. (1966, p. 80) there were 64 with hypertrophie islets. These authors were able to easily distinguish between the two types of hypertrophy, recording observations similar to those made by Cecil (1909, 1911). They also noted the presence of irregular projections between the acini in the periphery of the islets. In their material, only a moderate number of islets were hypertrophied, although occasionally most were considerably enlarged. There was no correlation between the insular hypertrophy and either severity and duration of the diabetes or the age of the patients.

IX. Inflammatory Lesions of the Islets First described by Cecil (1909), inflammatory lesions in the pancreatic islets were called "insulitis" by von Meyenburg (1940) and have been encountered in two-thirds of young diabetics dying after clinically manifest diabetes of one year or less (Gepts, 1965); they have not been observed in chronic juvenile diabetics, older diabetic persons, or nondiabetics (Deconinck et al., 1972a). These lesions occur in only a few islets and mainly display small lymphocytes along with occasional polymorphonuclear leukocytes and histiocytes. The affected islets may be atrophie but often consist of well-recognizable A and B cells (Gepts, 1965). Fibrosis may at times be present, too (Cecil, 1909; Gepts, 1965; LeCompte, 1958; Warren, 1927; Stansfield and Warren, 1928). LeCompte et al. (1966; LeCompte, 1958) suggested t h a t lymphocytic infiltration of the islets is possibly a response to toxic injury or a subdued reaction to islet cell necrosis. Most authors favor an infectious etiology for this lesion. Thus, Gundersen (1927) related t h a t there is an increased rate of mortality from diabetes two to four years after an epidemic of mumps, while Barboni and Manocchio (1962) described

The Pancreas in Idiopathic Diabetes

353

inflammatory lesions in the pancreas of cows that had become diabetic shortly after the onset of foot and mouth disease. Furthermore, high titers for Coxsackie B virus, especially of type B4, were found in insulin-dependent diabetics within three months after the onset of the disease (Gamble et al., 1969). These findings appear to receive support from experimental data indicating that insulitis may result from an immunologie process. For instance, cows repeatedly injected with pork or beef insulin exhibit biological signs of immunization against insulin as shown by Renold et al. (1964). In such animals, LeCompte and coworkers (1966) saw considerable infiltration of the islets by lymphocytes and other mononuclear cells associated with moderate fibrosis and a diminution in the number of B cells. In the islets of rabbits immunized against beef insulin, Toreson et al. (1968) encountered severe inflammatory lesions, while the animals became diabetic.

X. Regeneration and Ductal Proliferation Both diabetic and nondiabetic pancreases commonly display proliferation of ductules (Lazarus and Volk, 1962, p. 212) that either may lie within the islets or may be contiguous with or encircle the latter. The proliferating duct epithelium may resemble, in general appearance and shape, an islet of Langerhans ("pseudoislet") or may imitate the features of a regenerating or hyperplastic islet. Special staining techniques have to be applied in an effort to avoid confusion between proliferating duct epithelium and pseudoislets with insular tissue (Lazarus and Volk, 1962, p. 212). Islet regeneration has been thought to take place in infants and children in particular (Cecil, 1909; Nakamura, 1924; Hüth, 1936; Gutman, 1903; Schmidt, 1959). Weichselbaum (1908) saw these changes in 58 of 183 organs, especially in the pancreatic head, while Cecil observed regeneration and hypertrophy of the islets in 100 diabetic individuals (Cecil, 1911); no such alterations existed in 17 cases of pancreatic neoplasia and in 33 instances of chronic pancreatitis.

XI. Fat Deposits in B Cells After Dogiel's (1893) initial description, several other authors (Seyfarth, 1920b; Nakamura, 1924; von Meyenburg, 1940; Weichselbaum and Stangl, 1902; Symmers, 1909; Wilder, 1926; Gibb and Lo-

354

Bruno W. Volk and Klaus F. Wellmann

gan, 1929) reported the occurrence of fat droplets in the islets of Langerhans; they were at times present in large numbers. Weichselbaum and Stangl (1902) believed t h a t the pancreatic islets of diabetics contained more fat than those of nondiabetics. N a k a m u r a (1924) saw large and small fat vacuoles in the islets of all of his diabetic patients. Both Symmers (1909) and Seyfarth (1920b) confirmed the presence of fat in diabetic pancreases, but also noted its occurrence in a number of other conditions, especially in alcoholism. Wilder (1926) encountered fatty deposits more often in diabetic t h a n in nondiabetic persons and surmised them to be the result rather than the cause of diabetes; the changes were severe in 11 and moderate in 15 of his cases. Hartroft (1960) also noted a higher incidence of fat droplets in diabetic individuals and emphasized their importance. Lipochrome (ceroid) granules were encountered in the B cells of nondiabetic persons by Like (1967), while Deconinck et al. (1971) in ultrastructural investigations found lipid inclusions of varying sizes and shapes, as well as some vacuoles in the B cells of nondiabetic adults but not in those of newborn infants (Deconinck et al, 1972b).

XII. Degranulation of B Cells In a series of 995 diabetic pancreases examined by Bell (1953b), there was complete or partial degranulation of the B cells in all patients less t h a n 20 years old, in 79.5% of those between the ages of 20 and 40, in 48.2% of those 40 to 60 years old, and in 33.6% of persons over 60 years of age; in contrast, only two degranulated islets were noticeable in a control group of 250 cases without diabetes. Bell concluded t h a t the presence of degranulation is highly suggestive of the diabetic state. Wrenshall and co-workers (1952) established t h a t the insular B cell granules and the extractable pancreatic insulin correlated on a one to one basis. They determined t h a t the extractable insulin averages approximately 3 units/gm of wet tissue weight. Even though the B cell granules autolyze fairly rapidly after death, the extractable insulin does not disappear simultaneously. Warren et al. (1966, p. 62) examined 223 consecutive cases of diabetes and saw poor B cell granulation in 14 instances. These authors believed t h a t B cell degranulation is not sufficiently characteristic for diabetes since there occurs a progressive loss of granules in these pancreases t h a t parallels the time elapsed between death and postmortem examination.

The Pancreas in Idiopathic Diabetes

355

XIII. Quantitative Changes of the Islets of Langerhans The evaluation of quantitative changes in the endocrine pancreas, often attempted since the turn of the century, is not an easy task because of technical difficulties. Also, the number of islets is not the same in all portions of the organ. All quantitative examinations of the islets utilize more or less numerous fragments of tissue taken from different parts of the pancreas, but representing only a minute fraction of the whole organ. There is, however, agreement that in diabetes there is usually a decline in the number as well as in the area of the islets. In a study of 18 diabetic pancreases, Weichselbaum and Stangl (1901) encountered only a few islets. Ssobolew (1902) recorded the absence of islets in four diabetic patients and a considerable diminution in 13. Sauerbeck (1902) also noted a marked decrease in the number of islets in diabetic as compared to nondiabetic persons, and so did Cecil (1909) in 20 cases. Other investigators endeavored to determine the amount of insular tissue by calculating the number of islets present in the pancreas. Kraus (1929) noted that the number of islets per 50 cm2 of pancreas in 20 diabetic patients ranged from 8.7 to 128, with an average value of 54.1; there was a ratio of 2.4:1 between the normal and the diabetic organ. He remarked that in many diabetics islet decrease is sufficiently pronounced to become obvious even without counting. Other authors (Schwartzman et al, 1947; Bell, 1953b; Herzog, 1902; Moore, 1936; Dieckhoff, 1895; Bence, 1907; Potter and Milne, 1911) have also described the total or subtotal absence of islets in diabetic patients. In a 13 year old girl with a history of diabetes of at least six years, Moore (1936) observed an aplasia of the islets in which only a few small cell groups suggestive of insular elements were present. Gepts (1971) recorded a considerable decrease of the number of islets in young diabetics. In view of these findings it would seem that most cases characterized by an absence of islets, as reported by the earlier workers, actually represent examples of juvenile diabetes.

XIV. Estimation of Islet Volume Most investigators concerned with this subject attempted to relate the measured area of the islets to that of the pancreas as a whole. Heiberg (1906) was among the first to employ planigraphic measurements in this context. Neumann (1929) found the volume of the

356

Bruno W. Volk and Klaus F. Wellmann

pancreatis islets markedly decreased in most diabetic individuals. The islet tissue comprised 0.9-3.5% of the total pancreatic mass in 145 nondiabetic persons studied by Susman (1942); in 60% of 55 diabetics, it was less than 0.9%. Employing a microscopic grid for insular area measurements, Maclean and Ogilvie (1955) calculated the weight of the islets in 30 diabetic patients as ranging from 0.019 to 1 gm, with a mean of 0.45 gm, as compared to a range of 0.57 to 1.89 gm and a mean of 1.06 gm in age- and sex-matched controls. The same authors (Maclean et al., 1959) found the mean islet weight of eight young diabetic individuals who had died within eight weeks after the onset of their disease to be 0.70 gm, while that of 19 chronic cases in whom death had occurred nine months to 19 years after the first signs of diabetes amounted to 0.21 gm; the corresponding value for 22 nondiabetic controls was 1.19 gm. The islets in the acute group were larger than those in the chronic cases; the mean proportion of islet tissue was 1.54% in the acute instances, 0.67% in the chronic group, and 2.45% in the controls. The proportion of B cells was greater in the acute than in the chronic examples in the few cases in which it was determined. Gepts (1957) used planimetric measurements of islets in 200 sections per pancreas and calculated the total weight of the islets in 28 diabetic patients 50 years and older to be 0.765 gm, while that of 31 nondiabetics amounted to 1.385 gm; the difference is significant in spite of a considerable overlap in individual values between the two series. Employing WickselFs formula, Hellman (1961) found that in both diabetic and nondiabetic persons there is a uniform mathematical relationship between the total islet volume and the islet diameter, so that a symmetrical volume distribution curve ensues and a balance between the number of large and small islets is being retained. Hellman and Angervale (1961) obtained a symmetrical volume distribution curve in maturity-onset diabetics; slight asymmetry was observed in two cases and was attributed to hyalinized islets. Asymmetry also pertained in instances of insulinoma, in acromegaly, and in two patients with juvenile diabetes in whom it was ascribed to the smallness and scarcity of the islets.

XV. Differential Counts of the Islets of Langerhans Although A and B cells have been known since the turn of the century, estimates of their relative proportions in the diabetic patient have been attempted only in fairly recent times. Such studies have

The Pancreas in Idiopathic Diabetes

357

been facilitated by the development of improved staining techniques permitting the differentiation of the various islet cell types. The methods commonly employed were the chrome alum hematoxylin and phloxine stain of Gomori and the silver impregnation procedure of Gros-Schultze. However, both have their pitfalls, so that attempts to clearly delineate the various constituents of the islets are fraught with error. For instance, Creutzfeldt (1953, 1956; Creutzfeldt and Theodossiou, 1957) showed that different results could be obtained with the Gros-Schultze procedure by slight variations in technique and that some B and D cells may also stain with it, in addition to the A cells. He attributed some of the conflicting results of other authors to the difficulty of counting cells in thick frozen sections, as are used for the Gros-Schultze method, and to differences in staining technique. Similarly, both A and D cells may be stained by the phloxine of the Gomori procedure so that A cell counts based on it become inaccurate. In differential counts on 59 nondiabetic and 11 diabetic pancreases, Gomori (1943,1945), using his chrome alum hematoxylin and phloxine method, observed a definite reduction of the B : A ratio in several of the diabetic cases. With the Gros-Schultze silver stain applied to frozen sections of the diabetic pancreas, Ferner (1938, 1942a,b, 1947, 1951, 1952) noted an increase in the proportions of A cells, ranging from 35 to 100%. In a four year old diabetic child he encountered many islets as well as islet "buds" (Inselsprossen) composed almost exclusively of silver-positive cells considered by him to be A cells. Therefore, he proposed that diabetes, especially the juvenile variety, is basically due to a failure of the silver cells, which he also considered to constitute immature (unreife or inselpotente) elements, to ripen into B cells. He furthermore surmised that in most juvenile diabetics a higher proportion of A cells exists than in older diabetics and that the increase in A cells is accompanied by a simultaneous decline in the number of B cells. Ferner concluded that A cells are precursors of B cells and that human diabetes results from a maturation arrest somewhere during transformation of A into B cells. Hess (1946) supported Ferner's conclusions and, with the Gros-Schultze method, found a mean B : A ratio of 8.6:9.9 in nine normal pancreases and 2.1:4.8 in ten diabetic organs. In the latter the figures appeared closely grouped, while in the former a greater spread of distribution was obtained. Von Meyenburg (1946) observed a low B:A ratio in diabetic pancreases. In 16 nondiabetic pancreases, Hultquist et al. (1948) found 27-40% silver cells, while there were 41-50% of these elements in two diabetic organs. Terbrüggen (1947, 1948) primarily using a modification of Bensley's acid fuchsin methyl-green technique obtained a B : A ratio of 3:5 for most nondiabetics and one of less than 3:5 in 80% of the diabetic

358

Bruno W. Volk and Klaus F. Wellmann

pancreases. In juvenile diabetics, he described a definite decrease of B cells and a concomitant increase of A cells. With the same method, Seifert (1954) recorded an A : B cell ratio of 1:1.3 to 1:4.6 in nondiabetic control pancreases and 1:1 to 2:5 in diabetic organs. He felt that the lower ratio in the latter is not specific since there was a considerable overlap of the individual figures between the two groups. Both Bürkl (1951; Bürkl and Kovac, 1951) and Creutzfeldt (1953, 1956) could not confirm the observations of Ferner and of other authors using the silver impregnation technique. Gepts (1957, 1958) employed a modification of Holmes' silver impregnation procedure on thin sections of paraffin-embedded tissue and compared the cell count obtained with the results derived from Gomori's chrome alum hematoxylin and phloxine method. He found consistently higher values for the A cells with the silver method and therefore, in agreement with Creutzfeldt, concluded t h a t some of the B cells were also impregnated. Gepts (1971) emphasized t h a t the apparent increase of A cells in the diabetic is relative rather t h a n absolute and can be attributed to a decrease in the number of B cells. In juvenile diabetics dying less than six months after the manifestation of their disease, he found 100-1000 B cells per cm2 as compared with 1000-10,000 in age-matched nondiabetic controls. With specific granule stains, Warren et al. (1966, p. 97) obtained a B : A ratio of 1.8 in diabetic and 8.2 in nondiabetic pancreases.

XVI. Quantitative Estimation of Islet Tissue Several authors determined the weights of the islets as well as those of their cellular components and found, in general, a distinct decrease in islet weight in diabetic patients even though there is considerable overlap in the figures between diabetic and nondiabetic organs. Ogilvie (1964) noted an average number of 7130 islets per gram of pancreas in 30 diabetic persons as compared with 14,000 in as many control organs. He obtained mean weights of islet tissue, A cells, and B cells of 0.1, 0.04, and 0.05 gm, respectively, in a group of growth-onset diabetics, while the corresponding figures in maturity-onset diabetic patients were 0.5, 0.17, and 0.25 gm, respectively. Thus, the mean weight of the B cells appeared considerably reduced in both growth- and maturity-onset diabetes, especially so in the former. Ogilvie also observed t h a t the A and B cell weights were suggestively lower in patients aged 40-60 years t h a n in those over 60 years old.

The Pancreas in Idiopathic Diabetes

359

In another study, Ogilvie (1933) determined the proportion of the islet tissue and the size of the islets in 19 lean subjects as well as in 19 persons with obesity, a frequent antecedent to diabetes. The percentage of islet tissue ranged from 0.74 to 5.71 and averaged 3.19 in the obese individuals, whereas it varied from 0.80 to 3.84, with a mean of 2.05, in the control group. Average islet size, expressed in cm2 at a magnification of 120, ranged from 1.46 to 5.78 (mean: 2.59) in obese and from 0.99 to 2.22 (mean: 1.57) in lean persons. Thus, the proportion of islet tissue and the size of the islets were greater in the obese than in the control group by 56 and 65%, respectively. In 31 nondiabetic pancreases, Gepts (1957) calculated the mean total weight of the islet tissue as 1.358 gm, that of the B cells as 0.754 gm, and that for the A cells as 0.341 gm. The corresponding values for 28 diabetic pancreases were 0.765, 0.301, and 0.319 gm, respectively.

XVII. Pathogenetic Considerations of Idiopathic Diabetes Soon after von Mering and Minkowsky (1889-1890) first demonstrated in the dog that the removal of the pancreas induces diabetes, an insufficient secretion of insulin has been thought to cause the idiopathic form of the disease as well. However, attempts to confirm this hypothesis have not been conclusive. Thus, while some diabetic patients have shown a decrease in the number of their B cells and B cell granules, as well as diminished values for extractable insulin and for circulating insulinlike substances, others have not (Bornstein, 1953a,b; Wrenshall et al, 1952, 1954; Hartroft and Wrenshall, 1955; Bell, 1953a). The fact that the sulfonylureas that stimulate the secretion of pancreatic insulin are effective in lowering the blood sugar in many maturity-onset diabetics also favors the view that many of these patients have adequate insulin reserves (Bertram et al., 1955; Beaser, 1956). Of various extrahepatic factors, among them trauma, emotional stress, obesity, and infection, that have been implicated at one time or another, none appears to be of primary significance in the pathogenesis of diabetes. It has also not been possible to show that any of the pituitary or adrenal hormones are causative agents in cases of idiopathic diabetes. Other views proposed were that diabetics inherit a defect that creates forces antagonistic to insulin, or that they elaborate substances that bind, or have cells that are insensitive to, insulin, or that an

360

Bruno W. Volk and Klaus F. Wellmann

autoimmune mechanism damages the islet cells so t h a t diabetes will ensue. Thus far there is insufficient evidence to support any of these hypotheses (Par et al, 1963; Chetty and Watson, 1965; Deckert, 1967; Mancini et al., 1965). While in most patients with idiopathic diabetes the blood insulin values are not decreased, close monitoring of the secretory response in such patients has shown t h a t the initial rise in serum insulin during a continuous infusion of glucose is lacking or considerably diminished, and t h a t the total response falls short of normal. Cerasi and Luft (1967) concluded from these observations t h a t the secretory deficiency of the B cells is the inherited factor but t h a t in the majority of cases diabetes will occur only under the effect of added diabetogenic factors with which the genetically deficient B cells cannot cope. The results of recent studies suggest t h a t the second islet cell hormone, glucagon, plays a more important role in the pathogenesis of diabetes t h a n had hitherto been thought. Since endogenous hyperglycemia has rarely been observed in the absence of glucagon, the diabetic abnormalities in glucose homeostasis have been looked upon as the result of a bihormonal disorder in which a relative or absolute deficiency of insulin and a relative or absolute excess of glucagon are both present (Unger and Orci, 1975). The postulated role of glucagon is underscored by observations related to the action of somatostatin, which inhibits the release of growth hormone as well as of both glucagon and insulin; thus hyperglycemia fails to occur unless glucagon levels are restored to normal by the concomitant infusion of exogenous glucagon (Goodner et al., 1974; Sakurai et al., 1935; Alford et al., 1974). However, certain additional experiments elucidating the action of glucagon have caused Levine (1976), in a recent editorial, to conclude that "glucagon is a potent diabetogenic factor in the absence of insulin, but t h a t physiologic amounts of insulin can overcome or prevent the effects of appreciably increased glucagon levels, at least in man." There is no doubt t h a t recent work has contributed appreciably to a better understanding of the disturbance in glucose homeostasis t h a t characterizes the diabetic state. Whether the opposing effects of insulin and glucagon constitute the final answer with regard to the pathogenesis of diabetes remains to be seen. While the newer clinical observations would seem to lessen the significance of the earlier morphologic studies related to the pancreas, the fact remains that this organ is frequently markedly decreased in size, particularly in juvenile diabetes, and t h a t both qualitative and quantitative changes exist in the pancreatic A and B cells of diabetic individuals, which cannot be properly explained by purely physiologic and pathophysiologic studies.

The Pancreas in Idiopathic Diabetes

361

Attempts will have to be made to correlate the endocrinological abnormalities with the morphologic alterations occurring in the islets of Langerhans in diabetes with the help of newer, refined biochemical, histochemical, and immunocytochemical techniques and by investigating the genetic potential of B cell division and the kinetics of islet cell proliferation. References Ahronheim, J. H. (1943). Am. J. Pathol. 19, 873. Alford, F. P., Bloom, S. R., and Nabarro, J. D. N. (1974). Lancet 2, 974. Allen, F. M. (1922). J. Metab. Res. 1, 5. Antoniades, H. N., Gundersen, K., Beigelman, P. M., Pyle, H. M., and Bougas, J. A. (1962). Diabetes 1 1 , 2 6 1 . Arey, J. B. (1943). Arch. Pathol. 36, 32. Barboni, E., and Manocchio, I. (1962). Arch. Vet. Ital. 13, 477. Bastenie, P. (1956). "Cortico-surrénale et Diabète Humain," Masson, Paris. Beaser, S. B. (1956). Metabolism 5, 933. Bell, E. T. (1952). Diabetes 1, 341. Bell, E. T. (1953a). Diabetes 2, 125. Bell, E. T. (1953b). Diabetes 2, 376. Bell, E. T. (1959). Am. J. Pathol. 35, 801. Bell, E. T. (1960). "Diabetes Mellitus," Thomas, Springfield, Illinois. Bence, J. (1907). Wien. Klin. Wochenschr. 20, 721. Bertram, F., Bendfeldt, E., and Otto, H. (1955). Drsch. Med. Wochenschr. 80, 1455. Bloom, F. (1937). N. Engl. J. Med. 217, 395. Blumenthal, H. T., and Probstein, J. G. (1959). "Pancreatitis: A Clinical-Pathologic Correlation." Thomas, Springfield, Illinois. Bornstein, J. (1953a). J. Endocrinol. 7, 59. Bornstein, J. (1953b). Diabetes 2, 23. Boyd, J. D. (1946). In "Practice of Pediatrics" (J. Brennemann, ed.) Prior, Hagerstown, Maryland. Bürkl, W. (1957). ActaAnat. 12, 358. Bürkl, W., and Kovac, W. (1951). Mikroskopie 6, 283. Cecil, R. L. (1909). J. Exp. Med. 11, 266. Cecil, R. L. (1911). J. Exp. Med. 14, 500. Cerasi, E., and Luft, R. (1967). Ada. Endocrinol. 55, 330. Chetty, M. P., and Watson, K. C. (1965). Lancet 1, 67. Comfort, M. W., Gambill, E. E., and Baggenstoss, A. H. (1946). Gastroenterology 6, 239. Conroy, M. J. (1922). J. Metab. Res. 2, 367. Creutzfeldt, W. (1953). Beitr. Pathol. Anat. 113, 133. Creutzfeldt, W. (1956). Z. Inn. Med. 37, 217. Creutzfeldt, W., and Theodossiou, A. (1957). Beitr. Pathol. Anat.Allg. Pathol. 117, 235. Danowski, T. S. (1957). "Diabetes Mellitus, With Emphasis on Children and Young Adults." Williams & Wilkins, Baltimore, Maryland. Deckert, T. (1967). Ada Med. Scand. Suppl. 476, 30. Deconinck, J. F., Potvliege, P. R., and Gepts, W. (1971). Diabetologia 7, 266.

362

Bruno W. Volk and Klaus F. Wellmann

Deconinck, J., Potvliege, P. R., and Gepts, W. (1972a). In "Handbook of Physiology" (R. O. Greep, E. B. Astwood, D. F. Steiner, N. Freinkel, and S. R. Geigen, eds.), Sect. 7, Vol. 1, p. 295. Williams & Wilkins, Baltimore, Maryland. Deconinck, J. F , Van Assche, F. A., Potvliege, P. R., and Gepts, W. (1972b). Diabetologia 8, 326. Dieckhoff, C. (1894). Beiträge zur path. Anatomie des Pankreas. Med. Inaug.-Diss., Rostock. Dieckhoff, G. (1895). Beitr. z. Wissensch. Med., Festschrift. Leipzig, Theodor Thierfelder, p. 97. Dogiel, A. S. (1893). Arch. Anat. Entwicklungsgesch. 2, 117. Eder, M. (1955). Beitr. Pathol. Anat. Allg. Pathol. 115, 157. Ehrlich, J. C., and Ratner, I. M. (1961). Am. J. Pathol. 38, 49. Ferner, H. (1938). Z. Mikrosk. Anat. Forsch. 44, 451. Ferner, H. (1942a). Virchows Arch. A 309, 87. Ferner, H. (1942b). Dtsch. Z. Verdau. Stoffwechselkr. 6, 21. Ferner, H. (1947). Dtsch. Med. Wochenschr. 72, 540. Ferner, H. (1951). Virchows Arch. 319, 390. Ferner, H. (1952). "Das Inselsystem des Pankreas," p. 74. Thieme, Stuttgart. Fischer, B. (1915). Frankfurt. Z. Pathol. 17, 218. Gamble, D. R., Kinsley, M. L., Fitzgerald, M. G., Bolton, R., and Taylor, K. W. (1969). Brit. Med. J. 3, 627. Gellerstedt, N. (1938). Beitr. Pathol. Anat. Allg. Pathol. 101, 1. Gepts, W. (1957). Ann. Soc. Roy. Sei. Med. Natur. Bruxelles 10, 5. Gepts, W. (1958). Endokrinologie 36, 185. Gepts, W. (1964). In "The Structure and Metabolism of the Pancreatic Islets" (S. Brolin, B. Hellman, and A. Knutson, eds.), p. 513. Macmillan, New York. Gepts, W. (1965). Diabetes 14, 619. Gepts, W. (1971). In "Handbook of Diabetes Mellitus" (E. F. Pfeiffer, ed.), Vol. 2, pp. 3-39. Lehmann, Munich. Gibb, W. F , and Logan, V. W. (1929). Arch. Int. Med. 43, 376. Gomori, G. (1941). Am. J. Pathol. 17, 395. Gomori, G. (1943). Arch. Pathol. 36, 217. Gomori, G. (1945). Bull. N.Y. Acad. Med. 2 1 , 99. Goodner, C. J., Ensinck, J. W., Chideckel, E., Palmer, J., Koerker, D. J., Ruch, W., and Gale, C. C. (1974). J. Clin. Invest. 53, 28a. Gruber, G. (1929). In "Handbuch der spez. Path. u. path. Anat." (F. Henke and O. Lubarsch, eds.), Vol. 5/2, p. 211. Springer-Verlag, Berlin. Gundersen, E. (1927). J. Infect. Dis. 4 1 , 197. Gutman, C. (1903). Virchows Arch. 172, 493. Hansemann, D. (1894). Z. Klin. Med. 26, 191. Hartroft, W. S. (1950). Proc. Am. Diabetes Assoc. 10, 46. Hartroft, W. S. (1956). Diabetes 5, 98. Hartroft, W. S. (1960). "Diabetes" (R. H. Williams, ed.), p. 350. Harper (Hoeber), New York. Hartroft, W. S., and Wrenshall, G. A. (1955). Diabetes 4, 1. Heiberg, K. A. (1906). Anat. Anz. 29, 49. Heiberg, K. A. (1911). Centralbl. Allg. Pathol. Pathol. Anat. 221, 532. Hellman, B. (1961). A eta Pathol. Microbiol. Scand. 5 1 , 95. Hellman, B., and Angervale, L. (1961). Ada Pathol. Microbiol. Scand. 53, 230. Herxheimer, G. (1906). Virchows Arch. 183, 228.

The Pancreas in Idiopathic Diabetes

363

Herxheimer, G. (1925). In "Handb. d. inneren Sekretion" (M. Hirsch, ed.). Springer, Berlin. Herxheimer, G. (1932). Verh. Dtsch. Ges. Verdau. Stoffwechselkr. 1 1 , 112. Herzog, M. (1902). Virchows Arch. 168, 83. Hess, W. (1946). Schweiz. Z. Pathol. Bakteriol. 9, 46. Hoppe-Seyler, G. (1904). Dtsch. Arch Klin. Med. 8 1 , 119. Houssay, B. A. (1937). Am. J. Med. Sei. 193, 581. Hranilovich, G. T., and Baggenstoss, A. H. (1953). Arch. Pathol. 55, 443. Hüttl, T. (1936). Beitr. Klin. Chir. 163, 206. Hultquist, G. T., and Olding, B. B. (1975). Lancet 2, 1016. Hultquist, G., Dahlen, M., and Helander, C. G. (1948). Schweiz. Z. Pathol. Bakteriol. 11, 570. Karakascheff, K. I. (1904-1905). Dtsch. Arch. Klin. Med. 82, 60. Kraus, E. J. (1923). Virchows Arch. 247, 1. Kraus, E. J. (1929). In "Handbuch der speziellen pathologischen Anatomie und Histologie" (F. Henke and O. Lubarsch, eds.), Vol. 5/2, pp. 662-727. Springer-Verlag, Berlin. Lacy, P. E. (1964). In "Aetiology of Diabetes and Its Complications" (M. P. Cameron and M. O'Connor, eds.), Vol. 15, p. 84. Little, Brown, Boston, Massachusetts. Lang, F. J. (1925). Virchows Arch. 257, 246. Lazarus, S. S., and Volk, B. W. (1961). Arch. Pathol. 7 1 , 44. Lazarus, S. S., and Volk, B. W. (1962). "The Pancreas in H u m a n and Experimental Diabetes." Grune & Stratton, New York. LeCompte, P. M. (1958). Arch. Pathol. 66, 450. LeCompte, P. M. (1960). "Diabetes" (R. H. Williams, ed.), p. 309. Harper (Hoeber), New York. LeCompte, P. M., Steinke, J., Soeldner, J. S., and Renold, A. E. (1966). Diabetes 15, 586. Legg, M. A. (1966). Quoted in: Warren, S., LeCompte, P. M., and Legg, M. A.: "The Pathology of Diabetes Mellitus" p. 60. Lea & Febiger, Philadelphia, Pennsylvania. Levine, R. (1976). N. Engl. J. Med. 294, 494. Like, A. A. (1967). Lab. Invest. 16, 937. Long, C. N. H., and Lukens, F. D. W. (1936). J. Exp. Med. 6 3 , 465. Maclean, N., and Ogilvie, R. F. (1955). Diabetes 4 , 367. Maclean, N., Robertson, F., and Ogilvie, R. F. (1959). Diabetes 8, 83. Maimon, S. N., Kirsner, J. B., and Palmer, W. L. (1948). Arch. Int. Med. 8 1 , 56. Mallory, F. B. (1914). "The Principles of Pathologic Histology." Saunders, Philadelphia, Pennsylvania. Mancini, A. M., Zampa, G. A., Vecchi, A., and Costanzi, G. (1965). Lancet 1, 1189. Marble, A. (1971). i n "Joslin's Diabetes Mellitus" (A. Marble, P. White, R. Bradley, and L. P. Krall, eds.), p. 1. Lea & Febiger, Philadelphia, Pennsylvania. Martius, K. (1915). Frankfurt. Z. Pathol. 17, 276. Milne, L. S., and Peters, H. L. (1912). J. Med. Res. 26, 405. Moore, R. A. (1936). Am. J. Dis. Child 52, 627. Moritz, A. R., and Oldt, M. B. (1937). Am. J. Pathol. 13, 679. Moschcowitz, E. (1957). Ann. Int. Med. 34, 1Ï37. N a k a m u r a , N. (1924). Arch. Pathol. Anat. 253, 286. Neumann, F. (1929). In "Handbuch der spez. Path. u. path. Anat." (F. Henke and O. Lubarsch, eds.), Vol. 5/2, p. 689. Springer-Ver lag, Berlin. Ogilvie, R. F (1933). J. Pathol. Bacteriol. 37, 473. Ogilvie, R. F (1944). Edinburgh. Med. J. 5 1 , 460.

364

Bruno W. Volk and Klaus F. Wellmann

Ogilvie, R. F. (1964). In "Aetiology of Diabetes and Its Complications" (M. P. Cameron and M. O'Connor, eds.), Vol. 15, p. 69. Little, Brown, Boston, Massachusetts. Ohlmacher, J. C. (1904). Am. J. Med. Sei. 128, 287. Opie, E. L. (1900-1901a). J. Exp. Med. 5, 397. Opie, E. L. (1900-1901b). J. Exp. Med. 5, 527. Opie, E. L. (1910). "Diseases of the Pancreas," p. 317. Lippincott, Philadelphia, Pennsylvania. Otani, S. (1927). Am. J. Pathol. 3, 1. Pav, J., Jezkova, Z., and Skrha, F. (1963). Lancet 3, 221. Popper, H. (1952). Gastroenterology 19, 183. Porta, E. A., Yerry, Y. R., and Scott, R. F. (1962). Am. J. Pathol. 4 1 , 623. Potter, N. B., and Milne, L. S. (1911). Am. J. Med. Sei. 143, 46. Priesel, R. (1922). Frankfurt. Z. Pathol. 26, 453. Renold, A. E., Soeldner, J. S., and Steinke, J. (1964). In "Aetiology of Diabetes and Its Complications" (M. P. Cameron and M. O'Connor, eds.), p. 122. Little, Brown, Boston, Massachusetts. Robbins, S. L. (1957). "Textbook of Pathology." p. 870. Saunders, Philadelphia, Pennsylvania. Sakurai, H., Dobbs, R. E., and Unger, R. H. (1974). J. Clin. Invest. 54, 1395. Saltykow, O. (1909). Cor.-Bl. f. Schweiz. Aerzte 39, 625. Sauerbeck, E. (1902). Erg. Pathol. Pathol. Anat. 8, 538. Sauerbeck, E. (1904). Virchows Arch. 177, 1. Schmidt, M. B. (1959). Munch. Med.Wchschr. 49, 51. Schwartz, P. (1965). Trans. N.Y. Acad. Sei. 27, 393. Schwartzman, J., Crusius, M. E., and Beirne, D. P. (1947). Am. J.Dis. Child. 74, 587. Seifert, G. (1954). Virchows Arch. 325, 379. Seifert, G. (1959). Verh. Dtsch. Ges. Pathol. 18, 50. Seyfarth, C. (1920a). München. Med. Wchnschr. 67, 617. Seyfarth, C. (1920b). Verh. Kongr. Inn. Med. 32, 178. Simmonds, F. (1912). Dtsch. Med. Wochenschn. 38, 1020. Ssobolew, L. W. (1902). Virchows Arch. 168, 91. Stansfield, O., and Warren, S. (1928). N. Engl. J. Med. 198, 686. Susman, W. (1942). J. Clin. Endocrinol. 2, 97. Symmers, D. (1909). Arch. Int. Med. 3, 379. Terbrüggen, A. (1947). Klin. Wochenschr. 24-25, 434. Terbrüggen, A. (1948). Virchows Arch. A 315, 407. Terplan, K. (1966). Personal communication. Thoinot, L., and Delamare, G. (1907). Arch. Med. Exp. Anat. Pathol. 19, 176. Toreson, W. E. (1951). Am. J. Pathol. 27, 327. Toreson, W. E., Lee, J. C , and Grodsky, G. M. (1968). Am. J. Pathol. 52, 1099. Unger, R. H., and Orci, L. (1975). Lancet 1, 14. Vallance-Owen, J. (1964). Diabetes 13, 241. Van Beek, C. C. (1939). Ned. Tijdschr. Geneesk. 83, 646. Vartiainen, I. (1944). Ada Med. Scand. 118, 536. Volk, B. W., and Lazarus, S. S. (1963). Diabetes 12,162. Volk, B. W., and Lazarus, S. S. (1964). Diabetes 13, 60. Volk, B. W., and Wellmann, K. F., eds. (1977). "The Diabetic Pancreas." Plenum, New York. von Halasz, A. (1909). Wien Klin. Wochenschr. 22, 1481. von Mering, J., and Minkowski, O. (1889-1890). Arch. Exp. Pathol. Pharmakol. 26, 371. von Meyenburg, H. (1940). Schweiz. Med. Wochenschr. 70, 554.

The Pancreas in Idiopathic Diabetes

365

von Meyenburg, H. (1946). Schweiz. Med. Wochenschr. 76, 207. Warren, S. (1927). J. Am. Med. Ass. 88, 99. Warren, S. (1938). "The Pathology of Diabetes Mellitus." Lea & Febiger, Philadelphia, Pennsylvania. Warren, S., and LeCompte, P. M. (1952). "The Pathology of Diabetes Mellitus." Lea & Febiger, Philadelphia, Pennsylvania. Warren, S., LeCompte, P. M., and Legg, M. A. (1966). "The Pathology of Diabetes Mellitus." Lea & Febiger, Philadelphia, Pennsylvania. Weichselbaum, A. (1908). Sitzungsber. Akad. Wiss. Wien. Math.-Naturwiss. KL, 117, 211. Weichselbaum, A. (1910). Sitzungsber. Akad. Wiss. Wien. Math-Natur wiss. KL, 119, 73. Weichselbaum, A. (1911). Wien. Klin. Wochenschn. 24, 153. Weichselbaum, A., and Stangl, E. (1901). Wien. Klin. Wochenscher. 14, 968. Weichselbaum, A., and Stangl, E. (1902). Wien. Klin. Wochenscher. 15, 969. White, P. (1932). "Diabetes in Childhood and Adolescence." Lea & Febiger, Philadelphia, Pennsylvania. White, P. (1960a). Diabetes 9, 345. White, P. (1960b). In "Diabetes" (R. H. Williams, ed.), p. 381. Harper (Hoeber), New York. Wilder, R. M. (1926). South. Med. J. 19, 241. Wrenshall, G. A., Bogoch, A., and Ritchie, R. C. (1952). Diabetes 1, 87. Wrenshall, G. A., Hartroft, W. S., and Best, C. H. (1954). Diabetes 3, 444. Williamson, J. R., and Lacy, P. E. (1961). Arch. Pathol. 72, 637. Wright, A. W. (1927). Am. J. Pathol. 3 , 461. Young, F. G. (1937). Lancet 2, 372.

SUAD EFENDIC,* TOMAS HOKFELT,* and ROLF LUFT*

Somatostatin I. Introduction II. Isolation and Characterization III. Localization and Action A. Hypothalamus B. Release of Growth Hormone from the Hypophysis C. Release of Other Hormones from the Anterior Pituitary D. Posterior Pituitary E. Extrahypothalamic Brain Areas F. Spinal Cord, Spinal Ganglia, and Peripheral Nerves G. Endocrine Pancreas H. Gastrointestinal Tract I. Thyroid Gland J. Liver K. Other Tissues IV. Mode of Action V. Clinical Applications A. Acromegaly B. Nelson's Syndrome C. Spontaneous Hyperinsulinemia D. Glucagon-Producing Tumor E. Gastrin-Producing Tumor F. Diabetes Mellitus VI. Analogs of Somatostatin VII. Side Effects and Toxicity of Somatostatin VIII. Conclusions References

368 368 369 370 377 378 378 380 381 381 392 397 397 401 401 404 404 405 405 405 406 406 408 410 411 412

* Departments of Endocrinology and Histology, Karolinska Hospital and Institute, and the Research Department of the Kabi Group, Stockholm, Sweden. 367

368

Su ad Efendic, Tomas Hokfelt, and Rolf Luft

I. Introduction The secretion of hormones by the hypophysis is regulated by specific hypothalamic hormones, the releasing and inhibitory hormones or factors. Out of these hypophysiotropic hormones, those regulating the secretion of thyrotropin (thyrotropin-releasing hormone, TRH) and luteinizing hormone (luteinizing-hormone-releasing hormone, LH-RH) have been isolated and synthesized. The evidence for the presence of growth-hormone-releasing hormone (GH-RH or GRF) in semipurified hypothalamic extracts is now substantial (Malacara et al., 1972). The chemical nature of GRF has not been established. It was while searching for the real GH-RH t h a t Guillemin and coworkers noted the powerful and consistent inhibitory activity on the secretion of GH of some of their crude hypothalamic extracts. They realized the significance of a possible GH release-inhibiting factor (GIF) or activity, an activity t h a t had previously been observed in the median eminence region of rats and sheep by McCann and co-workers (Krulich et al., 1968, 1971; Dhariwale* al., 1969; Krulich and McCann, 1969). Using the in vitro secretion of GH by pituitary monolayer cultures as an assay system, they isolated GIF from ovine hypothalamus (Brazeau et al., 1973). It turned out to be a tetradecapeptide, and it was named somatostatin. This unexpected finding attracted even more attention when it was found t h a t somatostatin—or somatostatin-like peptides—was present outside the hypothalamus: in other parts of the central nervous system, in the pancreas, gastrointestinal tract, and thyroid gland.

II. Isolation and Characterization Using ovine hypothalamic fragments, Brazeau et al. (1973) found t h a t alcohol-chloroform extracts in minute amounts significantly inhibited the secretion of GH from primary cultures of dissociated r a t anterior pituitary cells. The measurements of the GH secreted were done on the incubation fluids by a radioimmunoassay. The purification procedure for somatostatin is shown in Table I (taken from Burgus et al., 1973a). The original purification sequence was started on about 500,000 lyophilized ovine hypothalamic fragments. As little as 0.0015 fragments per milliliter of the initial extract inhibited GH release by approximately 50%. This dose corresponded to 6 ^g dry weight per milli-

369

Somatostatin TABLE I

SEQUENCE OF PURIFICATION OF SOMATOSTATIN0

Stage Lyophilized ovine hypothalami Alcohol-chloroform extraction Ultrafiltration diaflow membrane Partitioning in 11:5:3, 0.1% HOAc:w-BuOH:pyridine Partitioning in 4:1:5, n-BuOH:HOAc:H20 Ion exchange chromatography on CMC Gel filtration, Sephadex G-25, 0.5 M HOAc Partition chromatography, Sephadex G-25, 4:1:5, n-BuOH:HOAc:H20

1. 2. 3. 4. 5. 6. 7. 8.

Number of fragments

Weight

Units/mg*'0

Total units x 103c

490,000 490,000 269,000

36.8 kg 2.2 kg 126.5 g

ND 0.2 ND

ND 490 ND

490,000

17.2 g

5.3

56

490,000

9.1g

7.6

490,000

30.9 mg

214

7

490,000

12.4 mg

421

5

490,000

8.5 mg

505

4

6.9

a

Burgus et al. (1973a). A unit of somatostatin activity is defined as the biological activity of one ovine hypothalamic fragment at stage 2. c ND, not determined. b

liter of medium. After stage 6 in the purification sequence, a fraction inhibited GH secretion at 20 ng/ml by more than 50%. The last step gave a peptide that accounted for 77% of the weight of the preparation. This peptide was submitted to stepwise Edman degradation for determination of its sequence. The primary structure of isolated ovine somatostatin was found to be H-Ala-Gly-Cys-Lys-Asn-Phe-Phe-Trp-Lys-Thr-Phe-Thr-Ser-Cys-OH

The linear tetradecapeptide was synthesized by solid-phase methodology and purified by gel filtration in the presence of 2-mercaptoethanol. After purification, the synthetic peptide had the biological activity of the native somatostatin.

III. Localization and Action Our present knowledge of the cellular localization of somatostatin in tissues is based on immunohistochemical studies. Several research groups have been involved in such studies, partly using the antisera

370

Suad Efendic, Tomas Hökfelt, and Rolf Luft

from the same laboratory. Thus, the American, Canadian, Hungarian, English, and our own Swedish groups have worked with Arimura's antiserum (Arimura et al., 1975a) whereas the French and Swiss groups have used an antiserum prepared by Dubois (1975). The immunohistochemical work on somatostatin localization has recently been reviewed (Parsons et al., 1976; Pelletier et al., 1976) and a more detailed analysis of the distribution patterns in the central nervous system will appear elsewhere (Johansson et al., in preparation). The results are generally in good agreement, although differences exist as to the extent of the distribution of somatostatin-positive cell systems. Such differences may, however, be explained by the different immunohistochemical techniques employed (Johansson and Eide, in preparation). The immunohistochemical results are also in general agreement with radioimmunological determinations of somatostatin (Brownstein et al., 1975; Arimura et al., 1975b; see also Vale et al., 1975; Eide et al, 1977). The immunoreaction for somatostatin described by the various authors is said to be specific in the sense that it is abolished by absorption of the antiserum with somatostatin. It must, however, be emphasized that the antiserum may cross-react with unknown somatostatin-like peptides or larger protein molecules containing somatostatin-like amino acid sequences. It may be added that this situation is, of course, also true for radioimmunological studies. Therefore, in many instances the expression somatostatin-like immunoreactivity (SLI) may be preferable.

A. Hypothalamus In the hypothalamus both somatostatin-positive cell bodies (Dubois and Kolodziejczyk, 1975; Eide and Parsons, 1975; Hökfelt et al, 1975a; Alpert et al, 1976) and fiberlike structures (Dubois et al, 1974; Hökfelt étal, 1974, 1975a,f; Oubé et al., 1975; King et al, 1975; Pelletier et al, 1975; Sétâlo et al., 1975) were found. The neuronal nature of these fibers has definitely been proven in an elegant ultrastructural immunohistochemical analysis by Pelletier et al. (1974), clearly demonstrating that somatostatin is localized to granular vesicles in nerve endings. We observed fluorescent cell bodies in the peri ventricular area at the level of the suprachiasmatic nucleus extending at least 0.5 mm in the

Somatostatin

371

caudal direction (Figs. 1 and 2) (Hökfelt et al., 1975a). The cell bodies were seen on both sides of the ventricle along its entire height but did not enter the paraventricular nucleus. The number of fluorescent cell bodies was approximately 25-50 per section as a maximum. Similar results were obtained by Eide and Parsons (1975) and Alpert et al. (1976). Parsons et al. (1976) have reported somatostatinimmunoreactive cell bodies also in the suprachiasmatic nucleus, the preoptic suprachiasmatic nucleus, and the arcuate nucleus. Dubois and Kolodziejczyk (1975) found somatostatin-positive cell bodies in the nucleus paraventricularis and nucleus supraopticus. Strongly fluorescent fibers were present in the external layer of the median eminence extending from its very beginning all the way down into the stalk (Figs. 3-5, 7, and 8). There was a general impression of an increase in the number of fibers in the caudal direction. The positive fibers were distributed both in the lateral and in the medial parts of the median eminence. In the stalk, the entire external layer was covered with a dense plexus of fluorescent fibers (Fig. 5), whereas at this level the overlying floor of the basal hypothalamus contained only single positive fibers. Also in the internal layer, especially in the medial part, numerous positive fibers were observed, often distinctly surrounding blood vessels (Fig. 8). Rather dense plexuses of positive fibers were observed in the ventromedial (Fig. 9) and the arcuate (Fig. 10) nuclei extending caudally into the ventral premammillary nucleus. The suprachiasmatic nucleus contained a dense plexus of fluorescent fibers. Single positive fibers were also seen in the periventricular region and along the basal surface of the hypothalamus. A more detailed description of SLI in the hypothalamus will be given elsewhere (Johansson et al., in preparation). The somatostatin-positive cell bodies in the periventricular region of the anterior hypothalamus are the origin of the somatostatin-positive fiber plexuses in the median eminence, whereas the cell bodies of the terminal networks, e.g., in the arcuate and ventromedial nuclei, seem to have a different localization. This has been established in recent studies with hypothalamic lesions (Eide et al., in preparation). Brownstein et al. (1975) measured the immunoreactive somatostatin and LH-RH in various nuclei of the rat hypothalamus. The concentrations of immunoreactive somatostatin in the extracts of each nucleus are shown in Table II. The highest concentrations of somatostatin, as in the immunofluorescence studies, were found in the median eminence and arcuate nucleus. Considerable amounts were also obtained in the ventromedial, periventricular, and suprachiasmatic nuclei.

372

Suad Efendic, Tomas Hökfelt, and Rolf Luft

FIGS. 1 AND 2. Immunoftuorescence micrographs of the anterior periventricular area of the hypothalamus after incubation with somatostatin antiserum. Numerous somato statin-positive cell bodies are seen close to the third ventricle (asterisk). In the higher magnification (Fig. 2) varicose cell processes (arrows) can also be seen. Magnification 160x and 400x, respectively.

Somatostatin

373

FIGS. 3 AND 4. Immunofluorescence micrographs of consecutive sections of the median eminence after incubation with somatostatin (Fig. 3) and LH-RH (Fig. 4) antisera. Somatostatin-positive nerve terminals are seen in the entire external layer and in the medial parts of the internal layer, whereas the LHRH-positive fibers are seen in the lateral parts of the external layer. Asterisks indicate third ventricle. From Hôkfelt et al. (1975f). Magnification 160x.

374

Suad Efendic, Tomas Hökfelt, and Rolf Luft

FIGS. 5 AND 6. Immunofluorescence micrographs of consecutive sections of the basal hypothalamus and the stalk after incubation with somatostatin (Fig. 5) and LH-RH (Fig. 6) antisera. A dense plexus of somatostatin-positive fibers is seen in the entire external layer of the stalk, whereas only a few LH-RH-positive fibers are seen in this area. Note the numerous LH-RH-positive fibers in the basal hypothalamus below the third ventricle (asterisk). From Hökfelt et al. (1975f). Magnification 160x.

FIGS. 7-10. Immunofluorescence micrographs of the hypothalamus after incubation with somatostatin antiserum. Figures 7 and 8 show a higher magnification of the lateral and medial part of the median eminence, respectively. Somatostatin-positive fibers are present in the ventromedial (Fig. 9) and arcuate nuclei (Fig. 10). Asterisks indicate third ventricle. From Hökfelt et al. (1975a). Magnifications 330x (Figs. 7 and 8) and 400x (Figs. 9 and 10).

376

Suad Efendic, Tomas Hökfelt, and Rolf Luft TABLE II DISTRIBUTION OF SOMATOSTATIN AMONG N U C L E I OF THE HYPOTHALAMUS 0 - 0

ng/mgprotein ± SEM Medial preoptic nucleus Periventricular nucleus Suprachiasmatic nucleus Supraoptic nucleus Anterior hypothalamic nucleus Lateral anterior nucleus Paraventricular nucleus Arcuate nucleus Ventromedial nucleus Dorsomedial nucleus Perifornical nucleus Lateral posterior nucleus Ventral premamillary nucleus Dorsal premamillary nucleus Posterior hypothalamic nucleus Median eminence a b

10.4 23.7 8.0 3.2 8.6 4.9 4.4 44.6 14.6 5.4 3.8 3.5 17.3 4.3 3.8 309.1

± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ±

2.5 9.0 0.6 0.6 1.5 1.1 1.8 6.1 2.1 2.1 0.7 0.7 4.4 0.7 0.8 60.8

Brownstein et al. (1975). Means and standard errors (SEM) of six separate determinations are presented.

The presence of a dense somatostatin innervation of the median eminence is in good agreement with the postulated role of somatostatin as the inhibitory hormone of growth hormone release and as a humoral factor transported in the portal vessels to the anterior pituitary. On the other hand, the existence of a rather dense plexus of somatostatin containing probable nerve endings in the ventromedial, arcuate, and ventral premammillary nuclei deserves some comment. The localization of these nerve endings does not seem to indicate a release into blood vessels but sooner a release from nerve endings at synapses, and could thus indicate a transmitter or modulator role for somatostatin. In fact, recently Renaud et al. (1975) have been able to demonstrate a depressant action of somatostatin on the activity of neurons in various regions of the brain. The same authors discuss the very interesting possibility that one and the same hypothalamic neuron may send axon collaterals both to the external layer of the median eminence and to some other hypothalamic and extrahypothalamic loci (Martin et al., 1975; Renaud and Martin, 1975). Thus the possibility exists that one neuron may liberate a certain substance, e.g., somatostatin, in one case as a hormone into the blood, and in the other case as a neurotransmitter (see Nicoll and Barker, 1971; Renaud and Martin 1975; Renaud et al., 1975). Such a mechanism would hypothetically imply the possibil-

Somatostatin

377

ity of a double control of growth hormone secretion: first, by somatostatin released into the portal vessels acting at the level of the pituitary gland, and second, by an inhibitory action of somatostatin released at synapses on hypothetical GH-RH producing cell bodies, e.g., in the ventromedial nucleus. This hypothesis is, however, not directly supported by our recent results mentioned above that the median eminence and ventromedial nucleus terminals seem to have a different origin (Eide et al., in preparation). The distribution of somatostatin in the rat differs clearly from that of LH-RH (see Hökfelt et al., 1975f). Whereas somatostatin occurs both in the lateral and medial part of the median eminence (Fig. 3), LH-RH is in the major part confined to the lateral median eminence (Fig. 4). Furthermore, in the stalk only a few LH-RH-positive fibers were seen (Fig. 6), whereas SLI covered the entire surface of the stalk (Fig. 5). TRH is localized mainly in the medial parts of the median eminence (Hökfelt et al, 1975c,d). B. Release of Growth H o r m o n e from the Hypophysis The effect of somatostatin on several aspects of GH secretion has been studied in animals and man. Thus, in vivo somatostatin suppressed stimulated GH secretion in rats (Brazeau et al., 1973, 1974a; Kato et al., 1974; Martin, 1974), in dogs (Lovinger et al., 1974), in baboons (Ruch et al., 1973), and in man (Hall et al., 1973; PrangeHansen etal., 1973; Silerei al., 1973; Parker etal., 1974; Peracchi et al., 1974; Yen et al, 1974; Giustina et al., 1975; Köbberling et al., 1976). The stimuli used for this purpose were as different as arginine, L-dopa, barbiturates, morphine, chlorpromazine, insulin-induced hypoglycemia, exercise, sleep, meals, electrical stimulation of the hypothalamus or amygdala, or catecholamine infusion in the third ventricle. Somatostatin probably also suppresses basal GH release, since there is usually a rebound on discontinuation of somatostatin infusion (Hall et al., 1973; Prange-Hansen et al., 1973; Siler et al., 1973; Gerich et al., 1974b; Mortimer et al., 1974; Leblanc et al., 1975a). The inhibitory effect of somatostatin on growth hormone release was also demonstrated in in vitro studies on rat pituitary cell cultures (Vale et al., 1972; Borgeat et al., 1974a) and on perifused rat pituitary explants (Stachura, 1975). In diseases associated with elevated basal growth hormone levels, such as acromegaly (see p. 404), diabetes (Prange-Hansen et al., 1973; Gerich et al., 1975a; Ward et al., 1975), protein-caloric malnutrition

378

Suad Efendic, Tomas Hökfelt, and Rolf Luft

(Pimstone et al., 1975a), and renal and liver disease (Pimstone et al., 1975b), suppression has been readily achieved with somatostatin. C. Release of Other Hormones from the Anterior Pituitary In their studies on the action of somatostatin, Vale et al. (1974) observed in rats that somatostatin also inhibited in vivo and in vitro the secretion of TSH (thyroid-stimulating hormone) stimulated by TRF. It did not affect the secretion of prolactin concomitantly stimulated by TRF. This was later confirmed in humans (Weeke et al., 1974; Siler et al, 1974; Carr et al, 1975). Furthermore, somatostatin did not affect the secretion of prolactin in rats stimulated by suckling (Chen et al., 1974) or chlorpromazine (Sawano et al., 1974), but potentiated prolactin release induced by periphenazine (Gala et al., 1976). Somatostatin did not affect basal TSH release in man in the experiments of Siler et al., which were performed during daytime when the TSH level is known to be low. On the other hand, Weeke et al. (1975) significantly suppressed the high night levels of serum tyrotropin with somatostatin. Furthermore, the high TSH levels in subjects with primary hypothyroidism were lowered by the infusion of somatostatin (Lucke et al., 1975). These experiments, as well as those of Florsheim and Kozbur (1976) showing an enhanced release of TSH in rats when TRF was given together with somatostatin antibodies, suggest that somatostatin is involved in the control of TSH release. In normal man, somatostatin did not affect the secretion of LH (luteinizing hormone) and FSH (follicle-stimulating hormone) (Hall et al, 1973; Siler et al, 1974; Yen et al, 1974), nor of ACTH (Hall et al, 1973; Tyrell et al., 1975). On the other hand, it readily suppressed the increased ACTH levels in patients with primary adrenal insufficiency (Fehm et al., 1976). Atypical responses to somatostatin occurred in some patients with pituitary tumors: it lowered fasting prolactin levels in patients with acromegaly (Yen et al., 1974) and ACTH levels in patients with Nelson's syndrome (Tyrell et al., 1975). In contrast, Hall et al. (1973) did not observe any effect by somatostatin on prolactin secretion in their acromegalic patients. D. Posterior Pituitary A sparse plexus of somatostatin-positive varicose fibers was seen in the posterior pituitary (Figs. 11-13). They disappeared after hypothala-

Somatostatin

379

FIGS. 11-13. Immunofluorescence micrographs of the posterior pituitary (Fig. 11) and colon (Figs. 12 and 13) after incubation with somatostatin antiserum. Somatostatinpositive fibers are seen in the posterior pituitary (Fig. 11), around the basal p a r t s of the crypts (c) (Fig. 12), and in the myenteric plexus (arrows) (Fig. 13) of the colon, m, circular muscle layer. From Hökfelt et al. (1975a,b,e, respectively). Magnifications 400x, 300x, and 330x, respectively.

380

Suad Efendic, Tomas Hökfelt, and Rolf Luft

mic lesions (Hökfelt et al., unpublished data). This may indicate the existence of a hypothalamo-hypophyseal neurosecretory system in addition to the classical oxytocin and vasopressin containing neurons. However, quantitatively this system seemed to be considerably less extensive than the two classical ones and could, in fact, simply represent some stalk fibers extending into the proximal part of the neurohypophysis. E. Extrahypothalamic Brain Areas Somatostatin-positive fibers have been observed in several extrahypothalamic brain areas. Dubé et al. (1975) found SLI in the subfornical and subcommisural organs, in the organum vasculosum, and in the pineal gland, while Sétâlo et al. (1975) observed SLI in the periventricular gray. We have observed SLI in these same places except for the pineal gland and, in addition, in the amygdaloid complex, the tuberculum olfactorium, the nucleus accumbens, and in many other brain areas, including the lower brain stem (Johansson et al., in preparation). Numerous somatostatin-positive cell bodies are also present in extrahypothalamic brain areas including the entopenduncular nucleuszona incerta region, the amygdaloid complex, theprepiriform cortex, the hippocampus, the neocortex, and the caudate nucleus. These findings indicate that somatostatin, in contrast to many other peptides, distributes to cortical areas. This is of interest in view of some recent physiological studies linking somatostatin effects to neocortical and hippocampal areas (Rezek et al., 1976a,b). Using extracts of different zones of the rat brain and assaying them in vitro for somatostatin-like activity, Vale et al. (1975) found such activity in each of the five zones examined (Table III). TABLE III

NET SOMATOSTATIN-LIKE ACTIVITY (SLA) OF GENERAL BRAIN ZONES"

a

Zone

SLA (ng/mg wet weight)

SLA (ng/zone)

Hypothalamus Midbrain + thalamus Cerebral cortex Pons + medulla Cerebellum

4.7 2.4 2.4 2.5 0.2

166 540 2500 490 48

Vale et al. (1975).

381

Somatostatin F. Spinal C o r d , Spinal Ganglia, and P e r i p h e r a l

Nerves

There is evidence t h a t somatostatin or a somatostatin-like peptide was present in some primary sensory neurons (Hökfelt et al., 1975b, 1976). In spinal ganglia at the cervical, thoracic, and lumbar level several neuronal cell bodies exhibited a somatostatin-positive immunoftuorescence (Figs. 14 and 15). These cell bodies were almost exclusively of the small type (Andres, 1961). Very preliminary calculations indicated t h a t the somatostatin-positive cell bodies constituted less than 10% of all neuronal cell bodies of these ganglia. Also, in the trigeminal ganglion a few somatostatin-positive cell bodies were observed. In the spinal cord, a dense plexus of somatostatin fiber- and dotlike structures was seen, mainly in the substantia gelatinosa of the dorsal horns (Figs. 16 and 17) and in the adjacent part of the fasciculus lateralis. No such fibers were observed in the ventral horn or around the central canal. In peripheral tissues a sparse network of fluorescent fibers was observed in the various layers of the wall of the gut, especially the colon (Figs. 12 and 13) (see below). The somatostatin-positive fibers in the dorsal horn of the spinal cord probably arise from the somatostatin-positive cell bodies in the spinal ganglia, since rhizotomy markedly decreases the number of somatostatin-positive fibers in the superficial laminae. Further investigations are necessary to trace the origin of the SLI-containing fibers in the gut. These results indicate the existence of a population of primary sensory neurons containing somatostatin or a somatostatin-like peptide. In parallel experiments with antibodies to substance P it could be established t h a t the somatostatin-positive neurons and the previously described substance P-positive neurons (Hökfelt et al., 1975g,h) are not the same (Hökfelt et al., 1976). Thus up to now two subpopulations of neurons have been identified in spinal ganglia on a histochemical basis, one containing substance P and the second containing somatostatin. The majority of these cell bodies seem to belong to the small-sized cells in these ganglia (see Andres, 1961). G. Endocrine

Pancreas

It came as a surprising finding t h a t somatostatin also inhibited the secretion of insulin and glucagon in animals and m a n (see below). The fact t h a t very small quantities of the hormone were required for this purpose suggested a physiological role for somatostatin in the régula-

382

Suad Efendic, Tomas Hökfelt, and Rolf Luft

FIGS. 14 AND 15. Immunofluorescence micrographs of spinal ganglia after incubation with somatostatin antiserum. Some weakly fluorescent somatostatin-positive cell bodies are seen (arrows). Arrows in Fig. 15 point to some structures of the connective tissue capsule exhibiting unspecific fluorescence. From Hökfelt et al. (1976). Magnifications 160x and 400x, respectively.

Somatostatin

383

FIGS. 16 AND 17. Immunofluorescence micrographs of the dorsal horn of rat spinal cord after incubation with somatostatin antiserum. Figure 17 represents a higher magnification of part of Fig. 16, indicated by the rectangle. Densely packed, dotlike structures, probably representing nerve endings, are seen mainly in the substantia gelatinosa. Arrow in Fig. 16 points dorsally, double arrow laterally. From Hökfelt et al. (1976). Magnifications 160x and 400x, respectively.

384

Suad Efendic, Tomas Hökfelt, and Rolf Luft

tion of endocrine pancreatic functions (Efendic et al., 1974). For this reason we raised the question whether somatostatin might be produced in the pancreas. Soon afterwards it was established t h a t this was the case (Luft et al., 1974; Dubois, 1975; Dubois et al., 1975; Goldsmith et al., 1975; Hökfelt et al., 1975a; Orci et al., 1975a; Polak et al., 1975; Rufener et al., 1975a; Erlandsen et al., 1976; Parsons et al., 1976). The type of islet cell containing SLI was identified as Aj or D cells by electron microscopic immunohistochemistry (Goldsmith et al., 1975; Rufener et al., 1975a), by correlative immunofluorescence and silver staining (Hökfelt et al., 1975a; Polak et al. 1975), and by consecutive staining with antisera to somatostatin, glucagon, and insulin (Orci et al, 1975a). The somatostatin-positive structures were observed mainly in the peripheral parts of practically all islets (Figs. 18 and 21). Most immunoreactive structures could be identified as cells with a strong cytoplasmic fluorescence and a nonfluorescent nucleus, but fluorescent extensions of the cytoplasm were observed, often in direct connection with the cell bodies. In the rat, the positive elements were always localized to the periphery of the islets and occasionally formed a complete ring around the islets. In the guinea pig they were spread all over the islets. Single positive cells were seen, especially among the epithelial cells of the small ducts, in the wall of the large excretory duct, and occasionally also among the epithelial cells of this duct. The cellular elements described above were seen to the same extent after incubation with somatostatin antiserum pretreated with glucagon, but not after incubation with control serum. On the other hand, small structures in the connective tissue around the m a i n excretory duct contained some fluorescent elements, which were also observed after incubation with control serum or with FITC serum alone. Thus, they represented unspecifically stained structures. Positive cells were also observed in the periphery of islets after incubation with glucagon antibodies. Their distribution was similar to t h a t of the somatostatin-positive cells. However, there seemed to be more glucagon-positive t h a n somatostatin-positive cells. Identity of the two cell types could neither be stated nor excluded merely by comparing consecutive sections stained with antibodies to somatostatin and glucagon, respectively. By comparing photographs from the same section, stained consecutively with somatostatin and glucagon antisera, it could be clearly demonstrated t h a t a larger number of cells were stained with the two antisera t h a n with somatostatin or glucagon antiserum alone. (Fig. 19). To pursue the question of the identity of the somatostatin positive cells further, we combined the immunofluorescence procedure with

Somatostatin

385

FIGS. 18 AND 19. Immunofluorescence micrographs of the same section of a pancreatic islet, first incubated with somatostatin antiserum and photographed and thereafter with glucagon antiserum. The glucagon-positive cells are localized close to the somatostatinpositive ones in the periphery of the islet. Double arrows indicate some glucagon-positive cells, single arrows some somatostatin-positive ones. From Hökfelt et al. (1975a). Magnification 400x.

386

Suad Efendic, Tomas Hökfelt, and Rolf Luft

FIGS. 20 AND 21. Micrographs of the same section of a pancreatic islet processed according to the silver impregnation technique of Hellerström and Hellman (1969) (Fig. 20) after incubation with somatostatin antiserum and photography (Fig. 21). There is an absolute identity between somatostatin- and "silver"-positive cells. From Hökfelt et al. (1975a). Magnification 400x.

Sornatostatin

387

silver impregnation according to Heller ström and Hellman ( 1960). The latter technique allows a classification of the A cells of the islets into Ai and A2 cells, characterized by the presence and absence of argyrophilia, respectively. The two techniques were run consecutively on the same sections. As seen in Fig. 20, silver impregnation demonstrated numerous argyrophil cells with a distinctly blackened cytoplasm in the periphery of the islets. The argyrophilia was equally pronounced in untreated sections, or in sections previously incubated with sornatostatin antiserum (and FITC serum) alone or with the somatostatin-glucagon antisera (and FITC serum) combination. A careful comparison of photographs of immunopositive cells (Fig. 21) and silver-positive cells (Fig. 20) in one and the same section revealed that virtually all sornatostatin positive cells were silver positive. On the other hand, in most instances it could be decided that the glucagon cells lacked argyrophilia. The present findings of cells showing SLI in the pancreatic islets appears particularly noteworthy in view of the demonstration that sornatostatin exerts a direct inhibition of insulin and glucagon secretion (see below). Against this background it is tempting to suggest that the close morphological association between, on the one hand, the SLIcontaining cells [which were unequivocally identified as Aj cells (D cells)] and, on the other hand, the A2 (A cells) and B cells reflects a mechanism for local intraislet regulation of insulin and glucagon release. Indeed, a mechanism of this kind has been previously suggested by Hellman and Lernmark (1969a,b). These authors reported a marked in vitro inhibition of insulin release from isolated islets incubated in the presence of an extract obtained either from islets of obesehyperglycemic mice or from A r cell-rich islets from the duck. The authors suggested that gastrin was the active inhibitory principle, an assumption that at that time received support from observations by Potet et al. (1966) and Cavallero et al. (1967) of Αχ cells in gastrinproducing Zollinger-Ellison tumors of the pancreas. Later, Lomsky et al. (1969) and Greider and McGuigan (1970) reported the immunohistochemical identification of a gastrinlike substance in normal Ai cells of humans and other mammalian species. In addition, gastrinlike immunoreactivity was claimed to be present in pancreatic extracts (Greider and McGuigan, 1971; Nilssonei al., 1972). However, more recent studies by Greider et al. (1974) and Lostra et al. (1974) have failed to identify any gastrin cells either in ZollingerEllison tumors or in human islets, despite the use of the same immunofluorescent technique by which abundant gastrin-containing cells were identified in the antral mucosa. Also, in our laboratory antibodies to gastrin were tested in immunofluorescence studies on pan-

388

Suad Efendic, Tomas Hökfelt, and Rolf Luft

creatic sections, but no positive cellular elements were seen (Hökfelt and Nilsson, unpublished data). In addition, there is a marked disagreement as to the existence of any immunoreactivity toward gastrin in pancreatic extracts (Lostra et al., 1974), and nobody h a s so far been able to demonstrate gastrin in such extracts with bioassay (Gregory et al., 1960; Blair et al., 1969). In discussing the possibility of a local islet inhibition of the glucagon and/or insulin release, the distribution of the various cell types within the islets and the intraislet blood circulation must also be considered, and has recently also been discussed by Orci and collaborators (Orci and Unger, 1975; Orci et al., 1975c). In the rat, the A! and A2 cells appear so intimately juxtapositioned t h a t a transmitter-like mode of action of somatostatin seems possible. By contrast, the more centrally located B cells would be affected by somatostatin released into the islet capillaries. However, examination of the islet cell types in other vertebrate species indicates widely different distributions of the islet cells (Hellerström et al., 1964), which makes a local regulation of the kind discussed above less probable. For example, in man and guinea pig, the Ai, A2, and B cells are scattered all over the islets, while in the horse the A2 form a central core in the islets surrounded by A! and B cells (Hellerström et al., 1964). Furthermore, the microcirculation of the islets is probably unsuited for the conveyance of hormones from the islet periphery to its central part, since the afferent islet blood vessel penetrates into the center of the islet, where it breaks up into capillaries radiating toward the periphery (Fujita and Murakami, 1973). These capillaries subsequently form an islet-acinar portal system that, indeed, would favor the distribution of islet hormones among the exocrine rather than endocrine cells. Altogether therefore, the above suggestions of Orci et al. (1975c) of a local regulation of glucagon and insulin release by somatostatin or a somatostatin-like peptide must remain tentative. It seems more likely t h a t somatostatin is released diffusely into the interstitial fluid and thereby exerts its local or paracrine effects on the B cells as well as on the A2 cells. SOMATOSTATIN AND THE SECRETION OF INSULIN AND GLUCAGON

Somatostatin inhibited insulin release in man induced by glucose (Alberti et al., 1973; Mortimer et al., 1974; Efendic and Luft, 1975b; Efendic et al., 1976c) as well as by arginine, tolbutamide, and glucagon (Gerich et al., 1974a; Leblanc et al., 1975a; Efendic et al., 1976b,c). Furthermore, somatostatin suppressed glucose-, arginine-, tolbutamide-, isoproterenol-, and secretin-induced insulin release in baboons (Koerker et al., 1974; Chideckel et al., 1975).

389

Somatostatin

Figure 22 illustrates the effect in man of different glucose loads on insulin release in the presence and absence of somatostatin. Somatostatin in a low dose (0.5 μg/kg as a bolus followed by 0.5 μg/kg during 90 min) significantly inhibited insulin release accompanying the administration of the middle but not the low or high doses of glucose. On the basis of these experiments, the relationship between the blood glucose and plasma insulin levels was evaluated (Fig. 23). In the presence of somatostatin, the dose-response curve was shifted to the right, and the highest glucose load overcame the inhibitory action of somatostatin on the late insulin response. Thus, somatostatin in a very small

-40

0

40

80

120 -40

0

40

80

120 -40

0

40

80

120 min

FIG. 22. Effect of somatostatin on glucose-induced insulin release in man. Somatostatin was given between —30 and 60 min, glucose between 0 and 60 min. Three different doses of glucose were given. Filled circles and solid lines denote control experiments, open circles and broken lines denote those with somatostatin. Mean ± SEM of seven experiments. The significance of the paired differences between control and somatostatin experiments is denoted by *, p < 0.05; **, p < 0.01; and ***, p < 0.001. From Efendic et al (1976c).

390

Suad Efendic, Tomas Hokfelt, and Rolf Luft

0 -»

r-

100

1

200

i

400

1 ^

i

800 100

1

1

1

200 400 800 glucose mg/100ml

0 J

i

1

1

1

100 200 400 800 glucose (mgx minx mi')

FIG. 23. Modification of the glucose-insulin dose-response relationship by somatostatin. Early response corresponds to the 10 min values, late response to the 60 min values of Fig. 22. Total response relates the 0-60 min integrated incremental plasma insulin values to the integrated blood glucose concentrations. Symbols as for Fig. 22. From Efendic et al. (1976c).

dose seemed to exert a competitive inhibition of glucose-induced insulin release. In addition, somatostatin was shown to be a potent inhibitor of insulin release from the perfused isolated r a t and dog pancreas (Alberti et al., 1973; Efendic et al., 1974; Efendic and Luft, 1975a; Gerich et al., 1975b; Curry and Bennett 1974; Tamarit et al., 1974; Norfleet et al., 1975). It was effective on the r a t pancreas in as low a dose as 1 ng/ml of perfusate (Efendic et al., 1974). In contrast, considerably higher doses—500 ng/ml—were required for the inhibition of glucose-induced insulin release from collagenase-treated isolated rat islet (Efendic et al., 1975a). The sensitivity of such islets to somatostatin was restored by keeping the islets in culture for two days (Turcot-Lemay et al., 1975). In dose-response studies, the inhibition by somatostatin of the effect of glucose on insulin release from isolated rat islets could be overcome by a high concentration of the hexose (44.9 mM), suggesting competitive inhibition (Claro et al., 1977). Perfused rat islets were also rather insensitive to somatostatin (Norfleet et al., 1975). Fasting insulin levels were inhibited by somatostatin both in vivo and in the perfused pancreas (Christensen et al., 1974; DeVane et al., 1974; Mortimer étal., 1974; Leblanc et al., 1975a; Efendic et al., 1976a). In patients with acromegaly, usually with elevated basal insulin levels, somatostatin promptly and significantly inhibited insulin secretion (Besser et al., 1974a,b; Mortimer et al., 1974; Peracchi et al., 1974; Giustina et al., 1975).

Somatostatin

391

Similarly, somatostatin inhibited stimulated glucagon release in vivo in normal man (Gerich et al., 1974b; Mortimer et al., 1974; Leblanc et al., 1975a; Efendic et al., 1976a) and in experimental animals (Koerker et al., 1974; Chidekel et al., 1975), as well as from the perfused rat and dog pancreas (Gerich et al., 1975b; Iversen 1974; Norfleet et al., 1975; Efendic et al., 1976a). The stimuli used were hypoglycemia, arginine, isoproterenol, and theophyllamine. Basal glucagon release was also suppressed by somatostatin (Christensen et al., 1974; Mortimer et al., 1974; Gerich et al, 1975c; Efendic et al, 1976a). The inhibitory effect was as pronounced in insulin-dependent diabetic subjects (Gerich et al., 1974b; Mortimer et al., 1974; Ward et al., 1975; Lundbaek et al., 1976) and acromegalics (Besser et al., 1974b; Mortimer et al., 1974) as in normal subjects. In our hands, the inhibitory effect of somatostatin on glucagon and insulin release was of approximately the same magnitude (Efendic et al., 1976a). In the isolated perfused rat pancreas as little as 10 ng/ml of linear somatostatin significantly inhibited both insulin and glucagon release (Fig. 24). In divergence are the results by Gerich et al. (1975b) demonstrating a considerably stronger inhibitory action on glucagon than on insulin release. In man, about 200 μg during a 90 min infusion significantly suppressed arginine-stimulated release of insulin and glucagon, whereas 80 μg was without effect on either of the two pancreatic hormones (Fig. 25). At present, we have little information on the mechanisms controlling the release of somatostatin from the pancreatic A! cells. Recently, it was demonstrated that theophylline, 8-Br-cyclic AMP, norepinephrine (Barden et al., 1976a,b), as well as glucose and glucagon (Schauder et al., 1976) stimulated the release of immunoreactive somatostatin from isolated rat islets. The stimulatory effect of glucagon on somatostatin release was furthermore demonstrated in the perfused dog pancreas (Patton et al., 1976). Some data are available on the somatostatin content of the pancreas in diabetes. Thus, the somatostatin content was raised in rats (Orci et al., 1975a; Patel and Weir, 1976) and mice (Eide et al., unpublished) rendered diabetic with streptozotocin. On the other hand, somatostatin was decreased in the pancreas of spontaneously diabetic Chinese hamsters (Peterson et al., 1977) and mice (Patel et al., 1976). Also the stomach of the diabetic hamsters showed a decreased concentration of somatostatin. Although the precise relationship between the somatostatin-containing cells in the pancreas and those in the stomach is obscure, their similar alterations in the diabetic hamster favor the idea that these cells in some way may be functionally interrelated.

392

Suad Efendic, Tomas Hökfelt, and Rolf Luft

σ> 4000

-15-10-5 0 5 10 15 20 25 -15-10-5 0 5 10 15 20 25 -15-10-5 0 5 10 15 20 25min I

arginin« Smg/ml |

| arginirr 5mg/ml I llirwar somatostatin 1ng/mi~|

| arginine 5m g/rr ir somatostatin10ng/ml |

FIG. 24. Effect of somatostatin (1 and 10 ng/ml of perfusate) on arginine (5 mg/ml)induced insulin and glucagon release from the isolated perfused rat pancreas. Following isolation, the pancreases were equilibrated for 40 min with 0.8 mg/ml of glucose. The arginine stimulus was applied between 0 and 23 min, with somatostatin applied 7 min prior and during the arginine stimulation. Results are expressed as the mean ± SEM. From Efendic et al. (1976a).

H. Gastrointestinal Tract The gastrointestinal tract contains both somatostatin-positive cell bodies (Hökfelt et al., 1975a; Polak et al., 1975; Rufener et al., 1975b; Parsons et al., 1976) and nerve fibers (Hökfelt et al., 1975a,e, 1976). Somatostatin-positive cells were observed in the lamina propria between the gastric glands (Figs. 26 and 27). Mostly these cells had a characteristic elongated shape with short thick processes often running in parallel to the gland lumen. The cells were for the most part localized within the lower half of the mucosa of the antrum, the estimated ratio between gastrin-secreting and somatostatin cells being 7:1 (Polak, 1976).

393

Somatostatin 200

100 \

i

1

1

1

1

1

g 100 n.

Jv-^s

3 so H C

0

J

g 300 a c ET 200 "5> 100

0

J

-40 -20 I

0

20 40 60

largininé somatostatin

|

40 -20 0

20 40 60 min

|arginine somatostatin

|

FIG. 25. Effect of somatostatin on arginine-induced insulin and glucagon release in man. The left-hand side summarizes data with a small dose of somatostatin (80 μ% over 90 min), while the right-hand side gives data with a larger dose of somatostatin (200 ^ g over 90 min). Filled circles and solid lines denote control experiments, open circles and broken lines those with somatostatin. Mean ± SEM of seven experiments. From Efendic et al (1976a).

A sparse plexus of somatostatin-positive nerve fibers was observed in the lamina propria around the basal parts of the crypts and occasionally also around the ganglion cells of the myenteric but rarely of the submucosal plexus. In the large intestine, nerve fibers were also found around the ganglion cells of the myenteric but rarely of the submucosal plexus, and they sometimes penetrated into the muscle layer of the

394

Suad Efendic, Tomas Hökfelt, and Rolf Luft

FIGS. 26 AND 27. Immunoftuorescence micrographs of the stomach mucosa after incubation with somatostatin antiserum. Several somatostatin-positive cells (arrows in Fig. 26) are seen among the gland cells. From Hôkfelt et al. (1975a). Magnifications 160x and 400x, respectively.

395

Somatostatin

mucosa. These fibers seem to represent hitherto unknown neurons, in addition to the classical gut neuron systems comprised by excitatory cholinergic and inhibitory adrenergic neurons (Norberg and Hamberger, 1964; Jacobowitz, 1965; Kuntz, 1953), and the more recently introduced purinergic (see Burnstock, 1972) and the substance-Pcontaining neurons (Nilsson et al., 1975; Pearse and Polak, 1975; Hökfelt et al., 1976; see also Pernow, 1953). The significance of these axons is unclear but they may belong to primary sensory neurons (see above). In this connection, it is of interest t h a t somatostatin inhibited the release of acetylcholine, electrically induced in the myenteric plexuslongitudinal muscle of the guinea pig ileum (Guillemin, 1976). SOMATOSTATIN AND FUNCTION OF THE STOMACH, G U T , EXOCRINE

AND

PANCREAS

Somatostatin inhibited basal as well as food-stimulated gastrin release in healthy subjects, acromegalics, and patients with pernicious anemia (Bloom etal., 1974). Furthermore, somatostatin also suppressed gastrin release induced by insulin hypoglycemia and the raised basal gastrin levels in subjects with chronic liver and renal failure (Le Roith et al., 1975). In dogs, somatostatin inhibited basal as well as arginineinduced gastrin secretion (Ishida et al., 1976). In this connection, it is of interest t h a t somatostatin suppressed gastrin release induced in cats by electrical stimulation of the vagus nerve, whereas atropin was without such effect (Fig. 28) (Uvnäs-Wallensten et al, 1977a). Somatostatin also suppressed the secretion of gastric acid in man (Bloom et al., 1974; Lankisch et al., 1975), in cats (Gomez-Pan et al., 1975), dogs (Barros D'sa et al., 1975), and rats (Lippmann and Borella, 1976). In addition, an inhibitory action onpepsin release could be demonstrated (Gomez-Pan et al., 1975). The fact t h a t somatostatin inhibited the output of gastric acid induced by pentagastrin suggests t h a t the effect of somatostatin on the gastric parietal cells is a direct one (Barros D'sa et al. 1975). We recently demonstrated t h a t somatostatin was released into the antral lumen of the cat after vagal stimulation, especially when the perfusion medium was acid. The opposite was true for gastrin, which preferably was released when the perfusate was alkaline (UvnäsWallensten et al., 1977b). These findings suggest t h a t somatostatin as well as gastrin-producing cells release their hormones into the surrounding interstitial tissue from which the hormones diffuse in all directions, exerting their local or paracrine effects on neighboring cells. Furthermore, the reciprocal occurrence of gastrin and somatostatin in the antral lumen—depending on the pH—suggests t h a t the inhibitory

396

Suad Efendic, Tomas Hökfelt, and Rolf Luft GASTRIN 5 Hz

PG/MIN.

5 Hz

η

π

14000-

INE SOMATOSTATIN 0.5 pg / k g /min

10000

6000

2000 —i—

-20

-10

—r0

10

20

—i—

30

10

20

TIME

MIN.

FIG. 28. Effect of somatostatin on basal and vagally induced release of gastrin in anesthetized cat. The vagal nerves were isolated and cut in the cervical region and the distal end was stimulated electrically at 7 V, 2-5 msec duration, and 2-10 Hz. Gastrin was determined in gastric venous outflow. Somatostatin was infused from 15 min before the second stimulation until 10 min after it was ended. Uvnäs-Wallensten et al. (1977a).

effect of HCl on gastrin release might at least partially be due to a local effect of somatostatin on the G cells of the stomach. Therefore, it seems most likely that somatostatin is of physiological significance in the regulation of the secretions of gastrin as well as gastric acid. Of special interest in this connection is the finding by Polak et al. (1976) showing a 71% decrease in somatostatin cell number in the duodenal mucose of patients with duodenal ulcer when compared with healthy subjects, and a 99% increase in pernicious anemia patients. The role of somatostatin in the gut has only been partially explored. According to Sakurai et al. (1975), somatostatin suppressed the release of glucagonlike immunoreactivity (GLI) in dogs given glucose, casein hydrolysate, and fat intraduodenally. Furthermore, it is a very potent inhibitor of the endogenous release of cholecystokinin from the intestinal mucosa in dogs (Konturek et al., 1976). In man, somatostatin suppressed plasma motilin concentrations and retarded gastric emptying (Bloom etal.y 1975). Somatostatin also inhibits exocrine pancreatic functions. In dogs it suppressed hydrochloric-acid-stimulated release of secretin, pancreatic flow rate, and bicarbonate and protein secretion. It reduced nonstimu-

Somatostatin

397

lated pancreatic exocrine secretion but did not affect basal secretion concentration (Boden et al., 1975). Furthermore, it also competitively inhibited the action of secretion but not cholecystokinin on the exocrine pancreatic secretion (Konturek et al., 1976). In man, somatostatin inhibited pancreatic enzyme secretion stimulated by pure cholecystokinin-pancreozymin (Dollinger et al., 1976) and carbachol (Lankisch et al., 1975). The latter authors also demonstrated inhibition of gallbladder contraction. I. Thyroid Gland In the thyroid gland, specific cells exhibiting SLI could be demonstrated (Hökfelt et al., 1975a; Parsons et al., 1976) with a localization similar to that of the parafoUicular cells (Figs. 29-32). However, they seemed to be considerably more sparse as compared, e.g., to the number of parafoUicular cells demonstrated with formaldehydeinduced fluorescence by Dahlström and Eriksson (1972). Furthermore, the somatostatin-positive cells were also fluorescent after pretreatment of the antiserum with calcitonin (Fig. 31). Thus these cells, in all probability, constitute a minor fraction of the parafoUicular cells. Whether they constitute a separate cell population, or whether somatostatin and calcitonin occur in one and the same cell, can only be established by performing immunofluorescence studies with antibodies to calcitonin and somatostatin on the same sections, similar to our experiments with glucagon and somatostatin antisera on the pancreatic islets (see above). Somatostatin blocked the iodine turnover in the thyroid gland in man even under TSH stimulation (Birk et al., 1976). This effect seemed to be a direct one, since it inhibited the release of thyroxin and triiodothyronine from isolated human thyroid cells (Loos etal., 1976).

J. Liver No substantial amounts of somatostatin were found in the liver, either by immunoflourescence technique (Hökfelt et al., 1975a) or by extraction (Arimura et al., 1975b). On the other hand, short-term (1-2 hours) administration of somatostatin had a definite hypoglycémie effect in normal subjects (Alford et al., 1974; DeVane et al., 1974; Mortimer et al., 1974; Gerich et al., 1975c; Efendic et al., 1976c; Wahren

FIGS. 29-32. Immunofluorescence micrographs of the thyroid gland after incubation with somatostatin antiserum (Figs. 29 and 30), somatostatin antiserum pretreated with calcitonin (Fig. 31), or with somatostatin (control serum; Fig. 32). Figures 30-32 show consecutive sections. Somatostatin-positive cells are lying in a parafollicular position or seemingly among the epithelial cells. They are also present after absorption with calcitonin but disappear after absorption with somatostatin. From Hökfelt et al. (1975a). Magnifications 140x (Fig. 29) and 300x (Figs. 30-32).

399

Sornatostatin

et al., 1977), in insulin dependent diabetics (Gerich et al., 1974b; Ward et al., 1975; Lundbaek et al., 1976), and hypophysectomized patients (Gerich et al., 1975c). This hypoglycémie effect is due to diminished hepatic glucose output, as demonstrated in baboons (Koerker et al., 1974), man (Fig. 33) (Wahren et al., 1977), and dogs (Altszuler et al., 1976; Sherwin et al., 1977). During such short-term infusions, sornatostatin exerted little if any effect on peripheral glucose utilization in the baboons, humans, and dogs (Koerker et al., 1974; Wahren et al., 1977; Altszuler et al., 1976; Sherwin et al., 1977). The effect of sornatostatin on hepatic glucose output seems to be an indirect one, most likely mediated by suppression of glucagon release (see above). In favor of this suggestion are (a) the lack of effect on hyperglycemia induced either by glucagon (Gerich et al., 1974a; Efendic et al., 1976c) or epinephrine (Efendic and Luft, unpublished data); (b) the lack of effect on either glycogenolysis or gluconeogenesis in rat liver slices (Chideckel et al., 1975). Against this hypothesis might speak the finding that sornatostatin inhibited both glucagon-stimulated glyconeolysis and gluconeogenesis by 40-50% in isolated rat hepatocytes (Oliver and Wagle, 1975). However, the latter authors used such high concentrations of sornatostatin as 2 />tg/ml of incubation medium. Short-term infusion of sornatostatin to subjects after a 2-3 day fast Arterial Glucose 4.5

3.0 Splanchnic Glucose Output 1.0

=-0.6

ίθ.4 0.2 - 2 0 -10 0

FIG. 33. Influence of sornatostatin ( · ) and saline (O) infusions on arterial glucose concentration and splanchnic glucose output in healthy subjects (n = 8) in the postabsorptive state (12-14 hours fasting). During sornatostatin infusion, a 50-70 pg/ml fall in glucagon concentration was observed. The findings demonstrate t h a t in the basal state hypoglucagonemia is accompanied by inhibition of splanchnic glucose output and hypoglycemia. Mean values ± SE are indicated. From Wahren et al. (1977).

400

Suad Efendic, Tomas H'ôkfelt, and Rolf Luft

diminished glucose output from the liver (Fig. 34) (Wahren et al., 1977). Since under such conditions liver glycogen is totally depleted, this effect of somatostatin most probably could be ascribed to inhibition of gluconeogenesis. Prolonged (six hours) infusion of somatostatin, in contrast, induced a hyperglycémie response in normal subjects. This effect was apparent three hours after the start of the infusion and lasted throughout its course (Lins and Efendic, 1976; Sherwin et al., 1977). In our experiments the average maximal increase in glucose concentration was 25%. Hyperglycemia appeared in spite of simultaneous suppression of the glucagon levels, indicating that intact basal glucagon secretion is not a prerequisite for the development of fasting hyperglycemia. Thus, insulin deficiency, in spite of simultaneous decrease in glucagon, leads to a diabetic state. Under these experimental conditions hyperglycemia seems to be due both to increased hepatic glucose output and to decreased peripheral glucose utilization (Sherwin et al., 1977). The diabetogenic action of somatostatin was further demonstrated by the decrease in carbohydrate tolerance after both oral and intravenous glucose loads to healthy subjects (Alberti et al., 1973; Mortimer et al., 1974; Efendic and Luft, 1975b; Efendic et al., 1976a,c). This effect, again, most probably was due to inhibition of insulin release. 175

0 4

Arterial Glucagon

Splanchnic Glucose Output

0.3 0.2 0.1 10

20

30

40

50

6 0 min

Somatostatin Inf. 10ug/min

FIG. 34. Influence of somatostatin on arterial glucagon and glucose concentrations and on splanchnic glucose output after 60-64 hours fasting in five healthy subjects. Since hepatic glycogen stores are depleted in this situation, the findings indicate that hypoglucagonemia results in diminished hepatic gluconeogenesis. Mean values ± SE are indicated. From Wahren et al. (1977).

401

Somatostatin

In contrast, in insulin-requiring diabetics carbohydrate tolerance was improved by somatostatin (Gerich et al., 1974b, 1975d, 1976; Ward et al., 1975; Wahren and Feiig, 1976). This effect was ascribed to inhibition of glucagon release (Gerich et al., 1974b, 1975d, 1976) and to decreased and/or delayed carbohydrate absorption (Wahren and Felig, 1976).

K. O t h e r Tissues In lymph nodes (within the pancreas) and in the thymus, immunopositive cells were observed on the border between medulla and cortex, around blood vessels, and along trabecula and the capsule. In the spleen small groups of green fluorescent cells were observed mainly in the red pulp. The salivary gland contained single positive cells, mostly localized close to the striated ducts but occasionally also between acini. The striated ducts exhibited a fluorescence of modest intensity. In the ovary and oviduct single positive cells were observed. In all tissues belonging to this group, similar numbers of cells with a similar localization were present irrespective of whether incubation had been performed with somatostatin or with FITC serum alone. In a last group of tissues including the heart (left and right ventricle, lung, uterus, kidney, adrenal gland, testis, epididymis, urinary bladder, and anterior pituitary no cellular elements were identified t h a t exhibited a positive green immunofluorescence after incubation either with somatostatin antiserum or with control serum. In spite of the lack of somatostatin-positive cells in the kidney, it is of interest t h a t somatostatin significantly suppressed furosemidestimulated renin release, the effect being independent of sodium excretion, which remained unchanged (Rosenthal et al., 1976; Gomez-Pan et al, 1976). Somatostatin did not influence plasma concentrations of cortisol, epinephrine, and norepinephrine (Hall et al., 1973; Copinschi et al., 1974; Koerker et al., 1974; Christensen, N.J., et al, 1975).

IV. Mode of Action Some light has been shed on the mechanism of action of somatostatin during the last two years. As for the effect on the hypophysis, somatostatin inhibited the secretion of previously or acutely biosynthesized

402

Suad Efendic, Tomas Hökfelt, and Rolf Luft

immunoprecipitated GH (Vale et al., 1972). Its action was not interrupted by inhibition of protein synthesis by the addition of cyclohexamide, and cyclohexamide did not inhibit somatostatin-mediated suppression of TSH release (Vale et al., 1975). On the other hand, three days incubation of cultured pituitary cells with somatostatin resulted in inhibition in the rate of syntheses of TSH (Vale etal., 1975). Similar data regarding GH are lacking. The fact t h a t somatostatin inhibited TRH-mediated release of TSH opened the possibility to evaluated the possible action of somatostatin at the receptor level in the hypophysis. It was found t h a t the inhibition did not involve competition for TRH receptor sites, since somatostatin action exhibited noncompetitive kinetics in vitro, and inhibition occurred independently of the stimulus used (theophylline, prostaglandin, or elevation of K + ) (see Vale et al., 1975). At present, it seems likely t h a t the effect of somatostatin on the hypophysis is mediated by its interference with the effect of cyclic AMP as well as with the formation of cyclic nucleotides. Thus, somatostatin suppressed basal as well as prostaglandin-theophylline-TRHstimulated levels of cyclic AMP in pituitary glands and cells (Kaneko et al., 1973,1974; Borgeat et al., 1974a,b) but increased the cyclic GMP levels (Kaneko et al., 1974). In addition, somatostatin inhibited GH secretion induced by dibuturyl-cyclic AMP (Vale et al., 1972) and Br 8 cyclic AMP (Vale et al, 1975). The effect of somatostatin on the pancreatic islets also seems to involve the cyclic nucleotides. Somatostatin inhibited the action of all insulinogogues studied (see p. 388). In addition, somatostatin parallelly suppressed insulin release and cyclic AMP accumulation induced by glucose after 10-60 min of incubation (Efendic et al., 1975a; Claro et al., 1977). The effect seemed to be of competitive nature, probably being mediated through an inhibitory action on islet adenyl cyclase. This inhibitory effect on insulin release, unlike that of epinephrine, could not be counteracted by the α-adrenergic blocking agent phentolamine, indicating t h a t the effect of somatostatin is not mediated by the α-adrenergic receptors (Efendic and Luft, 1975b). On the other hand, increasing the calcium concentration in the perfusate counteracted the inhibitory effect of somatostatin on glu cose-induced insulin release in the perfused rat pancreas (Curry and Bennett, 1974). More recently, the same authors (Curry and Bennett, 1976) stated t h a t this effect of calcium was more prominent on late insulin release. Moreover, studies in vitro with isolated islets revealed t h a t somatostatin inhibited 45 Ca uptake of islets (Bhathena et al, 1976; Oliver, 1976). These data suggest t h a t the mechanism of somatostatin inhibition on insulin release may involve Ca 2+ uptake.

Somatostatin

403

There are three lines of evidence for the significance of somatostatin in the regulation of functions in the nervous system: 1. The presence of somatostatin or SLI in several neuron systems (p. 381). 2. The depressant action of somatostatin on the activity of neurons at several levels (cerebral and cerebellar cortex, brain stem, and hypothalamus) (Renaud et al., 1975). 3. The behavioral effect of somatostatin: it decreased spontaneous motor activity (Segal and Mandell, 1974), it interfered with strychnine-induced effects (Brown and Vale, 1975), it potentiated the behavioral effects of L-dopa (Plotnikoff et al., 1974), it decreased locomotor activity to the point of catalepsia and induced unusual rotational behavior in the rat (Cohn and Cohn, 1975), it prolonged pentobarbital-induced narcosis (Prange, 1974); it also induced a variety of behavioral, motor, and electrophysiological changes when injected into the hippocampus (Rezek et al., 1976a), the cortex (Rezek et al., 1976b), or intraventricularly (Havlicek et al., 1976). In addition, it has been suggested that somatostatin acts as partial agonistantagonist on opiate receptors in the CNS (Terenius, 1976). In this connection it is of interest that some other peptide hormones exert depressant (LH-RH, TRH) or excitatory actions (substance P, angiotensin II) on the activity of central neurons (Nicoll and Barker, 1971; Dyer and Dyball, 1974; Konishi and Otsuka, 1974; Krnjevic and Norris, 1974; Phillis and Limacher, 1974; Renaud et al., 1975; Täkahashi and Otsuka, 1975). These observations that only a certain population of neurons display peptide sensitivity argue for a degree of selectivity and specificity to effect with respect to the individual peptides. So far, no studies have been published regarding the mode of action of somatostatin in the nervous system. In analogy with the findings in the anterior pituitary and pancreas, indicating that the action of somatostatin involves changes in the levels of cyclic nucleotides (p. 402), it is conceivable that this mechanism may also contribute to the electrophysiological and behavioral effects of somatostatin in the central nervous system. This hypothesis is supported by some observations that demonstrate that other neurotransmitters can regulate the level of cyclic AMP in the central nervous system. Thus, nor adrenaline elevated the cyclic AMP level in the cerebellum (Hoffer et al., 1971; see also Bloom, 1975) and dopamine in the caudate nucleus (Kebabian and Greengard, 1971). In this connection, it is of interest that Guillemin (1976) recently demonstrated that somatostatin could inhibit the acute release in the gut of the synaptic transmitter acetylcholine.

404

Suad Efendic, Tomas Hökfelt, and Rolf Luft

V. Clinical Applications A substance like somatostatin, which suppresses the secretion of a number of hormones, might have potential significance in conditions characterized by overproduction of these hormones. Therefore, the discovery of somatostatin was soon followed by clinical studies in patients with such diseases, e.g., acromegaly and hyperinsulinism. A.

Acromegaly

The inhibitory action of somatostatin on GH release in acromegalic subjects has been amply demonstrated (Hall et al., 1973; Siler et al., 1973; Besser et al., 1974a,b;Milhaudétf al., 1974; Mortimer etal., 1974; Peracchi et al., 1974; Yen et al., 1974; Giustina et al., 1975). Since the half-life of somatostatin is only 2 - 4 min (Brazeau et al., 1973, 1974a), a sustained GH suppression should be achieved by constant infusion. Accordingly, Yen et al. (1975) almost completely inhibited GH release in acromegalics by the infusion of 2.5 ^tg/min of somatostatin for two hours. On the other hand, Christensen et al. (1976) only obtained a 65% decrease in the GH level by infusion of 25 μg/min for two hours. In our own studies, we have been unable to normalize GH levels in most acromegalics by 200 μg of somatostatin as a bolus followed by the infusion of 1.5 ^g/min for as long as six hours (Efendic ei at., to be published). Besser et al. (1974b) could only partially inhibit GH in blood by giving 1.3 /xg/min for 28 hours. Thus, it appears t h a t somatostatin in reasonable doses only partially inhibits GH release in acromegaly. Once suppressed, there has been no escape of GH from inhibition during continuous infusion for as long as 28 hours (Besser et al., 1974b). Furthermore, prolonged infusion of somatostatin (six hours) induced prominent hyperglycemia in all acromegalics studied (Efendic et al., to be published). In this connection, it is of interest t h a t somatostatin was unable to suppress TRH- and LH-RH-induced GH release in four acromegalic subjects when administered at a dose double the one sufficient to decrease plasma GH levels in such patients (Giustina et al., 1974). From available data it does not seem self-evident t h a t somatostatin as such, even if available in long-acting form, will occupy a place in the future treatment of acromegaly. Anyhow, it must be emphasized t h a t somatostatin would not be expected to inhibit the growth of the pituitary tumors in acromegalic patients.

405

Somatostatin B. Nelson's Syndrome

Somatostatin does not seem to influence ACTH secretion in normal humans (Hall et al., 1973; Efendic et al., unpublished). On the other hand, in a dose of 500 μg, it inhibited ACTH release in five subjects who had undergone total bilateral adrenalectomy for Cushing's syndrome and had developed enlargement of the sella turcica with elevated dexamethazone-nonsuppressive levels of plasma ACTH (Nelson's syndrome) (Tyrell et al., 1975). These findings indicate either that ACTH-producing pituitary tumor cells in Nelson's syndrome possess somatostatin receptors that are not functional in normal pituitary tissue, or that ACTH-producing tumor cells are much more sensitive to somatostatin than normal pituitary tissue. C. Spontaneous

Hyperinsulinemia

In light of the inhibitory effect of somatostatin on insulin release and the observation that somatostatin is produced in the pancreatic islets, the obvious question was whether it could also inhibit insulin secretion in patients with insulin hypersécrétion. Somatostatin, in a dose of about 500 ^g over 90 min, readily inhibited basal as well as glucosestimulated insulin release in one patient with /3-cell hyperplasia, and in patients with insulin-producing adenomas of the pancreas with insulin response to a glucose challenge (Efendic et al., 1976b; Fig. 35). Those who did not respond to glucose infusion appeared to be insensitive also to somatostatin. The inhibitory effect of somatostatin in patients with benign or even malignant islet tumors was also demonstrated by other authors (Christensen, S. E., et al., 1975; Curnow et al., 1975). In this connection, it is of interest that somatostatin failed to inhibit tolbutamide-induced insulin release in four such patients (Lorenzi et al., 1975). D. Glucagon-Producing Tumor Mortimer et al. (1974) were able to suppress glucagon release in a subject with a glucagon-secreting tumor with 1100 μg of somatostatin over 90 min.

406

Suad Efendic, Tomas Hökfelt, and Rolf Luft ■g 200

150 H « 3 0 0 H

200 A

100

50 H

100 H

oJ 0

20

40

| glucos» |

60

-10

i -20 20 somatostatin | glucose

40 min

FIG. 35. Effect of somatostatin on basal and glucose-induced insulin release in a patient with insulinoma. Somatostatin was given as a priming intravenous injection (3 μg/kg) followed by an intravenous infusion (4 μg/kg) over 60 min. From Efendic et al. (1976b).

E. G a s t r i n - P r o d u c i n g

Tumor

Somatostatin in a dose 700 μg over 20 min lowered plasma gastrin concentrations and almost totally suppressed gastric acid production in a patient with Zollinger-Ellison's syndrome (Bloom et al., 1974). In addition, somatostatin reduced elevated serum gastrin levels in one patient with pancreatic islet-cell carcinoma (Curnow et al., 1975). F. D i a b e t e s

Mellitus

The discovery of somatostatin attracted diabetologists mainly for three reasons:

the

attention

of

(a) Its hypoglycémie action might be a useful complement to insulin therapy especially in unstable diabetes. (b) Its inhibitory action on the release of glucagon and GH might be useful in the prevention and treatment of diabetic ketoacidosis. (c) The suppression of GH by somatostatin might open an avenue to the prevention and treatment of diabetic vascular disease.

Sornatostatin

407

The hypoglycémie effect of sornatostatin in diabetics is well established (p. 399). In ten insulin-dependent diabetic subjects given 1 mg of sornatostatin over two hours, a significant suppression of plasma glucagon and glucose was observed (Gerich et al., 1974b). A similar result was obtained in a hypophysectomized diabetic patient, suggesting that the effects of sornatostatin were mediated mainly by suppression of glucagon release. Furthermore, infusion of sornatostatin at a dosage of 500 μg per hour prevented glucagon responses to standard meals and diminished postprandial hyperglycemia by 60%. On the other hand, 15 units of insulin did not normalize the excessive glucagon responses, although postprandial hyperglycemia was reduced (Gerich et al., 1975d). In addition, the same authors observed that the combination of insulin and sornatostatin caused progressive fall in plasma glucose levels despite meal ingestion. Their interpretation was that sornatostatin, by inhibiting excessive glucagon release, may be useful as an adjunct to insulin in the treatment of diabetes. On the other hand, recently published data by Wahren and Felig (1976) suggest that the ameliorative effect of sornatostatin on postprandial hyperglycemia can largely be explained by its inhibitory action on gastrointestinal absorption of carbohydrates. In our opinion both these effects of sornatostatin are of significance for carbohydrate metabolism: the impact on absorption being more important in connection with oral carbohydrate loads, the suppression of glucagon release being more significant during prolonged administration of sornatostatin, especially in connection with withdrawal of insulin. The overall effect of sornatostatin infusion was of such a magnitude that the insulin requirement of diabetics was reduced to between 38 and 79% (Meissner et al, 1975). It has been suggested that enhanced glucagon and GH secretion in diabetic patients might be of significance for the development of diabetic ketoacidosis (Luft et al., 1959; Iversen, 1974; Unger and Orci, 1975). Therefore, it was considered appropriate to try sornatostatin in the prevention and treatment of this condition. Acute withdrawal of insulin from seven patients with juvenile-type diabetes was accompanied by mild ketoacidosis occurring 10 hours after insulin was stopped. Administration of sornatostatin (500 μg per hour) over 18 hours prevented the development of ketoacidosis for this period of time. Furthermore, plasma /3-hydroxybuturate, glucose, and free fatty acids were all markedly lower during sornatostatin infusion (Gerich et al., 1975a). The authors attributed this effect of sornatostatin to inhibition of glucagon release and stated that glucagon by means of its gluconeogenic, ketogenic, and lipolytic actions is a prerequisite for the development of ketosis.

408

Suad Efendic, Tomas Hökfelt, and Rolf Luft

On the other hand, somatostatin did not exert antiketotic action in manifest diabetic ketoacidosis in spite of almost complete suppression of plasma pancreatic glucagon and GH and a considerable fall in plasma glucose (Lundbaek et al., 1976). Thus, data up till now seem to indicate t h a t somatostatin given to diabetic patients might improve the carbohydrate tolerance and prevent the development of ketoacidosis, but might be of no use in the treatment of established ketoacidosis. In addition, we feel t h a t the experiments with somatostatin have shed some light on the role of glucagon in the development of hyperglycemia and ketoacidosis. It is the lack of insulin, and not excess of glucagon, t h a t is the major derangement leading to ketoacidosis. This is furthermore pertinent in the light of the findings of Barnes et al. (1975) t h a t there was no correlation between the plasma glucagon levels, on the one hand, and plasma glucose and ketone bodies, on the other, after withdrawal of insulin in 38 insulin-requiring diabetics. Furthermore, significant hyperglycemia was found in five pancreatectomized subjects although the mean fasting plasma glucagon level was not significantly elevated above zero (Barnes and Bloom, 1976). The discovery of somatostatin and its inhibitory action on GH release has again brought to the fore of discussion the significance of GH for the development of diabetic vascular disease (see Luft and Guillemin, 1974). If GH plays such a role in diabetes, somatostatin might represent a physiologically significant approach to replace the more drastic procedure of hypophysectomy proposed long ago by Luft et al. (1952, 1955) as a means of suppressing the secretion of GH in patients with retinopathy and other microangiopathies. Whether somatostatin—or somatostatin analogs—will be of use in this respect depends, of course, on further study and will not be known until several years of both fundamental and clinical investigations are completed.

VI. Analogs of Somatostatin At present, it can be envisaged t h a t somatostatin might find a place in the treatment of diabetes and of some diseases characterized by overproduction of hormones. Most of these will require life-long administration of somatostatin. This makes imperative the development of a long-acting preparation of the substance since, in its present form, it is degraded so rapidly by various enzymes that its effect persist only for some minutes. If given orally, somatostatin is degraded before it can

Somatostatin

409

ever reach the bloodstream. So far, Besser et al. (1974a,b) were able to prolong the action of somatostatin to four hours by administering it subcutaneously in oil or gelatin, or after coupling it to protamin-zink. According to an editorial statement in Science (88, 922, 1975), Guillemin and co-workers have demonstrated that such a complex of somatostatin, zink, and protamin, given by injection, released somatostatin into the blood for as long as six hours in humans. In an attempt to prolong the action of somatostatin as GH inhibitor a number of analogs were produced. It was stated that N-acylateddes-tAla'-Gly^-dihydrosomatostatin analogs exhibited extended duration of action (up to 72 hours) on GH release induced by pentobarbital in rats (Brazeau et al., 1974b; Brown et al., 1975). In variance with this finding in the rat is the observation that two acylated des[Ala'-Gly^-somatostatin analogs did not exert prolonged duration of action in man (Evered et al., 1975). Recently, a cyclic undecapeptide— Wy-40,770—was synthesized showing GH release inhibiting activity for four hours after subcutaneous injection in rats (Sarantakis et al., 1976b). Another analog, [D-Trp8]-somatostatin, was found to be six to eight times more potent than somatostatin in inhibiting the release of GH, glucagon, and insulin (Rivier et al., 1975). Recently, Marks and Stern (1975) reported that brain extracts contained endopeptidases that split somatostatin at several sites including the —Lys9—Trp8 peptide bond. It is possible, therefore, that [D-Trp8]-somatostatin is more potent because it is less easily degraded. It has to be taken into account that somatostatin is widely distributed in the body, e.g., in the nervous system, and probably exerts inhibiting actions in all these places. Therefore, the administration of longacting somatostatin, if available, for the treatment of one particular disease will necessarily influence the function of many tissues. For this reason, investigators have been searching for derivatives or analogs of somatostatin that might overcome these problems. It has been demonstrated that des-[Ala1-Gly2-Asn5]-somatostatin preferentially inhibited insulin release, whereas the effect on glucagon secretion was decreased (Efendic et al., 1975b; Sarantakis et al., 1976a). This analog still exerted some inhibitory action on GH release, although this seemed to be less pronounced than for somatostatin proper (Sarantakis et al., 1976a). Deletion of only Asn5 resulted in similar alterations of activity (Brown et al., 1976; Efendic et al., 1977). It addition, desAsnr)-[D-TrpH]-somatostatin was six to eight times as potent as des-Asn5somatostatin in inhibiting insulin secretion, while it did not affect glucagon secretion (Brown et al., 1976).

410

Suad Efendic, Tomas Hökfelt, and Rolf Luft

On the other hand, [D-Cys 14 ]-somatostatin was found to be more potent on the inhibition of glucagon and GH than on insulin (Brown et al., 1977; Meyers et al., 1977). At present, no analogs with potent and selective inhibition of GH release have been produced. An encouraging report by Grant et al. (1976) describes two analogs with weak but selective GH inhibition. These findings indicate t h a t analogs can be synthesized t h a t preferentially exhibit only one or some of the activities of somatostatin.

VII. Side Effects and Toxicity of Somatostatin During the last three years we have performed numerous studies on normal subjects and acromegalic patients given somatostatin synthesized by Kabi AB, Stockholm. Most were short-term studies lasting for 90 min to six hours, with a total dose of somatostatin of 80 to 700 μg. The only side effects noticed were moderate nausea and slight abdominal pain, mostly when a bolus of 300 to 500 /xg of somatostatin was given. Similar side effects have been observed by others (Gerich et al., 1974c; Parker et al., 1974; Siler et al., 1974). No alarming side effects have been reported in the literature or even observed by the groups working with somatostatin in man in spite of giving large doses for periods up to 72 hours. Koerker et al. (1975) reported that chronic treatment of baboons with somatostatin led to unexpected death of the animals, the autopsy displaying gross or microscopical pulmonary hemorrhage and increased hemosiderin in lungs and liver. These changes were ascribed to thrombocytopenia, since similarly treated baboons had decreased mean platelet counts. Besser et al. (1975) also suggested t h a t somatostatin might diminish platelet aggregation when infused in humans. On the other hand, it was demonstrated t h a t somatostatin infusion (500 μg per hour) for up to 18 hours in humans did not affect platelet count, leucocyte count, hematocrit, platelet adhesiveness and aggregation, bleeding time, partial thromboplastin time, protrombin time, and fibrogen levels (Mielke et al., 1975). Similar results were reported by Rasche et al. (1976), who also observed t h a t peptic ulcer bleeding in one patient stopped 60 min after the beginning of the somatostatin infusion. On the whole, some effect of somatostatin on the platelets cannot be excluded, but the clinical significance of this action does not seem to be of importance.

411

Somatostatin

VIII. Conclusions Somatostatin or somatostatin-like peptides are present not only in the hypothalamus but also in other parts of the central nervous system and in some peripheral tissues. The structures containing somatostatin are of two types and seem to represent endocrine or endocrinelike cells and neurons. Somatostatin-positive endocrine cells are observed in the pancreas, thyroid gland, stomach, and among the epithelial cells in the intestine. Somatostatin-positive neuronal cell bodies are found in the periventricular region, other areas of the hypothalamus, and many extrahypothalamic areas. Somatostatin-containing nerve fibers are present in the median eminence, in the ventromedial, arcuate, and suprachiasmatic nuclei, and in many extrahypothalamic brain regions. In addition, a dense plexus of somatostatin positive fibers is found in the substantia gelatinosa of the dorsal horns of the spinal cord and in the adjacent parts of the fasciculus lateralis. In the spinal ganglia, several cell bodies, almost exclusively of the small type, contain somatostatin. As for peripheral nerves, a sparse network of somatostatin-positive fibers is observed in the wall of the gut. Wherever present in endocrine cells, somatostatin seems to inhibit the secretion of hormones in the respective endocrine glands: GH and TSH in the anterior pituitary (somatostatin transported via the portal vessels), insulin and glucagon in the pancreas, gastrin in the stomach, thyroxin and triiodothyronine in the thyroid gland. The role of somatostatin in the nervous system is not clear, but it is possible that it may play a role as an inhibitory transmitter substance or modulator. As for the mode of action of somatostatin, available data suggest that interference with the cellular levels of cyclic nucleotides as well as Ca2+ uptake may be of importance. An extensive literature is already available regarding the basic effects of somatostatin. In addition, its possible use in clinical medicine has been explored, especially in diseases characterized by hypersécrétion of hormones known to be inhibited by somatostatin: acromegaly (GH), diabetes (glucagon), hyperinsulinism, Zollinger-Ellison's syndrome (gastrin), Nelson's syndrome (ACTH). No serious side effects have been observed. However, the widespread distribution of somatostatin and its multipotent inhibitory action calls for attention. For this reason, attempts are being made to synthesize analogs to somatostatin with specific actions. Such analogs have already been introduced.

412

Suad Efendic, Tomas Hokfelt, and Rolf Luft

References Alberti, K. G. M. M., Christensen, S. E., Iversen, J., Seyer-Hansen, K., Christensen, N. J., Prange-Hansen, A., Lundbaek, K., and Örskov, H. (1973). Inhibition of insulin secretion by somatostatin. Lancet 2, 1299-1301. Alford, F. P., Bloom, S. R., Nabarro, J. D. N., Hall, R., Besser, G. M., Coy, D. H., Kastin, A. J., and Schally, A. V. (1974). Glucagon control of fasting glucose in man. Lancet 2, 974-977. Alpert, L. C , Brawer, J. R., Patel, Y. C , and Reichlin, S. (1976). Somatostatinergic neurons in anterior hypothalamus: Immunohistochemical localization. Endocrinology 98, 255-258. Altszuler, N., Gottlieb, B., and Hampshire, J. (1976). Interaction of somatostatin, glucagon and insulin on hepatic glucose output in the normal dog. Diabetes 25, 116-121. Andres, K. H. (1961). Untersuchungen über den Feinbau von Spinalganglien. Z. Zellforsch. 55, 1-48. Arimura, A., Sato, H., Coy, D. H., and Schally, A. V. (1975a). Radioimmunoassay for GH-release inhibiting hormone. Proc. Soc. Exp. Biol. Med. 148, 784-789. Arimura, A., Sato, H., Dupont, A., Nishi, N., and Schally, A. V. (1975b). Somatostatin: Abundance of immunoreactive hormone in rat stomach and pancreas. Science 189, 1007-1009. Barden, N., Alvarado-Urbina, G., Côté, J . - P , and Dupont, A. (1976a). Cyclic AMPdependent stimulation of somatostatin secretion by isolated r a t islets of Langerhans. Biochem. Biophys. Res. Commun. 7 1 , 840-844. Barden, N., Côté, J. -P., Alvarado-Urbina, G., and Dupont, A. (1976b). Secretion of somatostatin by isolated r a t islets of Langerhans and its involvement in the control of islet function. Proc. Int. Congr. Endocrinol. 5th, 1976, p. 94 (abstr.). Barnes, E. J., and Bloom, S. R. (1976). Pancreatectomized man: A model for diabetes without glucagon. Lancet 1, 219-221. Barnes, A. J., Bloom, A., Crowley, M. F , lïïttlebee, J. W., Bloom, S. R., Alberti, K. G. M., Smythe, P., and Furnell, D. (1975). Is glucagon important in stable insulindependent diabetics? Lancet 2, 734-737. Barros D'Sa, A. A. J., Bloom, S. R., and Baron, J. H. (1975). Direct inhibition of gastric acid by growth-hormone release-inhibiting hormone in dogs. Lancet 1, 886-887. Besser, G. M., Mortimer, C. H., Carr, D., Schally, A. V., Coy,D.H., Evered,D.,Kastin, A. J., Funbridge, W. M. G., Thorner, M. O., and Hall, R. (1974a). Growth hormone release inhibiting hormone in acromegaly. Brit. Med. J. 1, 352-355. Besser, G. M., Mortimer, C. H., McNeilly, A. S., Thorner, M. O., Batistoni, G. A., Bloom, S. R., Kastrup, K. W., Hanssen, K. F , Hall, R., Coy, D. H., Kastin, A. J., and Schally, A. V. (1974b). Long-term infusion of growth hormone release inhibiting hormone in acromegaly: Effects on pituitary and pancreatic hormones. Brit. Med. J. 2, 622-627. Besser, G. M., Paxton, A. M., Johnson, S. A. N., Moody, E. J., Mortimer, C. H., Hall, R., Gomez-Pan, A., Schally, A. V., Kastin, A. J., and Coy, D. H. (1975). Impairment of platelet function by growth hormone release-inhibiting hormone. Lancet 1, 11661168. Bhathena, S. J., Perrino, P. V., Voyles, N. R., Smith, S. S., Wilkins, S. D., Coy, D. H., Schally, A. V., and Recant, L. (1976). Reversal of somatostatin inhibition of insulin and glucagon secretion. Diabetes 25, 1031-1040. Birk, J., Meyer, G., Rothenbuchner, G., Loos, U., Adam, W. E., and Pfeiffer, E. F. (1976). Inhibition by somatostatin of thyroid iodine turnover. J. of Clin. Endocrinol. Metab. (In press.)

Somatostatin

413

Blair, E. L., Falkmer, S., Hellerström, C , Östberg, H., and Richardson, D. D. (1969). Investigation of gastrin activity in pancreatic islet tissue. Acta Pathol. Microbiol. Scand. 75, 583-597. Bloom, F. (1975). The role of cyclic nucleotides in central synaptic function. Rev. Physiol. Biochem. Pharmacol. 74, 1-103. Bloom, S. R., Mortimer, C. H., Thorner, M. O., Besser, G. M., Hall, R., Gomez-Pan, A., Roy, V. M., Russell, R. C. G., Coy, D. H., Kastin, A. J., and Schally, A. V. (1974). Inhibition of gastrin and gastric-acid secretion by growth hormone releaseinhibiting hormone. Lancet 2, 1106-1109. Bloom, S. R., Ralphs, D. N., Besser, G. M., Hall, R., Coy, D. H., Kastin, A. J., and Schally, A. V. (1975). Effect of somatostatin on motilin levels and gastric emptying. Gut 16, 834 (abstr.). Boden, G., Sivitz, M. C , Owen, O. E., Essa-Koumar, N., and Landor, J. H. (1975). Somatostatin suppresses secretin and pancreatic exocrine secretion. Science 190, 163-164. Borgeat, P., Drouin, J., Bélanger, A., and Labrie, F. (1974a). Inhibitory effect of growth hormone-release inhibiting hormone on cyclic AMP accumulation in r a t anterior pituitary gland in vitro. Fed. Proc, Fed. Am. Soc. Exp. Biol. 263 A. Borgeat, P., Labrie, F., Drouin, J., Bélanger, A., Immer, H., Sestanj, K., Nelson, V., Götz, M., Schally, A. V., Coy, D. H., and Coy, E. J. (1974b). Inhibition of adenosine 3',5'monophosphate accumulation in anterior pituitary gland in vitro by growth hormone-release inhibiting hormone. Biochem. Biophys. Res. Commun. 56, 10521059. Brazeau, P., Vale, W., Burgus, R., Ling, N., Butcher, M., Rivier, J., and Guillemin, R. (1973). Hypothalamic polypeptide t h a t inhibits the secretion of immunoreactive pituitary growth hormone. Science 179, 77-79. Brazeau, P., Rivier, J., Vale, W., and Guillemin, R. (1974a). Inhibition of growth hormone secretion in the rat by synthetic somatostatin. Endocrinology 94, 184-187. Brazeau, P., Vale, W., Rivier, J., and Guillemin, R. (1974b). Acylated des (Ala'-Gly 2 )somatostatin analogs: Prolonged inhibition of growth hormone secretion. Biochem. Biophys. Res. Commun. 60, 1202-1207. Brown, M., and Vale, W. (1975). Central nervous system effects of hypothalamic pep tides. Endocrinology 96, 1333-1336. Brown, M., Rivier, J., Vale, W., and Guillemin, R. (1975). Variability of the duration of inhibition of growth hormone release by N a -acylated-des-[Ala'-Gly 2 ]-H 2 somatostatin analogs. Biochem. Biophys. Res. Commun. 65, 752-756. Brown, M., Rivier, J., and Vale, W. (1977). Somatostatin analogs with selected biological activities. Metabolism 25, Suppl 1, 1501-1503. Brownstein, M., Arimura, A., Sato, H., Schally, A. V., and Kizer, J. S. (1974). The regional distribution of somatostatin in the rat brain. Endocrinology 96, 1456-1461. Burgus, R., Brazeau, P., and Vale, W. (1973a). Isolation and determination of the primary structure of somatostatin (a somatostatin release inhibiting factor) of ovine hypothalamic origin. Tra "Advances in H u m a n Growth Hormone Research" (S. Raiti, ed.), pp. 144-158, DHEW Publ. No. (NIH) 74-614. US Govt. Printing Office, Washington, D.C. Burgus, R., Ling, N., Butcher, M., and Guillemin, R. (1973b). Primary structure of somatostatin, a hypothalamic peptide that inhibits the secretion of pituitary growth hormone. Proc. Natl. Acad. Sei. U.S.A. 70, 684-688. Burnstock, G. (1972). Purinergic nerves. Pharmacol. Rev. 24, 509-581. Carr, D., Gomez-Pan, A., Weightman, D. R., Roy, V. C. M., Hall, R., Besser, G. M., Thorner, M. O., McNeilly, A. S., Schally, A. V., Kastin, A. J., and Coy, D. H. (1975).

414

Suad Efendic, Tomas Hökfelt, and Rolf Luft

Growth hormone release inhibiting hormone: Actions on thyrotrophin andprolactin secretion after thyrotropin-releasing hormone. Brit. Med. J. 2, 67-69. Cavallero, C , Solcia, E., and Sampietro, R. (1967). Cytology of islet tumours and hyperplasias associated with the Zollinger-Ellison syndrome. Gut 8, 172-177. Chen, H. J., Mueller, G. P., and Meites, J. (1974). Effects of L-dopa and somatostatin on suckling-induced release of prolactin and GH. Endocrinol. Res. Commun. 1, 2 8 3 291. Chideckel, E. W., Palmer, J., Koerker, D. J., Ensinck, J., Davidson, M. B., and Goodner, C. J. (1975). Somatostatin blockade of acute and chronic stimuli of the endocrine pancreas and the consequences of this blockade on glucose homeostasis. J. Clin. Invest. 55, 754-762. Christensen, N. J., Christensen, S. E., Prange-Hansen, A., and Lundbaek, K. (1975). The effect of somatostatin on plasma noradrenaline and plasma adrenaline concentrations during exercise and hypoglycemia. Metabolism 24, 1267-1972. Christensen, S. E., Prange-Hansen, A. A., Iversen, J., Lundbaek, K., Örskov, H., and Seyer-Hansen, K. (1974). Somatostatin as a tool in studies of basal carbohydrate and lipid metabolism in man: Modifications of glucagon and insulin release. Scand. J. Clin. Lab. Invest. 34, 321-325. Christensen, S. E., Prange-Hansen, A. A., Lundbaek, K., Örskov, H., and Seyer-Hansen, K. (1975). Somatostatin and insulinoma. Lancet 1, 1426. Christensen, S. E., Nerup, J., Hansen, A. P., and Lundbaek, K. (1976). Effects of somatostatin on basal levels of plasma growth hormone and insulin in acromegalics: Doseresponse studies and attempted total growth hormone suppression. J. Clin. Endocrinol. Metab. 42, 839-845. Claro, A., Grill, V., Efendic, S., and Luft, R. (1977). Studies on the mechanism of somatostatin action on insulin release. IV. Effect of somatostatin on cyclic-AMP levels and phosphodiestarase activity in isolated r a t pancreatic islets. Ada Endocrinol. (Copenhagen) 85, 379-388. Cohn, M. L., and Cohn, M. (1975). "Barrel rotation" induced by somatostatin in the non-lesioned rat. Brain Res. 96, 138-141. Copinschi, G., Virasoro, E., Vanhaelst, L., Leclerq, R., Golstein, J., and L'Hermite, M. (1974). Specific inhibition by somatostatin of growth hormone release after hypoglycaemia in normal man. Clin. Endocrinol. 3, 441-445. Curnow, R. T., Carwy, R. M., Taylor, A., Johanson, A., and Murad, F. (1975). Somatostatin inhibition of insulin and gastrin hypersécrétion in pancreatic islet-cell carcinoma. New Engl. J. Med. 292, 1385-1386. Curry, D. L., and Bennett, L. L. (1974). Reversal of somatostatin inhibition of insulin secretion by calcium. Biochem. Biophys. Res. Commun. 60, 1015-1019. Curry, D. L., and Bennett, L. L. (1976). Does somatostatin inhibition of insulin secretion involve two mechanisms of action. Proc. Natl. Acad. Sei. U.S.A. 73, 248-251. Dahlström, A., and Ericson, L. E. (1972). Monoamines in rat thyroid parafollicular cells and the effect of vitamin D 2 -induced degranulation. Z. Zellforsch. Mikrosk. Anat. 128, 406-425. DeVane, G. W., Siler, T. M., and Yen, S. S. C. (1974). Acute suppression of insulin and glucose levels by synthetic somatostatin in normal h u m a n subjects. J. Clin. Endocrinol. Metab. 38, 913-915. Dhariwal, A. P. S., Krulich, L., and McCann, S. M. (1969). Purification of a growth hormone-inhibiting factor (GIF) from sheep hypothalamus. Neuroendocrinology 4, 282-288. Dollinger, H. C , Raptis, S., and Pfeiffer, E. H. (1976). Effects of somatostatin on exocrine

Somatostatin

415

and endocrine pancreatic function stimulated by intestinal hormones in m a n . H o r m . Metab. Res. 8, 74-78. Dubé, D., Leclerc, R., Pelletier, G., Arimura, A., and Schally, A. V. (1975). Immunohistochemical detection of growth hormone-release inhibiting hormone (somatostatin) in the guinea-pig brain. Cell. Tissue Res. 161, 385-392. Dubois, M. P. (1975). Immunoreactive somatostatin is present in discrete cells of the endocrine pancreas. Proc. Natl. Acad. Sei. U.SA. 72, 1340-1343. Dubois, M. P., and Kolodziejczyk, E. (1975). Le système à somatostatine. Étude immunocytologique. J. Physiol. (Paris) 70, 14 B. Dubois, M. P., Barry, J., and Leonardelli, J. (1974). Mise en évidence p a r immunofluorescence et repartition de la somatostatine (SRIF) dans l'éminence médiane des vertébrés (mammifères, oiseaux amphibiens, poissons). C.R. Acad. Sei., Ser. D 279, 1899-1902. Dubois, P. M., Paulin, C , Assan, R., and Dubois, M. P. (1975). Evidence for immunoreactive somatostatin in the endocrine cells of h u m a n foetal pancreas. Nature (London) 256, 731-732. Dyer, R. G., and Dyball, R. E. J. (1974). Evidence for a direct effect of LRF and TRF on single unit activity in the rostral hypothalamus. Nature (London) 253, 486-488. Efendic, S., and Luft, R. (1975a). Studies on the inhibitory effect of somatostatin on glucose induced insulin release in the isolated perfused rat pancreas. Acta Endocrinol. (Copenhagen) 78, 510-515. Efendic, S., and Luft, R. (1975b). Studies on the mechanism of somatostatin action on insulin release in man. I. Effect of blockade of α-adrenergic receptors. Acta Endocrinol. (Copenhagen) 78, 516-523. Efendic, S., Luft, R., and Grill, V. (1974). Effect of somatostatin on glucose induced insulin release in isolated perfused rat pancreas and isolated rat pancreatic islets. FEBSLett. 42, 169-172. Efendic, S., Grill, V., and Luft, R. (1975a). Inhibition by somatostatin of glucose induced 3' : 5'-monophosphate (cyclic AMP) accumulation and insulin release in isolated pancreatic islets of the rat. FEBS Lett. 55, 131-133. Efendic, S., Luft, R., and Sievertsson, H. (1975b). Relative effects of somatostatin a n d two somatostatin analogues on the release of insulin glucagon and growth hormone. FEBS Lett. 58, 302-305. Efendic, S., Claro, A., and Luft, R. (1967a). Studies on the mechanism of somatostatin action on insulin release. III. Effect of somatostatin on arginine induced release of insulin and glucagon in man and perfused r a t pancreas. Acta Endocrinol. (Copenhagen) 8 1 , 753-761. Efendic, S., Lins, P. E., Sigurdsson, G., Ivemark, B., Granberg, P. O., and Luft, R. (1976b). Effect of somatostatin on basal and glucose induced insuline release in five patients with hyperinsulinemia. Acta Endocrinol. (Copenhagen) 8 1 , 525-529. Efendic, S., Luft, R., and Claro, A. (1976c). Studies on the mechanism of somatostatin action on insulin release in man. II. Comparison of the effects of somatostatin on insulin release induced by glucose, glucagon and tolbutamide. Acta Endocrinol. (Copenhagen). 8 1 , 743-752. Efendic, S., Lins, P. E., Luft, R., Sievertsson, H., and Westin-Sjödahl, G. (1977). Studies on the mechanism of somatostatin action on insulin release. V. Effect of somatostatin analogues on arginine induced release of insulin and glucagon from the perfused r a t pancreas. Acta Endocrinol. (Copenhagen) 85, 579-586. Eide, R. P., and Parsons, J. A. (1975). Immunocytochemical localization of somatostatin in cell bodies of the r a t hypothalamus. Am. J. Anat. 144, 541-548.

416

Suad Efendic, Tomas Hökfelt, and Rolf Luft

Eide, R. P., Efendic, S., Hökfelt, T., Johansson, O., Luft, R., Parsons, J. A., Roovete, A., and Sorensen, R. L. (1977). Production of antibodies to somatostatin for radioimmunoassay and immunohistochemistry. Ada Endocrinol. (Copenhagen) (In press). Erlandsen, S. L., Hegre, O. D., Parsons, J. A., McEvoy, R. C , and Eide, R. P. (1976). Pancreatic islet cell hormones. Distribution of cell types in the islet and evidence for the presence of somatostatin and gastrin within the D-cell. J. Histochem. Cytochem. 24, 883-897. Evered, D. C., Gomez-Pan, A., Tunbridge, W. M. G., Hall, R., Lind, T., Besser, G. M., Mortimer, C. H., Thorner, M. O., Schally, A. V., Kastin, A. J., and Coy, D. H. (1975). Analogues of growth-hormone release-inhibiting hormone. Lance t 1, 1250. Fehm, H. L., Voigt, K. H., Lang, R., Beinert, K. E., Raptis, S., and Pfeiffer, E. F. (1976). Somatostatin: A potent inhibitor of ACTH-hypersecretion in adrenal insufficiency. Klin. Wochenschr. 54, 173-175. Florsheim, W. H., and Kozbur, X. (1976). Physiological modulation of thyrotropin secretion by somatostatin and thyroliberin. Biochem. Biophys. Res. Commun. 72, 6 0 3 609. Fujita, T., and Murakami, T. (1973). Microcirculation of monkey pancreas with special reference to the insuloacinar portal system. A scanning electron microscope study of vascular casts. Saibokaku Byorigaku Zasshi 3 5 , 255-263. Gala, R. R., Subramanian, M. G., Peters, J. A., and Jaques, S., J r . (1976). The influence of somatostatin in drug-induced prolactin release in the monkey. Experientia (Basel) 32, 941-943. Gerich, J. E. (1976). Metabolic effects of long-term somatostatin infusion in man. Metabolism 25, Suppl. 1, 1505-1507. Gerich, J. E., Lorenzi, M., Schneider, V., and Forsham, P. H. (1974a). Effect of somatostatin on plasma glucose and insulin responses to glucagon and tolbutamide in man. J. Clin. Endocrinol. Metab. 39, 1057-1060. Gerich, J. E., Lorenzi, M., Schneider, V., Karam, J. H., Rivier, J., Guillemin, R., and Forsham, P. H. (1974b). Effects of somatostatin on plasma glucose and glucagon levels in h u m a n diabetes mellitus: Pathophysiologic and therapeutic implications. N. Engl. J. Med. 291, 544-547. Gerich, J. E., Lorenzi, M., Schneider, V., Kwan, C. W., Karam, J. H., Guillemin, R., and Forsham, P. H. (1974c). Inhibition of pancreatic glucagon responses to arginine by somatostatin in normal man and in insulin-dependent diabetics. Diabetes 23, 8 7 6 880. Gerich, J. E., Lorenzi, M., Bier, D. M., Schneider, V., Tsalikian, E., Karam, J., and Forsham, P. H. (1975a). Prevention of human diabetic ketoacidosis by somatostatin. Evidence for an essential role of glucagon. N. Engl. J. Med. 292, 985-989. Gerich, J. E., Lovinger, R., and Grodsky, G. M. (1975b). Inhibition by somatostatin of glucagon and insulin release from the perfused r a t pancreas in response to arginine, isoproterenol and theophylline: Evidence for a preferential effect on glucagon secretion. Endocrinology 96, 749-754. Gerich, J. E., Lorenzi, M., Hane, S., Gustafson, G., Guillemin, R., and Forsham, P. H. (1975c). Evidence for a physiologic role of pancreatic glucagon in h u m a n glucose homeostasis: Studies with somatostatin. Metabolism 24, 175-182. Gerich, J. E., Lorenzi, M., Karam, J. H., Schneider, V., and Forsham, P. (1975d). Abnormal pancreatic glucagon secretion and postprandial hyperglycemia in diabetes mellitus. J. Am. Med. Ass. 234, 159-165. Giustina, G., Reschini, E., Peracchi, M., Cantalamessa, L., Cavagnini, F , Pinto, M., and Bulgheroni, P. (1974). Failure of somatostatin to suppress thyrotropin releasing factor and luteinizing hormone releasing factor-induced growth hormone release in acromegaly. J. Clin. Endocrinol. Metab. 38, 906-909.

Somatostatin

All

Giustina, G., Peracchi, M., Reschini, E., Panerai, A. E., and Pinto, M. (1975). Doseresponse study of the inhibiting effect of somatostatin on growth hormone and insulin secretion in normal subjects and acromegalic patients. Metabolism 24, 807-815. Goldsmith, P. C , Rose, J. C , Arimura, A., and Ganong, W. F (1975). Ultrastructural localization of somatostatin in pancreatic islets of the rat. Endocrinology 9 7 , 1 0 6 1 1064. Gomez-Pan, A., Albinus, M., Reed, J. D., Shaw, B., Hall, R., Besser, G. M., Coy, D. H., Kastin, A. J., and Schally, A. V. (1975). Direct inhibition of gastric acid and pepsin secretion by growth-hormone release-inhibiting hormone in cats. Lancet 1,888-890. Gomez-Pan, A., Snow, M. H., Piercy, D. A., Robson, V., Wilkinson, R., Hall, R., Evered, D. C., Besser, G. M., Schally, A. V., Kastin, A. J., and Coy, D. H. (1976). Actions of growth hormone-release inhibiting hormone (somatostatin) on the renin aldosterone system. J. Clin. Endocrinol. Metab. 4 3 , 240-243. Grant, N., Clark, D., Garsky, V., J a u n a k a i s , I., McGregor, W., and Sarantakis, D. (1976). Dissociation of somatostatin effects. Peptides inhibiting the release of growth hormone but not glucagon or insulin in rats. Life Sei. 19, 629-632. Gregory, R. A., Tracy, H. J., French, J. M., and Sircus, W. (1960). Extraction of gastrinlike substance from a pancreatic tumor in a case of Zollinger-Ellison syndrome. Lancet 1, 1045-1048. Greider, M. H., and McGuigan, J. E. (1970). A comparison of ulcerogenic tumors of the pancreas with the gastrin cell of normal h u m a n pancreas and dog antrum. Am. J. Pathol. 59, 76A. Greider, M. H., and McGuigan, J. E. (1971). Cellular localization of gastrin in the h u m a n pancreas. Diabetes 20, 387-396. Greider, M. H., Rosai, J., and McGuigan, J. E. (1974). The h u m a n pancreatic islet cells and their tumors. II. Ulcerogenic and diarrheogenic tumors. Cancer 33,1423-1443. Guillemin, R. (1976). Somatostatin inhibits the release of acetylcholine induced electrically in the myentric plexus. Endocrinology 99, 1653-1654. Hall, R., Schally, A. V., Evered, D., Kastin, A. J., Mortimer, C. H., Tunbridge, W. M. G., Besser, G. M., Coy, D. H., Goldie, D. J., McNeilly, A. S., Phenekos, C , and Weightman, D. (1973). Action of growth-hormone-release-inhibitory hormone in healthy men and in acromegaly. Lancet 2, 581-584. Havlicek, V., Rezek, M., and Friesen, H. (1976). Somatostatin and thyrotropin releasing hormone: Central effect on sleep and motor system. Pharmacol., Biochem. Behav. 4, 455-459. Hellerström, C , and Hellman, B. (1960). Some aspects of silver impregnation of the islets of Langerhans in the rat. Ada Endocrinol. (Copenhagen) 35, 518-532. Hellerström, C , Hellman, B., Peterson, B., and Alm, G. (1964). Wenner-Gren Center Int. Symp. Ser. 3, 117. Hellman, B., and Lernmark, Â. (1969a). Evidence for an inhibitor of insulin release in the pancreatic islets. Diabetologia 5, 22-24. Hellman, B., and Lernmark, Â. (1969b). Inhibition of the in vitro secretion of insulin by an extract of pancreatic alpha-1 cells. Endocrinology 84, 1484-1488. Hökfelt, T., Efendic, S., Johansson, O., Luft, R., and Arimura, A. (1974). Immunohistochemical localization of somatostatin (growth hormone releasing-inhibiting factor) in the guinea-pig brain. Brain Res. 80, 165-169. Hökfelt, T., Efendic, S., Hellerström, C , Johansson, O., Luft, R., and Arimura, A. (1975a). Cellular localization of somatostatin in endocrine-like cells and neurons of the rat with special reference to the Aj-cells of the pancreatic islets and to the hypothalamus. Acta Endocrinol. (Copenhagen) Suppl. 80, 200. Hökfelt, T., Eide, R., Johansson, O., Luft, R., and Arimura, A. (1975b). Immunohis-

418

Suad Efendic, Tomas Hökfelt, and Rolf Luft

tochemical evidence for the presence of somatostatin, a powerful inhibitory peptide, in some primary sensory neurons. Neurosci. Lett. 1, 231-235. Hökfelt, T., Fuxe, K., Johansson, 0., Jeffcoate, S. L., and White, N. (1975c). Distribution of thyrotropin releasing hormone (TRH) in the central nervous system as revealed with immunohistochemistry. Eur. J. Pharmacol. 34, 389-392. Hökfelt, T., Fuxe, K., Johansson, O., Jeffcoate, S. L., and White, N . (1975d). Thyrotropin releasing hormone (TRH)- containing nerve terminals in certain brain stem nuclei and in the spinal cord. Neurosci. Lett. 1, 133-139. Hökfelt, T., Johansson, O., Efendic, S., Luft, R., and Arimura, A. (1975e). Are there somatostatin-containing nerves in the rat gut? Immunohistochemical evidence for a new type of peripheral nerves. Experientia 3 1 , 852-854. Hökfelt, T., Johansson, O., Fuxe, K., Löfström, A., Goldstein, M., Park, D., Ebstein, R., Fraser, H., Jeffcoate, S. L., Efendic, S., Luft, R., and Arimura, A. (1975f). Mapping and relationship of hypothalamic neurotransmitters and hypothalamic hormones. Proc. Int. Congr. Pharmacol. 6th. 3, 93-110. Hökfelt, T., Kellerth, J.-O., Nilsson, G., and Pernow, B. (1975g). Experimental immunohistochemical studies on the localization and distribution of substance P in cat primary sensory neurons. Brain Res. 100, 234-252. Hökfelt, T., Kellerth, J. -O., Nilsson, G., and Pernow, B. (1975h). Substance P: Localization in the central nervous system and in some primary sensory neurons. Science 190, 889-890. Hökfelt, T., Eide, R., Johansson, O., Luft, R., Nilsson, G., and Arimura, A. (1976). Immunohistochemical evidence for separate populations of somatostatin-containing and substance P-containing primary afferent neurons. Neuroscience 1, 131-136. Hoffer, B., Siggins, G. R., Oliver, A. P., and Bloom, F. (1971). Cyclic AMP mediation of norepinephrine inhibition in rat cerebellar cortex: A unique class of synaptic responses. Ann. N.Y. Acad. Sei. 185, 531-549. Hughes, J., Smith, T. W., Kosterlitz, H. W., Fothergill, L. A., Morgan, B. A., and Morris, H. R. (1975). Identification of two related pentapeptides from the brain with potent opiate agonist activity. Nature (London) 258, 577-579. Ishida, T., Takahara, J., Kawanishi, K., Kawamura, K., Goto, A., Ogawa, N., and Ofuji, T. (1976). Direct inhibitions of basal and arginine-induced plasma gastrin secretion by somatostatin in anesthetized dogs. Endocrinol. Jpn. 23, 203-206. Iversen, J. (1974). Inhibition of pancreatic glucagon release by somatostatin in vitro. Scand. J. Clin. Lab. Invest. 33, 125-129. Jacobowitz, D. (1965). Histochemical studies of the autonomie innervation of the gut. J. Pharmacol. Exp. Ther. 149, 358-364. Kaneko, T., Oka, H., Saito, S., Munemura, M., Musa, K., Oda, T., Yanaihara, N., and Yanaihara, C. (1973). In vitro effects of synthetic somato-tropin-release inhibiting factor of cyclic AMP level and GH release in rat anterior pituitary gland. Endocrinol. Jpn. 20, 535-538. Kaneko, T., Oka, H., Munemura, M., Suzuki, S., Yasuda, H., and Oda, T. (1974). Stimulation of guanosine 3',5'-cyclic monophosphate accumulation in r a t anterior pituitary gland in vitro by synthetic somatostatin. Biochem. Biophys. Res. Commun. 6 1 , 53-57. Kato, Y., Chihara, K., Ohgo, S., and Imura, H. (1974). Effects of hypothalamic surgery and somatostatin on chlorpromazine-induced growth hormone release in rats. Endocrinology 95, 1608-1613. Kebabian, J. W., and Greengard, P. (1971). Dopamine-sensitive adenyl cyclase: Possible role in synaptic transmission. Science 174, 1346-1349. King, J. C , Arimura, A., Gerall, A. A., Fischback, J. B., and Elkind, K. E. (1975).

Somatostatin

419

Growth hormone-release inhibiting hormone (GH-RIH) pathway of the r a t hypothalamus revealed by the unlabeled antibody peroxidase-antiperoxidase method. Cell. Tissue Res. 160, 423-430. Köbberling, J., Jüppner, H., and Hesch, R. D. (1976). The stimulation of growth hormone release by ACTH and its inhibition by somatostatin. A eta Endocrinol. (Copenhagen) 8 1 , 263-269. Koerker, D. J., Ruch, W., Chi deckel, E., Palmer, J., Goodner, C. J., Ensinck, J., and Gale, C. C. (1974). Somatostatin: Hypothalamic inhibitor of the endocrine pancreas. Science 184, 482-484. Koerker, D. J., Harker, L. A., and Goodner, G. J. (1975). Effects of somatostatin on hemostasis in baboons. N. Engl. J. Med. 293, 476-479. Konishi, S., and Otsuka, M. (1974). The effects of substance P and other peptides on spinal neurons of the frog. Brain Res. 65, 397-410. Konturek, S. J., Tasler, J., Obtulowicz, W., Coy, D. H., and Schally, A. V. (1976). Effect of growth hormone-release inhibiting hormone on hormones stimulating exocrine pancreatic secretion. J. Clin. Invest 58, 1-6. Krnjevic, K., and Morris, M. E. (1974). An excitatory action of substance P on cuneate neurons. Can. J. Physiol. Pharmacol. 52, 736-744. Krulich, L., and McCann, S. M. (1969). Effect of GH-releasing factor and GH-inhibiting factor on the release and concentration of GH in pituitaries incubated in vitro. Endocrinology 85, 319-324. Krulich, L., Dhariwal, A. P. S., and McCann, S. (1968). Stimulatory and inhibitory effects of purified hypothalamic extracts on growth hormone release from rat pituitary in vitro. Endocrinology 83, 783-790. Krulich, L., Illner, P., Fawcett, C. P., Quijada, M., and McCann, S. M. (1971). Dual hypothalamic regulation of growth hormone secretion. Excerpta Med. Found. Int. Congr. Ser. 244, 306-316. Kuntz, A. (1953). "The Autonomie Nervous System." Lea & Febiger, Philadelphia, Pennsylvania. Lankisch, P. G., Arnold, R., and Creutzfeldt, W. (1975). Wirkung von Somato-statin auf die betazol-stimulierte Magensekretion und die carbachol-stimulierte Pankreassekretion und Gallenblasenkontraktion des Menschen. Dtsch. Med. Wochenschr. 100 (36), 1797-1800. Leblanc, H., Anderson, J. R., Sigel, M. B., and Yen, S. S. C. (1975a). Inhibitory action of somatostatin on pancreatic a and ß cell function. J. Clin. Endocrinol. Metab. 40, 568-572. Leblanc, H., Rigg, L. A., and Yen, S. S. C. (1975b). The response of pancreatic and pituitary hormones to pulses and constant infusion of somatostatin. J. Clin. Endocrinol. Metab. 4 1 , 1105-1109. Le Roith, D., Vinik, A. I., Epstein, S., Baron, P., Olkonitzky, M. N., and Pimstone, B. L. (1975). Somatostatin and serum gastrin in normal subjects and in patients with pernicious anaemia, chronic liver and renal disease. S. Afr. Med. J. 49,1601-1604. Lins, P. E., and Efendic, S. (1976). Hyperglycemia induced by somatostatin in normal subjects. Horm. Metab. Res. 8, 497-498. Lippmann, W., and Borella, L. E. (1976). Inhibition of gastric acid secretion by somatostatin and analogs in the rat. Pharmacol. Res. Commun. 8, 445-454. Lomsky, R., Langer, F., and Vortel, V. (1969). Immunohistochemical demonstration of gastrin in mammalian islet of Langerhans. Nature (London) 223, 618-619. Loos, U., Voigt, K. H., Kampshoflf, H., Birk, J., Raptis, S., and Rotenbuchner, G. (1976). Secretion of T 3 and T4 by isolated h u m a n thyroid cells and its inhibition by somatostatin. Eur. Thyroid Assoc. Meet. 77th, 1976.

420

Suad Efendic, Tomas Hökfelt, and Rolf Luft

Lorenzi, M., Gerich, J. E., Karam, J . H., and Forsham, P. (1975). Failure of somatostatin to inhibit tolbutamide-induced insulin secretion in patients with insulinomas: A possible diagnostic tool. J. Clin. Endocrinol. Metab. 40, 1121-1124. Lostra, F., van der Loo, W., and Gepts, W. (1974). Are gastrin-cells present in mammalian pancreatic islets? Diabetologia 10, 291-302. Lovinger, R., Boryczka, A. T., Shackelford, R., Kabblan, S. L., Ganong, W. F., and Grumbach, M. M. (1974). Effect of synthetic somatotropin release inhibiting factor on the increase in plasma growth hormone elicited by L-dopa in the dog. Endocrinology 95, 943-946. Lucke, C , Höffken, B., and von Zur Mühlen, A. (1975). The effect of somatostatin on TSH levels in patients with primary hypothyroidism. J. Clin. Endocrinol. Metab. 4 1 , 1082-1084. Luft, R., and Guillemin, R. (1974). Growth hormone and diabetes in man: Old concepts—new implications. Diabetes 23, 783-787. Luft, R., Olivecrona, H., and Sjögren, B. (1952). Hypofysektomi pa människa. Acta Med. Scand. 47, 351-358. Luft, R., Olivecrona, H., andSjögren, S. (1955). Hypophysectomy in man: Experiences in severe diabetes mellitus. J. Clin. Endocrinol. Metab. 15, 391-408. Luft, R., Ikkos, D., Gemzell, C. A., and Olivecrona, H. (1959). The effect of human growth hormone in hypophysectomized h u m a n diabetic subjects. Acta Endocrinol. (Copenhagen) 32, 330-340. Luft, R., Efendic, S., Hökfelt, T., Johansson, O., and Arimura, A. (1974). Immunohistochemical evidence for the localization of somatostatin-like immunoreactivity in a cell population of pancreatic islets. Med. Biol. 52, 428-430. Lundbaek, K., Prange-Hansen, A., örskov, H., Christensen, S. E., Iversen, J., SeyerHansen, K., Alberti, K. G. M. M., and Whitefoot, R. (1976). Failure of somatostatin to correct manifest diabetic ketoacidosis. Lancet 1, 215-218. Malacara, J. M., Valverde-R, C., Reichlin, S., and Bollinger, J. (1972). Elevation of plasma radioimmunoassayable growth hormone in the rat induced by porcine hypothalamic extract. Endocrinology 9 1 , 1189-1198. Marks, N., and Stern, F. (1975). Inactivation of somatostatin (GH-RH) and its analogs by crude and partially purified r a t brain extracts. FEBS Lett. 55, 220-224. Martin, J. B. (1974). Inhibitory effect of somatostatin (SRIF) on the release of growth hormone (GH) induced in the r a t by electrical stimulation. Endocrinology 9 4 , 4 9 7 502. Martin, J. B., Renaud, L. P., and Brazeau, P. (1975). Hypothalamic peptides: New evidence for "peptidergic" pathways in the C. N. S. Lancet 2, 393-395. Meissner, C., Thum, C , Beischer, W., Winkler, G., Schröder, K. E., and Pfeiffer, E. F. (1975). Antidiabetic action of somatostatin—assessed by the artificial pancreas. Diabetes 24, 988-996. Meyers, C , Arimura, A., Gordin, A., Fernandez-Durango, R., Coy, D. H., Schally, A. V., Drouin, J., Ferland, L., Beaulien, M., and Labrie, F. (1977). Somatostatin analogs which inhibit glucagon and growth hormone more than insulin release. Biochem. Biophys. Res. Commun. (In press.) Mielke, C. H., Jr., Gerich, J. E., Lorenzi, M., Tsalikian, E., Rodvien, R., Forsham, P. (1975). The effect of somatostatin on coagulation and platelet function in man. N. Engl. J. Med. 293, 480-483. Milhaud, G., Ribeiro, F , and Rivaille, P. (1974). Effect de la somatostatine sur la secretion d'hormone somatotrope (STH) chez l'acromégale. Ann. Endocrinol. (Paris) 35, 671-673. Mortimer, C. H., Carr, D., Lind, T., Bloom, S. R., Mallinson, C. N., Schally, A. V.,

Somatostatin

All

Tunbridge, W. M. G., Yeomans, L., Coy, D. H., Kastin, A., Besser, G. M., and Hall, R. (1974). Effects of growth-hormone release-inhibiting hormone on circulating glucagon, insulin, and growth hormone in normal, diabetic, acromegalic, and hypopituitary patients. Lancet 1, 697-701. Nicoll, R. A., and Barker, J. L. (1971). The pharmacology of recurrent inhibition in the supraoptic neurosecretory system. Brain Res. 35, 501-511. Nilsson, G., Simon, J., Yalow, R. S., and Berson, S. A. (1972). Plasma gastrin and gastric acid responses to sham feeding and feeding in dogs. Gastroenterology 63, 51-59. Nilsson, G., Larsson, L. I., Hàkansson, R., Brodin, E., Sundler, F., and Pernow, B. (1975). Localization of substance P-like immunoreactivity in mouse gut. Histochemie 4 3 , 97-99. Norberg, K. A., and Hamberger, B. (1964). The sympathetic adrenergic neuron. Some characteristics revealed by histochemical studies on the intraneuronal distribution of the transmitter. Ada Physiol. Scand. 63, Suppl. 238, 1-42. Norfleet, W. T., Pagliara, A. S., Haymond, M. W., and Matschinsky, F (1975). Comparison of alpha- and beta-cell secretory responses in islets isolated with collagenase and in the isolated perfused pancreas of rats. Diabetes 24, 961-970. Oliver, J. R. (1976). Inhibition of calcium uptake by somatostatin in isolated r a t islets of Langerhans. Endocrinology 99, 910-913. Oliver, J . R., and Wagle, S. R. (1975). Studies on the inhibition of insulin release, glycogenolysis and gluconeogenesis by somatostatin in the r a t islets of Langerhans and isolated hepatocytes. Biochem. Biophys. Res. Commun. 62, 772-777. Orci, L., and Unger, R. H. (1975). Functional subdivision of islets of Langerhans and possible role of D-cells Lancet 2, 1243-1244. Orci, L., Baetens, D., Dubois, M. P., and Rufener, C. (1975a). Evidence for the D-cell of pancreas secreting somatostatin. Horm. Metab. Res. 7, 400-402. Orci, L., Baetens, D., Rufener, C , Amherdt, M., Ravazzola, M., Studer, P., MalaisseLagae, F , and Unger, R. (1975b). Réactivité de la cellule à somatostatine de l'îlot de Langerhans dans le diabète expérimental. C.R. Acad. Sei., Ser. D 281, 1883-1885. Orci, L., Malaisse-Lagae, F , Amherdt, M., Ravazzola, M., Weisswange, A., Dobbs, R., Perrelet, A., and Unger, R. H. (1975c). Cell contacts in human islets of Langerhans. J. Clin. Endocrinol. Metab. 4 1 , 841-844. Parker, D. C , Rossman, L. G., Siler, T. M., Rivier, J., Yen, S. C. C , and Guillemin, R. G. (1974). Inhibition of the sleep-related peak in physiologic human growth hormone release by somatostatin. J. Clin. Endocrinol. Metab. 38, 496-499. Parsons, J., Erlandsen, S., Hegre, O., McEvoy, R., and Eide, R. P. (1976). Central and peripheral localization of somatostatin: Immunoenzyme immunocytochemical studies. J. Histochem. Cytochem. 24, 872-882. Patel, J. C , and Weir, G. C. (1976). Increased somatostatin content of islets from strep tozo toe in-diabetic rats. Clin. Endocrinol. 5, 191-194. Patel, J. C , Orci, L., Bankier, A., and Cameron, D. P. (1976). Decreased pancreatic somatostatin (SRIF) concentration in spontaneously diabetic mice. Endocrinology 99, 1415-1418. Patton, G. S., Dobbs, R., Orci, L., Weil, W., and Unger, R. H. (1976). Stimulation of pancreatic immunoreactive somatostatin (1RS) release by glucagon. Metabolism, Suppl. 25, 1449-1500 Pearse, A. G. E., and Polak, J. (1975). Immunocytochemical localization of substance P in mammalian intestine. Histochemistry 4 1 , 373-375. Pelletier, G., Labrie, F., Arimura, A., and Schally, A. V. (1974). Electron microscopic immunohistochemical localization of growth hormone-release inhibiting hormone (somatostatin) in the r a t median eminence. Am. J. Anat. 140, 445-450.

422

Suad Efendic, Tomas Hökfelt, and Rolf Luft

Pelletier, G., Leclerc, R., Dubé, D., Labrie, F., Puviane, R., Arimura, A., and Schally, A. V. (1975). Localization of growth hormone-release-inhibiting hormone (somatostatin) in the r a t brain. Am. J. Anat. 142, 397-401. Pelletier, G., Leclerc, R., and Dubé, D. (1976). Immunohistochemical localization of hypothalamic hormones. J. Histochem. Cytochem. 24, 864-871. Peracchi, M., Reschini, E., Cantalamessa, L., Giustina, G., Cavagnini, F., Pinto, M., and Bulgheroni, P. (1974). Effect of somatostatin on blood glucose, plasma growth hormone, insulin, and free fatty acids in normal subjects and acromegalic patients. Metabolism 23, 1009-1015. Pernow, B. (1953). Studies on substance P: Purification, occurrence and biological actions. Ada Physiol. Scand., Suppl. 29, 1-90. Petersson, B., Eide, R., Efendic, S., Hökfelt, T., Johansson, O., Luft, R., Cerasi, E., and Hellerström, C. (1977). Somatostatin in the pancreas, stomach and hypothalamus of the diabetic Chinese hamster. Diabetologia (In press.) Phillis, J. W., and Limacher, J . J. (1974). Substance P excitation of cerebral cortical Betz cells. Brain Res. 69, 158-163. Pimstone, B. L., Becker, D., and Kronheim, S. (1975a). Disappearance of plasma growth hormone in acromegaly and protein-caloric malnutrition after somatostatin. J . Clin. Endocrinol. Metab. 40, 168-171. Pimstone, B. L., Le Roith, D., Epstein, S., and Kronheim, S. (1975b). Disappearance rates of plasma growth hormone after intravenous somatostatin in renal and liver disease. J. Clin. Endocrinol. Metab. 4 1 , 392-395. Plotnikoff, N. P., Kastin, A. J., and Schally, A. V. (1974). Growth hormone release inhibiting hormone: Neuropharmacological studies. Pharmacol., Biochem. Behav. 2, 693-696. Polak, J. M. (1976). Localization of substance P, somatostatin, pancreatic polypeptide and skin peptides. J. Endocrinol. 70, 13P-14P. Polak, J. M., Grimelius, L., Pearse, A. G. E., Bloom, S. R., and Arimura, A. (1975). Growth-hormone release-inhibiting hormone in gastrointestinal and pancreatic D cells. Lancet 1, 1220-1222. Polak, J. M., Bloom, S. R., McCrossan, M. V., Arimura, A., and Pearse, A. G. E. (1976). Morphology of somatostatin in gastrointestinal health and disease. Gut 17, 816 (abstr.). Potet, F., Martin, E., Thiery, J. P., Bader, U. P., Bonfils, S., and Lambling, A. (1966). Étude histologique et cytologique du pancréas endocrine, tumoral et non tumoral dans le syndrome de Zollinger-Ellison. Rev. Int. Hepatol. 16, 737-761. Prange, A. J., Breese, G. R., Cott, J. M., Martin, B. K., Cooper, B. R., Wilson, I. C , and Plotnikoff, N. P. (1974). Thyrotropin releasing hormone: Antagonism of pentobarbital in rodents. Life Sei. 14, 447-455. Prange-Hansen, A. A., 0rskov, H., Seyer-Hansen, K., and Lundbaek, K. (1973). Some actions of growth hormone release inhibiting factor. Brit. Med. J. 3 , 523-524. Rasche, H., Raptis, S., Scheck, R., and Pfeiffer, E. F. (1976). Coagulation studies and platelet function after somatostatin infusion. Klin. Wochenschr. (In press.) Renaud, L. P., and Martin, J. B. (1975). Electrophysiological studies of connections of hypothalamic ventromedial nucleus neurons in the rat: Evidence for a role in neuroendocrine regulation. B rain R es. 93, 145-151. Renaud, L. P., Martin, J. B., and Brazeau, P. (1975). Depressant action of TRH, LH-RH and somatostatin on activity of central neurons. Nature (London) 255, 233-235. Rezek, M., Havlicek, V., Hughes, K. R., and Friesen, H. (1976a). Central site of action of somatostatin (SRIF): Role of hippocumpus. Neuropharmacology 15, 499-504.

Somatostatin

423

Rezek, M., Havlicek, V., Hughes, K. R., and Friesen, H. (1976b). Cortical administration of somatostatin (SRIF): Effect on sleep and motor behaviour. Pharmacol., Biochem. Behav. 5, 73-77. Rivier, J., Brown, M., and Vale, W. (1975). D-Trp H -somatostatin: An analog of somatostatin more potent t h a n the native molecule. Biochem. Biophys. Res. Commun. 65, 746-751. Rosenthal, J., Raptis, S., Escobar-Jimenez, F., and Pfeiffer, E. F. (1976). Inhibition of Furosemide induced hyperreninemia by growth-hormone release-inhibiting hormone. Lancet (In press.) Ruch, W., Koerker, D., Carins, M., Johnsen, S., Webster, B., Susinck, J., Goodner, C , and Gale, C. (1973). Studies on somatostatin (somatotropin release inhibiting factor) in conscious baboons. In "Advances in H u m a n Growth Hormone Research" (S. Raiti, ed.), p. 271-294. DHEW Publ. No. (NIH)74-612. U.S. Govt. Printing Office, Washington, D.C. Rufener, C., Amherdt, M., Dubois, M. P., and Orci, L. (1975a). Ultrastructural immunocytochemical localization of somatostatin in r a t pancreatic monolayer culture. J. Histochem. Cytochem. 2 3 , 866-869. Rufener, C , Dubois, M. P., Malaisse-Lague, F., and Orci, L. (1975b). Immuno-ftuore scent reactivity to anti-somatostatin in the gastro-intestinal mucosa of the dog. Diabetologia 11, 321-324. Sakurai, H., Dobbs, R. E., and Unger, R. H. (1975). The effect of somatostatin on t h e response of GLI to the intraduodenal administration of glucose, protein and fat. Diabetologia 11,427-430. Sarantakis, D., McKinley, W. A., J a u n a k a i s , J., Clark, D., and Grant, N. H. (1976a). Structure activity studies on somatostatin. Clin. Endocrinol. Suppl. 5, 275s-278s. Sarantakis, D., Teichman, J., Lien, E. L., and Fenichel, R. L. (1976b). A novel cyclic undecapeptide, Wy-40,770, with prolonged growth hormone release inhibiting activity. Biochem. Biophys. Res. Commun. 73, 336-342. Sawano, S., Baba, Y., Kokubu, T., and Ishizuka, Y. (1974). Effect of synthetic growth hormone-release inhibiting factor on the secretion of growth hormone and prolactin in rats. Endocrinol. Jpn. 2 1 , 399-405. Schauder, P., Mclntosh, C , Ebert, R., Aronds, J., Arnold, R., Frerichs, H., and Creutzfeldt, W. (1976). Insulin and somatostatin release and cAMP content of rat pancreatic islets stimulated by glucose and glucagon. Diabetologia 12,417 (abstr.). Segal, D.S., and Mandell, A. J. (197 4). In "The Thyroid Axis, Drugs and Behaviour" (A. J. Prange, ed.), p. 129. Raven, New York. Sétâlo, G., Vigh, S., Schally, A. V., Arimura, A., and Flerko, B. (1975). GH-RIHcontaining neutral elements in the rat hypothalamus. Brain Res. 90, 352-356. Sherwin, S., Hendler, R., DeFronzo, R. A., Wahren, J., and Felig, P. (1977). Fasting hyperglycemia during prolonged infusion of somatostatin in normal subjects. Proc. Natl. Acad. Sei. U.SA. (In press.) Siler, T. M., VandenBerg, G., Yen, S. S. C , Brazeau, P., Vale, W., and Guillemin, R. G. (1973). Inhibition of growth hormone release in humans by somatostatin. J. Clin. Endocrinol. Metab. 37, 632-635. Siler, T. M., Yen, S. S. C , Vale, W., and Guillemin, R. (1974). Inhibition by somatostatin on the release of TSH induced in m a n by thyrotropin-releasing factor. J. Clin. Endocrinol. Metab. 38, 742-745. Stachura, M. E. (1975). Influence of synthetic somatostatin upon growth hormone release from perifused r a t pituitaries. Endocrinology 99, 678-683. Takahashi, T., and Otsuka, M. (1975). Regional distribution of substance P in the spinal

424

Suad Efendic, Tomas Hökfelt, and Rolf Luft

cord and nerve roots of the cat and the effect of dorsal root section. Brain Res. 87, 1-11. Tamarit, J., Tamarit-Rodriguez, J., Goberna, R., and Lucas, M. (1974). Inhibition of insulin secretion in vitro by somatostatin. Rev. Espan. Fisiol. 30, 299-302. Terenius, L. (1976). Somatostatin and ACTH are peptides with partial antagonist-like selectivity for opiate receptors. Eur. J. Pharmacol. 38, 211-213. Turcot-Lemay, L., Lemay, A., and Lacy, P. E. (1975). Somatostatin inhibition of insulin release from freshly isolated and organ cultured rat islets of Langerhans in vitro. Biochem. Biophys. Res. Commun. 63, 1130-1138. Tyrrell, J. B., Lorenzi, M., Gerich, J. E., and Forsham, P. H. (1975). Inhibition by somatostatin of ACTH secretion in Nelson's syndrome. J. Clin. Endocrinol. Metab. 40, 1125-1127. Unger, R. H., and Orci, L. (1975). The essential role of glucagon in the pathogenesis of diabetes mellitus. Lancet 1, 14-16. Uvnäs-Wallensten, K., Efendic, S., and Luft, R. (1977a). Inhibition of vagally induced gastrin release by somatostatin in cats. Horm. Metab. Res. 9, 120-123. Uvnäs-Wallensten, K., Efendic, S., and Luft, R. (1977b). Release of somatostatin into the antral lumen of cats. Ada Physiol. Scand. 99, 126-128. Vale, W., Brazeau, P., Grant, G., Nussey, A., Burguss, R., Rivier, J., Ling, N., and Guillemin, R. G. (1972). Premières observations sur le mode d'action de la somatostatin, un facteur hypothalamique qui inhibe la sécrétion de l'hormone de croissance. C.R. Acad. Sei., Ser. D 275, 2913-2916. Vale, W., Rivier, C , Brazeau, P., and Guillemin, R. (1974). Effects of somatostatin on the secretion of thyrotropin and prolactin. Endocrinology 95, 968-977. Vale, W., Brazeau, P., Rivier, C , Brown, M., Boss, B., Rivier, J., Burgus, R., Ling, N., and Guillemin, R. (1975). Somatostatin. Recent Progr. Horm. Res. 3 1 , 365-397. Wahren, J., and Felig, P. (1976). Influence of somatostatin on carbohydrate disposal and absorption in diabetes mellitus. Lancet 2, 1213-1216. Wahren, J., Efendic, S., Luft, R., Felig, P., Hagenfeldt, L., and Björkman, 0 . (1977). Influence of somatostatin on splanchnic glucose metabolism in postabsorptive and briefly fasted man. J. Clin. Invest. 59, 299-307. Ward, F. R., Leblanc, H., and Yen, S. C. C. (1975). The inhibitory effect of somatostatin on growth hormone, insulin and glucagon secretion in diabetes mellitus. J. Clin. Endocrinol. Metab. 4 1 , 527-532. Weeke, J., Prange-Hansen, A. A., and Lundbaek, K. (1974). The inhibition by somatostatin of the thyrotropin response to thyrotropin releasing hormone in normal subjects. Scand. J. Clin. Lab. Invest. 33, 101-103. Weeke, J., Prange-Hansen, A. A., and Lundbaek, K. (1975). Inhibition by somatostatin of basal levels of serum thyrotropin (TSH) in normal men. J. Clin. Endocrinol. Metab. 4 1 , 168-171. Yen, S. S. C., Soler, T. M., and DeVane, G. W. (1974). Effect of somatostatin in patients with acromegaly: Suppression of growth hormone, prolactine, insulin and glucose levels. N. Engl. J. Med. 290, 935-938. Yen, S. S. C., Lasley, B. L., Wang, C. F., Leblanc, H., and Silver, T. M. (1975). The operating characteristics of the hypothalamic-pituitary system during the menstrual cycle and observations of biological action of somatostatin. Recent Progr. Horm. Res. 3 1 , 321-363.

RUNE FILIPSSON,* KERSTIN HALL,f and JAN LINDSTENÎ

Dental Maturity as a Measure of Somatic Development in Children I. Introduction II. The Eruption Curve of the Permanent Teeth III. Correlation between Height Development and Dental Maturity in Healthy Girls IV. Correlation between Dental Development and Sexual Maturation in Normal Girls V. Correlation between Somatomedin A in Serum, Dental Maturity, and Height Development VI. Genetic Regulation of Dental Maturity and Height Development VII. Clinical Application of the Tooth Eruption Curve of the Permanent Teeth A. Diagnosis of Growth Disturbances B. Prediction of Adult Height and Age of Sexual Maturation in Girls with Growth Disturbances VIII. Conclusions Appendix References

425 426 431 437 438 440 442 442 444 446 449 450

I. Introduction Information on somatic development is of prime importance for the diagnosis and treatment of children with growth disturbances. Somatic development, or in other terms biological age, normally demonstrates a wide variation among children of the same chronological age. Four variables have mainly been used for the assessment of biological age: skeletal age, height age, secondary sex character age, and dental age. * Department of Orthodontics, Karolinska Institute, Stockholm, Sweden. f Department of Endocrinology and Metabolism, Karolinska Hospital, Stockholm, Sweden. Φ Department of Clinical Genetics, Karolinska Hospital, Stockholm, Sweden. 425

426

Rune Filipsson, Kerstin Hall, and Jan Lindsten

A relatively high correlation has been found between skeletal age, height development, and time of sexual maturation during puberty, while the correlation between these variables before puberty has been reported low or moderate (Tanner, 1962). The correlation between dental maturity on the one hand and skeletal age, sexual maturation, peak height velocity, and actual height attained at various chronological ages on the other, has generally been reported low or moderate (Steggerda, 1945; Lamons and Gray, 1958; Bamba and van Natta, 1959; Meredith, 1959; Lewis and Garn, 1960; Nanda, 1960; Green, 1961; Lauterstein, 1961; Garn et al, 1965; Björk and Helm, 1967). However, some authors have observed that there is an earlier eruption of the deciduous teeth in children exhibiting a rapid growth rate during the first years of life (Lysell et al, 1962; Ninomiya et al, 1966-1968). Moreover, among children of the same chronological age, the taller ones often demonstrate an advanced dental development (Spier, 1918; Cattell, 1928; Talmers, 1952; Garn et al, 1965). These observations were of importance for our decision to study the correlation between dental maturity and height development in healthy children and children with certain growth disturbances in some greater detail. The aim of the present review is to summarize the results that we have obtained in studies on this topic during recent years.

II. The Eruption Curve of the Permanent Teeth Dental maturity was estimated by a newly developed method based on the individual curve of the number of erupted permanent teeth (Filipsson, 1975). The tooth eruption curve constitutes a function with the number of erupted teeth as the dependent variable and the chronological age at each registration as the independent variable (Fig. la). As seen from the figure, the age at the eruption of the first permanent teeth occurs at different ages in different children, and the time required for the completion of the permanent dentition is longer the later the teeth start to erupt. Therefore, a logarithmic scale was introduced on the age axis. The curves for different subjects then become almost parallel and similar in shape (Fig. lb). On this basis, a standard tooth eruption curve was constructed using the mean time of eruption of each tooth (Fig. 2). The values for the first 12 teeth were calculated from the figures obtained in a longitudinal investigation of the time of eruption of the incisors and first molars in Swedish children

Dental Maturity and Somatic Development

427

NUMBER OF ERUPTED PERMANENT TEETH 28-,

5

6

7

8

10

9

11

12

13

15

H

16

17

18

19

20

CHRONOLOGICAL AGE, YEARS (LIN) (a) NUMBER OF ERUPTED PERMANENT TEETH 28 Ί

•

o

1 1

2k

1 1

r 18

1 1

?

! ·"

I

/

i

I

1\

1

I

6

' I

i I

/

I

I n

7

x

/

OOOOX

i

! '*

1

/

I

i

I

/ d

1

8

1 1

X 1 1 1 1

/·

f 12

/

1 rx

T

I"

I

^

I

I

9 10 11 12

I

I

U

I

CHRONOLOGICAL AGE, YEARS

I

I

I

I

I

16 18 20

(LOG)

(b) FIG. 1. Tooth eruption curves from three healthy girls (broken lines) and one girl with pituitary dwarfism (solid line), (a) Linear age scale, (b) Logarithmic age scale. Note that the curves in (b) are more parallel and similar in shape t h a n in (a). (From Filipsson, 1975.)

428

Rune Filipsson, Kerstin Hall, and Jan Lindsten NUMBER OF ERUPTED PERMANENT TEETH 28 1Ù 20· 16 1210 8 6 5 U 3 2

1

5

6 7 8 9 10 11 12 13 U CHRONOLOGICAL AGE, YEARS

FIG. 2. The mean eruption curve of the permanent teeth for boys (solid line) and girls (broken line). (From Filipsson, 1975.)

(Lysell et al., 1969). The values for the remaining part of the curve were obtained from another Swedish study (Dahlberg and Maunsbach, 1948). The eruption curve of a single individual can be predicted by the use of the standard curve even if only a few registrations of the number of erupted permanent teeth are available. The chronological age at which a specific point on this curve occurs can then be used as a measure of dental maturity. Obviously other measures such as the age at the eruption of a specific tooth can also be used for determination of dental maturity. The drawback with such methods is that they require frequent, prospective observations, which must be carried out in all children in order to have access to this information in those relatively few cases where an evaluation of dental maturity might turn out to be of clinical importance. In contrast, the present method can be applied even if only a few, retrospective or prospective registrations are available. In Sweden an eruption status is made annually in most children

Dental Maturity and Somatic Development

429

by the school dentist, and this information is stored for ten years after the child has finished school. Local factors in the bite, such as extraction of deciduous molars as well as crowding and retention of teeth, are known to influence the time of eruption of the permanent teeth (Butler, 1962; Fanning, 1962; Moorrees et al, 1963; Sleicter, 1963; Posen, 1965; Thilander and Jakobsson, 1968; Filipsson, 1974). However, these factors will largely affect the eruption of the premolars and canines, which most often erupt later than the first 12 teeth. Therefore, only the first part of the mean curve up to and including the plateau at the level of 12 erupted teeth (Fig. 2) was used for the determination of dental maturity. It should be pointed out that the shape of the lower part of the curve can occasionally also be disturbed, for instance, by trauma against the deciduous incisors or by the presence of supernumerary teeth (Selliseth, 1970). The variability of the standard curve was tested in 30 randomly chosen subjects in whom at least three registrations before the plateau were available. In 15 of these subjects there was a perfect fit between the registrations and the standard curve. The standard deviation of the difference in the whole material of 30 subjects was 0.0106 In units, which corresponds to approximately 0.1 years at 6 years of age and 0.2 years at 10 years of age. In order to determine the displacement of the individual eruption curve along the age axis when comparing different individuals, a reference point had to be selected on the mean curve. The reference point was chosen at the plateau of the curve, more exactly 0.4 years after the eruption of the 12th tooth. A relatively late reference point on the curve was chosen because information on height at the chronological age of this point was desired for the analysis of the correlation between dental maturity and height development. Since data on height rarely are available before the children have attained elementary school, i.e., at about seven years of age in Sweden, information on height at this point would have been lacking in most children if an earlier reference point had been chosen. The chronological age at the reference point for the individual subject was determined graphically as illustrated in Fig. 3. The number of erupted permanent teeth registered at different ages was plotted on the standard chart. The mean eruption curve with its reference point, traced on a transparent sheet, was then moved horizontally over the chart until the best fit was obtained with the actual observations. This movement of the curve was the only one necessary in most

430

Rune Filipsson, Kerstin Hall, and Jan

Lindsten

NUMBER OF ERUPTED PERMANENT TEETH 12 10 8 6 5

3

2-

1 5

6

7

8

9

CHRONOLOGICAL AGE,

10 11 12 YEARS

FIG. 3. Determination of the age at the reference point by the mean tooth eruption curve. The mean curve with its reference point traced on a transparent sheet (shaded area) is moved along the x axis to the best fit with the registration available (crosses) in the individual case. (From Filipsson, 1975).

cases, since the plateau was generally located at the level of 12 teeth. In a few subjects (4.8%), the plateau of the tooth eruption curve occurred at a lower level, and accordingly the mean curve had to be lowered in order to obtain the best fit with the actual observations (Fig. 4). However, vertical displacement of the standard curve for adjustment of the plateau, still keeping the fit with the earlier registrations, could only be performed with a logarithmic scale also on the y axis. It should be pointed out t h a t at least two observations at the level of the plateau are needed in order to demonstrate t h a t it is lowered, and t h a t the determination of the age at the reference point only was based on the eruption of the incisors and first molars, which does not allow an evaluation of elevation of the eruption curve. The chronological age at the reference point, in the following abbreviated A , was determined in 133 boys and 137 girls. These children were selected at random among apparently healthy school children in the town of Umeâ in Sweden. The children were examined annually with respect to the number of erupted permanent teeth. As seen from Fig. 2 the mean curve is almost identical in shape in boys and girls. The boys showed a slightly later tooth eruption than the girls, which is in agreement with previous studies (Dahlberg and Maunsbach, 1948; Clements et al., 1953; Carlos and Gittelsohn, 1965; de Boer, 1970). The

431

Dental Maturity and Somatic Development NUMBER OF ERUPTED PERMANENT TEETH

1

1

I

I

5 6 7 8 9 10 11 12 CHRONOLOGICAL AGE, YEARS

FIG. 4. Determination of the age at the reference point in a subject with a tooth eruption curve which is lowered to the level of 10 teeth along they axis. (From Filipsson, 1975.)

mean age at A was 8.75 years (S.D. = 0.73) for the boys and 8.44 years (S.D. = 0.68) for the girls. The difference between these means was statistically significant (p < 0.001). The distribution of the age at the reference point did not differ significantly from a normal distribution when tested for goodness of fit by χ2 analysis, 0.80 < p < 0.90 for the boys and 0.30