Taking on TIVA: Debunking Myths and Dispelling Misunderstandings [1 ed.] 1316609367, 9781316609361

Table of contents :
Cover
Taking on TIVA:

Debunking Myths and Dispelling Misunderstandings
Copyright
Contents
Contributors
Foreword
Introduction: Power to the People: the Rationale
of a Practical Text
1 Why Bother?
2 You Say ‘PK’ and I Say ‘No Way!’; You Say
‘Keo’ and I Say ‘Time to Go!’
3 TCI and TIVA: What a Good Idea!
4 Milk of Amnesia
5 A Catwalk with a Difference
6 Let’s Get Started
7 Let’s Get Pumped!
8 ‘But I’m Used to MAC!’
9 Be Aware, Unaware and Confusion
Everywhere
10 Do You Want Fries with That?
11 Intra- and Post-operative Analgesia for TIVA
12 Wakey Wakey!
13 Under Pressure
14 Ankle Biters
15 Old Timers
16 Big Can Be Beautiful!
17 A Bun in the Oven
18 Saving the Whales by Taking on TIVA
19 TIVA Drugs for Sedation
20 Skiing Off-Piste and Other Assorted Goodies
Index

Citation preview

Taking on TIVA

Taking on TIVA Debunking Myths and Dispelling Misunderstandings Edited by

Michael G. Irwin University of Hong Kong

Gordon T. C. Wong University of Hong Kong

Shuk Wan Lam University of Hong Kong

University Printing House, Cambridge CB2 8BS, United Kingdom One Liberty Plaza, 20th Floor, New York, NY 10006, USA 477 Williamstown Road, Port Melbourne, VIC 3207, Australia 314–321, 3rd Floor, Plot 3, Splendor Forum, Jasola District Centre, New Delhi – 110025, India 79 Anson Road, #06–04/06, Singapore 079906 Cambridge University Press is part of the University of Cambridge. It furthers the University’s mission by disseminating knowledge in the pursuit of education, learning, and research at the highest international levels of excellence. www.cambridge.org Information on this title: www.cambridge.org/9781316609361 DOI: 10.1017/9781316659069 © Cambridge University Press 2020 This publication is in copyright. Subject to statutory exception and to the provisions of relevant collective licensing agreements, no reproduction of any part may take place without the written permission of Cambridge University Press. First published 2020 Printed in the United Kingdom by TJ International Ltd, Padstow, Cornwall A catalogue record for this publication is available from the British Library. Library of Congress Cataloging-in-Publication Data Names: Irwin, Michael G., editor. Title: Taking on TIVA : debunking myths and dispelling misunderstandings / [edited by] Michael Irwin, Gordon Wong, Shuk Wan Lam. Description: New York : Cambridge university press, 2020. | Includes bibliographical references and index. Identifiers: LCCN 2019029316 (print) | LCCN 2019029317 (ebook) | ISBN 9781316609361 (hardback) | ISBN 9781316659069 (epub) Subjects: LCSH: Intravenous anesthesia. Classification: LCC RD85.I6 T35 2020 (print) | LCC RD85.I6 (ebook) | DDC 617.9/62–dc23 LC record available at https://lccn.loc.gov/2019029316 LC ebook record available at https://lccn.loc.gov/2019029317 ISBN 978-1-316-60936-1 Paperback Cambridge University Press has no responsibility for the persistence or accuracy of URLs for external or third-party internet websites referred to in this publication and does not guarantee that any content on such websites is, or will remain, accurate or appropriate.

Every effort has been made in preparing this book to provide accurate and up-to-date information that is in accord with accepted standards and practice at the time of publication. Although case histories are drawn from actual cases, every effort has been made to disguise the identities of the individuals involved. Nevertheless, the authors, editors, and publishers can make no warranties that the information contained herein is totally free from error, not least because clinical standards are constantly changing through research and regulation. The authors, editors, and publishers therefore disclaim all liability for direct or consequential damages resulting from the use of material contained in this book. Readers are strongly advised to pay careful attention to information provided by the manufacturer of any drugs or equipment that they plan to use.

Contents List of Contributors vii Foreword ix Introduction: Power to the People: the Rationale of a Practical Text

1 Why Bother? The Advantages of TIVA 1 Michael G. Irwin and Gordon T. C. Wong 2 You Say ‘PK’ and I Say ‘No Way!’; You Say ‘Keo’ and I Say ‘Time to Go!’: Pharmacokinetics for TIVA 5 Frank Engbers and Michael G. Irwin 3 TCI and TIVA: What a Good Idea! 14 Michael G. Irwin and Gavin N. C. Kenny 4 Milk of Amnesia: Propofol for TIVA Kane O. Pryor and Paul S. Myles

22

5 A Catwalk with a Difference: What Distinguishes TIVA Models? 31 Lim Thiam Aun 6 Let’s Get Started: The Know-Hows and NoHows of Setting Up TIVA 40 Michael G. Irwin and Gordon T. C. Wong 7 Let’s Get Pumped! The Nitty Gritty of TIVA Syringe Pumps 46 Nigel J. Huggins 8 ‘But I’m Used to MAC!’: How Do I Get the Dose Right with TIVA? 52 Nicholas Peter Sutcliffe and Michael G. Irwin 9 Be Aware, Unaware and Confusion Everywhere: TIVA and Awareness 63 Pablo Martinez-Vazquez, Claus Lindner, Umberto Melia, Jaideep J. Pandit and Erik Weber Jensen

x

10 Do You Want Fries with That? The Role and Effects of Adjuvant Medications in TIVA 73 Vivian Man Ying Yuen 11 Intra- and Post-operative Analgesia for TIVA 80 Talmage Egan, Byron Bankhead and Greg Smith 12 Wakey Wakey! Smooth Recovery from TIVA 95 Peter J. van der Mast and Anthony R. Absalom 13 Under Pressure: TIVA in Emergency Surgery 106 Anup Biswas, Gordon T. C. Wong and Michael G. Irwin 14 Ankle Biters: How to Use TIVA in Children 111 Brian Anderson and James Houghton 15 Old Timers: How to Use TIVA in the Elderly 124 Fernando Neira-Reina, J. Luisa OrtegaGarcía and Luis Miguel Torres 16 Big Can Be Beautiful! How to Use TIVA in the Obese 132 Luis Ignacio Cortínez and Brian Anderson 17 A Bun in the Oven: How to Use TIVA in Obstetrics 139 Pamela Flood and Jessica Ansari

v

Contents

18 Saving the Whales by Taking on TIVA: The Environmental Impact of TIVA 146 Vincent K. F. Kong

20 Skiing Off-Piste and Other Assorted Goodies: Advanced TIVA 162 Gordon T. C. Wong and Michael G. Irwin

19 TIVA Drugs for Sedation 154 Stefan Schraag Index

vi

170

Contributors

Anthony R. Absalom, MBChB, FRCA, MD Department of Anesthesiology, University of Groningen, The Netherlands Brian Anderson, MBChB, PhD, Dip Obstet, FANZCA, FCICM Department of Anaesthesiology, University of Auckland, New Zealand Jessica Ansari, MD Department of Anesthesiology, Stanford University Medical Center, CA, USA Byron Bankhead, MD Acute Pain Service, University of Utah, UT, USA Anup Biswas, MBBS, FRCA Department of Obstetric/Adult Anesthesiology, Sidra Women’s and Children’s University Medical and Research Center, Doha, Qatar Luis Ignacio Cortínez, MD Department of Anesthesiology, Clinical Hospital Catholic University, Santiago, Chile Talmage Egan, MD Department of Anesthesiology, University of Utah, UT, USA Frank Engbers, MD, FRCA Leiden University Medical Centre, The Netherlands Pamela Flood, MD, MA Perioperative and Pain Medicine (OB), Stanford University Medical Center, CA, USA James Houghton, BSc, MBChB, FANZCA Department of Paediatric Anaesthesia, Starship Children’s Hospital, Auckland, New Zealand Nigel J. Huggins, FRCA Department of Neuro-Anaesthesia, Queen Elizabeth Hospital Birmingham, UK

Michael G. Irwin, MBChB, MD, FRCA, FCAI, FANZCA, FHKCA, FHKAM Department of Anaesthesiology, University of Hong Kong, Hong Kong Erik Weber Jensen, PhD R&D Department Quantium Medical/Fresenius Kabi and Automatic Control and Informatic (ESAII) Department, Centre for Biomedical Research (CREB) UPC-Barcelonatech, Systems Pharmacology Effect Control and Modeling (SPEC-M) Research Group, Anesthesiology Department, Hospital Clinic de Barcelona, Spain Gavin N. C. Kenny, MSc, MBChB, MD Glasgow University Department of Anaesthesia, Glasgow, UK Vincent K. F. Kong, MBBS, MSc, PGDipEcho, FANZCA, FHKCA, FHKAM Department of Anaesthesiology, Gleneagles Hospital Hong Kong, Hong Kong Claus Lindner, PhD R&D Department, Quantium Medical/Fresenius Kabi, Barcelona, Spain Pablo Martinez-Vazquez, MSc, PhD R&D Department, Quantium Medical/Fresenius Kabi, Barcelona, Spain German Primate Center, Department of Cognitive Neuroscience, Goettingen, Germany Umberto Melia, PhD R&D Department, Quantium Medical/Fresenius Kabi, Barcelona, Spain Paul S. Myles, MBBS, MPH, MD, DSc, FCAI, FANZCA, FRCA Department of Anaesthesia and Perioperative Medicine, Alfred Hospital and Monash University, Melbourne, VIC, Australia

vii

List of Contributors

Fernando Neira-Reina, PhD, MD Department of Anaesthesia, Cádiz University, Andalusia, Spain J. Luisa Ortega-García, PhD, MD Department of Anaesthesia, Cádiz University, Andalusia, Spain Jaideep J. Pandit, MA, BM, DPhil, FRCA, FFPMRCA, DM Nuffield Department of Anaesthetics, Oxford University Hospitals Foundation Trust, Oxford, UK Kane O. Pryor, MD Clinical Anesthesia, and Academic Affairs, Weill Cornell Medicine, New York, NY, USA Stefan Schraag, MD, PhD, FRCA, FFICM Department of Anaesthesia, Golden Jubilee National Hospital, Clydebank, Glasgow, UK Greg Smith, MD University of Utah, UT, USA Nicholas Peter Sutcliffe, MD, FRCA Deputy Chairman, Information and Technology, Department of Anesthesiology, ICU and

viii

Perioperative Medicine, Hamad Medical Corporation, Doha, Qatar Lim Thiam Aun, MBBS, MD, FRCA, FFARCSI, FAMM Department of Anaesthesiology, Faculty of Medicine and Health Sciences, University Putra, Malaysia Luis Miguel Torres, MD, PhD Department of Anesthesia, University of Cádiz, Spain Peter J. van der Mast, MD Department of Anesthesiology, Ommelander Hospital Group, Scheemda, The Netherlands Gordon T. C. Wong, MBBS, BSc, MD, FANZCA, FHKCA, FHKAM Department of Anaesthesiology, Hong Kong University, Hong Kong Vivian Man Ying Yuen, MBBS, MD, FHKCA, FHKAM, FANZCA Department of Anaesthesiology and Perioperative Medicine, Hong Kong Children’s Hospital and Queen Mary Hospital, Hong Kong University, Hong Kong

Foreword

Having been personally convinced, in 1990, of the merit of the technique of target-controlled infusion (TCI) by path-finding experts in Germany, Holland, Belgium, France, Switzerland, the UK and the USA, it took a further couple of years before ICI Pharmaceuticals Division (now Zeneca) agreed to begin the development of what became the ‘Diprifusor’ TCI system for the administration of propofol. I became involved in the organisation of the clinical validation of this system with the Marsh pharmacokinetic (PK) model and took part in many discussions with infusion pump manufacturers, and drug and device regulatory authorities, before its approval in most countries of the world from 1996 onwards – one significant exception being our failure to gain approval in the USA. As cheaper generic propofol began to appear, in 2002, there was a demand for ‘open’ TCI systems, which did not require an electronically tagged presentation to confirm the presence and concentration of propofol. These systems provided a choice of PK model for propofol and allowed control of blood or effect-site concentrations, the latter being achieved in some cases with a choice of keo values. The introduction of Diprifusor TCI systems had been accompanied by an extensive programme of lectures and demonstrations and Gavin Kenny travelled the world with missionary zeal to describe and advocate the use of this system. With the arrival of the open systems, drug delivery with the different models, particularly at induction of anaesthesia, can be quite different. While different models and modes of administration provide opportunities for studies by enthusiasts, in my view they also make the TCI technique appear more complex to less experienced anaesthetists. I was disappointed to read that more than 20 years after the introduction of propofol TCI, with its benefits of rapid, clear-headed recovery and a low frequency of post-operative emetic events, the NAP5 activity survey on awareness found that propofol TIVA was used in only 8% of general anaesthetics in the UK.

Michael Irwin, Professor of Anaesthesia at Hong Kong University, gained experience of total intravenous anaesthesia (TIVA) and TCI with Gavin Kenny during his training in Glasgow Royal Infirmary. In Hong Kong, he and his colleagues have continued to study and publish on TIVA techniques and in a recent large survey of anaesthetists, the results of which are summarised in this book, have identified the barriers considered to be important by infrequent and frequent TIVA users. A prominent finding was that many of the factors considered to be barriers by infrequent users were not considered to be significant barriers once more experience had been gained. The aim of the present book is to present a practical approach to TIVA to encourage those with little experience of the technique to begin to use TIVA in healthy patients or in patients desiring sedation during procedures conducted with regional anaesthesia. In this way experience can be gained in titrating the target concentration to achieve the depth of anaesthesia or sedation desired – taking on more complex cases where TIVA may be required. With open TCI pumps, new users can use a saline-filled syringe to play with the pump, inputting different targets and patient characteristics, and observing the amounts of drug given at induction and the changing infusion rate over time. Becoming familiar with setting up a pump and the appropriate infusion sets before using it on a patient should help to reduce the clinical set-up time. The authors have gathered an international collection of experts in each of the areas covered and many practical approaches for example patients are provided. Instructed to write in a clear and easily readable style, I believe the authors and editors have accomplished their objective. With the approach adopted in this book, I am confident that more infrequent users of TIVA will be encouraged to join the TIVA train. Iain Glen, BVMS, PhD, MRCVS, DVA, FRCA

ix

Introduction: Power to the People: the Rationale of a Practical Text

No one throws a new-born baby onto the street with the advice ‘take care of yourself kid’. On the other hand, they wouldn’t use Plato’s The Republic for their child’s bedtime reading. You nurture children and give them the knowledge and confidence to explore the world. We appreciate that anaesthetists are probably the most knowledgeable and experienced clinicians in pharmacology (applied pharmacology is what we do!) but many have not been properly trained in the use of TIVA. Whatever you think about TIVA, it is essential for us to be able to use this important modality (see Chapter 1). A useful analogy is that you can drive a car without knowing exactly how the engine works. Most anaesthetists using inhalational agents will not be able to tell you the saturated vapour pressure or oil:gas solubility, although they should know the minimum alveolar concentration (MAC) and have a rough idea of the relative blood:gas solubility and the basic workings of a vaporiser. We believe therefore that there is a need for a practical book that enables anaesthetists to begin propofol-based TIVA safely and with confidence. It complements more detailed textbooks and review articles and is intended more as a guidebook than a definitive text. TIVA is not complicated. The PK models that are used for stable drug delivery are, of course, complex but modern infusion pumps take care of this for us. In fact, these TCI pumps can be thought of as ‘vaporisers’ for TIVA. Most inhalational agents have both analgesic and anaesthetic properties but with TIVA we control anaesthesia/consciousness with one drug (propofol) and analgesia with another (a titratable opioid such as remifentanil). This is an important distinction but one that may also have advantages now that there is increasing concern over the adverse effects of excessively deep

x

anaesthesia. It also means that we need to change our way of thinking about anaesthesia when using TIVA as inhalation and intravenous (IV) skills are not directly transferable. As well as pharmacokinetics (what the body does to the drug) and pharmacodynamics (what the drug does to the body) we also have pharmacogenetics. We are all different (fortunately!) and the same applies to our hypnotic and analgesic requirements. With inhalational drugs there is MAC. This is a statistical concept introduced more than 50 years ago and is the concentration of a vapour in the lungs that is needed to prevent movement in 50% of subjects in response to a surgical (pain) stimulus. By definition it will be either too high or too low for most patients. We have a similar concept with propofol (effective concentration 50 or EC50), which can guide us to the average dose requirement for loss of consciousness. However, with TIVA it is much easier to titrate the effectsite concentration of propofol to individual requirements (clinical response) as we start and maintain anaesthesia with the same drugs rather than guessing as, essentially, happens with MAC. We have a geographically diverse collection of authors who have helped us to present the most important topics in TIVA in a clear and easily readable style. Our vision is to empower readers to start or continue using this technique. Ideally this should be in young, healthy patients undergoing relatively straightforward surgery without the use of muscle relaxants. As your experience grows so then can your range of TIVA practice. Eventually, like us, you will use it for everything! Michael G. Irwin

Chapter

1

Why Bother? The Advantages of TIVA Michael G. Irwin and Gordon T. C. Wong

Why Total Intravenous Anaesthesia? Like many of you, we’re sure, we were trained to use IV anaesthetic agents for induction of anaesthesia but inhalational for maintenance – a sensible and seemingly safe combination that has been used for decades. So why change? The initial attraction of TIVA was the extremely rapid, smooth and clear-headed recovery of patients when using propofol as the hypnotic component of an anaesthetic. This is particularly apparent when the drug is used for cases of short to intermediate duration, for example in day-case surgery with earlier discharge from the post-anaesthetic care unit.[1] Clearly in modern practice, which is moving towards shorter in-patient stays, this represents a major advantage. In addition, improved levels of patient satisfaction occur with TIVA, presumably due to the favourable recovery profile.[2] Certainly, desflurane and sevoflurane allow rapid recovery but it is not as smooth, there may be more emergence delirium and quality indicators are not as good.[3] A study using psychomotor tests to compare the performance of volunteers at different blood alcohol concentrations with performance after anaesthesia with propofol and remifentanil, showed that about 40 minutes after TIVA, patients were sufficiently recovered to be able to drive in continental Europe with a blood alcohol concentration of 50 mg.100 ml−1 and after about one hour they were considered suitable to drive in Sweden with its lower legal alcohol level of 20 mg.100 ml−1.[4] There are many systems in the body that are affected adversely by inhalational anaesthetics; these include the lungs, liver, kidneys and heart. In addition, rare conditions such as malignant hyperthermia can be triggered in susceptible individuals by inhalation, but not by IV, agents. Neurotoxicity has also been reported in transgenic mice with the use of sevoflurane.[5] Inhalational anaesthetics have been assessed for use as sedatives in the intensive care

unit (ICU) but were universally rejected because of concerns about toxicity. Recently there has been considerable interest in the role of the peri-operative period in recurrence after cancer surgery and the influence of anaesthetic techniques.[6] A study examined long-term survival for patients undergoing inhalation versus IV anaesthesia for cancer surgery. The mortality was reported to be 50% greater with inhalation than with IV anaesthesia. Although this was retrospective data, propofol displays anti-tumour properties in both in vivo and in vitro laboratory settings, enhances cytotoxic T lymphocyte activity, inhibits cyclo-oxygenase (COX) in vitro[7] and hypoxia-inducible factor (HIF) 1-α activation.[8] Prospective research in this field is on-going. Post-operative nausea and vomiting (PONV) has been considered a relatively unimportant consequence of anaesthesia but its occurrence has a significant impact on post-anaesthetic morbidity and increases overnight admission rates. For many patients it was reported as being a more unpleasant adverse effect than post-operative pain.[9] Propofol also reduces pain after surgery compared to inhalational anaesthesia.[10] The final benefit of IV drugs is that this will be the future course of development of all new agents. New anaesthetic drugs will be developed at the molecular level, rather than relying on serendipity, and will be administered intravenously. TIVA and therefore TCI are likely to be the future for anaesthesia.

So If TIVA Is So Good Why Aren’t More People Using It? We too were interested in knowing why more people aren’t using TIVA – although we had our theories.[11] Given our bias, we performed a study to collect objective data. We surveyed an international audience of anaesthetists[12] and divided the respondents into

Downloaded from https://www.cambridge.org/core. IP address: 5.189.207.239, on 02 Dec 2019 at 21:40:01, subject to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316659069.003

1

Chapter 1: Why Bother? The Advantages of TIVA

Table 1.1 Reasons for not choosing to use TIVA on a particular occasion. Values are the number of respondents considering the respective reasons ‘extremely important’ or ‘very important’ over the total number of recipients in that group (%). Modified from Wong et al.[12]

Reasons for not using TIVA

Infrequent users (50%) (n = 124)

Additional effort

164 (52%)

48 (15%)

17 (14%)

Lack of real-time monitor of propofol concentration

127 (40%)

56 (17%)

5 (4%)

Risk of missing drug delivery failure

108 (34%)

16 (5%)

5 (4%)

Institutional preference

106 (34%)

34 (11%)

14 (11%)

IV access invisible or inaccessible

104 (33%)

21 (7%)

4 (3%)

Increase turnover time

99 (31%)

25 (8%)

2 (2%)

Volatile is better

96 (30%)

29 (9%)

7 (6%)

Large inter-patient variability in dose requirements

93 (30%)

57 (18%)

9 (7%)

Difficult to predict wake-up

93 (30%)

23 (7%)

6 (5%)

Unavailability of depth-ofanaesthesia monitoring

91 (29%)

25 (8%)

9 (7%)

Additional expense

89 (28%)

25 (8%)

2 (2%)

Difficult to titrate doses to clinical needs

77 (25%)

47 (15%)

7 (6%)

No outcome benefits with TIVA

74 (24%)

113 (35%)

14 (11%)

Increased incidence of awareness

69 (22%)

27 (8%)

2 (2%)

Unavailability of TCI pumps

68 (22%)

43 (13%)

4 (3%)

Pharmacokinetic models not accurate

54 (17%)

56 (17%)

8 (6%)

Greater likelihood of cardiovascular instability

54 (17%)

91 (28%)

27 (22%)

Difficulty in titrating analgesia on emergence

53 (17%)

76 (24%)

12 (10%)

Creates crowded conditions around patient

48 (15%)

48 (15%)

3 (2%)

Complicated pharmacokinetic models

48 (15%)

79 (24%)

16 (13%)

Long induction time

44 (14%)

69 (21%)

7 (6%)

Difficult IV access

42 (13%)

94 (29%)

27 (22%)

Unavailability of non-reflux/oneway valves

38 (12%)

50 (15%)

11 (9%)

Insufficient training in the use of TIVA

34 (11%)

62 (19%)

30 (24%)

three groups depending on their frequency of TIVA usage. Infrequent users are those using TIVA for less than 5% of cases; those who use TIVA more than 50% of the time were designated as frequent users; while those using between 5 and 50% were considered intermediate users. We asked them to consider a list of factors and indicate whether they thought each was extremely, very, moderately, not very or not at all

important to their decision for not choosing to use TIVA on a particular occasion. Table 1.1 summarises some of the findings. We refer the readers to the paper for detailed discussion but several interesting points can be seen from a cursory look at the table. The most striking is the difference in the perception of importance to the same factor depending on how frequently one uses

Downloaded from https://www.cambridge.org/core. IP address: 5.189.207.239, on 02 Dec 2019 at 21:40:01, subject to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316659069.003

2

Chapter 1: Why Bother? The Advantages of TIVA

TIVA. For example, presumably anaesthetists are aware of the difference in reported rate of accidental awareness under anaesthesia between inhalational and intravenous techniques. Yet there is a ten-fold difference in the response of those who perceived it to be important depending on whether they were a frequent TIVA user or not. A similar observation is applicable to their perception of having a lack of real-time monitoring of propofol concentration. A second point is the importance of non-technical factors in their decision not to use TIVA, with ‘additional effort’, ‘institutional preference’ and ‘increased turnover time’ ranked high on the list. Interestingly, a contemporaneous survey to ours of 1000 anaesthetists from the Australian and New Zealand College of Anaesthetists,[13] 18% of whom use TIVA in the majority of cases, indicated that 41% would use TIVA more often if set-up were easier. What do we conclude from all this? We think that non-technical factors play a significant role in our choice to use TIVA or not and decisions are not only based on ‘evidence’. An example is the testimony of Dr Nick Sutcliffe, the author of Chapter 8, whose conversion to TIVA enthusiast was based on an entirely different motivation. As seen in the next section, rather than being deterred by the additional effort, Nick was in fact motivated by the excitement of using more thought and effort!

So Tell Me Again, Why Bother? In most people’s anaesthetic career, there will come a time when one will have to use TIVA – like it or not, this is an essential skill for the modern anaesthetist. If your patients are immune to PONV or malignant hyperthermia, or do not require surgery for cancer, or it is always technically possible to use the inhalational route, then you may be a little less motivated to develop and maintain competency in TIVA. However, for the rest of us: will you be confident when required to use TIVA? Murphy’s Law would probably have it that such occasions occur in less than optimal circumstances. Intuitively we know that we are more inclined to make errors when dealing with an unfamiliar technique and this may be a major reason why the incidence of accidental awareness is reported to be higher with TIVA. So, if not for any other reasons than for patient safety, you should be

competent and maintain competency in TIVA by regular use and an understanding of the principles. Are there any drawbacks of TIVA? Of course! At present, we do not have a reliable and convenient way of detecting disconnection or non-delivery of the drug, although research and technology is always advancing. With practice, vigilance and following the tips and tricks outlined in this text, drawbacks can be reduced and you should feel completely comfortable with this technique. Humankind has always struggled, and will continue to struggle, with the problem of not doing something despite knowing it is ‘good’. An example is the simple health advice of consuming fewer calories and exercising more. Most people in the affluent world know that this is good advice and yet how many people follow that consistently? There are a multitude of benefits with TIVA. If you were a patient, would you want your anaesthetist to make the effort and use a technique that would confer these benefits? If so, don’t you owe it to your patients to do the same?

References 1. M.M. Sahinovic, M. Struys, A.R. Absalom. Clinical pharmacokinetics and pharmacodynamics of propofol. Clin Pharmacokinet 2018; 57: 1539–58. 2. C.K. Hofer, A. Zollinger, S. Buchi, et al. Patient well-being after general anaesthesia: a prospective, randomized, controlled multi-centre trial comparing intravenous and inhalation anaesthesia. Br J Anaesth 2003; 91: 631–7. 3. D.D. Wong, C.R. Bailey. Emergence delirium in children. Anaesthesia 2015; 70: 383–7. 4. J.A. Murdoch, S.A. Grant, G.N. Kenny. Safety of patient-maintained propofol sedation using a target-controlled system in healthy volunteers. Br J Anaesth 2000; 85: 299–301. 5. Y. Lu, X. Wu, Y. Dong, Z. Xu, Y. Zhang, Z. Xie. Anesthetic sevoflurane causes neurotoxicity differently in neonatal naive and Alzheimer disease transgenic mice. Anesthesiology 2010; 112: 1404–16. 6. T.J. Wigmore, K. Mohammed, S. Jhanji. Long-term survival for patients undergoing volatile versus IV anesthesia for cancer surgery: a retrospective analysis. Anesthesiology 2016; 124: 69–79. 7. T. Inada, K. Kubo, T. Kambara, K. Shingu. Propofol inhibits cyclo-oxygenase activity in human monocytic THP-1 cells. Can J Anesth 2009; 56: 222. 8. H. Zhao, M. Iwasaki, J. Yang, S. Savage, D. Ma. Hypoxiainducible factor-1: a possible link between inhalational

Downloaded from https://www.cambridge.org/core. IP address: 5.189.207.239, on 02 Dec 2019 at 21:40:01, subject to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316659069.003

3

Chapter 1: Why Bother? The Advantages of TIVA

anesthetics and tumor progression? Acta Anaesthesiol Taiwan 2014; 52: 70–6. 9. L.H.J. Eberhart, A.M. Morin, H. Wulf, G. Geldner. Patient preferences for immediate postoperative recovery. Br J Anaesth 2002; 89: 760–1. 10. Q. Qiu, S.W. Choi, S.S.C. Wong, M.G. Irwin, C. W. Cheung. Effects of intra-operative maintenance of general anaesthesia with propofol on postoperative pain outcomes – a systematic review and meta-analysis. Anaesthesia 2016; 71: 1222–33.

11. M.G. Irwin, G.T.C. Wong. Taking on TIVA. Why we need guidelines on total intravenous anaesthesia. Anaesthesia 2018. 12. G.T.C. Wong, S.W. Choi, D.H. Tran, H. Kulkarni, M. G. Irwin. An international survey evaluating factors influencing the use of total intravenous anaesthesia. Anaesth Intensive Care 2018; 46: 332–8. 13. A. Lim, S. Braat, J. Hiller, B. Riedel. Inhalational versus propofol-based total intravenous anaesthesia: practice patterns and perspectives among Australasian anaesthetists. Anaesth Intensive Care 2018; 46: 480–7.

Downloaded from https://www.cambridge.org/core. IP address: 5.189.207.239, on 02 Dec 2019 at 21:40:01, subject to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316659069.003

4

Chapter

2

You Say ‘PK’ and I Say ‘No Way!’; You Say ‘Keo’ and I Say ‘Time to Go!’ Pharmacokinetics for TIVA Frank Engbers and Michael G. Irwin

What the Drug Does to the Body and What the Body Does to the Drug. . . Anaesthesia is possibly the most pharmacology oriented of all clinical medical specialities. What we do every day is, effectively, applied pharmacology. Yet with all this practical experience, some aspects, particularly pharmacokinetics (PK), can appear dauntingly complex. Indeed, mathematical modelling and developing targetcontrolled infusions is difficult as is the design of modern vaporisers for inhalational drugs but, as an analogy, we don’t need to be able to design a car in order to know how to drive it. However, there are certain basic PK features that will enhance your understanding and improve your ability to use these drugs appropriately.

What Are These Compartments? After an IV drug bolus, the plasma concentration follows an exponential decline in three distinct phases. There is a rapid initial drop, then the decrease slows and, finally, there’s a steady but stable decrease. These observations can be explained by the distribution (movement) of the drug between a central Three-compartment model Drug input

Peripheral compartment V2

Central compartment V1

Effect site

compartment (called V1; this is principally the plasma) and two further compartments, which equilibrate rapidly with V1 (Figure 2.1): tissue with high blood flow such as muscle (V2); and tissue with less blood flow but which has a huge capacity to absorb the drug, such as fat and bone (V3). Of course, these are not distinct anatomical compartments but are a very good way to conceptualise what is happening. The compartments can also be visualised as a ‘hydraulic model’ (Figure 2.2). The central compartment (V1) will equilibrate with the effect site (in the case of anaesthetic and opioid drugs, this is the brain) depending on how quickly it can cross the blood–brain barrier. As a student, I was told that this is one ‘arm–brain’ circulation time but this is clearly nonsense. The speed will depend on the physicochemical properties of the drug, how high its concentration is (a higher concentration obviously has a faster onset) and subtler factors such as membrane transporter proteins and even the patient’s genetic make-up (another reason for titration). Some drugs will have a slower onset (e.g. propofol, fentanyl) than others (e.g. thiopentone, remifentanil). This is covered in more detail below where we describe keo. Figure 2.1 A three-compartment model divides the body into a central and two peripheral compartments. The central compartment (V1) consists of the plasma and tissues where the distribution of the drug is practically instantaneous. The peripheral compartments (V2 and V3) consist of tissues where the distribution of the drug is slower compared to V1 but the capacity is large.

Peripheral compartment V3

Elimination

Downloaded from https://www.cambridge.org/core. IP address: 193.56.73.28, on 02 Dec 2019 at 21:40:43, subject to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316659069.004

5

Chapter 2: Pharmacokinetics for TIVA

Intravenous infusion

Plasma concentration

3 volumes of distribution 3 clearances

Larger container but smaller entrance pipe

SLOW

V3 Metabolism and elimination

Figure 2.2 A three-compartment hydraulic model. The same concept as Figure 2.1 but an easier way to conceptualise it.

FAST

Cl3 Cl1

V1

CI2 V2

Smaller container but larger entrance pipe

Figure 2.3 shows what happens in the compartments after a short infusion and Figure 2.4, after a longer infusion. The drug’s effect (brain concentration) will depend on the concentration in the central compartment (V1) and this is also the compartment from which the drug is finally eliminated from the body because it also equilibrates with the kidneys, liver and lungs. The other compartments act as a huge drug reservoir with movement backwards and forwards that can be described by complex mathematical equations that, fortunately, we don’t need to know as they are programmed into the TCI pumps. Context sensitive half-time is defined as the time taken for the blood plasma concentration of a drug to decline by one half after an infusion designed to maintain a steady state (i.e. a constant plasma concentration) has been stopped. The ‘context’ is the duration of infusion and, almost always, increases with time. You can see above that as the peripheral compartments become full (saturated), the cessation of drug effect will depend more and more on metabolism and elimination from the central compartment. This is what causes the half-time to, generally, increase with the duration of infusion and is clearly illustrated with different opioids (Figure 2.5 and Chapter 20).

What is Keo? Early pharmacological research focused mainly on models and equations that described the dose–blood concentration relationship of intravenously administered drugs. There are multiple ways to describe this

relationship but one of the models comprised compartmental analyses where the drug is assumed to be administered in a central (blood) compartment, then distributed and redistributed to one or more peripheral compartments and, finally, eliminated from this central compartment. In these early days of pharmacology research[1] it was assumed that the blood concentration was immediately relatable to an effect. Targetcontrolled infusion uses such a compartmental model to reverse this dose–blood concentration relationship to calculate the infusion rate required to obtain, maintain and adjust a theoretical calculated blood concentration. Similar to the assumption made in early pharmacology research the effect was assumed to be directly, albeit not linearly, related to the concentration when using TCI. This assumption soon appeared to be incorrect. Early experimental TCI systems were developed to maintain a selected blood concentration in order to study the effects of the anaesthetics in a controlled situation.[2] It was during these studies that the delay between changing the blood concentration and the resulting change in clinical effect and its importance were noticed. Manual dosing can be seen as a time variant system, which means that the response to the same input varies with time. Consider the fact that a dose given at induction of anaesthesia has a lesser and shorter impact than at the end of an anaesthetic procedure due to differences in distribution and redistribution over time. Target-controlled infusion makes this, in theory, a time invariant system: with an ideal model the response to a change of target

Downloaded from https://www.cambridge.org/core. IP address: 193.56.73.28, on 02 Dec 2019 at 21:40:43, subject to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316659069.004

6

Chapter 2: Pharmacokinetics for TIVA

Figure 2.3 Drug levels in different compartments following a short infusion.

Still large drug reservoir Ten minute intravenous infusion

C3 C2

C1

K10 Elimination

not so important for cessation of drug effect - redistribution -

Figure 2.4 Drug levels in compartments following a prolonged infusion.

Peripheral compartment full Prolonged intravenous infusion

C2

C3

C1

K10 Elimination

more important for cessation of drug effect - less redistribution -

Figure 2.5 Context sensitive half-time for common opioids given by intravenous infusion. Relationship between keo (/min) and its half-life (/min).

Time to 50% Decrease in Blood Concentration (Minutes) 100

262.5 min after 240 Fentanyl

75 58.5 min after 240 Alfentanil 50 33.9 min after 240 Sufentanil 25

3.7 min after 240 Remifentanil 0 100 200 300 400 500 600 Minutes Since Beginning of Continuous Infusion

Downloaded from https://www.cambridge.org/core. IP address: 193.56.73.28, on 02 Dec 2019 at 21:40:43, subject to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316659069.004

7

Chapter 2: Pharmacokinetics for TIVA

Why is Keo Called Keo? The terminology is somewhat misleading. One of the first studies that used an effect compartment to explain the delay between blood and effect[4] had a set-up of three time-constants: for the transport from the central compartment to the effect compartment (an effect-transport compartment) ‘k1e’; one for the reverse ‘ke1’; and a ‘keo’ for the drug eliminated from the final effect compartment. In its concept the effect compartment has an infinitely small volume to prevent it having an influence on the PK model – it is only there to explain the delay in drug onset. To comply with this, ke1 has to be

infinitely small and k1e then becomes equal to keo (drug in equals drug out as no drug is assumed to go back). In the common PK/PD studies in anaesthesia the complex concentration-effect relation is simplified to assume that there is only one effect compartment for each effect and, hence, only one timeconstant is required: keo. The keo should be pronounced as ‘ke-oh’ where o stands for outside not ‘ke-zero’.

How is the Keo Measured? Initially keo was measured using hysteresis loops.[1] This is comparable to pressure volume loops on a ventilator where, at inspiration, first the pressure increases and, after a short delay, this is followed by an increase in the volume; while at expiration a decrease in pressure precedes a decrease in volume. With IV drugs these loops were constructed by giving the drug as a constant infusion over a short period while simultaneously measuring blood concentration and effect. When starting to run the infusion, blood concentration will increase but the effect lags behind. After stopping the infusion, the situation is reversed and the concentration decreases before the effect decreases. The concentration in the effect compartment is now modelled in such a way that this hysteresis curve is collapsed to an almost single line (Figure 2.6). This line represents the changing concentration in the effect compartment. This methodology has some practical shortcomings: it requires

100

Effective (100-BIS)

will always be the same irrespective of the time at which the target is changed. Already before the development of TCI, it had been noticed by some researchers that the change in the drug effect after a change in blood concentration was delayed.[3] Many reasons can be postulated for this delay but the most obvious notion was that the drug did not cause the effect in the blood but at the site of the receptor that was contained in another part of the body, which was called the bio-phase,[4] as opposed to the blood-phase. One could hypothesise that the concentration in one of the distribution compartments would be better correlated to the effect but even direct tissue sampling did not explain the effect delay in an experimental study on morphine in rats.[5] So the idea of relating the effect to one of the other PK compartments has now been abandoned. Instead a delay was built in between the blood concentration and the resulting effect. To stay in line with compartmental modelling, an effect compartment was constructed between the blood compartment and the resulting effect. This effect compartment belongs neither to the PK model nor to the PD model. It is considered to be a link model between the two. Although the delay in clinical onset for anaesthetic drugs is usually not more than a few minutes, it has great impact on the concept and usability of TCI and the understanding of IV anaesthesia (it is also important in inhalational anaesthesia, although not addressed in their current delivery systems). This time-delay constant has been given a name: keo. Every drug effect has its own keo and, between patients, there is usually a considerable variability in this blood-effect equilibration constant.

75

50

25

0 0

2

3

5

Figure 2.6 A hysteresis loop between the concentration in the blood or effect compartment versus the effect. Data from a recent developed PK/PD model are used.[25] The residual hysteresis in the effect versus effect-site concentration is caused by an age-related delay in the effect response.

Downloaded from https://www.cambridge.org/core. IP address: 193.56.73.28, on 02 Dec 2019 at 21:40:43, subject to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316659069.004

8

6

Blood / Effect site concentration (μg/ml)

Chapter 2: Pharmacokinetics for TIVA

a frequent and, preferably, continuous effect measurement (usually an EEG derivative). The concentration range in which the effect is studied is clinically relevant for hypnotics when using an EEG derivative, but for other drugs such as opioids the concentration at which a measurable effect on the brain becomes apparent[6,7] will lie outside the normal clinical dose range seen when the drug is used as an analgesic. It is not clear if the effect on the brain and the analgesic effect share the same keo. Furthermore, the measured concentrations may (and probably will) not fully coincide with the predicted concentrations from the PK model. The measured concentrations can be replaced by concentrations predicted by the model, in which case the derived keo is said to be parametric. When, on the other hand, measured concentrations are used, the keo is not relying on a specific PK model and it is then called non-parametric.[8] A parametric keo will be influenced by mis-specification of the (early phase of the) PK model.[9] Many other methodologies have been published to specify the blood-effect delay of anaesthetic drugs. A widely used approach to determine the keo is the time to peak effect (TTPE) method. It is based on the fact that, when a drug is delivered as a bolus, the peak effect will occur after the peak concentration in the blood.[10] As with the hysteresis-loop method, the measured values can be related to real concentrations (non-parametric) or predicted model concentrations (parametric). It is not difficult to understand that the PK model, which assumes immediate mixing of the drug in the central compartment, will not correctly predict the early-phase pharmacokinetics and, hence, will cause an even bigger mis-specification of the keo value when this TTPE method is used to specify a keo to be used with TCI. To summarise: the determination of the keo is dependent on many factors including: • how the effect is measured (EEG derivatives such as the bispectral index (BIS), auditory evoked potentials, canonical univariate analysis, spectral edge frequency etc.) • the methodology and input mode (hysteresis loop/ infusion, time to peak effect/bolus, multiple targets with TCI)[11] • the PK model (when predicted concentrations are used) This explains why there are so many different keos for the same drug and the chaotic situation that currently exists in commercial TCI devices.[12]

Target-Controlled Infusion and Keo Target-controlled infusion will allow the user to start, maintain and change a selected theoretical concentration in the blood of the patient. The TCI system will deal with distribution and redistribution of the drug but the real interest of the anaesthetist is the expected effect of the drug and the timing of the effect, which is also dependent on the keo when target drug concentrations are changed. This-time constant keo can be expressed as a half-life (ln(2)/keo) and, in the case of TCI, the evolution of the effect-site concentration will follow the simple half-life rules: after one half-life the concentration in the effect compartment will be 50% of the TCI target; at two half-lives, 75%; at three half-lives, 87.5%; and so on (Figure 2.7). This is not how the effect will evolve. The relation between effect-site concentration and effect is, for a large part, not linear and the patient may be unconscious after three keo half-lives, or not. This depends, obviously, on the level of the set target. If the appropriate target for an effect, say loss of consciousness (LOC), with propofol is correctly selected and the PK model does not deviate from real patient pharmacokinetics, it will still take some time before the effect is established. The keo will only indicate the speed at which the effect will be obtained. In order to speed up this process one can choose effect compartment control (TCIB) in which case the blood concentration will overshoot the set target in order to create an effect-site concentration as fast as possible. The smaller (slower) the keo the larger the dose that will be delivered and the larger will be the

Figure 2.7 Relationship between keo (/min) and its half-life (/min).

Downloaded from https://www.cambridge.org/core. IP address: 193.56.73.28, on 02 Dec 2019 at 21:40:43, subject to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316659069.004

9

Chapter 2: Pharmacokinetics for TIVA

overshoot in the blood compartment (Figure 2.8).[13] Unfortunately, a multitude of keos and PK models have found their way into commercially available TCI systems. As the keo is an important parameter in effect-site control (TCIE), the different systems will choose different induction doses for the same patient,[13] even if the PK models are comparable or the same. Also, after stopping a drug infusion the effect wear-off will be influenced by the keo. The

Blood conc. TCI Target

Effect conc. Infusion rate(mg/hr) × 100

Keo 2 times smaller

10 Keo: 0.146 / 2 Peak conc: 10.8 μg/ml Induction: 1.1 μg/kg

3

0

5 minutes

10 minutes

Original keo from model

10 Keo: 0.146 Peak conc: 16.2 μg/ml Induction: 1.8 mg/kg

3

0 5 minutes

10 minutes

10 Keo 2 times bigger

Keo: 0.146 × 2 Peak conc: 24.1 μg/ml

Patient Parameters Influencing Keo

Induction: 3 mg/kg

3

0 5 minutes

relationship between onset of effect during induction and offset after stopping infusion may only become apparent when effect-site concentrations at start of anaesthesia and end of anaesthesia are compared using appropriate models and keos.[14–18] All models are approximations of reality, and hence wrong, but some are more wrong than others! It is, however, obvious that in a clinical situation effectcompartment control with the correct keo has more to offer than blood control. The system response is faster with less overshoot and smaller step sizes in target adjustments. Consider a car where the rod that connects the steering wheel to the wheels of the car is not made of steel but of rubber. As long as you have to make slow bends you would not have a problem but when steering through faster bends is required, steering would become a real challenge. You would have to turn the wheel even before the bend starts and probably would then over-correct. If you know the road well you could train yourself to control the car but, unfortunately, every car is a bit different (PK variability) and the differences in the track of the course are even bigger (pharmacodynamic, or PD, variability). The tighter the connection between the steering wheel and the car wheels, the easier it will be to steer the car and, not unimportantly, your knowledge about the car’s behaviour, the track and how to travel would increase. You could even make a map so that others with the same equipment could have guidance. The benefits, such as better understanding of the anaesthetic drugs, better predictability and ease of training, require that all the TCI systems utilise the same model including the keo. TCI blood is controlling an invisible, virtual parameter that is irrelevant for the practitioner when there is no connection with the effect. The concept of keo is essential to tighten that connection.

10 minutes

Figure 2.8 Influence from keo on induction dose, peak effect concentration and duration of pump stopping in TCIE. In the middle panel original keo from a population PK/PD study is used.[25] In the top panel the keo is halved, whereas in the lower panel the keo is doubled.

The exact mechanism that is responsible for the delay between changing blood concentration and changing effect is multi-causal. Transport between the blood compartment and the receptor site is obviously one factor. Physical properties, such as lipophilicity and the ability to cross the blood–brain barrier (in the case of drugs working on the central nervous system), may be a factor in this transfer but explaining differences in the keo based on these properties is speculative. In many population PK/PD studies, age has been found

Downloaded from https://www.cambridge.org/core. IP address: 193.56.73.28, on 02 Dec 2019 at 21:40:43, subject to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316659069.004

10

Chapter 2: Pharmacokinetics for TIVA

to influence the keo. Younger patients tend to have a bigger or faster keo; the elderly, a smaller or slower keo. This may be the result of different body composition but as both the pharmacokinetics and pharmacodynamics also change with age it is difficult to isolate the age effect on the keo. The keo combines all factors that delay the blood-effect relationship. Gender was shown to have a large influence on the keo in a morphine study, whereas no gender influence was found on the keo in studies of alfentanil[19] and remifentanil.[20]

From Effect Compartment to Effect Finally, a relationship must be established between the drug concentration at the effect site and the actual effect. For most IV anaesthetics it is assumed that the effect is directly related to the number of receptors that are occupied. From this it follows that there is a spectrum from no effect, when no receptors are

occupied, to a maximum effect, when all receptors are occupied, and any further increase in concentration will not lead to an increase in the effect. A sigmoid Emax model is commonly used to describe this relationship between drug concentration at the effect site and the observed effect. This model is an empirical model with four parameters: E0 is the effect when no receptors (zero) are occupied (i.e. fully conscious, for example a BIS of 97); Emax is the effect when all receptors are occupied (maximum occupancy) (BIS of 0, burst suppression on EEG); EC50 is the effect-site concentration at 50% effect (for example a BIS of 50) and an exponential constant γ (gamma) that influences the steepness of the concentration–response curve. Receptor occupancy is not measured directly and one must realise that the drug effect is not a simple physical entity like temperature. Signal translation, processing, smoothing and interpretation will all influence the final number that is

Figure 2.9 Schematic representation of combined pharmacokinetic pharmacodynamic modelling: from dose to effect.

Downloaded from https://www.cambridge.org/core. IP address: 193.56.73.28, on 02 Dec 2019 at 21:40:43, subject to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316659069.004

11

Chapter 2: Pharmacokinetics for TIVA

visible on a monitoring device. Furthermore, the effects of IV drugs are usually multiple: some wanted, some not (side effects). Although these effects all share the same PK actuator (e.g. blood concentration), the PD model, including the keo, can be completely different. Propofol will not only render a patient unconscious but it will also cause vasodilation and, at higher concentrations, reduce the contractility of the heart.[21] These require different PD models and different keos. Anaesthesia is perceived to be complex and, by some, even considered to be more of an art than a science. However, more and more its complexity is being unravelled. A good example is a recent study on the pharmacokinetics[22] and dynamics[23,24] of dexmedetomidine: an alpha-2 agonist with a complex working mechanism and site effects. For each of the effects and site effects, bradycardia, hypo/hypertension, sedation and effect on the BIS, a PD model, including the different keos, was developed. When these data are implemented in simulation software (www.eurosiva.eu/ tivatrainerX/TTXinfo.html; accessed Jan 2019), they can help us to understand the multiple responses seen when such a drug is used in clinical practice. It explains that if the keo of the site is faster than the keo of the target effect, it is difficult to avoid the site effect (even with TCIE), a phenomenon well known to the practising anaesthetist.

References 1. C. Csajka, D. Verotta. Pharmacokinetic– pharmacodynamic modelling: history and perspectives. J Pharmacokinet Pharmacodyn 2006; 33: 227–79. 2. M.E. Ausems, J. Vuyk, C.C. Hug Jr., D.R. Stanski. Comparison of a computer-assisted infusion versus intermittent bolus administration of alfentanil as a supplement to nitrous oxide for lower abdominal surgery. Anesthesiology 1988; 68: 851–61. 3. L.B. Sheiner, D.R. Stanski, S. Vozeh, R.D. Miller, J. Ham. Simultaneous modeling of pharmacokinetics and pharmacodynamics: application to d-tubocurarine. Clin Pharmacol Ther 1979; 25: 358–71. 4. G. Segre. Kinetics of interaction between drugs and biological systems. Farmaco Sci 1968; 23: 907–18. 5. B.E. Dahlstrom, L.K. Paalzow, G. Segre, A.J. Agren. Relation between morphine pharmacokinetics and analgesia. J Pharmacokinet Biopharm 1978; 6: 41–53. 6. J.C. Scott, K.V. Ponganis, D.R. Stanski. EEG quantitation of narcotic effect: the comparative pharmacodynamics of fentanyl and alfentanil. Anesthesiology 1985; 62: 234–41.

7.

J.C. Scott, J.E. Cooke, D.R. Stanski. Electroencephalographic quantitation of opioid effect: comparative pharmacodynamics of fentanyl and sufentanil. Anesthesiology 1991; 74: 34–42.

8.

M. White, M.J. Schenkels, F.H. Engbers, et al. Effectsite modelling of propofol using auditory evoked potentials. Br J Anaesth 1999; 82: 333–9.

9.

K. Masui, M. Kira, T. Kazama, S. Hagihira, E. P. Mortier, M.M. Struys. Early phase pharmacokinetics but not pharmacodynamics are influenced by propofol infusion rate. Anesthesiology 2009; 111: 805–17.

10. C.F. Minto, T.W. Schnider, K.M. Gregg, T.K. Henthorn, S.L. Shafer. Using the time of maximum effect site concentration to combine pharmacokinetics and pharmacodynamics. Anesthesiology 2003; 99: 324–33. 11. Q. Wu, B. Sun, S. Wang, L. Zhao, F. Qi. Estimating the plasma effect-site equilibrium rate constant (Ke(0)) of propofol by fitting time of loss and recovery of consciousness. Biol Pharm Bull 2013; 36: 1420–7. 12. F.H.M. Engbers, A. Dahan. Anomalies in target-controlled infusion: an analysis after 20 years of clinical use. Anaesthesia 2018; 73: 619–30. 13. J.B. Glen, F.H. Engbers. The influence of target concentration, equilibration rate constant (ke0) and pharmacokinetic model on the initial propofol dose delivered in effect-site target-controlled infusion. Anaesthesia 2016; 71: 306–14. 14. A.G. Doufas, M. Bakhshandeh, A.R. Bjorksten, S.L. Shafer, D.I. Sessler. Induction speed is not a determinant of propofol pharmacodynamics. Anesthesiology 2004; 101: 1112–21. 15. T. Kazama, K. Ikeda, K. Morita, Y. Sanjo. Awakening propofol concentration with and without blood-effect site equilibration after short-term and long-term administration of propofol and fentanyl anesthesia. Anesthesiology 1998; 88: 928–34. 16. F. Engbers. Is unconsciousness simply the reverse of consciousness? Anaesthesia 2018; 73: 6–9. 17. J.H. Seo, E.K. Goo, I.A. Song, et al. Influence of a modified propofol equilibration rate constant (k(e0)) on the effect-site concentration at loss and recovery of consciousness with the Marsh model. Anaesthesia 2013; 68: 1232–8. 18. H. Iwakiri, N. Nishihara, O. Nagata, T. Matsukawa, M. Ozaki, D.I. Sessler. Individual effect-site concentrations of propofol are similar at loss of consciousness and at awakening. Anesth Analg 2005; 100: 107–10. 19. E. Olofsen, R. Romberg, H. Bijl, et al. Alfentanil and placebo analgesia: no sex differences detected in models of experimental pain. Anesthesiology 2005; 103: 130–9. 20. C.F. Minto, T.W. Schnider, S.L. Shafer. Pharmacokinetics and pharmacodynamics of

Downloaded from https://www.cambridge.org/core. IP address: 193.56.73.28, on 02 Dec 2019 at 21:40:43, subject to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316659069.004

12

Chapter 2: Pharmacokinetics for TIVA

remifentanil. II. Model application. Anesthesiology 1997; 86: 24–33. 21. C. Jeleazcov, M. Lavielle, J. Schuttler, H. Ihmsen. Pharmacodynamic response modelling of arterial blood pressure in adult volunteers during propofol anaesthesia. Br J Anaesth 2015; 115: 213–26. 22. L.N. Hannivoort, D.J. Eleveld, J.H. Proost, et al. Development of an optimized pharmacokinetic model of dexmedetomidine using target-controlled infusion in healthy volunteers. Anesthesiology 2015; 123: 357–67. 23. P.J. Colin, L.N. Hannivoort, D.J. Eleveld, et al. Dexmedetomidine pharmacokinetic-

pharmacodynamic modelling in healthy volunteers: 1. Influence of arousal on bispectral index and sedation. Br J Anaesth 2017; 119: 200–10. 24. P.J. Colin, L.N. Hannivoort, D.J. Eleveld, et al. Dexmedetomidine pharmacodynamics in healthy volunteers: 2. Haemodynamic profile. Br J Anaesth 2017; 119: 211–20. 25. D.J. Eleveld, P. Colin, A.R. Absalom, M. Struys. Pharmacokinetic-pharmacodynamic model for propofol for broad application in anaesthesia and sedation. Br J Anaesth 2018; 120: 942–59.

Downloaded from https://www.cambridge.org/core. IP address: 193.56.73.28, on 02 Dec 2019 at 21:40:43, subject to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316659069.004

13

Chapter

3

TCI and TIVA: What a Good Idea! Michael G. Irwin and Gavin N. C. Kenny

Know from whence you came. If you know whence you came, there are absolutely no limitations to where you can go. James Baldwin

Contemplations from an Enlightened Disciple (Gordon Wong) The introduction of TCI pumps into clinical practice invoked a sense of interest, and a sense of reservation stemming from the concern of getting the dose wrong and causing awareness. I experienced the frustration of not grasping all of the intricacies of the complicated pharmacokinetics. I experienced the annoyance of delayed emergence from getting the dose wrong from both TCI and manual infusions. There was only the rare TIVA enthusiast who could explain aspects of this dark art. Therefore, like a lot of my colleagues, I did not bother too much with using TIVA except for neurosurgery or for those with a history of severe PONV. It just felt like too much of a nuisance for not much gain. It was not until I began to work in a department where more people used TCI that I rekindled my interest in TIVA by which time there were more PK models and pumps to deal with. Then I heard a talk by ‘Jedi Master’ Gavin Kenny, who simplified important aspects of the pharmacokinetics and, more importantly, made one realise that one can practise TIVA with TCI safely without knowing every rate constant – and that boffins have taken care of a lot of this minutiae with their PK modelling. This enhanced my confidence in trying to use TCI and, with the availability of remifentanil, it turns out it is not hard to get it right. With a bit of practice and vigilance, even a mediocre anaesthetist such as me can look good when patients can move themselves off the operating table across to their bed after a few hours of surgery. As I contemplated why it took me so long to get to this point, I realised I was

influenced by well-meaning, but not necessarily bestinformed, colleagues who had not kept abreast of developments in TCI technology and thus developed misconceptions that hampered my progress. One thing about getting older is that you really start to learn the merits of knowing about history. As a young trainee, I used to dislike having to learn about the historical development of drugs and anaesthetic techniques. However, belonging now to an intermediate age group, I am old enough to see the evolution of TCI since it became commercially available and this first-hand experience has enabled me to see possible points of genesis of the misconceptions regarding this technique. I think an appreciation of the history of TCI development will help to clear up potential misunderstandings regarding its use.

Why Target-Controlled Infusion? The major problem with IV anaesthesia was that it was difficult to control accurately the concentration of anaesthetic agents, unlike inhalational drugs, and, therefore, the depth of anaesthesia. We were all trained to use inhalational anaesthesia and, consequently, this was a barrier to change. The aim of any dose regimen in anaesthesia is to titrate the administration of a drug to achieve the desired clinical effect for any individual patient while minimising the unwanted and potentially dangerous side effects. All anaesthetic drugs, by virtue of their PD effect, which is to induce coma, have a narrow therapeutic index. Too much will have potentially dangerous effects on the brain and other organs, while too little risks awareness. Titration is possibly more important therefore in anaesthesia than in any other branch of medicine. The need to have some understanding of pharmacokinetics was considered essential to use these IV agents successfully and the various manual infusion regimens or recipes, which aimed to guide new users

Downloaded from https://www.cambridge.org/core. IP address: 91.200.80.171, on 02 Dec 2019 at 21:42:02, subject to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316659069.005

14

Chapter 3: TCI and TIVA: What a Good Idea!

of this form of anaesthesia, were difficult to apply in practice. The main drawback of all manual administration schemes is the lack of knowledge of the predicted blood concentration, maintenance of a stable concentration and, therefore, of the level of anaesthesia at any point in time. To provide reliable anaesthesia, a drug delivery system has to allow more flexibility and accuracy than a manual system can provide. To increase the blood concentration quickly a bolus increment followed by a change in infusion rate is required but the calculations needed to select the correct bolus and new infusion rate are complex and certainly not practical during an anaesthetic. Reduction in dose has its hazards too, in that discontinuation of a manual infusion to reach a lower blood concentration could result in it inadvertently not being switched back on or being switched on at the wrong time. It was therefore necessary to develop an infusion system for TIVA that was as simple to use as a vaporiser for inhalational agents and that would enable the anaesthetist to transfer directly their skills learned with inhalational agents to the delivery of IV anaesthesia. This led to the development of TCI systems. These systems are pre-programmed with PK data for a large patient population. The anaesthetist just needs to input their patient’s physical characteristics (age, weight, height, sex) and multiple complex calculations can be carried out by a microprocessor to determine the dose required to achieve and maintain a given blood concentration. This makes the process of administering anaesthesia by TCI analogous to that of controlling inhalational anaesthesia with a vaporiser. A target level is selected by the anaesthetist on the infusion device for induction and is subsequently adjusted in response to clinical signs to maintain adequate anaesthetic depth. There is no need for a bolus induction dose as it is included in the machine’s calculations when the initial plasma concentration is selected. Depth of anaesthesia can be changed swiftly by simply selecting a new target blood concentration, similar conceptually to adjusting a vaporiser during inhalational anaesthesia. One of the first TCI systems was based on a standard desktop computer, which transmitted infusion rates to a computer-controlled infusion pump. This system proved that the technique could work well – but it was large and cumbersome. Replacing the desktop computer with a small Psion II hand-held computer enabled a device to be

produced with about the same size as a normal vaporiser.[1] More than one hundred of these systems were distributed to a variety of anaesthetists throughout Europe, with good clinical evaluations of the improved ability to titrate propofol to achieve the optimum level of anaesthesia for individual patients. Both the desktop computer and the Psion systems performed satisfactorily and were used for many clinical studies. However, they were based on nonmedical components.

Development of the Diprifusor™ The next development was to produce a device that was designed from the beginning as a medical-grade system to act as a flexible infusion platform.[2] The TCI system was produced with a double processor design to maximise safety and this complete control system was miniaturised into the ‘Diprifusor’ module. The microprocessors calculated the distribution and elimination of the drug about 25 times per second and altered the infusion rate appropriately. The Diprifusor module was incorporated into a range of syringe pumps that met the required infusion specifications. To prevent infusion of the incorrect drug, the pump operated in TCI mode only if the correct prefilled syringe was inserted into the TCI system. A magnetic recognition device informed the TCI system that either 1% or 2% Diprivan® (propofol was not a generic drug at that time, hence the name Diprifusor – infusor of Diprivan) was contained in the syringe and the software ensured that the appropriate rate was administered to achieve and maintain the selected target blood concentration. This Diprifusor has been licensed for use in most countries of the world, with the exception of the USA, and the product licence provides the only prescribing information for administration of propofol by TCI.

Accuracy of TCI: Predicted vs Measured Concentrations During the administration of TIVA the actual blood concentrations are not measured by the infusion system. The anaesthetist relies on the concentrations predicted and displayed on the infusion device to determine the concentration the patient is receiving. It is useful to know, then, how accurately these concentrations reflect the concentrations present actually in the plasma. Studies evaluating the predictive performance of the Diprifusor system have been

Downloaded from https://www.cambridge.org/core. IP address: 91.200.80.171, on 02 Dec 2019 at 21:42:02, subject to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316659069.005

15

Chapter 3: TCI and TIVA: What a Good Idea!

conducted. The measures used to quantify the predictive performance have been the median prediction error (MDPE) and the median absolute prediction error (MDAPE). The MDPE is a signed value and represents the direction of over- or under-prediction of concentration or the system bias. The MDAPE is a measure of the size of the error or system precision. A bias of 10–20% and a precision of 20–40% have been proposed as acceptable limits of error for a TCI system for clinical use.[3–5] Studies have shown that, for the Diprifusor system, the calculated concentration tends to be lower than the measured blood concentration and that the discrepancy is more marked on induction and on increasing the infusion rate, with the differences in concentration becoming minimal during maintenance. One study showed an MDPE of 16% and an MDAPE of 24% and another showed an MDAPE of 23%, both of which are within the acceptable limits outlined above. Useful comparisons can be made with the measurement of end-tidal inhalational anaesthetic agent concentration and the correlation of that measure with the actual blood concentration in terms of accuracy. One study comparing these two values for isoflurane found that end-tidal concentration was lower than arterial concentration by 20% and another study comparing inspired concentration to arterial isoflurane partial pressure showed an even greater bias.[6,7] Clearly, the predictive error of the Diprifusor system is no worse than the discrepancies observed with conventional systems used for inhalational agent monitoring.

Accuracy of TCI: Proportional Changes A major point is that, when using TCI to titrate the blood concentration, it is not the absolute values that are really important but the ability to make proportional changes. Anaesthetists make changes in drug concentrations by certain percentages in response to a patient’s clinical condition. The accuracy of these proportional changes is therefore more important that the raw predicted values. In a study by Glen and White, the average increase in target concentration of 40% requested, resulted in an increase of 53% in the measured value three minutes later.[8] Incomplete mixing in the central compartment probably contributed to this overshoot in the measured drug concentration. After decreasing the target

concentration by 46%, the measured value was on average 53% lower three minutes after reaching the new target.

Open TCI Systems Open TCI systems have been developed that allow IV agents to be administered by TCI. These systems allow non-tagged syringes (i.e. Diprifusor syringes are not a prerequisite) containing generic propofol to be used – and this has greatly decreased the cost associated with TIVA. However, they have introduced the question of the different PK models for propofol and this has led to considerable confusion as to which model provides the optimum benefit for the anaesthetist and their patient.

Blood vs Effect-Site Concentrations The site of action of general anaesthetics is not in the plasma but at a distant effect site in the brain. The increase in effect-site concentration lags behind the plasma concentration, which is the target blood concentration that the pump is programmed to achieve. This is because of the blood–brain barrier. To avoid the time delay due to effect-site equilibration, TCI systems have been modified to achieve a particular target concentration in the effect site itself instead of the plasma. Clinically, using these systems, the time to loss of consciousness is reduced by raising the plasma concentration beyond the intended effect-site concentration in order to raise the gradient between plasma and effect site and push the drug across the blood– brain barrier. EEG has been used in attempts to define the time course for the effect-site action of propofol and other drugs. However, the mathematical assumptions and calculations used to target the effect site introduce another potential source of error, which is difficult to quantify.

Effect-Site Control Open TCI systems offer the possibility of targeting the theoretical concentration in the effect site instead of the concentration in the blood (Figure 3.1). Effect-site control deliberately allows the blood concentration to overshoot the effect-site target concentration but calculates the exact moment when the blood concentration should be reduced to allow the effect concentration to be achieved with no, or only minimal, overshoot. There are relatively few studies that have used effect-site control but there are potential advantages in terms of the speed of response, i.e.

Downloaded from https://www.cambridge.org/core. IP address: 91.200.80.171, on 02 Dec 2019 at 21:42:02, subject to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316659069.005

16

Chapter 3: TCI and TIVA: What a Good Idea!

3

50 45 40 35

2

30 1.5

25 20

1

15

Marsh Schnider prediction BIS

0.5

10 5

0

0 0

2

4

6

8

10

12

14

100–BIS

Cep (µg.mI–1)

2.5

Figure 3.1 Propofol effect-site concentration vs Bispectral index in Marsh-driven TCI group (18 patients). The squares represent the propofol effect-site concentration as predicted by the Marsh model. The triangles represent the propofol effect-site concentration predicted by the Schnider model as calculated by the Tivatrainer when the same bolus and infusion rates are used. Predicted propofol concentration (Cep) is plotted on the primary y-axis. The circles represent the median Bispectral index on an inverted scale as calculated by 100–BIS plotted on a secondary y-axis.

16

Time (min)

a faster onset of effect. However, there are also potential disadvantages, such as higher peak blood concentrations, which may enhance adverse cardiovascular side effects, and further studies are required of this mode of TCI administration. In addition, it must be remembered that the PK model can only predict the various concentrations that will be achieved. If the patient’s pharmacokinetics and dynamics vary markedly from the model, then errors can arise.

Comparison of Pharmacokinetic Models In the Marsh model the only patient co-parameter that influences the model is the weight of the patient.[9] The implementation is undertaken in such a way that the size of the model is linearly scaled to weight. This reflects a common approach in pharmacokinetics to express the central volume and the clearance per kg without formal testing if weight is a co-parameter of the model with statistical significance. Since then, more elaborate methods to develop PK models and test the significance of co-variates have been developed. Despite these advances the clinical application of TCI has not benefited much from these developments. On the contrary, a confusing situation of different models and different implementation of these models has led to the situation where, at present, five different models for propofol are available for blood and effect TCI in commercially available systems. The differences between these models in terms of drug doses given are not trivial: in some situations, the difference between the dose of drug delivered can be as large as 100% in the same subject.

The implementation of the PK properties for TCI systems are based so far on two different models: the previously mentioned Marsh model and the Schnider model.[10] Both are open three-compartment models with elimination assumed from the central compartment. The central compartment is described by a volume (V1) and elimination rate constant k10. Two rate constants per added compartment are required, which describe drug transport in and out of the two additional compartments: k12, k21, k31 and k13. Therefore a total of six constants are required to specify a three-compartment model.

The Marsh Model The PK model implemented was published in 1991.[9] It is a refinement of a previously published dataset using a database of blood samples from 150 patients. The system was prospectively tested in TCI mode in a further 30 patients and many subsequent studies have confirmed its clinical value. In the Marsh model, V1 is expressed as ml per kg. The distribution and elimination are expressed in rate constants. Therefore changing the weight will linearly scale the model in total and hence linearly change the dose given to the patient. So a patient of 100 kg will at any time get twice the amount of drug compared with a patient of 50 kg whatever the shape or age of the patient. The model, as it is implemented in the Diprifusor, contains a blood–brain equilibration constant that was determined in a study of 15 male patients. Using a continuous infusion of propofol the auditory evoked potential was measured. The hysteresis between blood concentration and surrogate effect was modelled by assuming a virtual effect compartment. Also the

Downloaded from https://www.cambridge.org/core. IP address: 91.200.80.171, on 02 Dec 2019 at 21:42:02, subject to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316659069.005

17

Chapter 3: TCI and TIVA: What a Good Idea!

hysteresis between the calculated blood concentrations, using the Marsh model, and the effect was analysed. This resulted in a keo that connects an effect compartment specific to the Marsh model for male patients. This is important when using the Marsh model in female patients, who may not match the PK and PD values of the male study patients.

The Schnider Model The Schnider model is based on arterial blood samples from 24 volunteers. The study was designed to investigate possible non-linearity in the pharmacokinetics of propofol and the influence of population co-variates, bolus or infusion methods of administration, and EDTA supplementation on the pharmacokinetics of propofol.[10] In a separate publication, the PD properties were described.[11] The clinical effect of propofol was measured using a canonical univariate parameter (CUP), a mathematical analysis of the different frequencies in the EEG. Every subject received a bolus of propofol followed by an infusion for one hour. The PK model was developed according to the authors by using only the infusion data. The infusion model was superimposed on the calculated residual concentrations from the previous bolus. This ‘moderately complicated’ the analysis according to the authors. The model was not developed for TCI, nevertheless it performed well in studies that involved a comparable population.[12] However, the PD model parameters of time to peak effect (TTPE) and T1/2 keo were based on the bolus data. There is a combination of different problems with the Schnider model and its implementation in TCI systems. In particular, the doses administered are considerably

smaller than those administered by the Marsh model to achieve a given blood concentration. In addition, the proposed speed of drug movement from the blood into the brain is much faster in the Schnider model. The effect concentrations predicted by the original Schnider model were shown to have no relationship to the BIS or any other measure of propofol effect as illustrated in Figures 3.1 and 3.2.[13] In contrast, the effect concentrations predicted by the original Marsh model compared well with the measured values for BIS and the other surrogate markers of the propofol effect. A further issue of the Schnider model is that different TCI pump manufacturers have selected different methods of implementation of the model.[14] One implementation uses a fixed TTPE and hence a variable keo; whereas another implemented a fixed keo resulting in a variable TTPE. This leads to dramatically different induction doses in different patients.

White Model The White model was derived from a study in 113 patients who underwent surgery with propofol TCI, alfentanil and nitrous oxide.[8] In that study, the addition of co-variates reduced bias with the Marsh model from 12.6% to 3.2% and improved its predictive value, especially in older females. This model has not, so far, been incorporated into any commercial system. Five PK models are available in commercial TCI systems of which four were introduced without formal evaluation. Having five different implementations for one drug available in clinical practice is confusing and highly undesirable.[15,16] It should be

3

50 45 40 35

2

30 1.5

25 20

1

15 Marsh prediction Schnider BIS

0.5

10

100–BIS

Cep (µg.mI–1)

2.5

Figure 3.2 Propofol effect-site concentration vs Bispectral index in Schnider-driven TCI group (20 patients). The triangles represent the propofol effect-site concentration predicted by the Schnider model. The squares represent the propofol effect-site concentration as predicted by the Marsh model as calculated by the Tivatrainer when the same bolus and infusion rates are used. Predicted propofol concentration (Cep) is plotted on the primary y-axis. The circles represent the median Bispectral index on an inverted scale as calculated by 100–BIS plotted on a secondary y-axis.

5

0

0 0

2

4

6

8

10

12

14

16

Time (min) Downloaded from https://www.cambridge.org/core. IP address: 91.200.80.171, on 02 Dec 2019 at 21:42:02, subject to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316659069.005

18

Chapter 3: TCI and TIVA: What a Good Idea!

noted that the original Marsh model incorporated into the Diprifusor is the only PK model for which prescribing information is provided and is the only model licensed for use by drug regulatory authorities. Studies illustrating target concentrations for clinical use based on the original Marsh model cannot be transposed to another model and therefore we use the Marsh model for both blood and effect-site targeting. The display of predicted blood and effect-site propofol concentrations by TCI systems may provide users with a belief in greater accuracy than can be justified by the real data. It is important to recognise the possibility of major differences in pharmacokinetics and pharmacodynamics between different patients and the importance of titrating target concentrations to achieve a desired effect. The reliance on the accuracy of proportional changes to allow titration of the drug concentration to the clinical effects in individual patients is probably more important than the predicted or measured values.

Analgesic Supplements and TIVA A balanced anaesthetic technique includes the provision of analgesia. This may take the form of regional blockade, local anaesthetic infiltration of the painful area or co-administration of separate analgesic drugs. Propofol infusions can be supplemented with analgesics in a similar fashion to inhalational agents using pre- or intra-operative doses of non-steroidal antiinflammatory drugs (NSAIDs) or opioid drugs such as morphine or pethidine. NSAIDs and paracetamol have well documented opioid-sparing effects and can effectively extend analgesia into the post-operative period. Modern synthetic opioid agents can also be administered, either by intermittent injection or by infusion. Sufentanil, alfentanil and fentanyl are agents commonly given by infusion but care must be taken if infusing these drugs, particularly fentanyl, over a prolonged period as the time taken for concentrations to fall after infusion tends to increase markedly with prolonged duration of infusion. Remifentanil is metabolised very rapidly by nonspecific tissue esterases. This gives it a unique PK profile with a very high clearance and, even after very prolonged infusion, the time required for the concentration to decrease is short and consistent. It acts at the mu-receptor, in common with other opioid drugs, and it provides potent opioid effects including

analgesia. When used in combination with propofol, a lower maintenance dose of propofol can be used and while this is also true of other opioids none has such a consistently short duration of action as remifentanil. Remifentanil can provide all the advantages of high intra-operative opioid analgesia, such as cardiovascular stability, combined with a short duration of action, although this short duration of action will, of course, require alternative analgesia to be provided postoperatively.

Context Sensitive Half-Time A drug with a fast termination of effect, even after long infusion, is a desirable agent for TIVA. New measures have therefore been developed in an attempt to quantify drug offset. The context sensitive half-time (CSHT) is one such measure described by Hughes and colleagues.[17] How long it will take plasma concentrations to decrease after a drug has been administered is dependent on the PK profile of the drug and on the duration of the infusion that has taken place. Simply looking at the elimination half-life of drugs described by multi-compartment PK models is not sufficient in itself to give an indication of how long the drug concentration will take to decrease after infusion. This is because, as discussed earlier, reduction in plasma concentration depends on the drug distribution profile as well as, and possibly more than, the elimination half-life. The relative importance that the elimination and distribution processes play in drug offset are in turn affected by the duration of infusion that has taken place. The CSHT takes into account all of these factors. It is defined as the time it takes for plasma concentration to reduce by 50% after discontinuing an infusion of a specified duration: in other words, the 50% plasma reduction time in the ‘context’ of a particular duration of infusion. It is calculated by simulating drug infusions of varying duration using PK model-driven computer programs thus allowing comparisons of CSHT to be made between different drugs using the simulations. The longer a drug infusion is administered, the more the drug has moved into the peripheral compartments. This means that when the drug infusion is stopped, the reduction in concentration observed will be less dependent on redistribution to peripheral compartments and more dependent on metabolism and elimination. When a drug reaches a steady state

Downloaded from https://www.cambridge.org/core. IP address: 91.200.80.171, on 02 Dec 2019 at 21:42:02, subject to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316659069.005

19

Chapter 3: TCI and TIVA: What a Good Idea!

and all compartments have been filled by the drug, the CSHT then equals the elimination half-life. This occurs relatively quickly: for the drug remifentanil after approximately 15–20 minutes; and for alfentanil after around three hours.[18] At this point, the CSHT of these drugs can be said to have become context insensitive. For drugs such as fentanyl, the CSHT increases very rapidly after relatively short periods of infusion. In all cases, the CSHT for these drugs will also eventually reach a plateau level. Interestingly, when comparing CSHT, sufentanil has a CSHT shorter than alfentanil for infusions below three hours’ duration and fentanyl’s CSHT is shorter than those of both alfentanil and sufentanil up to an infusion duration of one hour. This reflects how, during short infusions, redistribution can account for an apparently shorter duration of drug action than might be anticipated if one focuses on drug elimination half-life values alone. The very short elimination half-life of remifentanil (due to esterase metabolism) combined with its small volume of distribution gives it its characteristic short and consistent off-set profile. Although a short CSHT might seem to infer a quick recovery following anaesthesia, in practice the CSHT is actually quite a poor predictor of recovery. The reason for this is that the time a patient takes to recover from a drug does not necessarily correlate simply with a decrease in plasma concentration by 50%. The plasma concentration at this point may not be one where recovery would be expected or alternatively plasma concentrations associated with recovery may be reached before decreasing to the 50% value. Drugs with a long CSHT can therefore be administered to achieve shorter recovery times than the CSHT may seem to suggest. Fentanyl administered intra-operatively at a concentration just above the concentration required for surgical analgesia and requiring a reduction of only 20% at the end of surgery might appear to have a shorter duration of action than alfentanil, given at a concentration that requires a 50% reduction for recovery. A further limitation of the CSHT in predicting recovery is that it does not take effect-site equilibration time into consideration. The intra-operative concentration must decrease by a certain percentage at the end of a procedure to levels where recovery would be expected. If a higher percentage decrease is required, this will take a longer

time. Importantly, the fall in concentration over time is not linear. In other words, there may be a steep and relatively fast initial decrement time for the first 10–20% decrease in drug concentration; whereas the time to reduce from 40–50% may be several times longer. In addition, the decrement time is very variable between different drugs, with the concentration of some decreasing much more quickly than others. To reduce the impact on recovery time and achieve optimum administration when using a drug with a long overall decrement time, it is best to minimise the actual percentage decrement required at the end of surgery. In this way the time required for the drug to fall to levels associated with recovery is shortened as far as possible. When used in the optimal way, the time to recovery is more dependent on the opioid selected than on the duration of infusion. Increased duration of infusion will increase the time to recovery but the selection of fentanyl compared with alfentanil will have a more significant impact on delaying recovery. In clinical practice, it is easy to use opioids in a less than optimal way and overdose with them. They are relatively safe and dosing to two times greater than the actual requirement during a procedure is not likely to cause any major adverse effects in a ventilated patient. However, the expected recovery time is likely to be more than twice as prolonged, which can result in apparently short-acting drugs having a long recovery time. The combination of propofol and remifentanil is a different situation. The decrement time for remifentanil is faster than propofol, which means that it is best to reduce the percentage decrement required for propofol at the end of anaesthesia by giving a low propofol–high remifentanil combination. The fact that the decrement time of remifentanil does not increase with infusion duration means that the recovery time is minimally affected by infusion duration. The recovery time for propofol combined with remifentanil is also faster than for the other opioids. Remifentanil, in contrast with fentanyl and alfentanil, is perhaps a more ‘forgiving’ drug when given suboptimally. It can be given in concentrations in excess of those required and, because of its extremely fast decay, the recovery time will only be marginally prolonged. Understanding these interactions can allow tailoring of drug combinations during TIVA to provide the most favourable recovery profile for different

Downloaded from https://www.cambridge.org/core. IP address: 91.200.80.171, on 02 Dec 2019 at 21:42:02, subject to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316659069.005

20

Chapter 3: TCI and TIVA: What a Good Idea!

durations of surgery and different combinations of drug. Once again, it is important to realise that absolute plasma concentrations cannot be applied to all individuals who will vary in their requirements but these principles can be used to optimise a TIVA regimen.

5. P.S. Glass, J.B. Glen, G.N. Kenny, J. Schuttler, S.L. Shafer. Nomenclature for computer-assisted infusion devices. Anesthesiology 1997; 86: 1430–1.

Conclusions

7. M.J. Landon, A.M. Matson, B.D. Royston, A.M. Hewlett, D.C. White, J.F. Nunn. Components of the inspiratory-arterial isoflurane partial pressure difference. Br J Anaesth 1993; 70: 605–11.

General anaesthesia is a dynamic balance between the level of hypnotic and analgesic and the effects of stimulation from surgery or instrumentation. The need to titrate is illustrated by the inter-patient variations in requirements for propofol and indeed most anaesthetic drugs. The ability to titrate is therefore an essential feature of any anaesthetic technique; and the simpler the technique makes for titration the easier it will be to use in the busy environment of an operating theatre. TCI systems enable the anaesthetist to make alterations in the target concentration of propofol in a simple and intuitive manner. The aim is to rapidly reach a steady concentration of propofol against which the anaesthetist can assess the adequacy of anaesthesia. There are only three choices available: to increase, to maintain the present level or to decrease the target concentration. The accurate proportional changes made with any TCI system are more important than the predicted values since this is how anaesthetists titrate the dose to the individual patient’s requirements during the reality of surgical anaesthesia.

References 1. G.N.C. Kenny, M. White. A portable computerised infusion system for propofol. Anaesthesia 1990; 45: 692–3. 2. J.M. Gray, G.N.C. Kenny. Development of the technology for ‘Diprifusor’ TCI systems. Anaesthesia 1998; 53: 22–7. 3. J. Schüttler, S. Kloos, H. Schwilden, H. Stoeckel. Total intravenous anaesthesia with propofol and alfentanil by computer-assisted infusion. Anaesthesia 1988; 43: 2–7. 4. P.S.A. Glass, J.R. Jacobs, L. Richard Smith, et al. Pharmacokinetic model-driven infusion of fentanyl: assessment of accuracy. Anesthesiology 1990; 73: 1082–90.

6. F.J. Frei, A.M. Zbinden, D.A. Thomson, H.U. Rieder. Is the end-tidal partial pressure of isoflurane a good predictor of its arterial partial pressure? Br J Anaesth 1991; 66: 331–9.

8. J.B. Glen, M. White. A comparison of the predictive performance of three pharmacokinetic models for propofol using measured values obtained during target-controlled infusion. Anaesthesia 2014; 69: 550–7. 9. B. Marsh, M. White, N. Morton, G.N. Kenny. Pharmacokinetic model driven infusion of propofol in children. Br J Anaesth 1991; 67: 41–8. 10. T.W. Schnider, C.F. Minto, P.L. Gambus, et al. The influence of method of administration and covariates on the pharmacokinetics of propofol in adult volunteers. Anesthesiology 1998; 88: 1170–82. 11. T.W. Schnider, C.F. Minto, S.L. Shafer, et al. The influence of age on propofol pharmacodynamics. Anesthesiology 1999; 90: 1502–16. 12. A.R. Absalom, V. Mani, T. De Smet, M.M. Struys. Pharmacokinetic models for propofol: defining and illuminating the devil in the detail. Br J Anaesth 2009; 103: 26–37. 13. A.R. Barakat, N. Sutcliffe, M. Schwab. Effect site concentration during propofol TCI sedation: a comparison of sedation score with two pharmacokinetic models. Anaesthesia 2007; 62: 661–6. 14. F.H. Engbers, N. Sutcliffe, G. Kenny, S. Schraag. Pharmacokinetic models for propofol: defining and illuminating the devil in the detail. Br J Anaesth 2010; 104: 261–2; author reply 62–4. 15. F. Engbers. Is unconsciousness simply the reverse of consciousness? Anaesthesia 2018; 73: 6–9. 16. F.H.M. Engbers, A. Dahan. Anomalies in target-controlled infusion: an analysis after 20 years of clinical use. Anaesthesia 2018; 73: 619–30. 17. M.A. Hughes, P.S. Glass, J.R. Jacobs. Context-sensitive half-time in multicompartment pharmacokinetic models for intravenous anesthetic drugs. Anesthesiology 1992; 76: 334–41. 18. A. Kapila, P.S. Glass, J.R. Jacobs, et al. Measured context-sensitive half-times of remifentanil and alfentanil. Anesthesiology 1995; 83: 968–75.

Downloaded from https://www.cambridge.org/core. IP address: 91.200.80.171, on 02 Dec 2019 at 21:42:02, subject to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316659069.005

21

Chapter

4

Milk of Amnesia Propofol for TIVA Kane O. Pryor and Paul S. Myles

History of Development Currently, propofol, while not perfect, has the most suitable PD and PK characteristics suitable for TIVA. Even for the unenlightened the drug is well established as an induction agent, so it is very likely that you will already have some experience of its use – but this chapter will further elucidate this fascinating drug. Propofol was initially synthesised in the mid-1970s as compound ICI 35868, the product of a programme by Imperial Chemical Industries to investigate the sedative properties of phenol derivatives. Although its potential as an anaesthetic was immediately recognised, its development was substantially prolonged by challenges in establishing a safe and effective formulation.[1] Propofol naturally exists as a pale yellow oil, which, unlike other lipophilic anaesthetic drugs, possesses no readily ionisable site enabling conversion to an aqueous salt. Consequently, it must be delivered within an organic solvent. The first formulation assessed in humans contained 2% propofol, 16% Cremophor EL, and 8% ethanol, but caused substantial pain on injection. A modified formulation of 1% propofol with 16% Cremophor EL and no ethanol was more acceptable, and was trialled in over 1000 European patients between 1977 and 1981. However, the emergence of anaphylactoid reactions and other adverse events, all attributed to the Cremophor EL vehicle, led to further development of this formulation being shelved.[2] The ultimate development of propofol to market was made possible by advances in emulsion science. The formulation that entered clinical trials contained the same components as Intralipid® (Fresenius Kabi AG, Bad Homberg, Germany), a fat emulsion used as a parenteral source of calories and essential fatty acids. Developed under the trade name Diprivan® (Fresenius Kabi USA, Lake Zurich, IL), it contained 1% propofol in 10% soybean oil, with 1.2% egg yolk lecithin as an emulsifier, 2.25% glycerol to maintain

isotonicity, and sodium hydroxide to maintain a pH of 7.0–8.5. Trials demonstrated the new formulation to have similar anaesthetic properties to the earlier Cremophor EL preparation, but without evidence of anaphylactoid reactions.[3] The Diprivan® preparation was approved in Europe in 1986, and in the United States in 1989, bringing into clinical practice one of the most welcomed and important pharmacologic developments in the recent history of anaesthesia.

Basic Pharmacology Structure–Activity Propofol is an alkylphenol. The core structure is a phenol, defined by the attachment of a hydroxyl (OH) group to an aromatic ring. Either side of the hydroxyl group is an isopropyl (—CH(CH3)2) group, which is a three-carbon chain attached to the ring at the middle carbon. As the hydroxyl group defines position 1 on the ring, the chemical name for propofol is 2,6-diisopropylphenol (see Figure 4.1). The activity of propofol at its key binding site, the GABAA receptor, appears to be critically related to the size and shape of the two alkyl groups, relative to the hydroxyl group that sits between them.[4]

Mechanism of Action The key target for propofol is the GABAA receptor, a ligand-gated ion channel that is the most abundant inhibitory post-synaptic receptor throughout the central nervous system. When the neurotransmitter GABA binds at the GABAA receptor, there is

Figure 4.1 The molecular structure of propofol.

Downloaded from https://www.cambridge.org/core. IP address: 95.181.217.50, on 02 Dec 2019 at 21:42:28, subject to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316659069.006

22

Chapter 4: Milk of Amnesia: Propofol for TIVA

a selective increase in chloride current through its pore, leading to hyperpolarisation of the cell. This hyperpolarisation decreases the ability of excitatory signals to trigger an action potential. The principal action of propofol is to increase the sensitivity of the receptor to endogenous GABA; in this sense, it is not acting as a direct GABA agonist as frequently described, but instead as a positive allosteric modulator. However, at higher concentrations, propofol does have a direct agonist effect, causing receptor activation in the absence of GABA.[5] There is substantial heterogeneity in the structure of the GABAA receptor which in part, explains why the actions of propofol differ from those of other GABAergic drugs such as the benzodiazepines. Every GABAA receptor is made up of five transmembrane subunits. Nineteen distinct subunits are known to exist (α1–6, β1–3, γ1–3, δ, ε, θ, π, ρ1–3),[6] and certain combinations of these subunits are aggregated in specific functional regions of the brain. Although incompletely understood, the GABAergic actions of propofol appear to dominantly involve the β2 and β3 subunits, with α and γ subunits involved to a lesser degree (in contrast to the dominant α/γ action of the benzodiazepines).[7] The effects of propofol therefore reflect the distribution and importance of the subunit combinations in different regions and systems of the brain. Propofol also reduces excitatory activity in the CNS through inhibitory effects at the N-methylD-aspartate (NMDA)-type glutamate receptor, which occur as a result of changes in Na+-channel gating.[8] The NMDA receptor plays a critical role in synaptic plasticity, including in long-term potentiation models of memory, but it is not established what role these actions play in the amnestic effects of propofol. Another potentially important target for analgesia, more recently described, is the hyperpolarisation-activated cation (HCN) channel, which has a critical role in pacing synchronised activity in networks connecting the thalamus and cortex.[9] Propofol also acts to potentiate the inhibitory action of glycine receptors, but the functional effects of this activity remain unclear.

Metabolism Propofol undergoes oxidative metabolism in the liver to form 1,4-diisopropyl quinol, which is

inactive. This is then conjugated to form a number of further inactive metabolites, which are excreted through the kidneys. The clearance of propofol (30 ml.min−1.kg−1) is substantially higher than that of other IV anaesthetics and exceeds hepatic blood flow, implying that extrahepatic metabolism occurs; indeed, propofol metabolism can be demonstrated during the anhepatic phase of liver transplantation. The major sites of extrahepatic metabolism are the kidneys and lungs[10] and, combined, may account for nearly half of total metabolism. Propofol inhibits the cytochrome P450 3A4 enzyme,[11] which is involved in the oxidation of a large diversity of drugs, including many of the opioids and benzodiazepines. Propofol may therefore reduce the metabolism of these drugs.

Clinical Pharmacology Pharmacokinetics Propofol is highly lipophilic and rapidly transitions across the blood–brain barrier. With an induction dose, unconsciousness is observed in 10–50 seconds at a plasma concentration of 2.0–2.5 µg.ml−1, but time to peak effect on EEG is about 90–120 seconds. Higher doses are required to prevent movement in response to pain and therefore it is more appropriate to use an adjuvant opioid analgesic to attenuate pain.[12] Propofol is very highly protein bound (98%) and therefore the unbound fraction, and thus the response, may be increased in patients with low serum protein levels. After a single bolus dose, plasma levels decline rapidly because of redistribution from the site of action to peripheral tissue compartments, with an initial distribution half-life of two to eight minutes.[13] Typical induction doses for different patient groups are shown in Table 4.1. The volume of distribution (VD), which describes the theoretical volume into which a drug has been dissolved to obtain the observed plasma concentration, is 2–10 l.kg−1, demonstrating that a high proportion of the drug is displaced into tissues outside the plasma due to its high lipophilicity. Because the concentration of propofol at the site of action is dominated by distribution phenomena, accumulations in peripheral tissue occurring during prolonged infusions dynamically shift the way the

Downloaded from https://www.cambridge.org/core. IP address: 95.181.217.50, on 02 Dec 2019 at 21:42:28, subject to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316659069.006

23

Chapter 4: Milk of Amnesia: Propofol for TIVA

Table 4.1 Induction dosing of propofol Adults ASA 1–2 ASA 3–4

2.0–2.5 mg.kg−1 1.0–1.5 mg.kg−1

Paediatrics 2 months–16 years, healthy

2.5–3.5 mg.kg−1

Geriatrics >60 years

1.0–1.5 mg.kg−1

Obesity Body mass index >35 kg.m−2

Use ideal body mass

Hepatic impairment

No dosage adjustment

Renal impairment

No dosage adjustment

drug moves between the various compartments of the body. This results in a CSHT, such that the effective half-time increases with the duration of infusion. After an eight-hour infusion, the context sensitive half-time of propofol is ~40 minutes.[14] The elimination half-time is four to seven hours, although after infusions of up to ten days, it may exceed 24 hours. In most clinical scenarios, the elimination kinetics only become dominant once the drug concentration has declined to largely subclinical levels, thus enabling rapid emergence from propofol even after extended infusions. The pharmacokinetics of propofol are altered at the extremes of age. This occurs largely because of changes in the ratio of the central (high perfusion) compartment relative to the peripheral (low perfusion) compartment. In the elderly, the size of the central compartment is reduced due to decreased cardiac output,[15] leading to substantially decreased dosage requirements to attain a target plasma concentration relative to younger adults. Conversely, young children have a larger central compartment, and clearance (50 ml.min−1.kg−1) >50% higher than in adults.[16] Dosage requirements remain weight based, but are significantly increased, with the ED95 approaching 3 mg.kg−1 in patients under the age of two years. Hepatic disease does not significantly alter the volume of distribution, clearance or half-life of propofol, and no adjustment in dosing is required.[17] Similarly, propofol pharmacokinetics are not altered in patients with end-stage renal disease.[18] Obesity significantly alters the pharmacokinetics of propofol; in patients with a body mass index (BMI) >35 kg.m−2, the use of their ‘ideal’ body weight preserves the effects equal to those seen in normal-weight controls.[19]

Pharmacodynamics Central Nervous System Induction doses of propofol rapidly cause unconsciousness. The transition to unconsciousness does not require that the cortex is inactive but, rather, that there is fragmentation or functional disconnection of cortical areas, leading to loss of information integration. Investigations using functional magnetic resonance imaging and electroencephalography have variously implicated the importance of thalamocortical relays and coherence[20] and fronto-parietal connectivity.[21] Propofol alters connectivity in large-scale resting networks, including the default mode network and external control network.[22] Propofol is a potent amnestic agent with effects on memory that do not rely on its sedative and hypnotic actions.[23] It prevents consolidation of new memory, similar to the benzodiazepines, but does not cause retrograde amnesia. These actions likely involve suppression of activity in the hippocampus.[24] Propofol reduces the regional cerebral metabolic oxygen rate (rCMRO2). However, because the coupling relationship between rCMRO2 and regional cerebral blood flow (rCBF) is preserved (in contrast to inhalational anaesthetics), decreases in rCMRO2 lead to reduced rCBF and regional cerebral blood volume (rCBV).[25] This will lead to decreased brain swelling and intracranial pressure acutely,[26] although no long-term outcome benefits have been demonstrated. Indeed, because vascular coupling is preserved, a judicious approach to the use of hyperventilation to cause additional reduction in rCBF is recommended. Relative to inhalational anaesthetics, at comparable depths of anaesthesia, propofol causes significantly less change in somatosensory evoked potential latency and amplitude[27] and is the optimal anaesthetic agent to use when it is necessary to measure motor evoked potentials.[28] Propofol has anti-convulsant activity and can be used in the treatment of status epilepticus but also has complex pro-convulsant action and is known to induce seizure.[29] As with other GABAergic anaesthetics, propofol may facilitate neuroprotection against primary ischaemic injury by reducing the rate of ATP consumption of neural tissue. However, despite studies in animal models showing reduced infarct volume, decreased release of pro-apoptotic factors and increased release of anti-apoptotic factors,[30] no clear outcome benefit has been demonstrated in

Downloaded from https://www.cambridge.org/core. IP address: 95.181.217.50, on 02 Dec 2019 at 21:42:28, subject to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316659069.006

24

Chapter 4: Milk of Amnesia: Propofol for TIVA

humans.[31] Several studies evaluating propofol as a neuroprotective agent in animal models of postischaemic secondary injury have shown promise.

Cardiovascular System Propofol has pronounced, dose-dependent effects on cardiovascular haemodynamics that are maximal around four to five minutes following loss of consciousness during the induction sequence. Vasodilation is mediated by direct changes in vascular smooth muscle calcium transport and nitric oxide production, as well as an upstream reduction in sympathetic vascular tone. However, while loss of systemic vascular resistance may contribute to the decreased systemic pressures seen with propofol,[32] a larger contribution may come from reduced left cardiac work index and stroke index. This may result from acute venodilation leading to decreased preload,[33] a direct myocardial depressant effect of propofol and an indirect depressant effect due to loss of tonic sympathetic output to the myocardium. Propofol attenuates the baroreflex response[34] and may be directly vagotonic. Therefore the loss of cardiac output due to decreased stroke volume is either augmented or inadequately compensated by changes in heart rate. The haemodynamic effects of propofol are, obviously, more pronounced in elderly patients or those with cardiovascular co-morbidity,[35] in part because of their increased reliance on sympatheticmediated venoconstriction and augmentation of myocardial contractility. The haemodynamic effects may be offset by avoiding hypovolaemia or by the co-administration of a vasoconstrictor such as phenylephrine. Higher doses of propofol will attenuate the response to α-adrenergic agonists. Propofol may possess cardioprotective properties, probably related to anti-oxidant effects on reactive oxygen species. High-dose infusions of propofol during cardiopulmonary bypass have been demonstrated to reduce levels of troponin T and I.[36] Recently, the PRO-TECT II trial reported reduced episodes of inhospital low cardiac output and heart failure in diabetic patients who received propofol infusions during cardiopulmonary bypass.[37] These benefits differ from the pre-conditioning seen with inhalational anaesthetics and opioids and may be most beneficial in patients with ischaemic heart disease since they are already pre-conditioned and will benefit from a different mode of cardioprotection.[38,39]

Respiratory System Propofol depresses the hypoxic ventilatory response[40] due to an effect on the chemosensitivity of the carotid body. It also depresses the ventilatory response to hypercapnia, which is due to action at central, but not peripheral, chemoreceptors.[41] Large induction doses (2.5 mg.kg−1) of propofol can cause apnoea that may last for more than three minutes;[42] the effect is exaggerated in the elderly and with the co-administration of opioids. This is not seen when induction is titrated with TCI and is one of the reasons that we recommend that technique, although infusions of propofol will also cause a decrease in tidal volume, with a less than compensatory increase in respiratory rate; the net effect is a decrease in minute ventilation. PaCO2 increases are attenuated because of metabolic depression.

Anti-emetic Action Propofol is an effective powerful anti-emetic at effectsite concentrations causing only minimal sedation. The use of propofol for induction and maintenance by infusion reduces the baseline risk of PONV compared to inhalational anaesthetics.[43] Small, minimally sedating bolus doses of propofol (10–20 mg) can be used as rescue therapy in the post-operative period in appropriately monitored settings.[44]

Adverse Effects and Clinical Issues Arising From Propofol Formulations Support of Microbial Growth Within three years of propofol’s release in the United States, the Centers for Disease Control and Prevention conducted epidemiologic investigations at seven hospitals that had reported unusual outbreaks of infection and febrile episodes in postsurgical patients, including a number that resulted in death. The investigators determined that the outbreaks were attributable to contamination of propofol occurring during handling and released a highly publicised report in the New England Journal of Medicine in 1995.[45] Controlled investigations subsequently demonstrated that the propofol emulsion was capable of supporting rapid growth of a number of pathogenic bacterial organisms, including Escherichia Coli, Staphylococcus aureus and Candida albicans.[46] In time, 0.005% disodium ethylenediaminetetraacetate (EDTA) emerged as the optimal anti-microbial additive

Downloaded from https://www.cambridge.org/core. IP address: 95.181.217.50, on 02 Dec 2019 at 21:42:28, subject to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316659069.006

25

Chapter 4: Milk of Amnesia: Propofol for TIVA

and it became a component of the Diprivan® formulation in 2000. EDTA exerts its anti-microbial action by chelating the essential trace metals required for the integrity of the bacterial cell wall. As a chelating agent, it is also able to sequester ionised calcium, magnesium and other metals in plasma, although any effect in patients receiving prolonged propofol infusions is likely of minimal clinical significance.[47] Some generic formulations of propofol do not contain EDTA and, instead, substitute 0.025% sodium metabisulphite as an anti-microbial agent. Sodium metabisulphite is a commonly used food preservative and industrial chemical, which exerts its anti-microbial action through the effects of released sulphur dioxide on bacterial cellular function. Other generic formulations of propofol substitute 0.1% benzyl alcohol. The FDA presently mandates that all propofol formulations in the United States contain an anti-microbial additive. However, even with the addition of EDTA or a substitute, propofol formulations can support some growth of microorganisms and do not meet United States Pharmacopeia (USP) standards for an anti-microbially preserved product. Therefore, in addition to the use of strict aseptic technique, the FDA advises that all propofol preparations are for single patient use, that administration must commence immediately after opening and that administration must be completed within six hours of opening. In the case of ICU sedation, administration must be completed within 12 hours of opening.[48] Some jurisdictions outside of the United States permit formulations without an anti-microbial additive but we do not recommend their use.

immunoglobulin (IgE) to egg, soy or peanut and found no evidence of propofol allergy.[49] Another evaluated 52 patients with eosinophilic oesophagitis and known sensitivity to egg, soy or peanut and, similarly, found no evidence of reaction to propofol.[50] One study of 28 children with egg allergy identified a single, nonanaphylactic reaction to propofol.[51] There is a sound rationale for why cross-reactivity may be absent. The allergenic protein underlying most soy allergies is removed in the refining process of soybean oil. Further, almost all egg allergies involve proteins found in the egg white, and not those found in egg lecithin, which is contained within the yolk. Nonetheless, as most regulatory bodies still list food allergies as a contra-indication to propofol, albeit inconsistently, caution is still advised.

Cross-Reactivity with Food Allergies

Hypertriglyceridaemia and Pancreatitis

As propofol emulsions contain egg lecithin and soybean oil, there is a basis for concern that administration could cause anaphylactic reactions in patients with allergies to egg, soy or peanuts; the latter are included because peanuts are phylogenetically and antigenically similar to soy. Empiric evidence to support these concerns is scant and recommendations differ between countries. There have been only six case reports linking reactions to propofol with food allergies, with only a single report providing confirmation with a positive skin test. More systematic investigations suggest that cross-reactivity is either absent or exceedingly rare. One recent study evaluated 99 patients with specific

Pain on Injection Mild to moderate pain on injection of propofol is a common side effect and is thought to be mostly due to the small quantity of propofol left in the aqueous phase of the emulsion, rather than to the other emulsion components. Pain has been markedly reduced by the introduction of mid-chain triglyceride vehicle formulations, which are now very popular and do alter PK/PD. With these, pain is seldom an issue even during sedation but, if required, the most effective intervention to prevent pain is the coadministration of lidocaine. In clinical practice, lidocaine is given either as a pretreatment (0.2 mg.kg−1), or mixed with the propofol solution (20 mg of lidocaine with 200 mg of propofol). However, because lidocaine does degrade the stability of the propofol emulsion over time, pre-treatment is theoretically the better approach.

The recommended maximum daily dose of parenteral lipid formulations (Intralipid® or equivalent) is 2.5 g.kg−1.day−1. To deliver an equivalent lipid dose via Diprivan® propofol formulation would require a continuous infusion of 175 μg.kg−1.min−1 for 24 hours, which greatly exceeds most practical infusion strategies. Nonetheless, several studies have demonstrated prolonged propofol infusions of several days to be associated with hypertriglyceridaemia in almost 20% of ICU patients,[52] although the clinical significance of these elevations is unclear.[53] Older patients, and critically ill patients with impaired hepatic lipid metabolism, may be especially

Downloaded from https://www.cambridge.org/core. IP address: 95.181.217.50, on 02 Dec 2019 at 21:42:28, subject to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316659069.006

26

Chapter 4: Milk of Amnesia: Propofol for TIVA

susceptible. Although there are very rare reports of clinical hypertriglyceridaemia occurring with the use of propofol infusions during surgery, cohort studies suggest that propofol infusions confined to surgery do not cause significant elevation of triglycerides.[54] A further concern that may in part be related to hypertriglyceridaemia is propofol-induced pancreatitis. A possible association between propofol and acute pancreatitis was first described in 1996[55] and has subsequently been reported in a number of cases that are notable for their heterogeneity. Hypertriglyceridaemia is a known cause of pancreatitis and propofol-induced hypertriglyceridaemia is postulated to explain most cases. However, idiosyncratic cases following only a single dose of propofol and the absence of hypertriglyceridaemia have also been reported, suggesting an alternative mechanism. Small prospective and retrospective series have not identified any patient group or procedure for which propofol should be specifically avoided. Notably, the use of propofol does not appear to increase the risk of pancreatitis following endoscopic retrograde cholangiopancreatography (ERCP).[56]

Propofol Infusion Syndrome Propofol infusion syndrome (PRIS) is a very rare but potentially lethal side effect of propofol that remains incompletely understood. In the early 1990s, case reports emerged of a fatal syndrome of metabolic acidosis and myocardial failure in paediatric ICU patients receiving prolonged propofol infusions at high doses (>4 mg.kg−1.h−1 for more than 24 hours), particularly those receiving catecholamines and glucocorticoids; subsequent reports of fatal events in adult patients emerged, featuring rhabdomyolysis. A seminal report in the Lancet in 2001[57] led to an FDA warning on the use of propofol for prolonged sedation in children, which was updated in 2006 to include an upper limit in dosing of 4 mg.kg−1.h−1. The clinical features of PRIS include metabolic acidosis, fever, rhabdomyolysis, hyperkalaemia, renal failure, arrhythmia and Brugada pattern on the ECG, hyperlipidaemia and, often, rapid myocardial failure. However, presentation is highly variable and, despite increased awareness and regulatory warnings, PRIS still occurs, with a mortality rate of ~35 to 50%.[58] Reports of PRIS in paediatric patients receiving high doses of propofol have declined but reports of atypical presentations in elderly adult patients receiving apparently normal doses have increased. The most important predictors of death are the cumulative

dose of propofol, the development of fever and the presence of traumatic brain injury. When using prolonged infusion of propofol in critically ill patients, it has been suggested to monitor creatinine phosphokinase (CPK) – if the value stays below 5000 ml, propofol may be safely continued. Understanding of the mechanism of PRIS has focused on the structural similarity of propofol to coenzyme Q, causing uncoupling of the mitochondrial respiratory chain through blocking electron transfer between cytochrome complexes.[59] Other studies have identified interference with fatty acid oxidation.[60]

References 1. M.T. Baker, M. Naguib. Propofol: the challenges of formulation. Anesthesiology 2005; 103: 860–76. 2. L.P. Briggs, R.S. Clarke, J. Watkins. An adverse reaction to the administration of disoprofol (Diprivan). Anaesthesia 1982; 37: 1099–101. 3. G.C. Cummings, J. Dixon, N.H. Kay, et al. Dose requirements of ICI 35,868 (propofol, ‘Diprivan’) in a new formulation for induction of anaesthesia. Anaesthesia 1984; 39: 1168–71. 4. M.D. Krasowski, A. Jenkins, P. Flood, A.Y. Kung, A.J. Hopfinger, N.L. Harrison. General anesthetic potencies of a series of propofol analogs correlate with potency for potentiation of gamma-aminobutyric acid (GABA) current at the GABA(A) receptor but not with lipid solubility. J Pharmacol Exp Ther 2001; 297: 338–51. 5. D. Ruesch, E. Neumann, H. Wulf, S.A. Forman. An allosteric coagonist model for propofol effects on alpha1beta2gamma2L gamma-aminobutyric acid type A receptors. Anesthesiology 2012; 116: 47–55. 6. J. Simon, H. Wakimoto, N. Fujita, M. Lalande, E. A. Barnard. Analysis of the set of GABA(A) receptor genes in the human genome. J Biol Chem 2004; 279: 41422–35. 7. M.D. Krasowski, K. Nishikawa, N. Nikolaeva, A. Lin, N.L. Harrison. Methionine 286 in transmembrane domain 3 of the GABAA receptor beta subunit controls a binding cavity for propofol and other alkylphenol general anesthetics. Neuropharmacology 2001; 41: 952–64. 8. R. Lingamaneni, M.L. Birch, H.C. Hemmings, Jr. Widespread inhibition of sodium channel-dependent glutamate release from isolated nerve terminals by isoflurane and propofol. Anesthesiology 2001; 95: 1460–6. 9. S.W. Ying, S.Y. Abbas, N.L. Harrison, P.A. Goldstein. Propofol block of I(h) contributes to the suppression of neuronal excitability and rhythmic burst firing in thalamocortical neurons. Eur J Neurosci 2006; 23: 465–80.

Downloaded from https://www.cambridge.org/core. IP address: 95.181.217.50, on 02 Dec 2019 at 21:42:28, subject to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316659069.006

27

Chapter 4: Milk of Amnesia: Propofol for TIVA

10. D. Takizawa, H. Hiraoka, F. Goto, K. Yamamoto, R. Horiuchi. Human kidneys play an important role in the elimination of propofol. Anesthesiology 2005; 102: 327–30. 11. T.L. Chen, T.H. Ueng, S.H. Chen, P.H. Lee, S.Z. Fan, C.C. Liu. Human cytochrome P450 mono-oxygenase system is suppressed by propofol. Br J Anaesth 1995; 74: 558–62. 12. H.B. Scott, S.W. Choi, G.T. Wong, M.G. Irwin. The effect of remifentanil on propofol requirements to achieve loss of response to command vs. loss of response to pain. Anaesthesia 2017; 72: 479–87. 13. J. Schuttler, H. Ihmsen. Population pharmacokinetics of propofol: a multicenter study. Anesthesiology 2000; 92: 727–38. 14. M.A. Hughes, P.S. Glass, J.R. Jacobs. Context-sensitive half-time in multicompartment pharmacokinetic models for intravenous anesthetic drugs. Anesthesiology 1992; 76: 334–41. 15. T. Kirkpatrick, I.D. Cockshott, E.J. Douglas, W.S. Nimmo. Pharmacokinetics of propofol (diprivan) in elderly patients. Br J Anaesth 1988; 60: 146–50. 16. I. Murat, V. Billard, J. Vernois, et al. Pharmacokinetics of propofol after a single dose in children aged 1–3 years with minor burns. Comparison of three data analysis approaches. Anesthesiology 1996; 84: 526–32. 17. F. Servin, J.M. Desmonts, J.P. Haberer, I.D. Cockshott, G.F. Plummer, R. Farinotti. Pharmacokinetics and protein binding of propofol in patients with cirrhosis. Anesthesiology 1988; 69: 887–91. 18. B. Ickx, I.D. Cockshott, L. Barvais, et al. Propofol infusion for induction and maintenance of anaesthesia in patients with end-stage renal disease. Br J Anaesth 1998; 81: 854–60. 19. D. Dong, X. Peng, J. Liu, H. Qian, J. Li, B. Wu. Morbid obesity alters both pharmacokinetics and pharmacodynamics of propofol: dosing recommendation for anesthesia induction. Drug Metab Dispos 2016; 44: 1579–83. 20. X. Liu, K.K. Lauer, B.D. Ward, S.J. Li, A.G. Hudetz. Differential effects of deep sedation with propofol on the specific and nonspecific thalamocortical systems: a functional magnetic resonance imaging study. Anesthesiology 2013; 118: 59–69. 21. U. Lee, S. Ku, G. Noh, S. Baek, B. Choi, G.A. Mashour. Disruption of frontal-parietal communication by ketamine, propofol, and sevoflurane. Anesthesiology 2013; 118: 1264–75. 22. M. Qiu, D. Scheinost, R. Ramani, R.T. Constable. Multi-modal analysis of functional connectivity and cerebral blood flow reveals shared and unique effects of

propofol in large-scale brain networks. Neuroimage 2017; 148: 130–40. 23. K.O. Pryor, R.A. Reinsel, M. Mehta, Y. Li, J.T. Wixted, R.A. Veselis. Visual P2-N2 complex and arousal at the time of encoding predict the time domain characteristics of amnesia for multiple intravenous anesthetic drugs in humans. Anesthesiology 2010; 113: 313–26. 24. K.O. Pryor, J.C. Root, M. Mehta, et al. Effect of propofol on the medial temporal lobe emotional memory system: a functional magnetic resonance imaging study in human subjects. Br J Anaesth 2015; 115 Suppl 1: i104–i13. 25. K.K. Kaisti, J.W. Langsjo, S. Aalto, et al. Effects of sevoflurane, propofol, and adjunct nitrous oxide on regional cerebral blood flow, oxygen consumption, and blood volume in humans. Anesthesiology 2003; 99: 603–13. 26. K.D. Petersen, U. Landsfeldt, G.E. Cold, et al. Intracranial pressure and cerebral hemodynamic in patients with cerebral tumors: a randomized prospective study of patients subjected to craniotomy in propofol-fentanyl, isoflurane-fentanyl, or sevoflurane-fentanyl anesthesia. Anesthesiology 2003; 98: 329–36. 27. E.H. Liu, H.K. Wong, C.P. Chia, H.J. Lim, Z.Y. Chen, T.L. Lee. Effects of isoflurane and propofol on cortical somatosensory evoked potentials during comparable depth of anaesthesia as guided by bispectral index. Br J Anaesth 2005; 94: 193–7. 28. D.B. Macdonald, S. Skinner, J. Shils, C. Yingling. Intraoperative motor evoked potential monitoring: a position statement by the American Society of Neurophysiological Monitoring. Clin Neurophysiol 2013; 124: 2291–316. 29. D. San-juan, K.H. Chiappa, A.J. Cole. Propofol and the electroencephalogram. Clin Neurophysiol 2010; 121: 998–1006. 30. K. Engelhard, C. Werner, E. Eberspacher, et al. Influence of propofol on neuronal damage and apoptotic factors after incomplete cerebral ischemia and reperfusion in rats: a long-term observation. Anesthesiology 2004; 101: 912–17. 31. G.W. Roach, M.F. Newman, J.M. Murkin, et al. Ineffectiveness of burst suppression therapy in mitigating perioperative cerebrovascular dysfunction. Multicenter Study of Perioperative Ischemia (McSPI) Research Group. Anesthesiology 1999; 90: 1255–64. 32. F. Boer, P. Ros, J.G. Bovill, P. van Brummelen, J. van der Krogt. Effect of propofol on peripheral vascular

Downloaded from https://www.cambridge.org/core. IP address: 95.181.217.50, on 02 Dec 2019 at 21:42:28, subject to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316659069.006

28

Chapter 4: Milk of Amnesia: Propofol for TIVA

resistance during cardiopulmonary bypass. Br J Anaesth 1990; 65: 184–9. 33. D.W. Green. Cardiac output decrease and propofol: what is the mechanism? Br J Anaesth 2015; 114: 163–4. 34. M. Sato, M. Tanaka, S. Umehara, T. Nishikawa. Baroreflex control of heart rate during and after propofol infusion in humans. Br J Anaesth 2005; 94: 577–81. 35. R. Larsen, J. Rathgeber, A. Bagdahn, H. Lange, H. Rieke. Effects of propofol on cardiovascular dynamics and coronary blood flow in geriatric patients. A comparison with etomidate. Anaesthesia 1988; 43 Suppl: 25–31. 36. Z. Xia, Z. Huang, D.M. Ansley. Large-dose propofol during cardiopulmonary bypass decreases biochemical markers of myocardial injury in coronary surgery patients: a comparison with isoflurane. Anesth Analg 2006; 103: 527–32. 37. D.M. Ansley, K. Raedschelders, P.T. Choi, B. Wang, R.C. Cook, D.D. Chen. Propofol cardioprotection for on-pump aortocoronary bypass surgery in patients with type 2 diabetes mellitus (PRO-TECT II): a phase 2 randomized-controlled trial. Can J Anaesth 2016; 63: 442–53. 38. S. Sayed, N.K. Idriss, H.G. Sayyedf, et al. Effects of propofol and isoflurane on haemodynamics and the inflammatory response in cardiopulmonary bypass surgery. Br J Biomed Sci 2015; 72: 93–101. 39. C.J. Jakobsen, H. Berg, K.B. Hindsholm, N. Faddy, E. Sloth. The influence of propofol versus sevoflurane anesthesia on outcome in 10,535 cardiac surgical procedures. J Cardiothorac Vasc Anesth 2007; 21: 664–71. 40. D. Nieuwenhuijs, E. Sarton, L. Teppema, A. Dahan. Propofol for monitored anesthesia care: implications on hypoxic control of cardiorespiratory responses. Anesthesiology 2000; 92: 46–54. 41. D. Nieuwenhuijs, E. Sarton, L.J. Teppema, et al. Respiratory sites of action of propofol: absence of depression of peripheral chemoreflex loop by low-dose propofol. Anesthesiology 2001; 95: 889–95. 42. N.W. Goodman, A.M. Black, J.A. Carter. Some ventilatory effects of propofol as sole anaesthetic agent. Br J Anaesth 1987; 59: 1497–503. 43. A. Gupta, T. Stierer, R. Zuckerman, N. Sakima, S. D. Parker, L.A. Fleisher. Comparison of recovery profile after ambulatory anesthesia with propofol, isoflurane, sevoflurane and desflurane: a systematic review. Anesth Analg 2004; 98: 632–41, table of contents. 44. H. Unlugenc, T. Guler, Y. Gunes, G. Isik. Comparative study of the antiemetic efficacy of ondansetron,

propofol and midazolam in the early postoperative period. Eur J Anaesthesiol 2004; 21: 60–5. 45. S.N. Bennett, M.M. McNeil, L.A. Bland, et al. Postoperative infections traced to contamination of an intravenous anesthetic, propofol. N Engl J Med 1995; 333: 147–54. 46. I. Wachowski, D.T. Jolly, J. Hrazdil, J.C. Galbraith, M. Greacen, A.S. Clanachan. The growth of microorganisms in propofol and mixtures of propofol and lidocaine. Anesth Analg 1999; 88: 209–12. 47. I.T. Cohen, R.S. Hannallah, D.B. Goodale. The clinical and biochemical effects of propofol infusion with and without EDTA for maintenance anesthesia in healthy children undergoing ambulatory surgery. Anesth Analg 2001; 93: 106–11. 48. FDA. Information for Healthcare Professionals: Propofol (marketed as Diprivan and as generic products). www.fda.gov/Drugs/DrugSafety/Postmark etDrugSafetyInformationforPatientsandProviders/uc m125817.htm [Accessed 20 January 2016]. 49. L.L. Asserhoj, H. Mosbech, M. Kroigaard, L.H. Garvey. No evidence for contraindications to the use of propofol in adults allergic to egg, soy or peanut. Br J Anaesth 2016; 116: 77–82. 50. J. Molina-Infante, A. Arias, D. Vara-Brenes, et al. Propofol administration is safe in adult eosinophilic esophagitis patients sensitized to egg, soy, or peanut. Allergy 2014; 69: 388–94. 51. A. Murphy, D.E. Campbell, D. Baines, S. Mehr. Allergic reactions to propofol in egg-allergic children. Anesth Analg 2011; 113: 140–4. 52. J.W. Devlin, A.K. Lau, M.A. Tanios. Propofolassociated hypertriglyceridemia and pancreatitis in the intensive care unit: an analysis of frequency and risk factors. Pharmacotherapy 2005; 25: 1348–52. 53. J.C. Devaud, M.M. Berger, A. Pannatier, et al. Hypertriglyceridemia: a potential side effect of propofol sedation in critical illness. Intensive Care Med 2012; 38: 1990–8. 54. P.S. Myles, M.R. Buckland, D.J. Morgan, A.M. Weeks. Serum lipid and glucose concentrations with a propofol infusion for cardiac surgery. J Cardiothorac Vasc Anesth 1995; 9: 373–8. 55. G.S. Leisure, J. O’Flaherty, L. Green, D.R. Jones. Propofol and postoperative pancreatitis. Anesthesiology 1996; 84: 224–7. 56. N. Li, A. Tieng, S. Novak, et al. Effects of medications on post-endoscopic retrograde cholangiopancreatography pancreatitis. Pancreatology 2010; 10: 238–42. 57. O.L. Cremer, K.G. Moons, E.A. Bouman, J.E. Kruijswijk, A.M. de Smet, C.J. Kalkman. Long-

Downloaded from https://www.cambridge.org/core. IP address: 95.181.217.50, on 02 Dec 2019 at 21:42:28, subject to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316659069.006

29

Chapter 4: Milk of Amnesia: Propofol for TIVA

term propofol infusion and cardiac failure in adult head-injured patients. Lancet 2001; 357: 117–18. 58. A. Krajcova, P. Waldauf, M. Andel, F. Duska. Propofol infusion syndrome: a structured review of experimental studies and 153 published case reports. Crit Care 2015; 19: 398.

59. A.V. Vanlander, J.G. Okun, A. de Jaeger, et al. Possible pathogenic mechanism of propofol infusion syndrome involves coenzyme q. Anesthesiology 2015; 122: 343–52. 60. A. Wolf, P. Weir, P. Segar, J. Stone, J. Shield. Impaired fatty acid oxidation in propofol infusion syndrome. Lancet 2001; 357: 606–7.

Downloaded from https://www.cambridge.org/core. IP address: 95.181.217.50, on 02 Dec 2019 at 21:42:28, subject to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316659069.006

30

Chapter

5

A Catwalk with a Difference What Distinguishes TIVA Models? Lim Thiam Aun

Introduction One of the first things that comes to mind, when we talk about models is fashion! Models walking down a catwalk can allow a designer to showcase his/her newest creation and a prospective purchaser can see what the clothes look like in use. Choosing a PK model is not much different. The anaesthetist looks at all the attributes of the models and decides which model is best suited for a particular purpose. Total intravenous anaesthesia involves giving IV medications to achieve a state appropriate for the degree of surgical stimulation. It is generally assumed that the degree of effect obtained is related to the concentration of the drug given. As such, the most rational approach to TIVA is to appropriately administer a drug to achieve and maintain a given target concentration, which can be easily adjusted up and down as appropriate. In order to do this, one must be able to describe how the body handles a drug once it is given and, from there, be able to predict what concentration of the drug will be achieved when given a specific dose. Of course, every individual is different and how the drug is distributed and eliminated differs from one person to another. On top of this, drug disposition is also affected to some extent by the stimulation or depression of the various organ systems during surgery. This makes it challenging to predict accurately drug concentrations over the course of a surgical procedure but it can be done with TCI.

Pharmacokinetic Models In order to describe the movement of the drug after it enters the body, mathematical models are used to calculate the concentration of the drug in different ‘apparent’ compartments. These models do not reflect the true anatomy of the body and,

at times, may describe volumes far in excess of any physiological structure, especially with drugs that are very lipid soluble. However, such PK models are used in TIVA to calculate the infusion rate needed to achieve and maintain the desired target drug concentration. The most common way to describe these models is by using a set of ‘apparent volumes’ and ‘clearances’. These same parameters can alternatively be expressed as a ‘central volume’ and a set of transfer rate constants (including the elimination rate constant). In addition, each parameter may itself be further modified by co-variates such as weight, height, age and sex; which may further refine its accuracy. Most models describing drugs given intravenously are three-compartment models as shown in Figure 5.1. Some of the PK sets of commonly used drugs are shown in Table 5.1. Because of inter- and intra-individual variability, the ability of any one model to predict every individual in the population is limited. The model, at best,

Table 5.1(a) Pharmacokinetic parameter sets for propofol TCI.

Propofol Marsh

Schnider

Central compartment (V1) (L)

0.228 × weight

4.27

k10 (min−1)

0.119

[1.89 + 0.0456 × (weight – 77) – 0.0681 × (LBM – 59) + 0.0264 × (height – 177)]/V1

k12 (min−1)

0.112

0.302 – 0.0056 (Age – 53)

k13 (min−1)

0.0419

0.196

k21 (min )

0.055

[1.29 – 0.024 (Age – 53)]/ [18.9 – 0.391 (Age – 53)]

k31 (min−1)

0.0033

0.0035

0.24 – 1.2

0.456

−1

−1 c

keo (min )

Downloaded from https://www.cambridge.org/core. IP address: 94.158.22.118, on 02 Dec 2019 at 21:43:41, subject to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316659069.007

31

Chapter 5: What Distinguishes TIVA Models?

Table 5.1(b) Pharmacokinetic parameter sets for remifentanil, alfentanil and sufentanil TCI.

Remifentanil

Alfentanil

Minto

Sufentanil a

Scott

Maitre

Gepts

Bovillb

Central compartment (V1) (L)

5.1–0.0201 (Age – 40) + 0.072 (LBM – 55)

2.19

M: 0.111 × weight F: 0.128 × weight

14.3

0.164 × weight

k10 (min−1)

[2.6 – 0.0162 × (Age – 40) + 0.0191 × (LBM – 55)]/V1

0.0894

M: 3.2/weight F: 2.8/weight

0.0645

0.089

k12 (min−1)

[2.05 – 0.0301 × (Age – 40)]/V1

0.654

0.104

0.1086

0.35

k13 (min−1)

[0.076 – 0.00113 × (Age – 40)]/V1

0.209

0.017

0.0229

0.077

k21 (min )

[2.05 – 0.0301 × (Age–40)]/[9.82 – 0.0811 × (Age–40) + 0.108 × (LBM – 55)]

0.118

0.0673

0.0245

0.161

k31 (min−1)

[0.076 – 0.00113 × (Age – 40)]/5.42

0.0177

0.0126

0.0013

0.01

0.595 – 0.007 × (Age – 40)

0.77

0.63

0.112

0.119

−1

−1 c

keo (min )

Some drugs may be described by more than one PK model. a Further adjustments needed for patients above the age of 40 years. b The Bovill model added for comparison – not usually used for TCI. c keo values may have been derived in other studies some time after the PK models named above were reported.

Bolus dose of drug Effect site Ke0 Central Compartment V1

K12 Rapidly equilibrating compartment V2

K21

K13 K31

Slowly equilibrating compartment V3

K10 Elimination/metabolism Figure 5.1 Three-compartment open mammillary model. k12 represents the first-degree rate constant relating concentration of the drug in the central compartment to the rate of transfer of drug from the central compartment to the rapidly equilibrating compartment. k13 represents the first-degree rate constant describing transfer from the central compartment to the slowly equilibrating compartment. k21 represents the first-degree rate constant describing transfer from the rapidly equilibrating compartment to the central compartment. k31 represents the first-degree rate constant describing transfer from the slowly equilibrating compartment to the central compartment. k10 is the elimination rate constant. keo is the rate constant relating the elimination of the drug from the effect compartment (effect site) to the concentration of the drug in the effect compartment. The effect compartment is assumed to be negligible in volume and has no significant effect on the distribution of drug within the three-compartment model.

represents an ‘average’ member of the population. Drug concentrations of individuals at extreme ends of the spectrum may not be well predicted. This is compounded by the fact that each model is derived

from a sample of a specific population defined by the inclusion criteria of the trial investigating the model. Such a model needs to be verified in order for it to be useful when applied to a different population.

Downloaded from https://www.cambridge.org/core. IP address: 94.158.22.118, on 02 Dec 2019 at 21:43:41, subject to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316659069.007

32

Chapter 5: What Distinguishes TIVA Models?

Naming Pharmacokinetic Models When the values of a PK model are first documented in a scientific publication, that particular model is usually named after the first author of the paper. At times, this may appear somewhat odd – as in the case of the Marsh model. The parameter values used for the calculation of the propofol infusion rate in adult patients first appeared in the article by Marsh and colleagues, although the article itself was on the pharmacokinetics of propofol in children.

In Come the Models Down the Catwalk It is not surprising, then, that many models have been derived by researchers over the years to describe the various drugs used in anaesthesia. One can imagine the many models of propofol moving down the catwalk, each showing off its own special characteristics. Like a fashion show where some clothes may not fit well, there will be models that do not predict concentrations well and which will be excluded from further investigation and development. On the flip side, there may not be a clear-cut winner. Two models may perform equally well (or equally badly). Just like certain clothes may look better on a tall person, or a muscular or thin person, one model may accurately predict concentrations in one patient better than another. Fashion models parading down a catwalk may show off costumes that may be intricate with many appendages, or may be simple without any frills. Similarly, PK models may simply be one set of numbers, or each parameter in the model may be an equation of its own. In either case, it is not how intricate the model is that matters, but how well it performs in practice. At times simplicity is more efficient than complexity.

The Models Taking Centre Stage on the Catwalk The hypnotic component of most TIVA is propofol and several models have been described for this drug. However, the two most commonly incorporated into commercially available TCI devices are the Marsh and the Schnider models.[1–3] Both have their supporters and are useful for calculating propofol delivery rates during TIVA. However, these models are as different as can be!

The Marsh Model This is an adapted version of a model originally described by Gepts.[4] It is relatively simple, consisting

of a central compartment, which is weight proportional, and five transfer rate constants, each having a fixed value. The simplicity of this model means that only the weight of the patient needs to be input to the TCI device. This model has been used in TCI pumps for many years, initially with the Diprifusor™, and has stood the test of time.[5] The model appears to work equally well in different populations[6,7] and, despite its simplicity, has acceptable bias and precision. The Diprifusor has been licensed for use in most countries of the world, with the exception of the USA, and the product licence provides the only prescribing information for administration of propofol by TCI. One feature of the Marsh model is that a value for the keo was not included until some time later. Different researchers attempted to derive a value for this rate constant as an add-on to the original set of parameters, which led to different values of keo being added to the model by different TCI device manufacturers. This will not affect the calculated plasma concentrations but has implications for the accuracy of the calculated effect-site concentration. Consequently, some TCI pumps will display a calculated or predicted effect-site value greater than the likely ‘actual’ effectsite concentration if the keo used is too fast. The actual effect-site concentration (i.e. in the brain) is impossible to measure and is estimated by using derived parameters such as EEG (see Chapter 2).

The Schnider Model This model is more complex than the Marsh model. Patient weight, height, age and sex are required to be entered to the model. Lean body mass (LBM), calculated from the above data, is also used in the calculation of the various PK parameters of the model. The rationale for the greater number of co-variates lies in the fact that two patients with the same weight (or height) may not have the same body structure and therefore will not handle the drug in the same way. By including more information on the patient, it is hoped that more accurate dosing can be achieved for any one particular subject. Despite the increased PK complexity there are, however, certain caveats and limitations. The model has a fixed central compartment volume of distribution, meaning the initial bolus will need to be small, otherwise small-sized patients would receive an initial overdose. On the other hand, in larger patients, there will need to be a faster rate of infusion following the initial bolus in order to raise the plasma propofol

Downloaded from https://www.cambridge.org/core. IP address: 94.158.22.118, on 02 Dec 2019 at 21:43:41, subject to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316659069.007

33

Chapter 5: What Distinguishes TIVA Models?

Problems with Models Predicting the Propofol Effect-Site Concentration As mentioned above, targeting the effect site requires one extra rate constant, the keo. The wide range of values reported by different investigators to be associated with the Marsh model testifies to the difficulty in obtaining a single value to describe this movement of drug between the central and effect compartments.[9–11] It may be that propofol affects its own pharmacokinetics as an increasing drug concentration may depress the cardiovascular system, hence delay the transfer of drug between compartments. The difficulty that anaesthetists face is deciding ‘Which is the correct keo to use?’. This issue has been addressed in an editorial and in Chapter 2.[12] On the other hand, the Schnider model, having incorporated the keo into the parameter set, may give less emphasis to the central compartment as long as the effect-site concentration matches the expected measured effect well. Here the anaesthetist may ask ‘How accurate is the predicted plasma concentration?’.

Comparison of the Two Propofol Models It all boils down to: ‘Should I use the Marsh model or the Schnider model for propofol TCI?’ Using the logic in the above section, it has been suggested to use the

Marsh model if you are targeting the plasma (or blood) concentration and use the Schnider model if you are targeting the effect site.[13] Unfortunately, this advice was taken by certain TCI pump manufacturers, such as B. Braun, who only incorporate these two options in their device. While this is intended to simplify administration by reducing choice, we do not agree with this interpretation because of the limitations of the Schnider model addressed elsewhere in this and other chapters. Essentially the Schnider model will give less propofol for a selected effect-site target than the Marsh. Such under-dosing will give the illusion of safety but it is actually, in our opinion, better to use the Marsh (more accurate) effect site but start at a low concentration and titrate up. Another important question is a bit trickier: ‘Do we target the same concentration when using the Marsh or Schnider models?’ Figure 5.2 shows the results using simulation to calculate the total dose of propofol at 30 minutes, needed in 40-year-old male patients of different weight and height, when given a TCI at 3 μg.ml−1. The target site was plasma for the Marsh model and effect site for the Schnider model. It appears that the dose is similar in patients weighing 54 kg with a height of 154 cm. At less than this, the Schnider model will lead to the calculation of a larger dose compared to the Marsh model, while the opposite is seen above this weight and height. Although it is not as straightforward as this, what is apparent is that the two models will use a different

600 Dose of propofol infused (mg)

concentration. From a practical perspective, as an example of this limitation, a 30 kg patient would receive approximately the same initial bolus dose of propofol as a 90 kg patient – which does not equate with our clinical experience. It has been shown, by measuring blood concentrations, that the Schnider model over-predicts the concentration and therefore under-delivers the dose required in the initial (induction) phase.[8] The Schnider model was developed during combined PK–PD studies, which meant that a value for keo was derived at the same time as the other parameters, although, from practical experience, the value is too fast leading to an over-estimate of effect-site concentrations during induction. Despite the differences between the Marsh and Schnider models, both can achieve and maintain reasonably stable plasma concentrations of propofol. To some extent it doesn’t really matter which model you choose to use as long as you understand its limitations. For this reason it is also best to get used to one particular model and stick with it.

500 400 300 Marsh model Schnider model

200 100 0 40 140

60 160

80 180

100 200

weight (kg) and height (cm)

Figure 5.2 Predicted dose of propofol infused after 30 minutes at TCI of 3 μg.ml−1. The curves were obtained using simulation assuming a standard subject is a 40-year-old male. The weight and height starts at 40 kg in a subject 140 cm tall, with each increase in weight by 1 kg being associated with an increase in height of 1 cm.

Downloaded from https://www.cambridge.org/core. IP address: 94.158.22.118, on 02 Dec 2019 at 21:43:41, subject to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316659069.007

34

Chapter 5: What Distinguishes TIVA Models?

dose for a specific target concentration. Will this lead to a risk of patients receiving too low (or too high) a dose if the target concentration selected does not match what is recommended for use with the model? The answer is simple: there is enough inter- and intrapatient variability to warrant the need to titrate the concentration in every patient receiving TIVA, irrespective of the chosen model. The target concentration should not depend solely on the model used, but on the overall clinical scenario including the patient’s response, clinical signs of consciousness and/or the effects on processed EEG. Now let us return to the initial suggestion of using a different model to target the different site. As the key to safe TIVA is the titration of drug to effect, there should be no reason for not using the Marsh model when targeting the effect site, or the Schnider model for targeting the central compartment. When targeting the effect site using the Marsh model, start low and gradually increase until the target concentration is at the level intended. This will avoid an excessive increase in the plasma concentration with the initial calculated bolus, thereby minimising cardiovascular side effects. A lower keo value programmed into the TCI device will require slower titration of the target concentration upwards. Conversely, if the Schnider model is used to target the plasma concentration, the patient will take a longer time to become unconscious as the dose administered is lower than with Marsh. With patience, however, the patient will reach the desired level of anaesthesia. With regard to the target concentration, the infusion rate appears higher when maintaining a plasma concentration using the Marsh model compared to targeting the effect-site concentration at a numerically similar value using the Schnider model. However, as discussed above, the target concentration depends on how the patient actually responds and will need to be adjusted for each individual patient.

Some Models on the Catwalk Come from Different Nations This parallel between fashion models and PK models extends to the fact that models can be derived for the different medications used for TIVA. Just like how fashion models may come from different nations with different national dress, different drugs come with different pharmacokinetics.

Models Describing Opioids Opioids are used together with propofol to provide the analgesic component of the anaesthetic. The most commonly used opioid is remifentanil, although alfentanil and sufentanil are suitable alternatives. Other opioids, such as morphine and fentanyl, are usually given as boluses for post-operative analgesia or when remifentanil, alfentanil or sufentanil are not available. Fentanyl can be used by infusion for shortterm administration but after an hour or so the context sensitive half-time increases substantially so that it will take much longer to wear off.

Remifentanil: the Minto Model Only one model is commonly used for remifentanil.[14] Essentially, remifentanil’s pharmacokinetics dictate that it will rapidly achieve a near steady state after any change in infusion rate and will rapidly decline when the infusion is turned off. It is metabolised by nonspecific esterases in the blood providing rapid and consistent organ-independent elimination (a short context sensitive half-time of three to five minutes). Using a properly derived model helps to achieve the target concentration in the shortest time and with minimal over- or under-dosage. The Minto model fulfils these criteria for non-obese patients. Like the Schnider model for propofol, the Minto model uses the lean body mass when calculating transfer rate constants. The advantage of using remifentanil is that it can be easily titrated to its endpoint, i.e. the blood pressure, heart rate or a combination of the two. Starting at one target concentration, be it in the plasma or effect site, one can titrate the target concentration up or down until the desired attenuation of surgical stimulation is achieved (usually measured by haemodynamic changes). On cessation of infusion, concentrations fall rapidly to levels consistent with spontaneous ventilation regardless of the target concentration set. As a result of the rapid onset and offset, remifentanil is very easy to administer as a manual infusion if you do not have access to TCI. Infusion rates of 0.15 to 0.35 µg.kg−1.min−1 are typical for surgical analgesia in most patients with boluses of 0.5 to 1 µg.kg−1 as required. When using TCI, values of 2 to 8 ng.ml−1 are typical.

Alfentanil and Sufentanil Like propofol, two separate models are used by TCI devices for alfentanil. Unlike propofol, both models

Downloaded from https://www.cambridge.org/core. IP address: 94.158.22.118, on 02 Dec 2019 at 21:43:41, subject to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316659069.007

35

Chapter 5: What Distinguishes TIVA Models?

are usually not found in the same pump. The Scott model gives a set of fixed values for each of the PK parameters.[15,16] There is no co-variate, meaning, regardless of the weight, height, age or sex of the patient, the same dose of drug will be infused. This may appear irrational but it works very well. Alfentanil has a pKa of 6.5, therefore a very high proportion is uncharged at physiological pH, which explains its rapid onset. The potency is around 20% of fentanyl. The context sensitive half-time (CSHT) is reasonably predictable but rises significantly with time (around 58 minutes after a four-hour infusion). Consequently, using TCI can allow ‘down’-titration towards the end of surgery (usually before propofol). The Maitre model requires weight, age and sex for the calculation of the infusion rates.[17] This seems to be the best model but under-estimates predicted concentrations. Typical doses with TCI are 30 to 150 ng.ml−1. Like alfentanil, sufentanil has also been described by a model using a set of fixed values (Gepts),[15] although other models with a correction for weight are available.[18] Sufentanil is a very potent fentanyl derivative (five to ten times more than fentanyl) with a shorter CSHT than alfentanil (around 30 minutes after a four-hour infusion) but slower onset. This may make it more suitable as a remifentanil substitute (it may also be less likely to induce acute tolerance). Typical doses with TCI (Gepts) are 0.2 to 1 ng.ml−1.

Models Describing Other Drugs Used During TIVA Every drug undergoes clinical trials during which the pharmacokinetics and pharmacodynamics of the drug are documented. Some drugs undergo further investigation, which leads to their pharmacokinetics being described in terms of the mammillary model above. TCI can be feasible for certain drugs but not particularly useful compared to manual regimes in drugs that have a fairly long half-life and greater therapeutic ratio (i.e. they don’t need such close control of concentration), e.g. dexmedetomidine. Apart from the IV hypnotics and analgesics, an infusion of a neuromuscular blocking agent may be given during TIVA. However, this is usually given as a constant-rate infusion and models describing these drugs are usually not used for TCI. The odd instance in which TCI may be considered for neuromuscular blocking agents is during closed-loop anaesthesia,

although this is a far cry from the current practice of TIVA.

Big Models and Young Models Show Up on the Catwalk The discussion, so far, has centred on non-obese, young adults. Just like one cannot use adult fashion models to show off children’s clothes, PK models derived for adults may not predict drug concentrations well in children. This is especially so for models with fixed parameter values (such as for alfentanil and sufentanil). The Minto model for remifentanil is more flexible as it allows the age, weight and height of the patient to be incorporated into the model. The models for propofol are much less flexible. Both models do not perform well in children as children have a higher weight-adjusted central volume and faster clearance when compared to adults. Researchers over the years have described models for propofol in children and some commercial TCI devices have these models incorporated. The two common models are the Kataria model and the Paedfusor model.[19,20] Both models report a higher central volume for children (Kataria: 0.52 l.kg−1; Paedfusor 0.46 l.kg−1) compared to the values for adults as given in Table 5.1. The Kataria model reported a clearance of 34 ml.kg−1. min−1 while the Paedfusor model gives a range of clearance from 27 to 35 ml.kg−1.min−1 in children whose weight ranges from 10 kg to 24 kg. These values are also higher than the values for adult patients. Their use is described in more detail in Chapter 14. The situation in obese patients is again different (see Chapter 16). Both the Schnider and Minto models require the lean body mass (LBM) for the calculation of the transfer rate constants and both rely on the James formula for calculating this.[21] Unfortunately, using this equation, the LBM decreases paradoxically in males when it is above a BMI of 43 kg.m−2 and in females when BMI is above 36 kg.m−2. Consequently, most TCI devices are programmed to reject the patient data if the calculated BMI exceeds these values. One method of getting around this problem is to use the maximum LBM for the patient’s height. This is easily obtained by calculating the weight using the patient’s height and the BMI values as above. Some TCI devices automatically calculate the maximum

Downloaded from https://www.cambridge.org/core. IP address: 94.158.22.118, on 02 Dec 2019 at 21:43:41, subject to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316659069.007

36

Chapter 5: What Distinguishes TIVA Models?

LBM while others require the user to key in the adjusted value for the weight.[13] However, as the adjusted weight is lower than the patient’s true weight, it will always bring along a worry that too little drug will be given. In addition, obese patients tend to have a significant amount of muscle (required to carry around the excess weight!) such that LBM is an under-estimate of ideal body weight. Generally, obese patients need a higher loading dose than patients of normal body weight but, interestingly, the maintenance doses are not much higher (an analogy is of a sponge soaking up more water). Consequently, some have suggested using a weight around 20 to 40% higher than LBM. However, whatever method you choose, the solution is, once again, titration to effect (you can see we emphasise this throughout the book!). With remifentanil, this can be done by observing the patient’s response to pain, e.g. airway manipulation or surgical stimulation and its effects on blood pressure and pulse rate. The implication is that no matter whether the anaesthetist wishes to use the Marsh model or the Schnider model, the principle is to start low and go slow. Start with a lower target concentration to ensure that the blood concentration does not increase too fast leading to undesirable haemodynamic side effects. The target concentration can then be slowly increased until the same endpoints as those in non-obese patients are reached. The use of a depth-ofanaesthesia monitor, such as processed EEG, may be useful to guide the management of these patients.

Down the Catwalk Comes a Non-Model! It is frequently assumed that TIVA will need to be given using two TCI devices: one for propofol and another for the opioid. A problem comes when only one, or no, TCI device is available. Which drug should we give using TCI? And does the model used help us to decide?

Propofol vs Remifentanil Models A re-look at the propofol models will reveal that the calculated elimination clearance is 26 to 27 ml.kg−1.min−1. The calculated clearance for remifentanil is much higher, ranging from less than 90 ml.kg−1.min−1 to more than 150 ml.kg−1.min−1, depending on the age and sex of the patient. This would suggest the concentration of remifentanil

declines faster when the infusion is switched off, making it easier to titrate. When faced with the choice of using only one TCI, some have suggested it is better to use propofol as a fixed-rate infusion and remifentanil as the TCI. One reason, as stated above, is that the concentration of remifentanil can be adjusted more easily. The second is that hypnosis is usually considered a yes-or-no phenomenon, meaning that once the ‘sleep’ concentration is reached, keeping the concentration constant should mean the patient will remain asleep. While we follow this argument, we think it is better to use the TCI for propofol. The reason is that propofol manual infusion is more complicated (PK) and making mistakes could lead to awareness or excessive depth (PD). On the other hand, remifentanil PK is quite simple and the µg.kg−1.min−1 regime is easy, and even recommended by the manufacturer, so that even inhalation anaesthetic users can manage it. The PD is also safer (assuming the patient is being ventilated) – too high is, obviously, not ideal but fairly safe; and too low likewise. PD can also be monitored with sympathetic responses. As both the Marsh and Schnider models report almost similar clearance, the concentration at near steady state for a given fixed infusion rate should be relatively similar. This means the expected fixed infusion rate of propofol for hypnosis should not vary much regardless of the model used to determine the infusion rate. The alternative to using a fixed constant rate infusion would be to incorporate the model into a handheld computer and periodically calculate the required infusion rate, then manually change the rate on the infusion pump – but this is laborious and complicated.

The Show (on the Catwalk) Finally Draws to an End Having considered all the information at hand, the anaesthetist has to decide which model is most suited for purpose. As in a fashion competition, each category may have a winner, and some categories may have more than one!

No Model Is Perfect Like most things, there are imperfections when it comes to using a model to predict drug concentrations. Considering that models are ‘mathematical’ rather than ‘physiological’ in nature, it is not surprising that such inaccuracies will occur. Added to this is

Downloaded from https://www.cambridge.org/core. IP address: 94.158.22.118, on 02 Dec 2019 at 21:43:41, subject to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316659069.007

37

Chapter 5: What Distinguishes TIVA Models?

the intra- and inter-individual variability that occurs in biological systems. Using the Marsh model as an example, three separate studies reported a wide range for the median absolute performance error (i.e. inaccuracy) during infusion when it was used in a TCI system. Swinhoe et al. reported a range of 6.3 to 84.1%, Short et al. reported a range of 6.8 to 49.7%, while Coetzee et al. gave a 10 to 90 percentile range of 8.3 to 52.8%.[5–7] This illustrates the difficulty of consistently achieving the target concentration. The Minto remifentanil model did not fare much better. The range of median absolute performance error reported by Mertens et al. was about 5 to 40%.[22] At the end of the day, there is still some art to the practice of anaesthesia. The machines can help us to provide stable blood concentrations of anaesthetic and analgesic drugs but we need to use our experience to titrate them appropriately.

How Do We Decide Which Model to Use? In most instances, PK models that perform poorly on evaluation will not be pre-installed in commercially available infusion pumps. However, the preference of which pre-installed model to use lies solely with the attending anaesthetist. The usual advice is for the anaesthetist to use the model that he or she is most comfortable with. One may predict the concentrations better in one patient than another. The same holds true when it comes to targeting the plasma vs effect site. Use a model that has been evaluated for targeting that particular compartment rather than a model that was developed to predict the concentration in a different compartment. At the end of the day, it boils down to ensuring the patient is receiving the expected amount of drug through the infusion system and that the monitored physiological parameters are within reasonable limits. Titration, titration, titration. . .

Postlude: When Two Models Get Married Usually more than one drug is used during TIVA and both drugs may be given via TCI. The usual practice is to assume the pharmacokinetics does not change but the pharmacodynamics may change because of drug interaction. As such, there is no need to alter the model used for calculating the infusion rate but adjustment may need to be made to the desired target concentration.

References 1. B. Marsh, M. White, N. Morton, G.N. Kenny. Pharmacokinetic model driven infusion of propofol in children. Br J Anaesth 1991; 67: 41–8. 2. T.W. Schnider, C.F. Minto, P.L. Gambus, et al. The influence of method of administration and covariates on the pharmacokinetics of propofol in adult volunteers. Anesthesiology 1998; 88: 1170–82. 3. T.W. Schnider, C.F. Minto, S.L. Shafer, et al. The influence of age on propofol pharmacodynamics. Anesthesiology 1999; 90: 1502–16. 4. E. Gepts, F. Camu, I.D. Cockshott, E.J. Douglas. Disposition of propofol administered as constant rate intravenous infusions in humans. Anesth Analg 1987; 66: 1256–63. 5. C.F. Swinhoe, J.E. Peacock, J.B. Glen, C.S. Reilly. Evaluation of the predictive performance of a ‘Diprifusor’ TCI system. Anaesthesia 1998; 53 Suppl 1: 61–7. 6. J.F. Coetzee, J.B. Glen, C.A. Wium, L. Boshoff. Pharmacokinetic model selection for target controlled infusions of propofol. Assessment of three parameter sets. Anesthesiology 1995; 82: 1328–45. 7. T.G. Short, T.A. Lim, Y.H. Tam. Prospective evaluation of pharmacokinetic model-controlled infusion of propofol in adult patients. Br J Anaesth 1996; 76: 313–5. 8. J.B. Glen, F. Servin. Evaluation of the predictive performance of four pharmacokinetic models for propofol. Br J Anaesth 2009; 102: 626–32. 9. N. Fabregas, J. Rapado, P.L. Gambus, et al. Modeling of the sedative and airway obstruction effects of propofol in patients with Parkinson disease undergoing stereotactic surgery. Anesthesiology 2002; 97: 1378–86. 10. M.M. Struys, T. De Smet, B. Depoorter, et al. Comparison of plasma compartment versus two methods for effect compartment–controlled targetcontrolled infusion for propofol. Anesthesiology 2000; 92: 399–406. 11. T.A. Lim, W.H. Wong, K.Y. Lim. Effect-compartment equilibrium rate constant (keo) for propofol during induction of anesthesia with a target-controlled infusion device. J Anesth 2006; 20: 153–5. 12. J.B. Glen, F.H. Engbers. The influence of target concentration, equilibration rate constant (ke0) and pharmacokinetic model on the initial propofol dose delivered in effect-site target-controlled infusion. Anaesthesia 2016; 71: 306–14. 13. A.R. Absalom, V. Mani, T. De Smet, M.M. Struys. Pharmacokinetic models for propofol: defining and illuminating the devil in the detail. Br J Anaesth 2009; 103: 26–37.

Downloaded from https://www.cambridge.org/core. IP address: 94.158.22.118, on 02 Dec 2019 at 21:43:41, subject to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316659069.007

38

Chapter 5: What Distinguishes TIVA Models?

14. C.F. Minto, T.W. Schnider, T.D. Egan, et al. Influence of age and gender on the pharmacokinetics and pharmacodynamics of remifentanil. I. Model development. Anesthesiology 1997; 86: 10–23. 15. J.C. Scott, D.R. Stanski. Decreased fentanyl and alfentanil dose requirements with age. A simultaneous pharmacokinetic and pharmacodynamic evaluation. J Pharmacol Exp Ther 1987; 240: 159–66. 16. A. Kapila, P.S. Glass, J.R. Jacobs, et al. Measured context-sensitive half-times of remifentanil and alfentanil. Anesthesiology 1995; 83: 968–75. 17. P.O. Maitre, S. Vozeh, J. Heykants, D.A. Thomson, D.R. Stanski. Population pharmacokinetics of alfentanil: the average dose-plasma concentration relationship and interindividual variability in patients. Anesthesiology 1987; 66: 3–12.

18. J.G. Bovill, P.S. Sebel, C.L. Blackburn, V. Oei-Lim, J.J. Heykants. The pharmacokinetics of sufentanil in surgical patients. Anesthesiology 1984; 61: 502–6. 19. B.K. Kataria, S.A. Ved, H.F. Nicodemus, et al. The pharmacokinetics of propofol in children using three different data analysis approaches. Anesthesiology 1994; 80: 104–22. 20. A. Absalom, G. Kenny. ‘Paedfusor’ pharmacokinetic data set. Br J Anaesth 2005; 95: 110. 21. W.P. James. Research on Obesity. London: Her Majesty’s Stationery Office, 1976. 22. M.J. Mertens, F.H. Engbers, A.G. Burm, J. Vuyk. Predictive performance of computer-controlled infusion of remifentanil during propofol/remifentanil anaesthesia. Br J Anaesth 2003; 90: 132–41.

Downloaded from https://www.cambridge.org/core. IP address: 94.158.22.118, on 02 Dec 2019 at 21:43:41, subject to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316659069.007

39

Chapter

6

Let’s Get Started The Know-Hows and No-Hows of Setting Up TIVA Michael G. Irwin and Gordon T. C. Wong

Introduction So you know the theory and you’re keen to use this technique. As with all modes of anaesthesia, careful preparation is mandatory. The arrival of versatile, easy-to-use, commercially available, target-controlled drug delivery systems has simplified TIVA making it as simple as using a vaporiser. Most have a choice of pharmacokinetic (PK) algorithms. The Marsh and Schnider models are the most commonly used for propofol and have various pros and cons. However, the important point about these models is that they can both make proportional changes in blood concentration allowing easy titration. New data is becoming available for more precise keo and PK algorithms that will improve accuracy – and, therefore, new models are likely to be developed. Remifentanil can be administered with TCI using the Minto model but, as the pharmacokinetics are relatively simple, can also be delivered as an ordinary infusion (µg.kg−1.min−1). The use of these techniques is discussed elsewhere in the book so here we will concentrate on how to physically set up your TIVA system. Infusion devices are used extensively in anaesthesia and intensive care and, unfortunately, historically there has been a high incidence of critical incidents related to their use. Most, however, are a result of human error and, consequently, a lot of effort has been directed to design better and more intuitive pumps, create and implement drug libraries, standardise techniques/drug dilution, and monitor pressures and infusion sites. TIVA is unique in that it allows accurate drug titration with separate control over sedation and analgesia. Awareness should be no more common than with inhalation anaesthesia with appropriate training and rational use, particularly with ‘smart’ pumps improving safety. An obvious advantage of inhalation anaesthesia is the ability to monitor vapour concentrations in the breathing

system but technology is also available to measure propofol in expired air and may well become commercially available in future clinical practice. In this chapter, we will first introduce you to the important nuances of different drug preparations and how to set up your drug infusions to minimise mishaps. Next, we will give you some advice on programming the TCI pumps and features to watch out for when considering the choice of pumps. When you are all set up, we will introduce the concept of titration with TCI, a concept that will be re-iterated time and again through the text by different authors – as once you have mastered this in your practice our job will be done.

Drug Preparation and Infusion Set-Up Propofol 0.5, 1 and 2% preparations are available on the market (similar to local anaesthetics, 1% = 10 mg.ml−1). 2% propofol was developed primarily for use in intensive care where the drug may be infused for long periods of time and a stronger concentration will reduce the amount of lipid drug vehicle being delivered. 0.5% was designed for paediatric use where smaller doses are being delivered and this may improve accuracy. However, 1% propofol is the mainstay concentration used in most operating theatres. Propofol emulsion is a highly opaque white fluid due to the scattering of light from the tiny (about 150 nm) oil droplets it contains; visually all these look the same so it is crucial not to mistake the concentration being delivered. It is prudent, therefore, to decide on one concentration in your workplace (usually 1%) and only stock that one. Only keeping one concentration in stock will reduce the risk of drug error. Lipid is potentially an excellent culture medium so, as with any IV drug, care should be taken in clean preparation and not opening ampoules until you need them. Preparations, however, all have some form of anti-

Downloaded from https://www.cambridge.org/core. University of Winnipeg, on 02 Dec 2019 at 21:44:27, subject to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316659069.008

40

Chapter 6: The Know-Hows and No-Hows of Setting Up TIVA

microbial agent. Diprivan®, the original version of propofol, contains EDTA, a chelation agent that is bacteriostatic; while newer generic formulations contain sodium metabisulphite or benzyl alcohol. Most generic formulations also use a mixture of mid- and long-chain triglycerides in the vehicle (MCT/LCT) to minimise the injection pain that was sometimes a problem with older formulations and, therefore, these are a good choice. None of these variations in drug vehicle or antimicrobial, however, make any material difference to the pharmacokinetics or pharmacodynamics of propofol. Opioids such as sufentanil, alfentanil and fentanyl come in standard universal concentrations. Remifentanil is available at 1 mg, 2 mg and 5 mg, in very similar looking vials as a dry powder for reconstitution. As with propofol, it is strongly advised to select a standard dilution for your workplace, e.g. 20, 25, 40 or 50 µg.ml−1. We use 50 µg.ml−1 with the only exception being paediatrics where the dilution may vary depending on the patient’s weight. Drugs are generally more dilute in paediatrics in order to ensure that the infusion rate is not too low that it takes a long time to overcome the dead space. For all drugs, colour-coded labels should be used with the concentration clearly marked. The syringes should not be labelled until after the drug has been added as there have been cases where the syringe was pre-labelled but the anaesthetist forgot to add the drug! Remifentanil has a very short duration of action and, as a controlled drug, is usually kept inside a locked cupboard, so remember to prepare a new syringe well in advance of the previous one finishing. Modern infusion pumps should have a pre-warning at least five minutes before the end of an infusion. To prevent cross-infection and to reduce any confusion associated with the presence of partially used drug vials, drugs should not be shared among different patients. Do not mix propofol and the opioid in the same syringe. The pharmacokinetics and pharmacodynamics of the two drugs are completely different.

Even if they are chemically compatible, there is no fixed ratio for delivery, and mixing the drugs will dilute the concentration. One of the advantages of TIVA is the ability to provide unconsciousness and analgesia separately. If you use a titrated induction, once the patient is unconscious the propofol concentration doesn’t need much adjustment during surgery, although your opioid concentration can be adjusted according to the extent of surgical stimulation (pain). Pre-mixing drugs makes this impossible. All infusion syringes and extension tubing should have luer lock (screw in) connections and be of a uniform brand that is compatible with the syringe pumps. As with drug dilutions, you can see that we like to ‘standardise’ things in our operation rooms to improve safety. Ensure all the connections are hand tight so they will not leak. On the other hand, do not over-tighten connections as this can occasionally cause the connectors to crack. A useful practice is to have two separate IV cannulas. A 20G cannula can be inserted before anaesthesia induction and used only for TIVA drugs (a dedicated line). At the end of surgery this line can be flushed and used for patient-controlled analgesia if required. In more major surgery, a larger bore cannula can be inserted after anaesthesia for IV fluid administration. It is also feasible to use one cannula for both drugs and fluids with a commercially available TIVA set or make your own using anti-reflux valves. The IV cannula should be sited, if possible, in a position where it can be easily secured and can be regularly inspected during surgery. Commercial TIVA administration sets (Figure 6.1) usually consist of two long infusion lines with integral anti-syphon/anti-reflux valves, and a side arm close to the patient through which IV fluid can be administered, which also has an anti-syphon and anti-reflux valve. All three arms end at one short infusion port and therefore only one cannula is needed. Because the lines and valves are integrated, the risk of disconnection is reduced.

TIVA dual administration set. Note the luer locks at each connection B

A C

The dead space at “A” should be as small as possible

B = Anti syphon valve

C = Non return valves

Figure 6.1 An example of a specifically designed TIVA administration set for two drugs. Downloaded from https://www.cambridge.org/core. University of Winnipeg, on 02 Dec 2019 at 21:44:27, subject to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316659069.008

41

Chapter 6: The Know-Hows and No-Hows of Setting Up TIVA

Table 6.1 Potential problems with drug delivery from IV anaesthesia pumps (adapted from [1]).

Problem

Prevention/Detection/Solution

IV cannula disconnect/out of vein

Venous access should be visible and accessible during procedure

Disconnection of infusion tubing from pump or cannula

Pump and tubing connections should be visible Use luer lock syringes

Pump power supply failure or pump paused

Ensure pump has an audible alarm

Occlusion of IV cannula or tubing

Pump high-pressure alarm

Occlusion alarm because of small cannula or long infusion tubing (e.g. PICC)

Ability to alter alarm threshold

Backtracking of propofol into IV fluid infusion tubing

Use of one-way valves

Drug disparity between settings and drug used (e.g. different concentration)

Keep only one concentration of propofol in hospital. Doublecheck drug dilution concentrations (or dispense from pharmacy premixed)

Wrong drug programmed into pump (remifentanil rather than propofol)

Prominent pump displays with the drug name Colour coding of the pump LCD displays and syringe labels Bar coding

Alternatively, if cost is an issue, self-made TIVA sets with three-way taps and freestanding valves can be used. However, due to the number of connections, the risk of disconnection is greater. If making up your own set, it is important to ensure that the three-way taps are compatible with propofol. Although very unusual, there are documented cases where propofol has damaged the plastic causing leakage from threeway taps after prolonged use. Use of these valves is very important. The antisyphon feature prevents drug in a syringe pump spontaneously moving under gravity if the syringe is not correctly seated in the pump. To reduce the risk of this, ideally syringe pumps should not be sited above the heart level. The anti-reflux feature will only allow forward movement of drug into the patient. For instance, if there is a blockage at the cannula level and no anti-reflux valves are being used, the drugs will flow back up the infusion lines with least resistance, or up the IV fluid line. If propofol and remifentanil are being used, propofol usually registers a higher line pressure as it tends to run at a higher rate and is more viscous, therefore propofol tends to flow back up the remifentanil line or the IV fluid line, and not trigger the high pressure alarm on the propofol pump. If this is not noticed by the anaesthetist, this will prevent drug delivery to the patient and may potentially cause anaesthesia that is too light. To reduce dead space, have the drugs infuse as close to the cannula as possible, or have a free-flowing IV fluid attached (with an anti-reflux valve of course!).

Use veins that are easily visible or accessible at all times if possible. Preferably, these should be forearm or hand veins. Avoid using antecubital fossa veins as ‘tissuing’ may not be easily detectable. If using cannulas/long lines that are already in situ, check that they are patent with a saline flush. Ensure all cannulas are well secured, loose infusion lines are secured to the patient and any connection points are tight. If the cannula site is not always in view, make sure they are secured in a way that enables regular checks to be made easily. Always have a standard set-up of syringe pumps on the drip stand, i.e. position your various syringe pumps in the same relative position. This reduces the risk of putting the propofol syringe inside the pump pre-programmed for remifentanil and vice versa. Ideally, the pumps should not be programmed until the appropriate syringe is actually inside. It may be prudent to label your TCI pumps for propofol and remifentanil for the same reason. Always programme them standing in front, not behind or above the pumps, so that you can see the numbers clearly. Table 6.1 summarises some potential problems that can occur with TIVA set-up and suggestions on how to avoid them.

Target-Controlled Infusion Pump Programming and Features Schnider or Marsh model? Plasma or effect-site target? Pharmacokinetic models are evolving but, in a way, it does not really matter which one you use as

Downloaded from https://www.cambridge.org/core. University of Winnipeg, on 02 Dec 2019 at 21:44:27, subject to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316659069.008

42

Chapter 6: The Know-Hows and No-Hows of Setting Up TIVA

long as you are familiar with and understand the limitations of your chosen model. There are pros and cons of each, although there are good reasons why the Schnider model in the plasma-targeting mode is not popular and can result in a very slow induction of anaesthesia. Likewise, the Schnider model over-estimates the effect-site concentration during induction. Ensure the patient demographics are correct. All patients should be weighed before surgery, of course. Sex and height are also required for the Schnider model to calculate the LBM – this is called allometric scaling. The Schnider model uses the James formula to calculate the LBM and, at high BMI, this can paradoxically result in an erroneously reduced LBM. It is important to know how your TCI machine corrects for this. Some machines will only allow a total body weight (TBW) to be used for a particular height which does not let the LBM fall on the declining part of the TBW/LBW curve; others will use the maximum LBM for any particular weight/height combination. Check the units being used. If a manual infusion of remifentanil is being used, check it is running at µg.kg−1.min−1 as some machines allow you to run at mg.kg−1.min−1 or even mg.kg−1.h−1 (if you have the resources, it is a good idea to label a syringe pump for a certain drug with the units fixed appropriately, e.g. dexmedetomidine at µg.kg−1.h−1). Does the dose in ml.h−1 look appropriate for your patient? The same goes for TCI propofol. Check, after programming the machine with your chosen model, that the calculated dose is an appropriate one for your patient. Check the initial bolus dose given and subsequent infusion rate in ml.h−1 and mg.kg−1.h−1. Similarly, do they look ‘appropriate’? If not, there could be a programming or data input error. It is handy to appreciate how your TCI machine pump makes its calculations. For instance, different brands of TCI pumps may both be pre-set with the Marsh model but may use completely different keo. Some Schnider models use fixed keo, whereas some use a fixed time to peak effect (TTPE) and a varying keo. As stated before, Schnider uses the James formula to calculate LBM and this can cause problems at high BMIs. Different machines prevent the calculation of a negative LBM by different methods. Understanding how it does this allows the user to circumvent this limitation. Always ensure the pumps are fully charged and plugged in, even when not in use. At high infusion rates, the battery can run out quickly, particularly

with more frequent use. TCI pumps will have a battery indicator stating how long the battery will last for if it continues to run at the present infusion rate. With most TCI pumps, if the power is lost and it switches off during infusion, it is not possible to continue on the same settings as all the preceding data will be lost. If such an event ever happens, an easy way to get around this is to, when you restart the TCI, disconnect the infusion to prevent the initial ‘loading dose’ from reaching the patient. Then reconnect for maintenance. It’s not entirely accurate but is a reasonable solution. Intermittently checking the infusion rate in ml.h−1 will enable you to see if the dose is similar in ml.h−1 in the event of a restart after pump failure. The machine allows the pressure in the system to be measured in mmHg. Most pumps operate at around 150 to 250 mmHg and alarm at 500 mmHg except when giving a bolus dose (some allow you to alter these settings). High pressure may indicate an obstruction to flow in the system, for example, from a misplaced cannula, accidentally closed three-way tap or kinked lines. Low pressure may indicate a disconnection or leak. Do not just silence the alarm without performing a check. Beware also of silencing the alarm during syringe changes as, if you fail to restart the pump correctly, it may not re-alarm to alert you. Having a syringe pump with a drug library with pre-set standard drug concentrations will reduce the risk of programming errors. As previously stated, it is best to keep all concentrations standardised. All machines have a visual indicator such as a moving icon or coloured lights to state that the machine is infusing. It is particularly important to check the visual indicator to see that the machine is infusing after relieving an obstruction, syringe change or dose adjustment. The secret to providing optimal anaesthesia is to titrate, titrate and you’ve guessed it, titrate. Titration is necessary to accurately gauge the patient’s response to propofol as individual patient requirements can vary quite widely. It is important to be patient during anaesthesia induction and make small changes in target concentration, especially in the elderly and the frail, in whom it can be easy to ‘overshoot’. Drug onset can be much slower than you think. Ensure IV access is reliable, visible at all times or easily accessible to be viewed. Keep the dead space short so rapid changes in drug dosages can be achieved quickly if necessary.

Downloaded from https://www.cambridge.org/core. University of Winnipeg, on 02 Dec 2019 at 21:44:27, subject to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316659069.008

43

Chapter 6: The Know-Hows and No-Hows of Setting Up TIVA

Keep the maintenance dose as close to the ‘sleep’ (where loss of consciousness occurs) dose as possible. Studies have shown that the ‘wake up’ dose is very similar to the ‘sleep’ dose. Excessive depth of anaesthesia may be detrimental to cerebral function. Although propofol is less so than inhalational anaesthetics, it is preferable to keep the maintenance dose of propofol relatively steady and vary the analgesic depending on the level of stimulation and stage of surgery. Remember that remifentanil and other opioids have little hypnotic action, and propofol is not an analgesic. Also remifentanil has a faster onset than propofol so it is possible that the patient will stop breathing before loss of consciousness if the dose is high during induction. Typically, we will start remifentanil at 2 to 3 ng.ml−1 effect-site concentration or 0.1 to 0.15 µg.kg−1.min−1 until the patient loses consciousness. If you are using a supraglottic airway device, then this dose is adequate for insertion. However, if you wish to intubate the trachea, the dose can be increased to 5 to 10 ng.ml−1 (depends on the age and physical status of the patient) or administer a bolus of 0.5 to 1.5 µg.kg−1. Following intubation, the dose of remifentanil can be decreased to the previous low maintenance dose and increased again for the start of surgery. During surgery, most patients will require 3 to 8 ng.ml−1 or 0.15 to 0.35 µg.kg−1.min−1. Obviously there will occasionally be patients who require more or less but that is unusual. The dose can be titrated against the sympathetic response to surgical stimulation, i.e. to avoid tachycardia and hypertension. In contrast, once unconscious, propofol doses should not need much adjustment during surgery. We recommend using positive pressure ventilation for all patients undergoing surgery unless a very short case. This is because opioids will always cause at least some degree of respiratory depression, even if there is still some spontaneous respiration, and lead to respiratory acidosis. Although the technology exists (mass spectrometry in expired air), at present there is no commercial system that measures propofol concentrations in real time, unlike with inhalational agents. The NIAA Health Services Research Centre National Audit Project (NAP5) found a higher risk of awareness in patients who underwent TIVA with a neuromuscular blocker (NMB), although the actual risk of awareness was still very low. In fact, the risk of awareness with any anaesthesia technique is increased with the concurrent use of an NMB as a paralysed patient is unable

to move and indicate consciousness with pain. It is logical, therefore, to avoid NMBs generally unless absolutely necessary and consider the use of a processed EEG monitor in patients with additional risk factors for awareness, or those in whom the IV cannula is not easily visible. An intubating dose of NMB is usually sufficient for many surgeries. Most TIVA anaesthetics using propofol and remifentanil +/– a working epidural usually provide adequate abdominal muscle relaxation without the use of NMBs. Even if a NMB is being used, the effects should be closely monitored and it is not always essential that the patient is fully paralysed. Studies have shown that the risk of awareness is very low if NMBs are not used. When gaining experience with TIVA, start with more simple cases where NMBs are not required. Safety may be enhanced by the use of processed EEG (pEEG) monitors such as BIS™, Conox™ and Sedline™. However, there are limitations to their use. These include: cost of the monitors and sensors; problems with data acquisition (poor sensor contact, not possible with all types of surgery, e.g. if the forehead is required for surgery); interference from EMG and diathermy; EEG changes not being the same with all drugs, e.g. use of pharmacological adjuncts such as ketamine; skill required for interpreting the EEG trace; the dimensionless number system reflecting anaesthetic depth; and interindividual variability. We find it useful to get familiar with the raw EEG trace at different levels of anaesthesia. If anything unexpected occurs, for example the patient goes to sleep much slower than expected, have a systematic system for trouble shooting. Common reasons include a mis-programmed pump, leaking drugs, back-flowing drugs up the IV infusion or the propofol syringe misplaced inside the pump, which has been programmed for remifentanil (yes. . . anything can happen). Regularly check your IV site for blockage, tissuing, etc., if possible. As you would with inhalational anaesthesia, give adequate doses of long-acting analgesic and use adjuvants such as ketamine and dexmedetomidine if applicable. The reason for using a longacting opioid (e.g. morphine, pethidine) is that it will be additive to your short-acting opioid infusion, thereby reducing the dose, reducing the small risk of tolerance and hyperalgesia, and providing a platform for post-operative analgesia once the short-acting drug effect dissipates. This is very important as

Downloaded from https://www.cambridge.org/core. University of Winnipeg, on 02 Dec 2019 at 21:44:27, subject to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316659069.008

44

Chapter 6: The Know-Hows and No-Hows of Setting Up TIVA

otherwise, after more extensive surgery, pain may develop very quickly. Towards the end of surgery, especially with longer procedures, propofol will wear off more slowly as the context sensitive half-time increases with duration of administration. You can start to titrate down the level of propofol during wound closure. If using remifentanil, keep the level reasonably high to maintain analgesia (remifentanil half-time is context insensitive and the effects will dissipate in around 10 to15 minutes, depending on the dose, irrespective of the duration of infusion). Other opioids such as alfentanil and sufentanil will also need to be titrated down. If the patient shows signs of awakening (clinically or with pEEG), then it is easy to increase the dose of propofol slightly and wait for surgery to almost finish before terminating the infusion. The aim is to have the patient recover consciousness while the remifentanil is still wearing off. The ventilation rate can be turned down to allow end-tidal CO2 to increase and, thereby, stimulate respiration. Once the patient responds to commands they can be asked to breathe until spontaneous respiration resumes. At this point, while still tolerating the endotracheal tube or starting to show signs of finding it uncomfortable, tracheal extubation can be performed smoothly after suctioning the oropharynx.

Summary You would not expect the pilot of your plane to take off without ensuring it is mechanically sound and he/she is prepared for any untoward event. Similarly, your patients expect the same from you. TIVA is the only anaesthesia technique where induction is titrated and then maintained with the same drugs. TIVA is, to some extent, more complicated than inhalation anaesthesia but has myriad advantages. Start with simple cases and follow our recommendations for safe administration while you

gain experience. You will soon become very comfortable with titration and have the pleasure of seeing your patients safely through surgery while waking up smoothly and comfortably at the end.

References This chapter is a fairly brief but, hopefully, readable overview of the important concepts and philosophies behind TIVA. Some points are addressed in more detail in other chapters. The list below includes articles where you will find more complete information. 1. A.F. Nimmo, A.R. Absalom, O. Bagshaw, et al. Guidelines for the safe practice of total intravenous anaesthesia (TIVA): Joint Guidelines from the Association of Anaesthetists and the Society for Intravenous Anaesthesia. Anaesthesia 2018. 2. A.F. Nimmo, T.M. Cook. Accidental awareness during general anaesthesia in the United Kingdom and Ireland. In: Pandit, J.J. and Cook, T.M., eds. NAP: Royal College of Anaesthetists and the Association of Anaesthetists of Great Britain and Ireland. 2014: 151–8. 3. M.G. Irwin, G.T.C. Wong. Taking on TIVA. Why we need guidelines on total intravenous anaesthesia. Anaesthesia 2018. 4. F.H.M. Engbers, A. Dahan. Anomalies in target-controlled infusion: an analysis after 20 years of clinical use. Anaesthesia 2018; 73: 619–30. 5. J.B. Glen, F.H. Engbers. The influence of target concentration, equilibration rate constant (ke0) and pharmacokinetic model on the initial propofol dose delivered in effect-site target-controlled infusion. Anaesthesia 2016; 71: 306–14. 6. L.I. Cortinez. What is the ke0 and what does it tell me about propofol? Anaesthesia 2014; 69: 399–402. 7. H.B. Scott, S.W. Choi, G.T. Wong, M.G. Irwin. The effect of remifentanil on propofol requirements to achieve loss of response to command vs. loss of response to pain. Anaesthesia 2017; 72: 479–87.

Downloaded from https://www.cambridge.org/core. University of Winnipeg, on 02 Dec 2019 at 21:44:27, subject to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316659069.008

45

Chapter

7

Let’s Get Pumped! The Nitty Gritty of TIVA Syringe Pumps Nigel J. Huggins

Introduction

A Look Underneath the Bonnet

This chapter will set out to describe the mechanical, electronic and programmed mechanisms of the pumps used for TIVA, from the first early prototypes to current build. Mention will be made of algorithms and of different patient populations in relation to these topics and to aspects of the safe delivery of TIVA, again in relation to the pumps themselves rather than the technique as a whole.

All syringe pumps, whatever their purpose, work via a stepper motor that drives a plunger down the barrel of the syringe at so many millimetres (mm) per hour. However, the ‘swept volume’ (the volume of drug given) of the plunger down the barrel at, for example, 10 mm per hour will depend upon the bore (internal diameter) of the syringe. This means that a 10 mm propulsion down a syringe of 20 mm bore will result in a smaller volume being administered than that from a 10 mm propulsion down a syringe of 30 mm bore. The stepper motor works by driving the syringe plunger in millimetres per hour, which is translated into millilitres (ml) per hour, by telling the syringe pump the make and size of the syringe from which the ‘syringe library’ will identify the bore and stroke volumes of the syringe loaded into the pump. The stepper motor will go faster or slower according to the make and size of the syringe, to deliver the calculated volume. It should be noted that a 50 ml syringe has a unique bore according to each manufacturer and so the pump needs to know what make of syringe is loaded and not just the volumetric size of the syringe. Consequently, basic syringe pumps could take things a step further by administering a drug at a constant infusion rate in ml.h−1 by consideration of the information referred to in the previous paragraph. However, infusing a drug at a constant rate will almost always lead to accumulation of a drug the longer the duration of the infusion. Remifentanil with its rapid, organ-independent metabolism is the notable exception. Clearly it is not feasible to manually change the infusion rate to compensate for the different compartments (in a three-compartment PK model) filling up, and elimination of the drug occurring. An accurate algorithm is required to describe the pharmacokinetics of a drug when it is injected intravenously. A three-compartment model best explains

A Snippet of Life Before Programmable Syringe Pumps The administration of anaesthetic drugs intravenously to maintain surgical anaesthesia is not a new technique. Prior to the development of algorithms (mathematical interpretations) to establish a steady plasma and, then later, a steady effect-site target concentration, anaesthetists would manually inject drugs such as phenoperidine and droperidol in what was referred to as neurolept anaesthesia – more accurately neurolept analgesia. This could be described more as an action of ‘stunning’ a patient rather than a technique of balanced anaesthesia. Consequently, recovery was somewhat hit and miss as drugs were titrated against effect and required depth of anaesthesia, rather than in a calculated manner with due consideration of the pharmacokinetics of the drugs administered. There can be no scientific mapping of synergism using drugs in this way and there are greater risks of patients being too deeply anaesthetised or in too light a plane of anaesthesia. The introduction of sodium thiopental in the 1930s led to its use for IV anaesthesia, although at this time less than 10% of the general anaesthetics administered were with IV barbiturates mainly because of the prolonged recovery and difficulties with administration (no syringe pumps). So much for simple manual titration via a collection of syringes containing drugs.

Downloaded from https://www.cambridge.org/core. University of Winnipeg, on 02 Dec 2019 at 21:47:17, subject to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316659069.009

46

Chapter 7: The Nitty Gritty of TIVA Syringe Pumps

the findings consisting of the central compartment, the vessel-rich compartment and the vessel-poor compartment referred to as V1, V2 and V3, respectively (see Chapter 5). In addition, there is an infinitesimally small effect-site volume Ve (the anatomical area where the drug works or has its effect). The concepts of the influx and outflow rate constants are described elsewhere along with other factors that determine flow in and out of the compartments. The mathematics will be unique to each drug and it became clear that four patient parameters (age, height, weight and sex) may be variously used to fine tune the pharmacokinetics of a drug in any given patient. These parameters are entered into a syringe pump’s computer and used by the drug algorithms to tell the pump stepper motor what speed to run at.

Pumping Out the Pumps Ultimately, the pumps we have today (the open TCI anaesthesia pumps) were developed. However, when the syringe pumps were initially used they had to overcome a fundamental design problem. It was the Ohmeda 9000 syringe pump[1] that was the first commercially available pump that could deliver a bolus at 1200 ml.h−1 that was necessary for volumetric loading of the drug into the central compartment. Without this facility the early algorithms for targeting propofol to plasma levels would not be possible. An initial bolus (or a temporarily faster infusion rate) of the drug was a fundamental requirement for the PK models (algorithms) to work in practice. A bolus is also needed to raise a required drug level to a new chosen setting – depending on whether the pump is set to effect-site targeting or plasma targeting, respectively. There were also other design issues. Early pumps had unrestrained stepper motors such that, if the delivery pump was mounted significantly above the cannula that was being used to deliver the drug to the patient, the syringe could syphon its entire contents via gravity into the patient as a slow but persistent bolus! Clearly this had to be addressed for reasons of patient safety and avoidance of overdosing. Anti-syphon valves are also fitted to commercial TIVA administration sets to ensure there is a complete safeguard against such an event. The first compact, transportable, programmable pump apparatus was based on an Ohmeda 9000 syringe pump (1990), with a first edition Psion Organiser ‘grafted’ on to the back bar.[2] The Psion contained an algorithm for propofol (launched in 1986) and the

pump could be programmed for Becton Dickinson (BD) Plastipak, Terumo or Monoject syringes and stated an infusion accuracy of its ‘stepper motor’ of +/–3% of the value stated (which was calculated from the plasma target chosen). It had a multi-modal alarm system for both high and low driving pressures, for ‘near empty’ syringe, low battery and for incompatible syringes. It was a significant advance in its time, utilising the Marsh algorithm and targeting to plasma level rather than effect site, which would become an option later with newer algorithms. The trials that were carried out were successful and so it was decided to develop a computer system within the back bar of the Ohmeda 9000 pump. The Psion system was both physically clumsy and with its ‘OPL’ (Organiser Programming Language and later Open Programming Language) it did not allow access to communication variables required for a more eloquent and patient-safe apparatus. The new pump contained a twin microprocessor system for quality control. The first processor ran the pump stepper motor and the second processor checked the accuracy of the first processor’s calculations and output. The main 16-bit processor therefore carried out a full drug-dosing calculation for a threecompartment model, while the second 8-bit processor was programmed with a mathematical solution to the algorithm based on the responses of the pump stepper motor. In other words, it was monitoring the speed of the stepper motor to check that the speed at which it was driving the syringe plunger would deliver the dose of drug that the first microprocessor said was required. For safety reasons the two microprocessors were made by different manufacturers and it was found that they correlated to within 0.5% accuracy.[3] Hence the ‘Diprifusor™’ module was born and it was subsequently built into the Graseby 3500 TCI pump, the Fresenius Vial Master TCI pump and the Alaris TCI pump. Each of these pumps required the Diprifusor module to respond to different programming characteristics and so the Diprifusor module (patented by Zeneca, a pharmaceutical company that subsequently became AstraZeneca) became a ‘multilingual’ mini-computer working three different pumps to achieve the same end point by monitoring its own performance. The final piece of the jigsaw was to capture the usage of propofol supplied by Zeneca and this was done by standardising a syringe, prefilled with either 1% or 2% propofol with an electronic tag that would communicate with the pumps to tell them

Downloaded from https://www.cambridge.org/core. University of Winnipeg, on 02 Dec 2019 at 21:47:17, subject to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316659069.009

47

Chapter 7: The Nitty Gritty of TIVA Syringe Pumps

which percentage propofol was being used. The familiar coloured tags on these prefilled syringes were, in fact, coils wound to transmit back differing frequencies to a code transmitter/reader in the pump, which, when the ‘near empty’ alarm was activated, would then transmit a signal to destroy the coil so that the syringe could not be re-used. As years went by, further interest was shown in developing algorithms that were more complex and that could define effect-site concentrations of propofol, and the soon to be launched remifentanil (1996). By the time this new drug was launched, the electronics as well as the PK profiles were ready to be programmed into what were to be called ‘open TCI’ systems for delivering TIVA using these two drugs (obviously in separate syringe pumps as their pharmacokinetics are very different), with each pump programmed with a choice of algorithms for propofol (Marsh or Schnider) and for remifentanil (Minto). How the algorithms were derived is addressed elsewhere, but the ‘new’ propofol algorithm developed by Schnider was designed to target the brain effect-site while Marsh remained as an algorithm for plasma targeting. Suffice to say here that Marsh may be used for either plasma or effect-site targeting, while Schnider can only be used for its intended purpose of effect-site targeting in practice. Minto may be used in either effect-site or plasma-targeting modes, but is best delivered by effect-site targeting, which will allow more rapid equilibration should the dose of remifentanil (analgesia) need to be increased according to surgical stimulation. There are some quite striking differences between the two algorithms for propofol in terms of how they perceive the pharmacokinetics, but as the mathematics is so complex the differences balance out on either side of the equations resulting in a very similar volume of drug being administered by each model. It can be thought of as comparing it to the answer to a sum being 18, and so the sum could be either 2 × 9 or 3 × 6, both arriving at the answer 18. The arrival of the open TCI facility brought with it the question of an appropriate equilibration constant (keo) for propofol, which would describe the rapidity of equilibration of the drug with its effect-site receptors or, strictly speaking, the rate of diffusion away from the effect site when a steady-state infusion is discontinued. Assumptions and calculations were based on simple observations of time to peak effect after

administration of a bolus as measured by EEG coupled with serial blood drug concentration measurements. This then gave an overall picture of the pharmacokinetics and pharmacodynamics of the drug. Subsequently, further questions have been raised, for example what is the most appropriate way to determine keo? The issue could be endless, but pump manufacturers want to know a suitable figure so that they can programme their pumps in the most appropriate manner. Currently a value of 1.2 is deemed to be the best match in the greatest number of clinical scenarios.[4] Armed with the knowledge of the patient’s age, height, weight and sex, the pumps will determine how much drug to give for the chosen target. It is worth mentioning at this point that the Marsh algorithm in the original Diprifusor, despite requesting age and weight, is only using weight to determine its drug delivery to plasma targeting. The age is only included to prevent use in under 16s whose handling of the drug in the Marsh algorithm is considered sufficiently different to prevent accurate usage. Clearly arguments will continue about those aged 15 years 11 months etc.! It may well be surprising to many that age is not a consideration with respect to the older population in the Marsh algorithm. Hence clinical status needs to be considered when using this algorithm in the elderly – as one would, if giving a manual bolus of propofol followed by an inhalational anaesthetic agent. The elderly, as we know, are sensitive to all drugs and we all know that the MACs of inhalational drugs are much lower in elderly patients, as is the induction dose of propofol. This may then influence, and indeed determine, the configuration of the pumps for anaesthesia in different patient populations – the sick, the elderly, the fit and those with chronic disease medication. This is why we emphasise the importance of titration throughout this book. As clinicians we have much more of an input than simply loading syringes, priming and programming pumps, and replacing syringes in the pumps. We need to have an idea of what the pumps do with the parameters we input to them and how we might need to modify ‘normal’ settings in extremes of age and health. Chapters 14 and 15, on extremes of age, explain these in detail. An intimate understanding of the mathematics and complex pharmacokinetics is not necessary but a conceptual knowledge of which particular parameters are considered by each algorithm is useful. Figure 7.1 summarises which algorithm takes into account which parameter.

Downloaded from https://www.cambridge.org/core. University of Winnipeg, on 02 Dec 2019 at 21:47:17, subject to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316659069.009

48

Chapter 7: The Nitty Gritty of TIVA Syringe Pumps

Who has done what with what? AGE

HEIGHT

LBM

MINTO MARSH SCHNIDER

WEIGHT SEX

X

X

X LBM

Figure 7.1 Figure showing which algorithm takes into account which parameter. LBM = lean body mass.

There are pumps available for paediatric populations – the Paedfusor and those with the Kataria algorithm as examples, but for obesity, educated changes have to be made to overcome the shortcomings of the extremes of weight. Different pump manufacturers address the issue in different ways, e.g. preventing the targeting of Schnider and Minto if the LBM (upon which their administration is dependent) starts to decrease due to obesity, or, by limiting the body weight entered for height and sex to a ‘maximal normal’ weight for those parameters. Neither solution is ideal, as one will affect steadystate drug levels the longer the time the infusion runs for and the other will encourage the seeking of ways around the lock out! These are discussed in more detail in other chapters.

Some of Us Are Not Like the Others Not all pumps are of the basic stepper motor design described above. There are some that run on a ‘volumetric roller pump’ design such as those infusion pumps used in critical care for accurately controlling the volume of fluid infused per hour. The same design of pump is adapted to deliver a set amount of drug by altering the speed of the motor and hence the speed of the rotating/compressing sequential rollers within the pump that ‘drive’ the fluid intravenously into the patient. Clearly the algorithm of the drug must be within a drug library in the pump and the same patient parameters must be entered into the computer module integral to the pump. These pumps need to have a dedicated giving set used with them, which has a known internal volume (cross-section/length ratio) so that a roller displacement equals a set volume of drug displaced forwards and, hence, infused into the patient. A choice of giving sets is not a cost-effective option for third-party manufacturers as the tubing is highly malleable to allow the rollers to propel the fluid by sequential compression of the giving set tubing and

there is a high chance of inter-manufacturer variation. Hence these delivery sets are significantly more expensive than the dedicated delivery sets used for the TIVA syringe pumps, which are of low compliance and this is much easier to achieve as a standard from different manufacturers. These roller volumetric pumps are ideal for establishing reasonably constant levels of sedation in patients in intensive care where nursing familiarity with equipment is important and bottles of, for example, 100 ml of generic propofol may be simply hung from a drip stand, rather than having the need to draw up repeated doses into syringes for delivery by syringe pump. Such a method of administration may allow for accurately guided sedation for patients in ICU but no trials of this have yet been evaluated or published. Of historical interest, it was actually an Imed 929 infusion pump of this roller type that was coupled with an Atari 1040ST ‘home’ computer via an RS232 interface that transmitted the first control programme written in GFA BASIC. This first-phase study included 33 patients who were successfully anaesthetised using the PK programme without untoward incident.[5] The second phase was to use the Atari with an Ohmeda 9000 pump and next came this pump with the attached Psion Organiser as described previously. Whichever type of pump is used to deliver the drugs, there are pre-programmed limits that can be set. These include such things as the maximum pump bolus infusion rate, the maximum peak plasma or effect-site level allowable at any time during bolus and the maximum infusion pressure generated by the pump. These are not necessarily intended to be altered at the individual patient programming phase and are independent of the patient’s characteristics such as age, height, weight and sex. In recent years there has been a plethora of open TCI pumps being launched into the market. Some look more robust than others and prices vary according to the costs of manufacture – including testing, safety, accuracy and materials used. It would not be appropriate to catalogue individual manufacturers’ products as prospective purchasers would always trial several of the examples that are available before making a purchase.

Safety Does Not Happen by Accident A mention of safety aspects relating to pumps for TIVA delivery is appropriate at this point. It should go without saying that the clinician should be familiar with the pumps being used. Programming is different

Downloaded from https://www.cambridge.org/core. University of Winnipeg, on 02 Dec 2019 at 21:47:17, subject to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316659069.009

49

Chapter 7: The Nitty Gritty of TIVA Syringe Pumps

for each pump type and, although quite intuitive, operators need to make sure that they also understand how to interpret and cancel alarms and how to restart the pumps after changing syringes. Does the loaded syringe match the manufacturer type that the pump is expecting? Some systems will alarm if the pump detects a different syringe diameter to that expected for the programmed syringe type. Does the anaesthesia assistant know the pumps too? Are the batteries good and have sufficient charge? Having the whole stack go dead when unplugged from the mains during a case, or when transferring a patient, is not recoverable! One needs to know what algorithms are available in different pumps and when to use them clinically. Having a departmental policy/protocol for the presets on the pumps (e.g. drug concentration, which drug is in the upper and which in the lower syringe pump, which algorithm for propofol and what target site for each drug) minimises errors on these fronts. All of these are programmable when setting up with the patient data but it is arguably safest to have settings that ‘work’ pre-set in each pump, which minimises the risk of programming errors. It is also best not to label the remifentanil syringe until the drug has actually been added. In terms of drug concentrations, try to use a concentration that is not going to work the stepper motor at either an extremely fast or extremely slow rate; as such, it is common to use 2% propofol and 50 µg.ml−1 of remifentanil in adult patients so that the ml per hour rate of each syringe is not at one extreme or another, outside of bolus dosing mode (1200–1500 ml.h−1). There are some ‘abuses’ of the chosen settings, which you should be wary of making. The classic example is as follows: consider an ASA 3 or lower patient, in other words a sick patient. If one were to administer 1% propofol by hand for induction from a 20 ml syringe, it would be given slowly and titrated against effect. This is done to avoid a strong PD ‘hit’ by rapid bolus of an unnecessarily large dose of the drug, creating a high peak plasma level and resulting in a significant fall in blood pressure. We similarly want to avoid such a phenomenon when administering TIVA. So if we set the targets to plasma targeting, the plasma level will never rise above this level. This set level will be the peak plasma level and so the PD ‘hit’ will be minimised (TCI can be very ‘gentle’ when titrated carefully like this). Consequently, as the concentration gradients between the central compartment, the peripheral compartments and the effect site are

smaller, the induction time will be longer. However, having thought this out, there have been episodes when anaesthetists experiencing that the induction will be ‘too slow’ have bumped the plasma target up to a higher level to counteract the safety of the slow, nonpharmacodynamic hit induction. It is very predictable that, now the plasma target that is reached is excessive for that patient, there will absolutely be a PD hit as the peak plasma level will now be similar to that achieved by setting the pumps to effect-site targeting, and the patient ends up getting the equivalent of a large bolus, which is what is being avoided by choosing plasma targeting in the first place. Needless to say, the avoidable PD hit ensues along with the consequent hypotension in a sick patient (i.e. don’t do this!). Effect-site targeting can be used but you must start with a low target (0.5–1 µg.ml−1), be patient and increase slowly in small increments. Pressure alarms are such that they will alarm for an occlusion as per ordinary syringe pumps, such as when a cannula is misplaced. It is important therefore not to set the high-pressure alarm to its maximum limit of about 1100 mmHg as significant extravasation of drug could occur (especially in children) before an occlusive alarm triggers. This could, theoretically, lead to awareness. An alarm state will also be triggered when a sudden drop in driving pressure is detected, which could indicate a disconnection. Due to the lubricating oiliness of propofol, however, a low pressure alarm may be triggered if the pump is running quite slowly (ml.h−1 equivalent); the syringe may be slightly sticky when first loaded and so the plunger will move jerkily down the bore, temporarily sticking and then ‘giving’ as the pressure builds up. While this will not alter the amount of propofol given overall, the sudden ‘give’ will indicate a driving pressure drop to the pump and a low-pressure alarm will be triggered. This should immediately be checked for an indication of leakage of propofol from the intended intravenous pathway.

Making the Connection: Considerations for TIVA Delivery Sets In terms of TIVA administration sets, several manufacturers make dedicated lines. These should be used where possible. They have luer lock ends for the luer lock syringes: non-return valves to prevent retrograde flow of fluids up the administration sets towards the syringes, hence inhibiting drug delivery – which would cause a high-pressure alarm. They also have

Downloaded from https://www.cambridge.org/core. University of Winnipeg, on 02 Dec 2019 at 21:47:17, subject to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316659069.009

50

Chapter 7: The Nitty Gritty of TIVA Syringe Pumps

anti-syphon valves at the cannula end luer lock connection to prevent inadvertent syphoning and administration of the drug from the syringe. There are some sets available with a side arm port for IV fluid administration. If you use these then ensure that the fluid flow is close to the cannula to minimise the chance of flushing anaesthetic drugs from the drug infusion lines. If the set does not have such a design, then the best practice is to have a dedicated TIVA line and a separate IV infusion line (see Chapter 6).

Loop de Loop: Taking the ‘Art’ Out of TIVA Dosing Attempts have been made to ‘close the loop’ by using depth of anaesthesia monitoring (brain stem auditory evoked potentials or processed EEG) to adjust the pump settings to maintain relatively steady state anaesthesia based on feedback from these measures. Although these systems have been produced commercially, they have very important limitations. There is great variability in the reproducibility of readings of all commercially available depth of anaesthesia monitors and this variability and lack of stability is not just due to the technical difficulty in picking up and interpreting cerebral microvoltages from cutaneous electrodes. In addition, the monitors only measure indices of consciousness, thereby feeding back information to adjust propofol delivery. However, adjustment of the analgesic component of TIVA (e.g. remifentanil) is equally important for stable anaesthesia. As a result, such technology has not been able to be proven to be fail safe. A feedback system that connects an open TCI pump to a bispectral index (BIS™) monitor and, by feedback, adjusts the target concentration of the propofol flies in the face of a propofol/remifentanil anaesthetic where the whole point is to give a low and

constant level of hypnotic and a higher and variable level of mu agonist – i.e. remifentanil (since the level of consciousness shouldn’t change much during surgery whereas the degree of pain stimulation may vary markedly). This, coupled with the variance of processed EEG (e.g. BIS) as a reliable depth of anaesthesia monitor, has prevented widespread acceptance.

Conclusion The chapter has covered historical to present-day development of TCI TIVA pumps and covered aspects of safety in their use along with programming options when faced with different populations. More details of appropriate and inappropriate algorithm choice and how to manage extreme patient physiology are covered in dedicated chapters.

References 1. D.N. Stokes, J.E. Peacock, R. Lewis, P. Hutton. The Ohmeda 9000 syringe pump: the first of a new generation of syringe drivers. Anaesthesia 1990; 45: 1062–6. 2. G.N.C. Kenny, M. White. A portable computerised infusion system for propofod. Anaesthesia 1990; 45: 692–3. 3. J.M. Gray, G.N.C. Kenny. Development of the technology for ‘Diprifusor’ TCI systems. Anaesthesia 1998; 53: 22–7. 4. M.M. Struys, T. De Smet, B. Depoorter, et al. Comparison of plasma compartment versus two methods for effect compartment–controlled targetcontrolled infusion for propofol. Anesthesiology 2000; 92: 399–406. 5. M. White, G.N.C. Kenny. Intravenous propofol anaesthesia using a computerised infusion system. Anaesthesia 1990; 45: 204–9.

Downloaded from https://www.cambridge.org/core. University of Winnipeg, on 02 Dec 2019 at 21:47:17, subject to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316659069.009

51

Chapter

8

‘But I’m Used to MAC!’ How Do I Get the Dose Right with TIVA? Nicholas Peter Sutcliffe and Michael G. Irwin

Don’t Panic, You Are Not Alone! The recent NAP5 report in the UK showed that many of the cases of awareness associated with TIVA were due to poor technique and inadequate dosing.[1] In short, we don’t teach TIVA as well as we do inhalational anaesthesia. Yes, we have those occasional lectures and workshops built into our curriculum for trainees, covering pharmacokinetics/pharmacodynamics (PK/PD) and TIVA. However, trainees learn from watching their seniors from day to day in the operating room (OR). The vast majority of this training is gas based; it’s easy to use, everything is set up, the vaporiser is full and you’re ready to go as soon as you reach the OR. Why bother going to all the work of setting up for TIVA: surely that’s just for specialised cases? Thus our trainees are rarely exposed to TIVA and, in turn, become the gaseous trainers of the future. No wonder that when called upon to use TIVA in those circumstances where there is no alternative, even experienced anaesthetists can have difficulty. It’s often our trainees who are called upon to provide sedation for patients transferring between locations, or provide anaesthesia and sedation in remote sites without the infrastructure to support inhalational anaesthesia. These same trainees have often not learned TIVA in a practical way and are then struggling to remember, was it mg.kg−1.h−1, or µg.kg−1.min −1 ? How can we transfer the knowledge and skills that we have learned, over many years of passing the gas, to the art of TIVA? Let me ask the reader this question: if faced with a new inhalational agent, what one piece of information would you require in order to make a reasonable job of dosing this new agent? Most anaesthetists would ask ‘what’s the MAC?’ This one parameter will allow you to select a reasonable number on the dial, making an allowance for the adjuvant agents you may be administering and then titrate according to the patient’s response.

My Epiphany: One Anaesthetist’s Journey to Professional Bliss My initial anaesthetic training was in Glasgow: an area many would link with expertise in TIVA. However, the vast majority of my training was the standard big syringe for hypnotic, medium syringe for relaxant, small syringe for opioid and turn on the vapour. Only a few enlightened anaesthetists dabbled in the dark art of TIVA. When I moved on to the dizzy heights of senior registrar in Liverpool, no one was using TIVA. So why did I start? The answer to be honest was boredom: having done virtually the same anaesthetic for countless operations the prospect of doing the same thing for the next 30 to 40 years was not appealing. I had seen in Glasgow the improved quality of recovery offered by TIVA, so why not give it a try? Well it certainly improved my attention in the OR, instead of slumbering next to the gas-delivery system I now had lots to keep me occupied. First, I had to find suitable equipment, then make sure the patients had their weight documented. Next, the calculations for a loading dose and infusion regimen, followed by edgeof-the-seat attention, looking for signs of light anaesthesia. Finally, titration down towards the estimated end of surgery; hoping for a quick wake-up, but not too quick before completion of surgery. A stressful time but, again and again as I got better, I would see the benefits in patients’ recovery and the appreciation of the recovery staff and patients’ relatives. I have since become an enthusiast, some would say a zealot, so much so that I haven’t used inhalational anaesthesia for 22 years, unless I have taken over a case from one of my colleagues. Things have improved somewhat since my early faltering steps in TIVA; we now have TCI that allows accurate dose titration in a manner similar to inhalational anaesthesia. The practitioner selects a number and then assesses the patient’s response, easily titrating up or down as required. TCI allows the anaesthetist to select a target blood concentration

Downloaded from https://www.cambridge.org/core. University of Winnipeg, on 02 Dec 2019 at 21:46:09, subject to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316659069.010

52

Chapter 8: How Do I Get the Dose Right with TIVA?

of both hypnotic and opioid; the system uses a PK/PD model to give an appropriate loading dose and a constantly adjusting infusion to maintain this concentration. The clinician can easily titrate up or down by simply selecting a new target. In the rest of this chapter, I will refer to TIVA dosing in terms of TCI drug concentrations, µg.ml−1 for propofol and ng.ml−1 for remifentanil and alfentanil.

EC50: a MAC Equivalent for TIVA but with a Twist Clearly MAC is not a valid concept for IV agents. Instead we use a similar concept, effective concentration (EC), referring to the blood concentration that achieves a certain effect. Thus EC50 is the blood concentration that prevents movement in response to a standard stimulus in 50% of patients. Similar to MAC studies, the assessment is performed at steady state to allow for drug equilibration between blood and brain or effect-site concentration (Ce), typically 15 minutes (this was the time used in the initial MAC studies). The EC50 for propofol was initially suggested to be 6 µg.ml−1 as a single agent, with a 25% reduction to 4.5 µg.ml−1 for patients breathing 67% nitrous oxide.[2] It’s important to remember that this is a measure of response to pain, not just loss of consciousness. It is even more important to remember that propofol is a hypnotic agent, not an analgesic, a point that has been emphasised throughout this book. The reason why this is important is that if you use propofol to produce loss of consciousness and attenuate response or movement to pain as a single agent, you will end up using an inappropriately high concentration and cause unnecessary haemodynamic perturbations and excessive suppression of the brain. As discussed elsewhere, it is best to use adjuvant agents, typically an opioid, to reduce propofol requirements[3] in a similar manner to balanced anaesthesia using inhalational agents. So, in some ways, EC50 is not as useful as MAC is with inhalational agents since inhalational anaesthetics do have analgesic properties (although their MAC can, and should, be reduced with concomitant analgesic administration). EC95 is the concentration at which we would expect only 5% of patients to move in response to surgery. We can determine the effect of differing supplementary doses of opioid, on both MAC for inhalational and EC for IV agents. We can also determine the response to different stimuli; a typical MAC study would use a standard surgical incision of set length

and depth. Similarly, we can determine the MAC and EC for prevention of response to, for example, airway instrumentation to guide the clinician during differing phases of anaesthesia. The problem with TIVA is the dosing. Using inhalational agents, the anaesthetist simply selects a number on the vaporiser and waits for equilibration. With TIVA, a loading dose and an infusion regimen needs to be calculated and implemented in order to achieve a stable blood concentration and time needs to be allowed for equilibration with the brain (effect-site concentration). Consequently, TIVA is not as easy as ‘passing the gas’ at the start of the learning curve and can be a daunting prospect for the inexperienced clinician. However, as with most techniques, practice makes perfect. The availability of TCI means that titration of dose can easily be achieved by changing one number in a manner analogous to changing the dial on a calibrated vaporiser. Before looking further at EC, we need to consider how we get to a stable blood or plasma concentration (Cp) of hypnotic, and consider the effect of drug interactions and the time for effectsite equilibration.

Dose Requirements for TIVA When we consider the dose requirements for TIVA, or inhalational anaesthesia for that matter, we can simply think in traditional MAC/EC50 terms, that is, one value related to a specific type of stimulus. However, we can also look at drug interactions and take an EC50, for a combination of drug effects, to further specify our dosing. We can go further still by adjusting the dosing strategy to different levels of stimulus through a surgical procedure. But first, let’s consider MAC/EC50 for a single agent. Anaesthetists assume that when they deliver an inhalational agent and achieve certain end-tidal (ET) concentration, they have also achieved the same blood concentration. Using TIVA with a PK model there is typically a 25 to 30% variation between measured and estimated concentrations.[2] Anaesthetists also know the MAC value of an agent to two decimal places. They can therefore confidently predict the response of their patients to a given ET agent concentration. However, if we look at studies simultaneously measuring blood and ET inhalational agent concentration, there is typically a 25 to 30% variation due to V/Q mismatch.[4] Further, if we closely examine MAC studies, we find that there is considerable inter-individual variation between subjects and also in the same subject at

Downloaded from https://www.cambridge.org/core. University of Winnipeg, on 02 Dec 2019 at 21:46:09, subject to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316659069.010

53

Chapter 8: How Do I Get the Dose Right with TIVA?

different points in time[5] (PD variability – as seen with most drugs we use). The bottom line is that, no matter which drug we use for anaesthesia, there is around a 30% PK variation and a larger interindividual, PD variation in response to a given dose. Why then should we waste our time worrying about accurate dosing and blood concentrations; should the anaesthetist not just titrate to the individual’s requirements? Although this last statement is sensible, we do require a starting point. We also need an idea about the interaction between co-administered agents and knowledge of what we can reasonably expect in a specific group of patients, subjected to specific types of surgery. The problem with titrating drugs, such as propofol, is the inherent delay in administering a dose, achieving a Cp and achieving an effect in the subject. This is where the EC studies can help inform our initial dosing for induction and maintenance of anaesthesia, as well as guide our titration to specific stimuli and the emergence from anaesthesia. I will limit my comments to the commonly used agents for TIVA: principally propofol using the Marsh PK[6] model and remifentanil using the Minto PK[7] model.

What Really Happens with a ‘Standard’ Induction Induction of anaesthesia using IV agents is something with which all anaesthetists are familiar. This is usually achieved with a bolus of hypnotic, with or without an adjuvant opioid, often together with a muscle relaxant if tracheal intubation is required. Observing some of my colleagues, these drugs are often administered in quick succession, without waiting to judge the response to the hypnotic agent. This does achieve a rapid induction of anaesthesia and facilitates speedy intubation, but is not always well tolerated by our patients and can lead to catastrophe in difficult airway cases. It also means that an

opportunity to assess the patient’s individual response to the hypnotic agent is lost. I will return to this issue later. The question you may want to ask is what blood concentration of propofol is required for induction of anaesthesia? As I have already indicated, there is considerable variation between patients; but this is also the wrong question. Anaesthetic agents act in the brain not in the blood, so we should therefore be asking, what hypnotic Ce is required for induction of anaesthesia? We have to consider both dose and time when dosing our patients. Let’s consider inducing anaesthesia with loading doses given at different infusion rates. Peacock and colleagues infused propofol at three different rates and measured the blood concentration at loss of consciousness (LOC).[8] Interestingly, the lowest infusion rate resulted in the longest induction time and the lowest induction dose. As can be seen from the data presented in Figure 8.1 and Table 8.1, the Cp of propofol at LOC varies from 4.3 µg.ml−1 at the 300 ml.h−1 induction rate, up to 9.2 µg.ml−1 at the 1200 ml.h−1 induction rate. However, the calculated Ce for LOC is similar for all three infusion rates at around 1 µg.ml−1 and is essentially rate independent when based on the Marsh PK model. Thus it is more appropriate to look at the Ce when considering the effect of our drugs for induction and maintenance of anaesthesia. All the simulations I will present in this chapter are performed using the PK simulation programme Tivatrainer © (www .eurosiva.org/TivaTrainer/tivatrainer_main.htm). This may all seem complex when compared to inhalational anaesthesia but the same is actually true of inhalational agents. There is a delay for equilibration between blood and brain concentrations but this is rarely appreciated because of the slow rise in blood concentration inherent with inhalational drug delivery. It is readily apparent with IV agents, because of the rapid rise in blood concentration associated with bolus dosing. Figure 8.2 compares a bolus dose with an

Table 8.1 Propofol concentrations at loss of consciousness.

Infusion rate ml.h−1

Cp LOC µg.ml−1

Ce LOC µg.ml−1

Dose at LOC mg

Time to LOC min:sec

300

4.3

1

86

00:50

600

6

0.9

110

01:06

1200

9.2

1.1

166

01:43

Cp = plasma concentration; Ce = effect-site concentration.

Downloaded from https://www.cambridge.org/core. University of Winnipeg, on 02 Dec 2019 at 21:46:09, subject to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316659069.010

54

Chapter 8: How Do I Get the Dose Right with TIVA?

Cp LOC 6 mcg/ml Ce LOC 0.9 mcg/ml Ce

a

Figure 8.1 Induction of anaesthesia with different rates of infusion of propofol. Data re-simulated from the infusion rate and induction times from Peacock and co-workers[8] using Tivatrainer with Marsh PK/PD. Panel (a) = 300 ml.h−1; (b) = 600 ml.h−1 and (c) = 1200 ml.h−1. White boxes and right y axis = infusion rate ml.h−1; left y axis = propofol concentration µg.ml−1; x axis = time in mins; Ce = effect-site concentration; Cp = plasma concentration; LOC = loss of consciousness.

0

10 600ml/hr b

1000 500 0

10

0 Cp

Cp LOC 9.2 mcg/ml Ce LOC 1.1 mcg/ml Ce

1200ml/hr c

1000 500 0

12 10 8 6 4 2 0

Cp

300ml/hr

0

10 8 6 4 2 0

C p

Cp LOC 4.3 mcg/ml Ce LOC 1 mcg/ml μg/ml Ce

1000 500

10 8 6 4 2 0

10

0

Cp TTPE Ce

0 9 8 7 6 5 4 3 2 1 0

10

20

30

Cp Ce 0

1000 800 600 400 200 0 1000 800 600 400 200 0

9 8 7 6 5 4 3 2 1 0

Figure 8.2 Bolus dose and infusion of propofol. White boxes and right y axis = infusion rate ml.h−1; left y axis = propofol concentration µg.ml−1; x axis = time in mins; Ce = effect-site concentration; Cp = plasma concentration; TTPE = time to peak effect.

10

infusion of propofol. We can see the clear mismatch between blood and effect-site concentration with a bolus dose, which is less apparent with the same dose given by slow infusion. What is also illustrated in the bolus graphic is the time to peak effect (TTPE); this is the time to the maximum Ce and thus the maximum effect on the patient. It is defined by both the decay in the blood concentration and the speed of equilibration between blood and brain concentrations. The latter is modelled in TCI systems using an equilibration constant (keo). The keo is different for different agents and is defined by the time taken for transport

across the blood–brain barrier and interaction with receptors to achieve the drug effect.

The Legacy of Misconceived Wisdom The delay in equilibration between blood and brain concentrations is in contrast to what I was taught as a young anaesthetist. I was told that anaesthetic agents pass instantaneously across the blood–brain barrier and that induction time was one ‘arm–brain’ circulation, following a bolus. This concept is clearly wrong, although it was the perceived wisdom at the time. The standard quoted propofol induction dose is

Downloaded from https://www.cambridge.org/core. University of Winnipeg, on 02 Dec 2019 at 21:46:09, subject to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316659069.010

55

Chapter 8: How Do I Get the Dose Right with TIVA?

1000

Cp

25

Thiopentone bolus TTPE = 1.5 min Ce

20 15

a

800 600 400

10

Figure 8.3 (a) Bolus dose of thiopentone compared with (b) propofol. Left y axis = propofol concentration µg.ml−1 and x axis = time in mins. Ce = effect-site concentration; Cp = plasma concentration; TTPE = time to peak effect.

200

5 0 Propofol bolus

Cp

10 8

TTPE = 3.6 min

6

b

Ce

4

0 1000 800 600 400 200

2

0

0 0

10

2 to 2.5 mg.kg−1 for adults, given in 40 mg aliquots every ten seconds until LOC. The implication is that the dose is titrated to response, in a similar technique to that used for thiopental. However, as can be seen from Figure 8.3, propofol has a slower TTPE after a bolus dose, compared to thiopental. This is due to a more rapid initial redistribution and effect-site equilibration seen with thiopental. Thus titrating a bolus dose of propofol is more difficult, takes longer and may lead to relative overdosing in an attempt to achieve a rapid induction. When you think about it, using this traditional IV induction with propofol followed by an inhalational agent, you are essentially guessing the correct propofol induction dose and then hoping to correctly match the waning IV drug effect to the rising inhalational effect. There is then with every case a potential for awareness if you guessed wrong or, worse still, get distracted and forget to turn the vaporiser on, say, after a difficult intubation. With procedures that involve a strong surgical stimulus shortly after induction, one is tempted to ‘over-pressure’ and dial up the vaporiser settings well above the MAC in order to achieve MAC as quickly as possible. Again, this runs the risk of temporarily overdosing the patient, a situation analogous to the inappropriate titration of propofol as described above.

One-Stop Shop for Induction and Maintenance of Anaesthesia A better option is to use TCI to induce and then maintain anaesthesia in a seamless manner, thus avoiding the pitfalls associated with overlapped inhalational agent following IV induction. Looking at Figure 8.4, we can see a TCI of propofol set at

a plasma target concentration of 4 µg.ml−1 using the Marsh PK model. The blood concentration rises rapidly to the target of 4 µg.ml−1 following the loading dose. If we were to assess our patient at this point, he is likely to be wide awake and ‘TIVA newbies’ may be tempted to increase the plasma target concentration. However, if we look at the estimated propofol Ce, we can see that it is still very low at this stage and far from equilibration. If we wait, we will see our patient pass through progressively deepening sedation until LOC. Typically this will result in a slow induction of anaesthesia over two to five minutes but with the maintenance of cardiovascular stability. Also it allows the anaesthetist to assess this individual’s response to the hypnotic. The practitioner can note the effect-site concentration at loss of consciousness, together with any response, thus calibrating this individual patient’s sensitivity to propofol. As an example, the data in Table 8.2 show actual values for LOC, together with demographics for three patients. As can be seen, there is some variation; the oldest patient losing consciousness at a propofol Ce of 0.7 µg.ml−1, which was achieved at 58 s and a dose of 1.09 mg.kg−1. However, the 30-year-old patient required a propofol Ce of 1.8 µg.ml−1, achieved at 2 min 26 s and a dose of 1.46 mg.kg−1. This may seem quite low compared to the recommended bolus dose of 2 to 2.5 mg.kg−1, but this dose is for LOC and does not equate to clinical anaesthesia. For the 68-year-old patient in our examples, the propofol Ce at which the anaesthetist felt the jaw was sufficiently relaxed for LMA insertion was 1.8 µg.ml−1 at 2 min 30 s and a dose of 1.57 mg.kg−1. However, for the younger patient the propofol Ce at LMA insertion was 3.2 µg.ml−1, achieved at 6 min

Downloaded from https://www.cambridge.org/core. University of Winnipeg, on 02 Dec 2019 at 21:46:09, subject to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316659069.010

56

Chapter 8: How Do I Get the Dose Right with TIVA?

Table 8.2 Propofol induction using plasma concentration TCI at 4 µg.ml−1 in three subjects.

Weight kg

Age years

Ce LOC µg.ml−1

Time to LOC min:sec

Dose at LOC mg.kg−1

Patient 1

82

68

0.7

00:58

1.09

Patient 2

86

57

1.4

01:46

1.27

Patient 3

75

30

1.8

02:26

1.46

Ce = effect-site concentration; LOC = loss of consciousness.

8 7 6

Ct Approximate time for equilibration between plasma and effect-site concentration

5

Figure 8.4 Propofol TCI induction of anaesthesia at 4 µg.ml−1 using plasma concentration targeting. Left y axis = propofol concentration µg.ml−1 and x axis = time in mins. Ct = target concentration; Ce = effect-site concentration; Cp = plasma concentration.

4 Cp 3 2

Ce

1 0 0

10

with a dose of 2.26 mg.kg−1. What these examples illustrate is that a TCI induction allows on-going assessment and titration of the induction dose. Once the appropriate effect has been achieved in a given patient, the target concentration can be titrated down to around the current Ce value. An appropriate opioid dosing strategy and/or regional anaesthesia can then be instituted to deal with the surgical stimulation. One drawback in using such a technique is the slow induction time, particularly with younger patients, as evidenced by the six-minute delay to insertion of LMA in our 30-year-old patient. Many anaesthetists find this unacceptable, especially when faced with an operating list of multiple short cases.

Hey Can You Speed Things Up a Little! There are a number of ways to speed up induction using TCI. One can start with a higher initial target propofol Cp, use effect-site concentration targeting on your pump (see Chapter 6) or add a supplementary agent, usually an opioid such as remifentanil. Figure 8.5 shows propofol TCI set at

8 µg.ml−1 reducing to 4 µg.ml−1 after induction of anaesthesia. A high initial Cp drives the drug into the effect site down its concentration gradient to achieve a rapid rise in Ce and therefore a rapid induction. Data were collected from a 28-year-old patient dosed with this regimen. In this case, propofol Ce LOC was 1.7 µg.ml−1, which occurred at 1 min 17 s and a dose of 2.3 mg.kg−1. Propofol Ce at LMA insertion was 3.2 µg.ml−1, which was achieved after 2 min 22 s and a dose of 2.6 mg.kg−1. I have observed some colleagues give a manual bolus dose of propofol for induction followed by a TCI, which achieves a similar rapid induction. This is not recommended as the TCI system is unaware of the manual administered propofol and gives a loading dose assuming there is no drug in the patient. If the TCI is started soon after the bolus, there is a risk of an initial large overdose and the Cp will be elevated above what is predicted throughout the procedure, see Figure 8.6. Another trick that can be used with a propofol TCI induction is to select a low target, say 1 µg.ml−1,

Downloaded from https://www.cambridge.org/core. University of Winnipeg, on 02 Dec 2019 at 21:46:09, subject to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316659069.010

57

Chapter 8: How Do I Get the Dose Right with TIVA?

1300

9 8

1100

Cp

7 900 6

Ct 700

5 4

Figure 8.5 Propofol TCI induction of anaesthesia at 8 µg.ml−1 reducing to 4 µg.ml−1 using plasma concentration targeting. White boxes and right y axis = infusion rate ml.h−1; left y axis = propofol concentration µg.ml−1; x axis = time in mins; Ct = target concentration; Ce = effect-site concentration; Cp = plasma concentration.

500 Ce

3

300

2

100

1 0 0

10

12

1000

11

900

10

800

9

Cp

700

8 7

600

6

500

5

400 Ce

4

Figure 8.6 Propofol manual bolus of 2 mg.kg−1 followed by a TCI of 4 µg.ml−1 using plasma concentration targeting. Note that the actual plasma concentration is above the target of 4 µg.ml−1 throughout the case. White boxes and right y axis = infusion rate ml.h−1; left y axis = propofol concentration µg.ml−1; x axis = time in mins; Ce = effect-site concentration; Cp = plasma concentration.

300

3 200

2

100

1 0

0 0

10

20

immediately after attaching monitoring lines, but while waiting for all parties to be ready for induction. This relaxes the patient and allows the propofol Ce to rise to a sedative level, which will reduce the time taken to achieve a Ce LOC once an induction target is selected. An alternative to the single-agent induction is to use a combination of hypnotic and opioid, continuing this as ‘balanced’ anaesthesia for the maintenance phase. Propofol and remifentanil are often used to maintain TIVA and the combination is also efficacious for induction. Consider a co-induction with propofol and remifentanil TCI target 4 µg.ml−1 and 4 ng.ml−1 respectively in a 42-year-old patient of mine, assessing for the clinical end points of LOC and insertion of LMA. In this case, LOC occurred after 1 min, at a propofol Ce of 0.8 µg.ml−1 and remifentanil Ce of 1.7 ng.ml−1, using 73 mg of

30

propofol and 36 µg of remifentanil. The time to insertion of LMA was 2 min 22 s at Ce concentrations of 1.8 µg.ml−1 propofol and 2.9 ng.ml−1 remifentanil, corresponding to a dose of 94 mg propofol and 60 µg remifentanil, respectively. The reader will appreciate that the effect-site concentration of remifentanil rises faster than propofol due to a ‘faster keo’. What is also apparent is the low volume of remifentanil infused at these two time points. If the syringe concentration of remifentanil is 20 µg.ml−1, only 1.8 and 3 ml are infused at these two time points, respectively. One can imagine that at the recommended dilution of remifentanil, of between 40 and 50 µg.ml−1, the infused volumes are even smaller. Therefore one has to be very vigilant regarding dead space in the infusion lines when using remifentanil. Milne and colleagues determined the values for calculated Cp and Ce of propofol, at LOC and recovery

Downloaded from https://www.cambridge.org/core. University of Winnipeg, on 02 Dec 2019 at 21:46:09, subject to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316659069.010

58

Chapter 8: How Do I Get the Dose Right with TIVA?

of consciousness (ROC), in three remifentanil groups: low, medium and high.[9] They found a calculated Ce LOC propofol of around 2 µg.ml−1, in the presence of a calculated Ce remifentanil of 2 ng.ml−1. This reduced to around 1.5 µg.ml−1 with an associated Ce remifentanil of 8 ng.ml−1. However, there was a large range with almost 300% variation between individual patients, again emphasising the need to titrate the dose to the individual. Scott and colleagues showed that propofol concentrations for LOC are only very modestly reduced by concurrent remifentanil TCI, whereas the dose requirement to prevent nociceptive reflexes is markedly decreased.[3] These observations highlight the need to be cognisant of the end points of titration both in clinical practice and when evaluating the literature. The EC50 and EC95 of propofol as a single agent for LOC are substantially lower than those required for obtunding response to pain. When all is said and done, in practice, induction of anaesthesia using TCI is easy. The anaesthetist selects an appropriate target (or targets when using two agents) for the patient; balancing the risk of cardiovascular compromise using high targets with the prolonged induction time associated with lower target concentrations. Young, healthy patients can tolerate a higher initial target, which is associated with faster induction. Elderly patients or those with co-morbidity are better induced slowly, with lower initial targets. In each case take note of the response of your individual patient to the rising calculated Ce displayed on the TCI device, thus gauging your patient’s sensitivity to the delivered drugs. This can be a guide to an individual’s maintenance requirements but remember surgical stimulation will increase the required analgesic dose. Although I have given some published EC50

LOC data for information, in practice the practitioner should look at the response of their patient during TCI induction, to get a much better idea of individual dosing requirements.

Maintenance of Anaesthesia It has been consistently shown that the EC50 for LOC is an effect-site concentration of around 2.5 µg.ml−1 with the Marsh model[10] and this is borne out in our clinical experience. It must be emphasised, again, that this must be supplemented with an analgesic to prevent movement in response to a painful stimulus and propofol alone is not suitable for that purpose. Milne and co-workers showed a synergistic relationship between calculated propofol and remifentanil concentrations, using a closed-loop system, in patients undergoing day-case surgery.[9] Three levels of remifentanil were used to represent low, medium and high doses given by TCI at 2, 4 and 8 ng.ml−1, respectively. An auditory evoked potential index (APEX) was used as an input signal for a closed-loop system, which then controlled a TCI of propofol to maintain an APEX value synonymous with unconsciousness of 35. Figure 8.7 shows the relationship between propofol and remifentanil for EC50 and EC95 for loss of response to intraoperative stimulus (IOP) and EC50ROC. The reader must also remember that these EC values are required to suppress movement in response to the more intense phases of surgery and will require downward titration during less stimulating phases of the procedure. This titration is best achieved by varying the target value of TCI remifentanil, which can produce rapid changes in the remifentanil Ce to match the varying surgical stimuli throughout the

Remifentanil calculated blood concentration (ng ml–1)

16 14

Adequate anaesthesia Recovery of consciousness

EC95lOP

12 10

EC50lOP

8

1.2

6

Figure 8.7 Drug interaction between propofol and remifentanil. Curves describe the effective concentration for 50% (EC50IOP) and 95% (EC95IOP) for lack of response to intra-operative stimulation; and effective concentration for 50% recovery of consciousness (EC50ROC) (from Milne et al. 2003).[10]

3.0

EC50ROC

4

1.5

3.5 4.9

2 1.7 0 0

1

2

3

4

5

6

7

8

9

10

11

12

13

14

15

Propofol calculated blood concentration (µg ml–1) Downloaded from https://www.cambridge.org/core. University of Winnipeg, on 02 Dec 2019 at 21:46:09, subject to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316659069.010

59

Chapter 8: How Do I Get the Dose Right with TIVA?

procedure. Once the patient is unconscious, propofol concentrations should need little, if any, adjustment as this is not what is providing anti-nociception (analgesia). Titrating our agents to the patient’s response during the procedure is part of the art of anaesthesia. It’s exactly the same for TIVA and inhalational-based anaesthesia but most anaesthetists have more experience, and therefore more confidence, with inhalational agents. In the past, drug delivery was also easier with inhalational agents. However, with the addition of TCI to the anaesthetist’s armamentarium, titration of TIVA is just as straightforward and, indeed, quicker, due to direct drug delivery to the blood compartment.

Anaesthetic Drug Modelling Systems There are two commercial systems available with anaesthetic and analgesic interaction modelling, covering both inhalational and TIVA-based scenarios: SmartPilot® from Dräger and Navigator from GE Healthcare. These systems have a number of drug interactions modelled in the software and give a graphical representation of both hypnotic and analgesic drug effect. They are based on EC and MAC data and, as such, do not consider the individual variability seen in clinical practice. They may have some utility when administering an inhalational anaesthetic with an opioid but are much less useful, if at all, with TIVA. This is because, unlike inhalational agents, propofol, and therefore propofol and opioids, are less interchangeable, i.e. low opioid/high propofol vs high propofol/low opioid. As we tend to emphasise, titrate propofol to produce LOC and then use an analgesic to attenuate nociception – simple! These PK/PD interaction prediction systems do not provide a full picture of clinical anaesthesia and cannot accurately predict either your individual patient’s anaesthetic requirements or their awakening time. The interindividual variability between patients means that any such ‘recipes’ can only be a guide. Thus dose always needs to be titrated to each individual subject. Some of these factors can be anticipated, for instance elderly patients tend to require less drug than younger individuals. The use of alcohol, or other drugs, tends to increase a subject’s tolerance to anaesthetic agents. An extreme example of this is opioid addicts, who require much larger opioid doses to achieve an adequate effect.

Recovery from TIVA The potential for a rapid clear-headed recovery after TIVA, with a low incidence of nausea and vomiting, is an attractive prospect for the anaesthetist, patient and recovery staff. However, this situation is often not achieved with ‘TIVA novices’. Clearly if remifentanil has been used as part of the technique then postoperative analgesia needs to be addressed. Large doses of opioid can counteract the normal antiemetic effect of propofol-based TIVA. The ideal technique would combine the use of TIVA with regional anaesthesia for the post-operative phase. The use of multi-modal analgesia, to limit the opioid dosage, can also be effective. The anaesthetist also needs to think about the drug concentrations as the end of surgery approaches. When remifentanil is used, a high opioid/ low hypnotic combination gives a more rapid recovery, due to the rapid metabolism of remifentanil. In my own practice, I titrate the propofol down towards the end of surgery leaving the remifentanil at a higher level (around 5 to 6 ng.ml−1), which often allows me to reduce the propofol down to around 2 µg.ml−1, to achieve a rapid awakening once surgery is over. I find the bispectral index (BIS™) is useful at this stage to alert me early to rising consciousness, thus avoiding movement or awakening before completion of surgery. Using this technique, I can usually produce a rapid clear-headed wakening after completion of surgery. As is always the case in clinical practice there will be a range of propofol CP ROC, but given adequate post-operative analgesia, the quoted values in the literature are around 1 to 1.5 µg.ml−1.[11]

A Snapshot of My Clinical Practice In my own practice I do the following: if there is likely to be a delay before induction, I select a low propofol TCI target of 0.5 to 1 µg.ml−1 for anxiolysis. When induction is required an appropriate target of propofol and remifentanil can be selected for this individual patient, balancing the risk of cardiovascular compromise against speed of induction. Typically, I start at around 4 to 5 µg.ml−1 TCI propofol blood concentration (note: you do not need such high concentrations with effect-site targeting) and 4 to 6 ng.ml−1 TCI remifentanil. The patient’s response is assessed during induction and airway instrumentation and, depending on this assessment, the propofol target can often be reduced at this stage. This is particularly the case if there is a period of low stimulation prior to

Downloaded from https://www.cambridge.org/core. University of Winnipeg, on 02 Dec 2019 at 21:46:09, subject to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316659069.010

60

Chapter 8: How Do I Get the Dose Right with TIVA?

commencement of surgery. Appropriate targets can then be selected prior to the start of surgery. Any increase in target concentrations of remifentanil, to deal with anticipated surgical stimulation, must be made prior to the stimulus to allow time for effectsite equilibration. To achieve a rapid effect, the target levels can be transiently increased to high values and then rapidly reduced to the anticipated required concentrations, in a similar manner to that which I described for a rapid induction of anaesthesia (this is not required if you are using effect-site targeting). There are also a number of depth of anaesthesia monitors available that can be used to guide dosing in both inhalational and TIVA techniques. The BIS is the most well established and a target BIS of between 40 and 60 is often used to avoid both under- and overdosing. The BIS is not perfect and its use cannot guarantee lack of awareness. However, I find it a useful adjuvant along with the normal clinical signs of depth of anaesthesia. I find the BIS particularly useful towards the end of surgery when I am titrating down the level of hypnotic.

My Thoughts on Pharmacokinetic Models A consideration of PK models is beyond the scope of this chapter (see Chapter 5 for more details) but I do need to make a few brief comments to avoid confusion. With regard to remifentanil, currently only the Minto model is available in commercial TCI systems. It is a model derived for adults and should not be used for children. Propofol is a little more confusing; TCI was first commercialised with the Marsh model for propofol. All the prescribing information and regulatory studies are based on this model. Marsh is adjusted for weight only and, as such, increases the dose proportionately as weight increases. Thus when dealing with obese patients, it can lead to overdosing. The anaesthetist can compensate by inputting a lower weight based on experience (see Chapter 16). Likewise, there is no adjustment for age; it is not suitable for children and a lower target should be selected for the elderly to avoid overdosing. The Schnider model for propofol was introduced in an attempt to address the shortcomings of Marsh.[12,13] However, there are errors in the implementation of this model, particularly related to the way LBM is calculated, making it unsuitable for obese patients.[13] In addition, the

model has been implemented differently by different manufacturers. This results in a discrepancy between the dose delivered by the two implementations, set at the same target concentration.[13,14] The PD modelling in Schnider predicts a much more rapid effect-site equilibration than Marsh. This doesn’t fit with the observed effect on patients in clinical practice. Barakat and colleagues studied patients receiving sedative doses of propofol. They found a marked mismatch between the time course of the propofol Ce predicted by the Schnider model and that of the BIS.[15] There was a similar disconnect between sedation scores and the Schnider Ce prediction. Both BIS and sedation scores correlated better with the Marsh propofol Ce prediction. Another issue is that the Schnider model uses a fixed central compartment volume of 4.2 l and adjusts the clearance for weight, height and LBM in a complex manner. This means that the initial doses it delivers, when set in plasma-control TCI, is very low at the target values we have been discussing. Patients will not go to sleep unless high target values are selected or prolonged induction times are allowed. For this reason, clinicians using this model tend to use it in effect-site TCI mode. However, because of the fixed central compartment volume and the rapid effect-site constant, Schnider effect-site TCI still delivers less drug than the Marsh model in plasma control set at the same target. Therefore effect-site TCI using Schnider will not produce a more rapid induction than plasma TCI using Marsh set at the same target. There is a modification of the Marsh model available for effect-site TCI; for a further explanation please refer to Chapter 5. There are no efficacy or safety studies published for effect-site TCI and no prescribing information. For these reasons, I use the Marsh model in clinical practice and all my comments, guidance and simulations are based on the Marsh model for propofol and the Minto model for remifentanil.

References 1. J.J. Pandit, J. Andrade, D.G. Bogod, et al. 5th National Audit Project (NAP5) on accidental awareness during general anaesthesia: summary of main findings and risk factors. Br J Anaesth 2014; 113: 549–59. 2. J.A. Davidson, A.D. Macleod, J.C. Howie, M. White, G.N. Kenny. Effective concentration 50 for propofol with and without 67% nitrous oxide. Acta Anaesthesiol Scand 1993; 37: 458–64. 3. H.B. Scott, S.W. Choi, G.T. Wong, M.G. Irwin. The effect of remifentanil on propofol requirements to

Downloaded from https://www.cambridge.org/core. University of Winnipeg, on 02 Dec 2019 at 21:46:09, subject to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316659069.010

61

Chapter 8: How Do I Get the Dose Right with TIVA?

achieve loss of response to command vs. loss of response to pain. Anaesthesia 2017; 72: 479–87. 4. R.L. Carpenter, E.I. Eger. Alveolar-to-arterial-to-venous anesthetic partial pressure differences in humans. Anesthesiology 1989; 70: 630–5. 5. E.I. Eger, L.J. Saidman, B. Brandstater. Minimum alveolar anesthetic concentration: a standard of anesthetic potency. Anesthesiology 1965; 26: 756–63. 6. B. Marsh, M. White, N. Morton, G.N. Kenny. Pharmacokinetic model driven infusion of propofol in children. Br J Anaesth 1991; 67: 41–8. 7. C.F. Minto, T.W. Schnider, S.L. Shafer. Pharmacokinetics and pharmacodynamics of remifentanil. II. Model application. Anesthesiology 1997; 86: 24–33. 8. J.E. Peacock, R.P. Lewis, C.S. Reilly, W.S. Nimmo. Effect of different rates of infusion of propofol for induction of anaesthesia in elderly patients. Br J Anaesth 1990; 65: 346–52. 9. H. Iwakiri, N. Nishihara, O. Nagata, T. Matsukawa, M. Ozaki, D.I. Sessler. Individual effect-site concentrations of propofol are similar at loss of consciousness and at awakening. Anesth Analg 2005; 100: 107–10

10. S.E. Milne, G.N. Kenny, S. Schraag. Propofol sparing effect of remifentanil using closed-loop anaesthesia. Br J Anaesth 2003; 90: 623–9. 11. J. Vuyk, T. Lim, F.H. Engbers, A.G. Burm, A.A. Vletter, J.G. Bovill. The pharmacodynamic interaction of propofol and alfentanil during lower abdominal surgery in women. Anesthesiology 1995; 83: 8–22. 12. J. Vuyk, M.J. Mertens, E. Olofsen, A.G. Burm, J.G. Bovill. Propofol anesthesia and rational opioid selection: determination of optimal EC50–EC95 propofol–opioid concentrations that assure adequate anesthesia and a rapid return of consciousness. Anesthesiology 1997; 87: 1549–62. 13. M.J. Mertens, F.H. Engbers, A.G. Burm, J. Vuyk. Predictive performance of computer-controlled infusion of remifentanil during propofol/remifentanil anaesthesia. Br J Anaesth 2003; 90: 132–41. 14. F. Engbers, N. Sutcliffe, G.N. Kenny, et al. The devil is not only in the detail. E-letter to the editor. Br J Anaesth 3 Nov 2009. 15. A.R. Barakat, N. Sutcliffe, M. Schwab. Effect site concentration during propofol TCI sedation: a comparison of sedation score with two pharmacokinetic models. Anaesthesia 2007; 62: 661–6.

Downloaded from https://www.cambridge.org/core. University of Winnipeg, on 02 Dec 2019 at 21:46:09, subject to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316659069.010

62

Chapter

9

Be Aware, Unaware and Confusion Everywhere TIVA and Awareness Pablo Martinez-Vazquez, Claus Lindner, Umberto Melia, Jaideep J. Pandit and Erik Weber Jensen

Scope of the Problem and Contributing Factors All anaesthetists need to be skilled in the administration of IV anaesthesia, yet evidence from the 5th National Audit Project (NAP5) of the Royal College of Anaesthetists and the Association of Anaesthetists of Great Britain and Ireland documenting cases of accidental awareness during general anaesthesia (AAGA) suggests that this might not be the case. In the UK, TIVA is used in only 7% of cases, with preponderance for certain surgical operations such as those on the airway where administration of an inhaled anaesthetic is not feasible. Conversely, TIVA has been almost completely avoided in other types of surgery such as caesarean sections, at least in the UK and Ireland. TIVA is also used for transfer of patients and in remote areas that do not have the equipment required to deliver inhaled anaesthesia. For the purpose of this discussion, it should be clarified from the outset that there are several modes of TIVA delivery. The majority (UK NAP5 data) is target-controlled infusion (TCI), carried out with dedicated PK pumps and used in ~5% of all cases or ~81% of TIVA cases. The most basic is manual boluses of propofol using a hand-held syringe: the anaesthetist simply injects the amount they judge necessary to maintain anaesthesia (0.7% of all UK cases, or ~10% of all TIVA cases). The third is non-TCI infusions, which constitute ~0.6% of all cases, or ~9% of TIVA cases. However, there is some international geographic variation in this practice, because in the United States TCI remains unlicensed and therefore seldom used. Earlier studies suggested a similar incidence of AAGA with TIVA as with an inhalational maintenance technique.[1,2] However, others have

suggested that the incidence may be higher with TIVA.[3,4] There are several reasons why, in theory, a higher incidence of AAGA with TIVA might be anticipated. Whereas with inhaled anaesthetic drugs the end-tidal anaesthetic gas (ETAG) concentration can be continuously measured, this is not currently possible for TIVA. Accidental disconnections to the delivery of propofol to the patient may go undetected, especially if the site where the infusion tubing connects to the IV line is hidden by drapes, as the infusion pump will continue to display ‘adequate’ delivery of drug. Even where the infusion line is visible, the line may be found later to have ‘tissued’. Other factors also include a poor understanding of the TCI models and the lack of model support in manual infusion regimes. [5] Table 9.1 lists some potential problems and how they may be prevented or managed. Guidelines on safe practice of TIVA are available.[6,7] Nevertheless, unintended awareness during surgery may occur under inhalational anaesthesia where end-tidal volatile concentration is not necessarily equivalent to the brain partial pressure. Multiple factors can affect this relationship, such as obesity, smoking, alcohol and drug abuse, pyrexia, hyperthyroidism and others.[8,9] Uncommonly, altered partial pressures of inhaled anaesthetic at high altitudes may increase the incidence of anaesthesia-related complications with inhaled anaesthesia techniques, as compared with TIVA.[10] Yet there is at least one potential and important advantage of TIVA for reducing awareness, in that it ensures a continuous delivery of anaesthetic from the moment of induction. By contrast, the common method of IV induction followed by inhalational anaesthetic for maintenance inevitably entails a period when there is a gap in anaesthetic delivery. The potential for this gap is compounded by induction in a separate anaesthetic room (as is common in

Downloaded from https://www.cambridge.org/core. University of Winnipeg, on 02 Dec 2019 at 21:47:20, subject to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316659069.011

63

Chapter 9: TIVA and Awareness

Table 9.1 Potential limitations of TIVA that might lead to under-dosing and AAGA. None of the prevention methods is absolute.

Problem

Prevention

IV cannula disconnection

Infusion line visible and accessible

Infusion tissuing (i.e. subcutaneous rather than intended IV infusion)

Checking IV line by flushing saline before use; pump highpressure alarm; response to patient discomfort during induction

Pump accidentally paused

Alarm for pause of infusion; pump asks operator to confirm before pausing

Power failure

Alarm for mains power disconnection; use only pumps previously fully charged for battery; mains disconnection and low-battery alarm

IV occlusion (including by closed tap)

Pump high infusion pressure alarm

Infusion failure due to high resistance of patent small cannula or long infusion tubing

Adjustable high infusion pressure alarm

Backtracking of propofol into IV fluid or other infusions

Use separate lines for TIVA and other infusions; use one-way valves to prevent backtracking

Mistakenly using 1% propofol with an algorithm for 2% propofol

Individual checks; stocking of only one propofol concentration

Mistakenly placing remifentanil infusion into pump programmed for propofol and vice versa

Colour coding of the pump or its screen

Under-dosing due to mixing of propofol and remifentanil or other opioid in same syringe

Using only one drug in each syringe

Table 9.2 Adapted data from NAP5. The first column shows the different types of techniques used to maintain anaesthesia. The next two columns show, respectively, the numbers in which that technique was used, and as %. The next two columns show the number of cases of AAGA detected by NAP5 for the given technique, and in %. The last represents the ratio of the third and fifth columns, a statistic that reflects the overall risk of the technique for AAGA. The Activity survey represents the methods in use in UK anaesthesia practice.

Activity survey Volatile agent Propofol infusion, TCI Propofol infusion, nonTCI Intermittent boluses Both volatile agent and propofol infusion Total

AAGA

Ratio

n

%

n

%

13, 479

93.1

112

82.4

0.89

764

5.3

14

10.3

1.94

82

0.6

2

1.5

2.50

106

0.7

1

0.7

1.00

48

0.3

7

5.1

17.00

14,479

100

136

100

the UK), which requires the patient to be disconnected from the anaesthetic and monitoring, transferred into theatre and then reconnected. Similarly, if inhalational anaesthesia maintenance has been used at the end of surgery and if transfer, to the ICU or for CT scan etc., is needed, this requires commencement of IV anaesthetic and, again, a gap in drug delivery can arise. NAP5 has now confirmed that, overall, TIVA does represent a higher risk for AAGA, in the UK especially, when using non-TCI techniques (Table 9.2). A particular risk is when an inhaled anaesthetic is used for maintenance but then anaesthesia is switched

–

to a non-TCI method, such as for transfer of the patient to the ICU or for a scan. Interestingly, manual boluses of propofol do not seem to constitute a risk but this may be because of the small numbers overall involved (Table 9.2). Because of the problems inherent in monitoring TIVA delivery, discussed above, the use of specific depth of anaesthesia (DOA) monitoring is often recommended when TIVA is used. Remarkably (or erroneously, perhaps, in the light of the NAP5 evidence), the National Institute for Health and Care Excellence (NICE) expressed the view that patients receiving TIVA were not at higher risk of AAGA but

Downloaded from https://www.cambridge.org/core. University of Winnipeg, on 02 Dec 2019 at 21:47:20, subject to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316659069.011

64

Chapter 9: TIVA and Awareness

nevertheless recommended that the use of DOA monitors should be an option in these patients (NICE, 2012). This last recommendation is now incorporated into the Association of Anaesthetics of Great Britain and Ireland (AAGBI) standards on minimum monitoring, which recognises the increased risk of accidental awareness with TIVA, and that a DOA monitor is perhaps the only way to detect if the drug is having its intended action on the brain if the patient is paralysed. That said, processed EEG (pEEG) monitors have their limitations[11] so their outputs should be interpreted with caution and, at the very least, combined with other evidence (e.g. clinical signs) to reach an opinion as to whether a patient is suitably anaesthetised or not.

Rationale behind the Use of Depth of Anaesthesia Monitoring TCI relies on different population PK and PD parametric models that estimate the plasma and effect-site concentrations (Ce), respectively.[12–14] The common TCI models[15–18] rely on the same compartmental principle of drug dissemination (for more details see Chapters 3 and 5) and there is no clinical evidence to support the use of one TCI model in preference to another.[19] These parametric models provide the practitioner with useful predictive estimates of plasma and effect-site concentrations with regard to specific controlled infusion profiles. However, since these models rely on population statistics, it is crucial to complement this information with patients’ clinical signs (in fact this applies to all modes of anaesthesia). Surrogate parameters, such as tachycardia, hypertension, sweating or movement, have shown to have low sensitivity and specificity in estimating sedation levels and detecting nociceptive stimuli. This limitation is important as it means that potential AAGA may not be predicted by TCI models. To overcome these limitations, non-invasive brain monitoring technologies based on EEG activity provide different derived indices with high sensitivity and specificity to estimate the patient sedation state[20–22] and, more recently, possibly the nociception state.[23–25] Since these different EEG-derived indices rely on the same physiological principles as the TCI models, they provide similar estimates. Regardless of the anaesthesia method used (intravenous or inhalation), EEGrelated technologies offer a valuable complementary real-time tool to avoid AAGA episodes and excessive

anaesthesia, individualising the administration of drugs to each patient.

What Do Depth of Anaesthesia Monitors Measure? General anaesthetic drugs, whether inhaled or intravenous, inhibit or block excitatory neurotransmission, acting on specific neurotransmitters and voltage-gated ion channels at the neuronal level. These agents produce a widespread depression of the central nervous system and can cause multiple responses, such as unconsciousness, analgesia, amnesia and immobility, depending on which specific neurotransmitters, synaptic receptors or neuronal pathways the drugs target. The EEG reflects the neuronal activity of hundreds of thousands of synchronised cortical pyramidal cells[26] recorded by a surface electrode on the scalp. The induced changes in neural neurotransmission by general anaesthetic drugs affect the synchronisation of the pyramidal neurons, inhibiting their communication with other neurons and brain areas such as the thalamus and hippocampus. Therefore the widespread inhibition at a micro level results in significant changes at the macro level of brain activity. In other words, the druginduced changes in the EEG wave patterns reflect brain activity changes, which in turn correlate with the anaesthetic effect. These changes depend on multiple factors such as, among others, the intrinsic mechanism of action of the particular drug, effect-site concentrations, drug interactions and inter-patient variability. Understanding the main qualitative changes of the EEG patterns is complex but the technology underlying different DOA monitors tracks the qualitative and quantitative changes of the complex EEG druginduced wave patterns and processes them into useful indices describing the patient state. As an example, we will describe the main EEG patterns under TCI TIVA with propofol and remifentanil. Figure 9.1(a) shows the effect-site (Ce) profile during the intra-operative period for propofol and remifentanil as calculated by the Schnider and Minto PK/PD models, respectively. Figure 9.1(b) shows the corresponding graphical trends of four indices produced by different DOA monitors: the BIS and qCON estimating measuring the depth of sedation; the qNOX estimating nociception; and the burst suppression ratio (BSR) measuring the degree of burst suppression in the EEG. The BIS parameter is provided

Downloaded from https://www.cambridge.org/core. University of Winnipeg, on 02 Dec 2019 at 21:47:20, subject to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316659069.011

65

Chapter 9: TIVA and Awareness

(a) 10 Ce μg/ml ng/ml

Ce Propofol Ce Remifentanil

5

0

(b)

0

1000

2000 3000 time (s)

4000

awake 100 qCON BIS qNOX BSR

60 50

0 LOC 0

BSR 1000

surgery 2000

3000

4000

time (s) Figure 9.1 (a) Effect-site (Ce) profile of a TCI TIVA intra-operative procedure with propofol and remifentanil; (b) EEG-derived indices for sedation, nociception and burst suppression (BIS, qCON, qNOX and BSR, respectively).

by the BIS™ monitor manufactured by Medtronic and indexes qCON and qNOX are included in the Conox® monitor by Quantium Medical/Fresenius Kabi. Both monitors include a measure of the BSR. Four events have been defined to illustrate the main changes in the EEG during a surgical procedure: when the patient is still conscious (awake); during the induction at LOC; at the beginning of the surgery; and during the induction when the EEG presents with the burst suppression pattern (BSR). Propofol mainly blocks the calcium ion channels, inhibits the N-methyl-D-aspartate and potentiates the GABA receptor activity of the neurons, producing a net inhibition that slows brain activity, which in turn produces a smoother EEG pattern in the time domain. Figure 9.2(a) shows examples of EEG signals while the patient is still awake, at LOC and during surgery. It can be seen that as the concentration of propofol increases the EEG signal becomes smoother. The change in the EEG signals, over time, has a corresponding effect in the frequency domain as seen in the power spectral density (Figure 9.2(b)). For

more details on how EEG time signals translate into frequency representation, see Rampil (1998).[27] Propofol shifts the EEG spectrum content to the lower frequencies, since it diminishes the frequency components in the beta and gamma bands. At certain concentrations, propofol induces periodic oscillations around 10 Hz, which correspond to the alpha band (see lower panel, Figure 9.2(a) (surgery) and 9.2(b)). Finally, higher concentrations of propofol induce major changes in the EEG waveform pattern. The EEG activities alternate between periods of highvoltage activities with periods of no activity. These patterns are called burst suppression (BS), where amplitude is consistently less than 5 μV for more than 500 ms. The amount of BS in the EEG signal is usually quantified by the BSR index, defined as the percentage of time that a EEG displays electrical suppression in a fixed period. Figure 9.2(c) shows 30 s of EEG with flat activity periods lasting around 32% of the time (BSR = 32). At higher concentrations of propofol, the BSR values increase. Some of the described changes of the EEG waveforms are not only due to propofol. Remifentanil, as a mu-receptor agonist, also has a net inhibitory effect that is analogous and complementary to propofol effects on the brain cortical activity,[28,29] with decreased activities for frequencies higher than 14 Hz (β and γ bands) and increase in α activity. Some of the druginduced EEG waveform changes can be assessed qualitatively by a trained practitioner; often these changes in EEG activity are very subtle, making it necessary to have a monitoring system that quantifies the diverse EEG changes combining them into a parameter that simplifies the assessment of the patient state.

Correlates with DOA Monitor Indices A number of different processed EEG monitors are available on the market with indices designed to estimate the depth of sedation (BIS-Bispectrum; qCON; SE-State Entropy; Narcotrend; PSI-Patient State Index) and an index estimating the EEG link to the nociception level (qNOX). These different indices were developed despite a lack of a gold standard as information regarding the complex mechanisms during the process of anaesthesia is still being uncovered. Despite some differences among the monitors,[30,31] there are similarities in the assessment of the patient’s state[32] among the available processed EEG monitors. The agreement across different indices can be evaluated with a Bland– Altman plot (Figure 9.3(a)), and the prediction

Downloaded from https://www.cambridge.org/core. University of Winnipeg, on 02 Dec 2019 at 21:47:20, subject to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316659069.011

66

Chapter 9: TIVA and Awareness

(a)

δ θ α

0

–20

–0.1 0.06 0.02 0 –0.02

γ

α-activity

Ce (propofol) = 2.2 μg/ml

LOC

β

–25 dB

mV

(b)

Ce (propofol) = 0 μg/ml

awake

0.1

Ce (propofol) = 3 μg/ml

surgery

–30

0.05 0 0

0.5

(c)

mV

0.05

freq (Hz)

1

1.5 2 time (s)

2.5

3

0

10

20 30 frequency (Hz)

40

Ce (propofol) = 5.8 μg/ml

BS

BSR = 32

0

–0.05 (d)

–35

α-activity ~100ms (10 Hz)

–0.05

0

10

awake

time (s)

20

30 dB

50

–24

40

–28

30

–32

20

–36

10 0

Figure 9.2 (a) Raw EEG activity during 3 s at different propofol concentrations, corresponding to three time-events (awake, LOC, surgery) selected from the intraoperative procedure given in Figure 9.1. (b) Power spectral density in decibels (dB), logarithm of power, of the selected raw EEG activities. The conventional waveforms’ subdivision into frequency bandwidths are shown on top of the axis (δ 28 Hz). (c) EEG activity (30 seconds) under high concentrations of propofol, with suppression periods of EEG activity. Burst suppression (BS) is induced from a high concentration of propofol. In this case, the percentage of time the activity is diminished (burst suppression rate, BSR) is around 32. (d) Density spectral array (DSA) for the whole intra-operative procedure, representing the power spectral density every second.

0

LOC BS

1000

2000

3000

4000

–40

time (s)

probability (pK) or the Spearman correlation (Rs). For instance, several comparison studies showed high agreement between BIS and qCON, Rs = 0.89[32] or pK = 0.92.[24] Brain monitoring during anaesthesia is important since the main target organ is the brain and, more importantly, brain-related indices show a strong correlation between anaesthetic concentrations and patient response. Figure 9.3(b) shows the relation between the qCON index and the propofol effect-site concentration (given by the Schnider TCI model) under TIVA anaesthesia with remifentanil as the analgesic drug, for n = 160 patients. The Spearman correlation is Rs = –0.63 (p