METHODS TO IMPROVE ANESTHETIC DRUG MANAGEMENT. Sandeep Choudary Manyam

Transcription

1 METHODS TO IMPROVE ANESTHETIC DRUG MANAGEMENT by Sandeep Choudary Manyam A dissertation submitted to the faculty of The University of Utah in partial fulfillment of the requirements for the degree of Doctor of Philosophy Department of Bioengineering The University of Utah December 2006

2 Copyright Sandeep Manyam 2006 All Rights Reserved

3 ABSTRACT Modern day anesthesia involves the use of multiple drugs simultaneously to maintain insensitivity to pain or analgesia, lack of awareness of the surgical procedure and suppression of autonomic responses. The sedative component of anesthesia is primarily provided by using a hypnotic drug (volatile or intravenously administered) and the analgesic component is provided by an opioid (primarily intravenously administered). The level of anesthetic effect produced by these drugs is assessed by the use of a multitude of physiologic responses such as heart rate, blood pressure, movement etc. The response dynamics of these indicators are typically non linear and change with the combination of anesthetics being used. The potency of drugs also vary among patients and across age groups. These factors make the accurate titration of anesthetic drugs challenging. Accurate titration of anesthetic drug such that the effect is just enough to cause unconsciousness and immobility in the patient helps to avoid adverse effects such as delayed emergence, awareness during the procedure, hyper variable cardiovascular state and memory loss that is thought to be associated with under or overdose. This work aims to improve anesthetic drug management through efficient drug delivery and real time monitoring. The first goal is to improve drug delivery and clinical outcomes for the average patient by identifying combinations of sedative and analgesic drugs that ensure fast recovery from anesthesia. Although the combinations are suitable to be applied in clinical practice they may not be effective when applied to individual

4 patients that are outliers (such as those who use chronic pain medication). The drug dose in such patients can be titrated by assessing the depth of anesthesia in real time. The second goal is to test the ability of emerging depth of anesthesia monitoring technologies to assess each patient s anesthetic state. Real time monitors of anesthetic effect can help the clinicians refine their dosing strategy and predict adverse events such as intraoperative awareness or patient responses to pain. v

5 CONTENTS ABSTRACT... iv ACKNOWLEDGEMENTS...viii 1. INTRODUCTION Goals References OPIOID-VOLATILE ANESTHETIC SYNERGY AND CONTEXT SENSITIVE TARGETS Abstract Introduction Materials and Methods Results Discussion Appendix A: The Logit Model for Pharmacodynamics Appendix B: Pharmacokinetic and Pharmacodynamic Simulations References CONTEXT SENSITIVE TARGETS FOR OPIOIDS AND INTRAVENOUS ANESTHETICS Abstract Introduction Materials and Methods Results Discussion References PROCESSED EEG TARGETS REQUIRED FOR ADEQUATE ANESTHESIA Abstract Introduction Materials and Methods Results

6 4.5 Discussion References PROCESSED EEG SIGNALS AS INDICATORS OF INADEQUATE ANESTHESIA Abstract Introduction Materials and Methods Results Discussion References SUMMARY AND CONCLUSIONS Summary Conclusions Impact Future Work vii

7 ACKNOWLEDGEMENTS I would like to acknowledge a number of people for their help and support during my doctoral work. Foremost, of course is my advisor, Dr. Dwayne Westenskow. Throughout my doctoral work he encouraged me to work on ideas that had practical applications in clinical anesthesia. Ever since I entered his laboratory he placed extreme confidence in me and provided me with limitless opportunities. He greatly assisted me with developing my scientific communication skills and translating my ideas into viable research grants. I am also grateful to a number of anesthesiologists who taught me all I know about clinical anesthesia and conducting clinical research. Dr. Talmage Egan, for spending countless hours in reviewing and helping me interpret my results and providing the direction to make my work clinically relevant and innovative. For the creative freedom he gave me while simultaneously insisting on the highest standards for both form and content. His simple words of encouragement -- keep up the good work when I had not shown him results for ages were an additional incentive for me to work harder. I am thankful for his efforts in providing me with the broad perspective with which I could relate any specific problem I was working on, to anesthesiology and patient care as a whole. I was deeply touched by his kindness and generosity with ideas and financial support.

8 Dr. Dhanesh Gupta, for his enthusiastic guidance and his step by step involvement in translating a paper napkin idea in to an exciting high-impact manuscript. Without his energy and emphasis on completion, I would be forever lost in refining my data analysis. For the numerous operating room breaks and weekends he decided to spend with me to ensure my simulations were meaningful. Dr. Ken Johnson, for his insightful comments on my results and constant encouragement. For the number of hours he spent helping me understand anesthetic dosing. Dr. Nathan Pace, for insisting on the right statistical methods at every stage. For allowing me to pick his brain at will and always helping me with a solution or pointing me in a direction in which I could find one. I am greatly appreciative of the committee members for their time and effort in not only clarifying my research ideas but also in ensuring that I receive a well rounded education. Dr. Richard Normann, who was incidentally the first professor whom I met in Utah, helped me continually in identifying my interests and helping me define my long term career goals. I am still in awe of his child-like enthusiasm when conducting laboratory research and hope that I am able to bring that level of energy in to my own experiments some day. I have always relied on his frank opinion and guidance throughout my graduate education and hope this will continue for years to come. Dr. Steve Kern, for his support with any engineering dilemmas and his stimulating discussions on pharmacodynamic models and methods. For the confidence he gave me by just being there for I knew that if I was stuck with a problem Dr. Kern could probably bail me out. x

9 Dr. Rob Macleod who, along with Dr. Patrick Tresco, taught a set of classes that formed the foundation of my graduate education. What made these two classes stand out was not just the content that was taught but their emphasis on the manner in which an engineer-scientist ought to approach a biological system. As I set out on my academic career their teaching style will always be the standard that I would try to achieve. I would also like to thank Rob for offering me a teaching assistantship. This enabled me to observe his teaching methods at close hand and also came at a time when I was faced with funding problems. Dr. Srikantan Nagarajan, who taught me so much about the basic principles of conducting research. So many of the concepts I learnt in his laboratory extend far beyond. Sri insisted in making the most out of any experiment. He would insist that every experiment whether a failure or a success needs to count toward my own as well as the society s learning process. His words focus on the science still ring in my ears and motivate me when I am frustrated with a research problem. Drs. Gregory Clark and Kenneth Horch, for their advice during the crucial days when I was faced with major decisions during my graduate studies. Dr. Clark for his particular emphasis on rigorous experimental techniques and personal attention to my experimental skills and writing techniques in the neural interfaces laboratory. Julia White, our research nurse, who was involved in all steps of planning the study, volunteer recruitment and data collection. Without her attention to detail these studies would have been monumentally difficult. xi

10 Noah Syroid and Jim Agutter at MedVis, for their continued support over the years. I will miss the informal discussions with Noah and his active participation in research conferences. The administrative staff at the anesthesiology and bioengineering departments. Specifically, Jeff Mann, Vicki Larsen, Karen Terry, Paul Dryden and Linda Twitchell among many for others their logistical and technical help. My past and present lab mates at the anesthesia bioengineering laboratory for their friendship and their insightful discussions. Finally, I deeply appreciate the unwavering support and encouragement my family and friends. My mother, Deepa Choudary, for her encouragement to explore and young age. For her numerous personal sacrifices to keep us oblivious to other problems. My sister, Kinnera Krishna, who has always encouraged me to pursue my interests no matter what the costs. My fiancée, Nirupama Ramkumar, who endured frustrating and good times with me through the various stages of graduate school. With great pleasure and gratitude I dedicate my work to them xii

11 CHAPTER 1 INTRODUCTION In modern clinical practice, anesthesia comprises of three main componentsinsensitivity to pain or analgesia, lack of awareness of the surgical procedure and suppression of autonomic responses. This is achieved by using different classes of drugs simultaneously. The analgesic component is most commonly provided by opioids which are primarily delivered intravenously. Lack of awareness or sedation is achieved by a hypnotic drug. The hypnotic agent may be administered through a vaporizer for volatile agents (ex. sevoflurane) or by using an infusion pump for intravenous drugs (ex. propofol). In addition to sedation and analgesia, muscle relaxants are used to suppress somatic motor responses. Certain hypnotic drugs alone can often produce surgically adequate anesthesia albeit at higher concentrations. 13 This approach, which was common in the past, is often associated with excessive hemodynamic depression 14 and other undesirable side effects of administration of high doses of the hypnotic drug for a long period of time (e.g., prolonged time to awakening from anesthesia, etc.). 15 Thus, for practical purposes, the current state of the art is to produce anesthesia with an opioid and a sedative in combination. 1 Interaction is observed among many drugs used in anesthetic practice. The addition of opioid reduces the concentration of the hypnotic drug required to produce

12 2 sedation. 3,16-29 Similarly the presence of a hypnotic drug enhances analgesia and reduces the opioid requirements. Although, anesthetic drug interactions were widely studied in the past, 17,22,27,28,30,31 it is only more recently that they have been quantified by the use of a mathematical model. 2,3,23,25,32-36 The pharmacodynamic interaction models relate the concentration of the two drugs to the level of effect they produce. These models can help clinicians determine if a certain dose combination of hypnotic and opioid will provide adequate sedation and analgesia. Response surface models allow the complete characterization of pharmacodynamic interactions over the entire spectrum of possible concentration pairs 32,33 instead of just a single level of drug effect such as a 50% probability of nonresponsiveness to surgical incision (e.g., Minimum Alveolar Concentration, MAC). Short, et al. describe a crisscross sampling method 37 which can be used to sample drug concentration pairs needed to construct a response surface. Response surface pharmacodynamic interaction methods provide a framework to define and explore these issues. However such methods have not been used to study the interaction between volatile anesthetics and opioids. These models can also form the basis for the development of a real-time pharmacokinetic-pharmacodynamic display system. 38 The choice of anesthetic drugs and their clinically effective concentrations is based on a number of factors. The opioid is selected based on a combination of the potency and the speed of decay of the drug at the effect site. 39 For shorter procedures a drug with rapid induction and a very short half life, such as remifentanil (t 1/2 = 0.9 min.) is ideal. For longer procedure a long acting drug such as, fentanyl (t 1/2 = 4.7 min.) or sufentanil (t 1/2 = 5.9 min.) may be preferred. The hypnotic drug is selected based on the patient s preexisting clinical conditions, the intensity and duration of procedure. The drug

13 3 dosage is computed using the patient parameters that influence uptake and delivery, such as age, weight, height, etc. The drug dose regimen that is determined based on knowledge of clinical testing of the drug is adapted intraoperatively to suit the patient. The accurate titration of drugs such that the level of drug is just enough to block responses is highly desirable. This enables the clinician to provide an adequate level of anesthesia within the operating room and facilitate rapid recovery once the procedure has ended. Several factors such as interpatient pharmacokinetic and pharmacodynamic variability make this task challenging. Pharmacokinetic variability can be described as the variation in the uptake and distribution of drug between patients. It is on the order of 70% (i.e. with an infusion rate of 10 mg/kg/hr of propofol the blood concentration may vary from 3 to 5 mg/l in patients). Differences in cardiac output, hepatic perfusion, enzyme activity and protein binding contribute to this variability. 6,40-46 Pharmacodynamic variability can be described as the variation of the potency of the drug in each patient. Several investigators have quantified this variation to be on the order of % (i.e., some patients may lose consciousness at a blood concentration of 1 mg/l while other s may need as much as 5 mg/l before they are sedated). The factors responsible for pharmacodynamic variability are still unclear although some investigators suspect the variability arises from genetic differences in receptor pharmacology. 55 Clinicians cope with this combined variability by adjusting the drug dose to suppress patient responses. These limitations necessitate the development of methodical schemes to determine the dose for combinations of anesthetic drugs that will work in all types of patients. To determine the level of anesthesia clinicians often depend on unreliable, nonspecific measures 56 such as hemodynamics, reflexes to stimuli, spontaneous

14 4 respiration rate, etc. to determine the level of anesthetic effect. To use these methods, the clinician is dependent on a number of factors such as training, experience and availability of intraoperative monitoring methods. It is difficult to monitor some measures such as blood pressure as a continuous signal intraoperatively. Hemodynamic responses are often affected by the presence of vasoactive and ionotropic drugs. 57 The lack of definite indicators for sedation and analgesia make the precise delivery of anesthetics drugs challenging. The use of patient responses to accurately titrate anesthetic drugs intraoperatively is not viable ethically, as eliciting patient responses may cause patient discomfort. Thus, many clinicians often chose to operate with a more than adequate level of drug to prevent patient awareness and responses. Even though there are no direct adverse effects with using this range of concentrations they may result in delayed emergence and higher operating costs. A real-time monitoring system may address may address many of these concerns. It is well understood that patterns within the electroencephalogram (EEG) are good correlates to clinical endpoints such as loss of consciousness. 58, 59 Despite this, EEG monitors have not been widely accepted intraoperatively by anesthesiologists. The primary reasons are (1) EEG is a data intensive signal and analysis in real time is tedious, (2) Large inherent variability in the signal (3) lack of clear guidelines to assess changing levels of sedation and (4) Not all drugs produce a similar effect on the EEG at a given clinical endpoint (loss of consciousness). These limitations are somewhat addressed by CNS effect monitors that extract salient features of the EEG waveform that correlate well with depth of anesthesia and quantify them in to a index.

15 5 The processed EEG has emerged as an important surrogate measure of CNS drug effect. 11, 12 Surrogate measures are employed when the clinical drug effect of interest is difficult or impossible to measure. The processed EEG has many characteristics of the ideal surrogate. In contrast to more clinically oriented measures of drug effect, it is an objective, continuous, reproducible, noninvasive, high resolution signal. It can also be used as an effect measure when an experimental subject is unconscious or apneic, whereas many of the more clinically oriented measurements require awake, cooperative subjects. The processed EEG signal has been commercialized by number of manufacturers. Preliminary studies validating the bispectral index (BIS), reported the concentration-bis relationship and examined the ability of the BIS monitor to track sedation. 12 A major limitation of several such studies is that they report the predictive performance of the BIS monitor when drugs are used in isolation. Since modern anesthesia calls for a balanced sedation and analgesia, opioids are almost ubiquitous in pain management. A study that evaluates such monitoring technologies must replicate the clinical environment in which they are intended for use. Although the ability of processed EEG monitors to track the sedative state has been extensively studied, the ability of these monitors to detect pain in patients who are undergoing a surgical procedure has not been reported. If processed EEG monitors correlate with patient responses to pain, they will be an invaluable tool to identify inadequate anesthesia in patients when traditional markers such as movement and heart rate are obscured by the presence of other drugs. Recent advances in drugs, monitoring technology and combined pharmacologic knowledge have shown that drugs can be improved in clinical anesthetic practice. Accurate knowledge of the drug disposition and a method of feedback of the analgesic

16 and sedative drug effect may eventually lead to the development of a closed loop 60, 61 computer controlled anesthesia system Goals This dissertation aims to improve anesthetic drug management in two steps. Pharmacokinetic and pharmacodynamic models can be used to predict the level of sedation and analgesia in a patient. The first step is to construct pharmacodynamic models for a commonly used opioid (remifentanil) and volatile hypnotic drug (sevoflurane). We can then use simulations based on these models, to identify certain factors which when applied to anesthetic practice will improve clinical outcomes. Specifically simulations will be used to identify a combination of opioid and hypnotic that will provide adequate anesthesia and enable the patient to regain sensation quickly after the procedure. Further, these models will help understand the combined effects of volatile anesthetics and opioids. Our second goal is to provide the clinician with a means for feedback of the patient s anesthetic state within the operating room. To achieve this we will test emerging technologies in their ability to monitor adequate anesthesia and their ability to detect patient responses. Understanding the operating characteristics of such monitors will improve intraoperative monitoring and enable more accurate drug administration. Chapters 2 and 3 of this dissertation describe pharmacodynamic models that estimate the interaction between commonly used hypnotic and opioid drugs. Chapter 2 in specific describes the interaction between a volatile agent and an opioid drug. Chapter 2 fills in an important void in our understanding of volatile anesthetic and opioid interactions. The quantitative description of analgesic and sedative effect caused by the

17 7 combinations of drugs can be extended to other volatile anesthetics and opioids. Chapter 2 also introduces an optimization technique used to estimate context sensitive optimal combinations that ensure adequate anesthesia by targeting drug doses that produce sedation and analgesia in a wide patient population and speed up emergence. After further validation, the clinical application of these results will lead to accurate anesthetic dosing in the general patient population. Chapter 3 extends the methods described in Chapter 2 to estimate optimal combinations of an intravenous hypnotic drug and an opioid. Chapter 3 introduces methods by which number of clinical endpoints (adequate sedation, analgesia and rapid emergence) can be ensured simultaneously through drug optimization. These techniques can be extended to wide range of anesthetic procedures that require a particular level of sedation and analgesia (e.g., outpatient procedures that are common in a gastroenterology clinic have specific sedation and analgesia requirements that differ from the typical surgical procedure). This technique can also be used to ensure other desirable clinical outcomes such as minimizing cost of anesthetics, minimal respiratory depression or preventing side effects such as nausea that are associated with a specific drug concentration. Chapters 4 and 5 examine the performance of two emerging processed electroencephalographic (EEG) monitors that can be used to determine the depth of anesthesia in real-time. In Chapter 4, the ability to monitor depth of sedation is studied. Processed EEG monitor targets that coincide with adequate analgesia and sedation are described. The manufacturers of processed EEG monitors recommend certain monitor indices that are associated with adequate sedation, the results presented in this chapter prove that the monitor index associated with adequate sedation varies as function of the

18 8 combination of drugs used to provide anesthesia. These limitations are addressed by the suggesting processed EEG monitor targets associated with adequate anesthesia. Chapter 5 examines the changes in processed EEG monitor indices in response to stimulation. The results of this exploratory study highlight the need for further algorithm development in the processed EEG monitors. Finally, Chapter 6 summarizes important conclusions from this work and suggests future work in this area of research References 1. Eger EI, 2nd, Saidman LJ, Brandstater B: Minimum alveolar anesthetic concentration: a standard of anesthetic potency. Anesthesiology 1965; 26: Zbinden AM, Petersen-Felix S, Thomson DA: Anesthetic depth defined using multiple noxious stimuli during isoflurane/oxygen anesthesia. II. Hemodynamic responses. Anesthesiology 1994; 80: Zbinden AM, Maggiorini M, Petersen-Felix S, Lauber R, Thomson DA, Minder CE: Anesthetic depth defined using multiple noxious stimuli during isoflurane/oxygen anesthesia. I. Motor reactions. Anesthesiology 1994; 80: Kissin I: General anesthetic action: an obsolete notion? Anesth Analg 1993; 76: Bouillon T, Schmidt C, Garstka G, Heimbach D, Stafforst D, Schwilden H, Hoeft A: Pharmacokinetic-pharmacodynamic modeling of the respiratory depressant effect of alfentanil. Anesthesiology 1999; 91: Brunner MD, Braithwaite P, Jhaveri R, McEwan AI, Goodman DK, Smith LR, Glass PS: MAC reduction of isoflurane by sufentanil. Br J Anaesth 1994; 72: Egan TD, Minto C: Common Pharmacodynamic Drug Interactions in Drug Practice. Anesthetic Pharmacology: Physiologic Principles and Clinical Practice 2004; Chap. 6: Glass PS, Gan TJ, Howell S, Ginsberg B: Drug interactions: volatile anesthetics and opioids. J Clin Anesth 1997; 9: 18S-22S 9. Katoh T, Kobayashi S, Suzuki A, Kato S, Iwamoto T, Bito H, Sato S: Fentanyl augments block of sympathetic responses to skin incision during sevoflurane anaesthesia in children. Br J Anaesth 2000; 84: 63-6

19 10. Katoh T, Nakajima Y, Moriwaki G, Kobayashi S, Suzuki A, Iwamoto T, Bito H, Ikeda K: Sevoflurane requirements for tracheal intubation with and without fentanyl. Br J Anaesth 1999; 82: Kazama T, Ikeda K, Morita K: Reduction by fentanyl of the Cp50 values of propofol and hemodynamic responses to various noxious stimuli. Anesthesiology 1997; 87: Kern SE, Xie G, White JL, Egan TD: A response surface analysis of propofol-remifentanil pharmacodynamic interaction in volunteers. Anesthesiology 2004; 100: Mertens MJ, Olofsen E, Engbers FH, Burm AG, Bovill JG, Vuyk J: Propofol reduces perioperative remifentanil requirements in a synergistic manner: response surface modeling of perioperative remifentanil-propofol interactions. Anesthesiology 2003; 99: Mertens MJ, Vuyk J, Parivar K, Engbers FH, Burm AG, Bovill JG: Pharmacodynamic interaction of eltanolone and alfentanil during lower abdominal surgery in female patients. Br J Anaesth 1999; 83: Minto CF, Schnider TW, Short TG, Gregg KM, Gentilini A, Shafer SL: Response surface model for anesthetic drug interactions. Anesthesiology 2000; 92: Nieuwenhuijs DJ, Olofsen E, Romberg RR, Sarton E, Ward D, Engbers F, Vuyk J, Mooren R, Teppema LJ, Dahan A: Response surface modeling of remifentanilpropofol interaction on cardiorespiratory control and bispectral index. Anesthesiology 2003; 98: Sebel PS, Glass PS, Fletcher JE, Murphy MR, Gallagher C, Quill T: Reduction of the MAC of desflurane with fentanyl. Anesthesiology 1992; 76: Smith C, McEwan AI, Jhaveri R, Wilkinson M, Goodman D, Smith LR, Canada AT, Glass PS: The interaction of fentanyl on the Cp50 of propofol for loss of consciousness and skin incision. Anesthesiology 1994; 81: 820-8; discussion 26A 19. Vuyk J, Engbers FH, Burm AG, Vletter AA, Griever GE, Olofsen E, Bovill JG: Pharmacodynamic interaction between propofol and alfentanil when given for induction of anesthesia. Anesthesiology 1996; 84: Katoh T, Ikeda K: The effects of fentanyl on sevoflurane requirements for loss of consciousness and skin incision. Anesthesiology 1998; 88:

20 21. Katoh T, Kobayashi S, Suzuki A, Iwamoto T, Bito H, Ikeda K: The effect of fentanyl on sevoflurane requirements for somatic and sympathetic responses to surgical incision. Anesthesiology 1999; 90: Minto C, Vuyk J: Response surface modelling of drug interactions. Adv Exp Med Biol 2003; 523: Greco WR, Bravo G, Parsons JC: The search for synergy: a critical review from a response surface perspective. Pharmacol Rev 1995; 47: Nieuwenhuijs D, Sarton E, Teppema LJ, Kruyt E, Olievier I, van Kleef J, Dahan A: Respiratory sites of action of propofol: absence of depression of peripheral chemoreflex loop by low-dose propofol. Anesthesiology 2001; 95: Dahan A, Nieuwenhuijs D, Olofsen E, Sarton E, Romberg R, Teppema L: Response surface modeling of alfentanil-sevoflurane interaction on cardiorespiratory control and bispectral index. Anesthesiology 2001; 94: Bouillon TW, Bruhn J, Radulescu L, Andresen C, Shafer TJ, Cohane C, Shafer SL: Pharmacodynamic interaction between propofol and remifentanil regarding hypnosis, tolerance of laryngoscopy, bispectral index, and electroencephalographic approximate entropy. Anesthesiology 2004; 100: Berenbaum MC: What is synergy? Pharmacol Rev 1989; 41: Short TG, Ho TY, Minto CF, Schnider TW, Shafer SL: Efficient trial design for eliciting a pharmacokinetic-pharmacodynamic model-based response surface describing the interaction between two intravenous anesthetic drugs. Anesthesiology 2002; 96: Syroid ND, Agutter J, Drews FA, Westenskow DR, Albert RW, Bermudez JC, Strayer DL, Prenzel H, Loeb RG, Weinger MB: Development and evaluation of a graphical anesthesia drug display. Anesthesiology 2002; 96: Shafer SL, Varvel JR: Pharmacokinetics, pharmacodynamics, and rational opioid selection. Anesthesiology 1991; 74: Bouillon T, Shafer SL: Does size matter? Anesthesiology 1998; 89: Egan TD, Huizinga B, Gupta SK, Jaarsma RL, Sperry RJ, Yee JB, Muir KT: Remifentanil pharmacokinetics in obese versus lean patients. Anesthesiology 1998; 89:

21 33. Ausems ME, Stanski DR, Hug CC: An evaluation of the accuracy of pharmacokinetic data for the computer assisted infusion of alfentanil. Br J Anaesth 1985; 57: Kuipers JA, Boer F, de Roode A, Olofsen E, Bovill JG, Burm AG: Modeling population pharmacokinetics of lidocaine: should cardiac output be included as a patient factor? Anesthesiology 2001; 94: Kuipers JA, Boer F, Olieman W, Burm AG, Bovill JG: First-pass lung uptake and pulmonary clearance of propofol: assessment with a recirculatory indocyanine green pharmacokinetic model. Anesthesiology 1999; 91: Maitre PO, Ausems ME, Vozeh S, Stanski DR: Evaluating the accuracy of using population pharmacokinetic data to predict plasma concentrations of alfentanil. Anesthesiology 1988; 68: Maitre PO, Vozeh S, Heykants J, Thomson DA, Stanski DR: Population pharmacokinetics of alfentanil: the average dose-plasma concentration relationship and interindividual variability in patients. Anesthesiology 1987; 66: Minto CF, Schnider TW, Egan TD, Youngs E, Lemmens HJ, Gambus PL, Billard V, Hoke JF, Moore KH, Hermann DJ, Muir KT, Mandema JW, Shafer SL: Influence of age and gender on the pharmacokinetics and pharmacodynamics of remifentanil. I. Model development. Anesthesiology 1997; 86: Bailey PL, Rhondeau S, Schafer PG, Lu JK, Timmins BS, Foster W, Pace NL, Stanley TH: Dose-response pharmacology of intrathecal morphine in human volunteers. Anesthesiology 1993; 79: 49-59; discussion 25A 40. Bouillon T, Bruhn J, Radu-Radulescu L, Andresen C, Cohane C, Shafer SL: Mixed-effects modeling of the intrinsic ventilatory depressant potency of propofol in the non-steady state. Anesthesiology 2004; 100: Drover DR, Lemmens HJ: Population pharmacodynamics and pharmacokinetics of remifentanil as a supplement to nitrous oxide anesthesia for elective abdominal surgery. Anesthesiology 1998; 89: Egan TD: Remifentanil pharmacokinetics and pharmacodynamics. A preliminary appraisal. Clin Pharmacokinet 1995; 29: Minto C, Schnider T: Expanding clinical applications of population pharmacodynamic modelling. Br J Clin Pharmacol 1998; 46: Ropcke H, Wirz S, Bouillon T, Bruhn J, Hoeft A: Pharmacodynamic interaction of nitrous oxide with sevoflurane, desflurane, isoflurane and enflurane in 11

22 surgical patients: measurements by effects on EEG median power frequency. Eur J Anaesthesiol 2001; 18: Schnider TW, Minto CF, Bruckert H, Mandema JW: Population pharmacodynamic modeling and covariate detection for central neural blockade. Anesthesiology 1996; 85: Somma J, Donner A, Zomorodi K, Sladen R, Ramsay J, Geller E, Shafer SL: Population pharmacodynamics of midazolam administered by target controlled infusion in SICU patients after CABG surgery. Anesthesiology 1998; 89: Kharasch ED, Jubert C, Senn T, Bowdle TA, Thummel KE: Intraindividual variability in male hepatic CYP3A4 activity assessed by alfentanil and midazolam clearance. J Clin Pharmacol 1999; 39: Schneider G, Sebel PS: Monitoring depth of anaesthesia. Eur J Anaesthesiol Suppl 1997; 15: Berne RM, Levy MN: Physiology. Fourth Edition, Mosby Rampil IJ, Lockhart SH, Eger EI, 2nd, Yasuda N, Weiskopf RB, Cahalan MK: The electroencephalographic effects of desflurane in humans. Anesthesiology 1991; 74: Rampil IJ: A primer for EEG signal processing in anesthesia. Anesthesiology 1998; 89: Gan TJ, Glass PS, Windsor A, Payne F, Rosow C, Sebel P, Manberg P: Bispectral index monitoring allows faster emergence and improved recovery from propofol, alfentanil, and nitrous oxide anesthesia. BIS Utility Study Group. Anesthesiology 1997; 87: Glass PS, Bloom M, Kearse L, Rosow C, Sebel P, Manberg P: Bispectral analysis measures sedation and memory effects of propofol, midazolam, isoflurane, and alfentanil in healthy volunteers. Anesthesiology 1997; 86: Locher S, Stadler KS, Boehlen T, Bouillon T, Leibundgut D, Schumacher PM, Wymann R, Zbinden AM: A new closed-loop control system for isoflurane using bispectral index outperforms manual control. Anesthesiology 2004; 101: Glass PS, Rampil IJ: Automated anesthesia: fact or fantasy? Anesthesiology 2001; 95:

23 CHAPTER 2 OPIOID-VOLATILE ANESTHETIC SYNERGY AND CONTEXT SENSITIVE TARGETS 2.1 Abstract Background Combining a hypnotic and an analgesic to produce sedation, analgesia, and surgical immobility required for clinical anesthesia is more common than administration of a volatile anesthetic alone. The aim of this study was to apply response surface methods to characterize the interactions between remifentanil and sevoflurane Methods Sixteen adult volunteers received a target controlled infusion of remifentanil (0-15 ng ml -1 ) and inhaled sevoflurane (0-6 vol %) at various target concentration pairs. After reaching pseudo-steady-state drug levels, the Observer's Assessment of Alertness/Sedation score and response to a series of randomly applied experimental pain stimuli (pressure algometry, electrical tetany, and thermal stimulation) were observed for each target concentration pair. Response surface pharmacodynamic interaction models were built using the pooled data for sedation and analgesic endpoints. Using computer Accepted for publication in Anesthesiology, February Reprinted with permission from Anesthesiology. Copyright 2006, American Society of Anesthesiologists. Original article titled: Opioid-volatile anesthetic synergy: A response surface model with remifentanil and sevoflurane as prototypes.

24 14 simulation, the pharmacodynamic interaction models were combined with previously reported pharmacokinetic models to identify the combination of remifentanil and sevoflurane that yielded the fastest recovery (Observer s Assessment of Alertness/Sedation score 4) for anesthetics lasting minutes Results Remifentanil synergistically decreased the amount of sevoflurane necessary to produce sedation and analgesia. Simulations revealed that as the length of the procedure increased, faster recovery was produced by concentration target pairs containing higher amounts of remifentanil. This trend plateaued at a combination of 0.75 vol % sevoflurane and 6.2 ng ml -1 remifentanil Conclusion Response surface analyses demonstrate a synergistic interaction between remifentanil and sevoflurane for sedation and all analgesic endpoints Acknowledgements Supported in part by a research grant from Alaris Medical Systems, Inc., San Diego, CA, U.S.A. (TDE) and by the National Institute of Biomedical Imaging and Bioengineering of the National Institutes of Health 8 RO1 EB00294 (SCM and DRW). Portions of this work have been presented at the 79 th Annual Clinical and Scientific Congress of the International Anesthesia Research Society in Honolulu, HI, March 15, 2005, (Poster S-405) and the 80 th Annual Clinical and Scientific Congress of the International Society of Anesthesia Research in San Francisco, CA, March 27, 2006.

25 15 The authors would like to thank Steve E. Kern, Ph. D. (Associate Professor, Departments of Pharmaceutics and Anesthesiology, University of Utah), for his insightful comments and feedback in the preparation of this manuscript. 2.2 Introduction In the modern era, anesthesia is at least a two drug process consisting of an opioid and a sedative. The sedative component is typically provided by a volatile anesthetic or the intravenous sedative propofol. The opioid component is most commonly provided by fentanyl or one of its congeners. Although it is possible to achieve anesthesia with high doses of the sedative alone (i.e., a volatile anesthetic or propofol), this approach is often associated with excessive hemodynamic depression 1 and other adverse effects such as prolonged time to awakening from anesthesia. 2 Thus, for practical purposes, the current state of the art is to produce anesthesia with an opioid and a sedative in combination. Opioid-hypnotic drug interaction studies have traditionally evaluated the effects of adding one or two fixed doses or concentrations of a drug to several defined concentrations of the second drug. 3-7 Analysis of this interaction data is most commonly performed utilizing an isobologram or demonstrating the shift of parallel dose-response curves. Studies designed to characterize the interaction between sedatives and opioids using these traditional methods confirm the synergistic nature of the pharmacodynamic interactions A significant drawback of the isobologram technique is that it describes the interaction at a single level of drug effect (e.g., the Minimum Alveolar Concentration, MAC- the end-tidal concentration of volatile anesthetic where there is a 50% probability of moving to a skin incision-among others). Recently, response surface methodology has been applied to the study of anesthetic drug interactions Response surface models

26 allow the complete characterization of pharmacodynamic interactions over the entire 16 spectrum of possible concentration pairs. 12,15 Isobolograms represent just a single slice through the response-surface, whereas the response surface approach provides information over the entire spectrum of drug effect. Response surface pharmacodynamic interaction methods provide a framework to define and explore opioid-hypnotic interactions. Information about whether the interaction between two drugs is supradditive (synergistic), additive, or antagonistic is easily determined by the morphology of the surface. Furthermore, through computer simulation, it is possible to combine these response surface pharmacodynamic models with pharmacokinetic models to identify combinations of drugs that produce the same probability of producing a therapeutic effect while optimizing some other desirable outcome, such as the speed of awakening from anesthesia. 8 Prior work in our laboratory created response surface pharmacodynamic models for remifentanil and propofol in combination. 13 The current study is intended to extend this work to the interaction between volatile anesthetics and opioids using sevoflurane and remifentanil as prototypes of their respective drug classes. The principle aim of this study was to characterize the pharmacodynamic interactions of remifentanil and sevoflurane in producing sedation and analgesia using response surface models. We hypothesized that sevoflurane and remifentanil would demonstrate synergistic interactions for all the analgesic and sedative endpoints. By quantitatively describing these interactions and utilizing previously described pharmacokinetic models, we hypothesized that we could determine, through simulation, those combinations of

27 sevoflurane and remifentanil that would provide clinically adequate anesthesia and result in the most rapid emergence from anesthetics of varying durations Materials and Methods Volunteer Recruitment and Instrumentation After approval by the Human Institutional Review Board at the University of Utah Health Sciences Center (Salt Lake City, Utah, U.S.A.), informed written consent was obtained from 16 healthy adult male and female volunteers. Eligible subjects had an American Society of Anesthesiologists Physical Status of I, were nonsmokers, were years of age, and deviated by no more than 25% from their ideal body weight. Volunteers who had a history of significant alcohol or drug abuse, a history of allergy to opioids, a family history of malignant hyperthermia, or a history of chronic drug use or medical illness that is known to alter the pharmacokinetics or pharmacodynamics of opioids or inhalation anesthetics were not eligible. After a period of overnight fasting, volunteers had an intravenous catheter placed for fluid and drug administration, and electrocardiogram, pulse oximetry, non-invasive blood pressure, expired carbon dioxide and expired anesthetic gas monitoring were applied. To measure the response to electrical tentanic stimulation, surface electrodes were placed at the posterior tibial nerve. Prior to administration of the study drugs, volunteers were treated with 0.2 mg glycopyrrolate to prevent bradycardia, and 1 mg pancuronium to prevent muscle rigidity due to the opioid infusion. Each volunteer received 30 ml of sodium citrate by mouth.

28 2.3.2 Study Design 18 The study was an open-label, randomized, parallel group study using a crisscross design as advocated by Short, et al. 16 to assess drug interactions. Similar methodology was used in our earlier report describing the interactions between propofol and remifentanil. 13 Each volunteer was randomized into one of two study groups. The primary drug for the first group was remifentanil ( ng ml -1 ) and for the second group the primary drug was sevoflurane (0.3-6 vol %). The primary agent was administered from a low to a high concentration in random steps determined a priori to allow characterization of the entire concentration range when all data were pooled (Figure 2.1). After obtaining pharmacodynamic measurements at the highest concentration of the primary agent, a washout period was observed during which time the primary agent decayed to predicted concentrations below the initial target concentrations. This was followed by the administration of the secondary drug at a stable background level. The primary agent was administered from low to high concentration in the same steps as in the initial period. Following another washout period, a higher background level of the secondary drug was administered before the primary agent was administered from low to high concentration in the same steps. Upon completion of this third set of data collection, all of the drugs were discontinued and the volunteer was allowed to recover Drug Delivery Remifentanil was administered to specific predicted effect site concentration targets using a computer assisted infusion pump (Pump 22, Harvard Apparatus, Limited, Holliston, MA ) utilizing the pharmacokinetic parameters described by Minto, et al., 17

29 19 Figure 2.1: A schematic summary of the infusion scheme. During each of the three study periods the primary drug is administered in a stepwise fashion (solid black line), while in the second and third study periods, the second drug (grey filled area) is held at a constant predicted effect site concentration or measured alveolar concentration. In between each study period there is a washout phase, during which the primary and secondary drugs are allowed to decay to predicted concentrations below that of the subsequent target concentration pair.

30 20 and controlled by STANPUMP software. Sevoflurane was administered in 2-10 L min -1 of oxygen by a tight fitting mask connected to a standard circle anesthesia circuit attached to an anesthesia machine (Drager Medical, Inc., Telford, PA ) Effect Measurements Five minutes after achieving the targeted effect-site concentration (or stable endtidal concentration) for a primary drug step, a battery of pharmacodynamic assessments were made. Effect measures included the Observer s Assessment of Alertness/Sedation score (OAA/S) 18 and three surrogates for surgical stimulus- pressure algometry and tetanic electrical stimulation, similar to those previously described by Kern, et al., 13 and thermal stimulation. All stimuli were applied until reaching supra-maximal levels-50 ma, 50 PSI, and 50 C for 5 seconds. The maximum intensity of the stimulation was decreased from those utilized by Kern, et al., 13 because intensity levels of 60 ma and 60 PSI were found to be well above the supra-maximal stimulus intensity. Sedation was measured first and then the experimental pain stimuli were measured in random order. In terms of sedation, volunteers were considered nonresponsive if the OAA/S was 1 (loss of response to shake and shout, Table 2.1). Once the volunteer became nonresponsive (OAAS 1), direct laryngoscopy was performed with a Macintosh #3 blade to achieve a Cormack Grade I view 19 at each target concentration pair. The volunteer was considered responsive to the noxious stimuli when the volunteer exhibited painful verbalization, withdrawal movement, or an increase in heart rate of 20% over the prestimulus level. With the exception of laryngoscopy, baseline measurements of the subject response to Available from Steven L. Shafer, M.D., at Posted April 29, Accessed October 18, 2005.

31 Table 2.1: Observer s Assessment of Alertness/Sedation Score (OAA/S)* 21 Responsiveness Score Responds readily to name spoken in normal tone 5 Lethargic response to name spoken in normal tone 4 Responds only after name is called loudly and/or repeatedly 3 Responds only after mild prodding or shaking 2 Does not respond to mild prodding or shaking 1 Does not respond to noxious stimulus 0 1 For the purposed of this study, an OAA/S 1 was considered nonresponsive, whereas an OAA/S 4 was considered awake.

32 22 each surrogate effect were made at the start of the study day in the absence of drugs. Two kinds of data were recorded as surrogate measurements to surgical stimulus- the level of tolerated stimulus (a continuous data variable) and a quantal response of whether the volunteer could tolerate the maximal stimulus level (e.g., no withdrawal, no increase in heart rate or blood pressure) 20. By convention, the maximum stimulation levels for the surrogate pain measures were 5 seconds of 50 ma for tetanic electrical pain, 50 PSI for pressure algometry, and 50 C for thermal stimulation Data Analysis Demographic data for the volunteers in each group were compared utilizing an unpaired, two-sided t-test using StatView version (SAS Institute, Inc., Cary, NC) with P < 0.05 considered significant. All demographic data were reported as means with standard deviations. Data points that revealed a hyperalgesic response to a noxious stimulation at low sevoflurane concentrations 21 were discarded in order to allow modeling of the drug response as a monotonic function Response Surface Models Response surface models were constructed for each pharmacodynamic response using the Logit model as shown below: 22 Effect 1 = β β Cs β Cr β Cs Cr ) 1+ e ( where C s and C r are the concentration of sevoflurane (alveolar end-tidal concentration, vol %) and remifentanil (effect site concentration, ng ml -1, as predicted by Stanpump ),

33 23 respectively, and ß i are the parameters describing the response surface. Additional details of the Logit model are provided in Appendix A. For each pharmacodynamic response, the data were combined and used to fit the three-dimensional response surface using a naïve pooled technique. Model coefficients and standard errors were estimated using MATLAB (MathWorks Inc., Natick, MA). Models were built by an iterative process in which the log likelihood (LL) between the observations and the model predictions was maximized. The contribution of each coefficient was evaluated by excluding it from the model and determining whether the model deteriorated significantly using the likelihood ratio test ( Likelihood Ratio 30%). The standard error of the model parameters was estimated using the bootstrap method for 5000 iterations. 23 Model performance was evaluated by assessment of Error Prediction (observed vs. predicted probability of effect for each dose combination) and the correlation coefficient. The Error Prediction is defined as the following: Error ediction = 100X Observed Pr edicted / Observed Pr The correlation coefficient of the regression parameter estimates was used to evaluate how well the nonlinear regression models described the observed data. A large value of the correlation coefficient ( 0.7) indicates that the responses predicted from the surface described the observed data well Determination of Synergy Using the response surfaces for surrogate surgical stimuli and sedation, it is possible to simulate two-dimensional concentration-effect relationship curves for sevoflurane at a variety of remifentanil concentrations. 9 Each of these curves represents

34 24 a vertical slice from the respective response surface. The synergistic effects of combining remifentanil and sevoflurane in producing sedation and analgesia are demonstrated by examining the change in the slope and the leftward shift of the sevoflurane concentrationeffect curves Combined Pharmacokinetic and Pharmacodynamic Simulations The time to regaining responsiveness from a single drug anesthetic is determined by the pharmacokinetics of the individual drug, the concentration-effect relationship, and the duration of administration of the drug. 2,25 For two-drug anesthetics, the time to awakening is not only dependent on the individual drug pharmacokinetics and the duration of the anesthetics, but also on the target concentrations achieved for each of the drugs administered. 8 To provide a clinically useful context for applying the response surface models to everyday anesthesia practice, the pharmacodynamic response surface models from this study were combined with pharmacokinetic models 17,26 using computer simulation as described by Vuyk, et al., 8 to identify target concentration pairs of remifentanil and sevoflurane that provided a high probability of nonresponsiveness to noxious stimulation and the most rapid emergence after cessation of anesthetic administration. Additional details are provided in Appendix B. The sevoflurane model described by Lerou, et al., 26 and the remifentanil model reported by Minto, et al., 17 were utilized to simulate a range of alveolar concentrations and effect site concentrations of the sevoflurane and remifentanil, respectively, that produced a 95% probability of nonresponsiveness to the maximal tetanic stimulus of 50 ma, as determined by the response surface. Electrical tetanic stimulation is a surrogate noxious stimulus that is thought to be similar to a skin incision. 27 These alveolar and

35 25 effect site concentrations were maintained at these levels for one hour, after which time the drugs were discontinued and the washout of the anesthetics was simulated. The shortest time during the washout until the drug interaction model predicted a 95% probability that OAA/S was 4 was found through iterative simulation utilizing a binary search algorithm. 28 The initial concentration pair was randomly picked from those target concentration pairs located along the EC 95 isobole for tetanic stimulation. After calculating the recovery time (OAA/S 4) for this initial target concentration pair, a fixed step of a 25% change in either the remifentanil concentration or the sevoflurane concentration in a random direction along the isobole was made and the time to awakening was calculated. If this time was higher than that of the previous concentration pair, the next concentration pair was picked half-way between the previous point and this point; otherwise, the next concentration pair was a picked to be the same size step change in concentration away from the previous point. This step-wise search was continued until a point was reached where recovery time was within 5% of the previously calculated recovery time at the previous concentration pair. The combination of sevoflurane and remifentanil that resulted in the quickest recovery (OAA/S 4) was determined for anesthetics of minutes in duration. 2.4 Results All 16 volunteers completed the study. The demographics of the two groups are shown in Table 2.2. There were no differences between the groups except that the remifentanil group was predominately male volunteers, whereas the sevoflurane group contained equal numbers of male and female volunteers.

36 Table 2.2: Demographics of Study Volunteers* 26 Group 1 Sevoflurane Group 2 Remifentanil Age [years] 25.0 ± ± 2.7 Weight [kg] 70.8 ± ± 9.3 Height [cm] ± ± 8.4 Sex [M:F] 4:4 7:1 1 All values are given as mean ± standard deviation, except for the ratio of males to females.

37 Response Surface Models and Determination of Synergy The parameters for all the response surface models were identifiable. The Logit model parameters estimated through nonlinear regression are shown in Table 2.3. The estimates of goodness of fit (e.g., Log Likelihood, Standard Errors, and Correlation Coefficient) suggest that the models describe the data well. Based on the drug concentrations required to achieve nonresponsiveness, thermal stimulation was the mildest and tetanic stimulation was the most noxious stimulus. All of the simulated concentration-effect relationship curves from the response surface models showed synergy for both analgesia and sedation. The response surface for sedation (OAA/S 1) of the unstimulated volunteers is shown in Figure 2.2. The response surface for tetanic stimulation is shown in Figure 2.3. The other pain stimuli surfaces (not shown) were of very similar shape. The raw data used to create these surfaces are shaded based on the residual error between the measured response and model prediction. Throughout most of the clinically relevant range of concentrations (sevoflurane 0-3 vol % and remifentanil ng ml -1 ) the residual error is below 10%. The OAA/S score and the tolerance to electrical tetanic stimulation are shown topographically in Figure 2.2b and Figure 2.3b, respectively. Figures 2.4a and 2.4b are two-dimensional concentration-response curves for sevoflurane at a variety of remifentanil concentrations that are based on the response surfaces for surrogate surgical stimuli and sedation. Each of these concentration-response curves was determined by taking a vertical slice through the respective response surface (Figure 2.2a and 2.3a, Table 2.4).

38 28 Table 2.3: Mean Model Parameters for the Logit Response Surface * ß 0 ß 1 ß 2 ß 3 Log Likelihood Correlation Coefficient Pressure algometry Tetanic Stimulation Thermal stimulation Laryngoscopy OAA/S * Model parameters are listed for all values. Standard errors for all parameters were < 0.01, as determined by the bootstrap method. OAA/S = Observer assessment of alertness and sedation score.

39 29 Figure 2.2: The remifentanil-sevoflurane interaction for sedation. The Logit response surface model prediction for sedation for unstimulated volunteers is presented in the top panel (Figure 2.2a). An Observer s Assessment of Alertness/Sedation (OAA/S) score 1 represents a sedated volunteer. A 0 indicates an OAA/S 2 and a 1 indicates an OAA/S 1. The symbols show measured responses and the surface predicted by the model is represented by the grid-lined surface. The raw data used to create this model is shaded based on the residual error. A topographic view of the 50% and 95% effect isoboles for probability of being sedated is presented in the bottom panel (Figure 2.2b). The OAA/S score at each target concentration pair is overlaid.

40 a) Figure b)

41 31 Figure 2.3: The remifentanil-sevoflurane interaction for electrical tetanic stimulation. The top panel (Figure 2.3a) shows the Logit response surface model prediction for tetanic stimulation of 50 ma. A 0 indicates a response (movement or a 10% increase in blood pressure or heart rate) to a 50 ma stimulus current and a 1 indicates no response to 50 ma stimulus current. The symbols show measured volunteer responses to 50 ma of stimulus current and the surface predicted by the model is represented by the grid-lined surface. The raw data used to create this model is shaded based on the residual error. The bottom panel (Figure 2.3b) shows a topographic view of the 50% and 95% effect isoboles for probability of tolerating a 50 ma stimulus current. The percentage of tolerated stimulus current at each target concentration pair is overlaid.

42 Figure a) b)

43 33 Figure 2.4: The effect of adding remifentanil on the concentration-effect relationships of sevoflurane for sedation (Figure 2.4a) and analgesia (Figure2.4b). Each curve represents the concentration-effect relationship for sevoflurane with a fixed effect site concentration of remifentanil simulated from the corresponding response surface model. The shift in the curves toward the left indicates that much less sevoflurane is needed when remifentanil is added, demonstrating the significant pharmacodynamic synergy between the sedative and the opioid. Note that the magnitude of the leftward shift decreases as the remifentanil concentration increases (i.e., there is a ceiling effect).

44 Figure a) b)

45 Table 2.4: Reduction in Sevoflurane Requirements by Remifentanil * 35 Remifentanil Remifentanil Sevoflurane Sevoflurane C e Infusion Rate EC 95% OAA/S 1 EC 95% Tetanic [ng ml -1 ] [mcg kg -1 min -1 ] [vol %] Stimulation [vol %] * The reduction in the alveolar concentration of sevoflurane that produces a 95% probability (EC 95% ) of an OAA/S score 1 or no movement or hemodynamic response to a 50 ma tetanic stimulation by the addition of remifentanil in doses ranging from mcg kg -1 min -1 (Effect Site Concentration, C e, ng ml -1 ) are reported. All infusion rates were calculated for a hypothetical 30 year old male who weighed 80 kg and was 183 cm tall utilizing Stanpump (

46 2.4.2 Combined Pharmacokinetic and Pharmacodynamic Simulations 36 For shorter procedures the target concentration pairs that resulted in the most rapid return to responsiveness approached the maximally synergistic combination-a combination that lies on the point of the response surface where the surface curves maximally towards the origin. (Figure 2.5a) At this combination, the plasma concentrations of the drugs are both relatively low and therefore the plasma concentrations of the drugs decline to sub-clinical levels quickly (Figure 2.5b). As the duration of the anesthetic increases, the target concentration pairs with the shortest recovery time must be adjusted to be weighted towards the drug with the shorter acting kinetic profile, in this case remifentanil. By avoiding a large increase in the accumulation of sevoflurane in the body, the kinetics of washout of these combinations would allow rapid emergence from anesthesia. This trend plateaued at 0.75 vol % sevoflurane and 6.2 ng ml -1 remifentanil (Figure 2.6, Table 2.5). 2.5 Discussion In this study we utilized response surface models to characterize the pharmacodynamic interactions between a potent volatile agent, sevoflurane, and a synthetic opioid, remifentanil, across a wide range of concentration pairs. With these pharmacodynamic models, we determined that the addition of remifentanil to sevoflurane anesthesia not only synergistically decreases the response to painful stimulation but also synergistically potentiates the sedative effects of the volatile anesthetic. Furthermore, utilizing these pharmacodynamic models and previously described pharmacokinetic models, 17,26 we performed simulations to identify the target concentration pairs of remifentanil and sevoflurane that produced clinically adequate anesthesia (e.g., 95%

47 37 Figure 2.5: The results of computer simulations designed to identify optimal target concentration pairs of remifentanil- and sevoflurane that minimize the time to responsiveness. The top panel (Figure 2.5a) shows the predicted decline in effect site and alveolar concentrations for remifentanil and sevoflurane after stopping drug administration regimens targeted to reach the EC 95 isobole for tetanic stimulation for one hour. The EC 95 isobole is on the floor of the cube; the vertical axis represents time elapsed since stopping the administration of the drugs. The isobole representing a 95% probability of the return of responsiveness (Observer s Assessment of Alertness/Sedation score 4) is shown by a dotted line that is superimposed on the concentration decay curves. The highlighted curve is the sevoflurane and remifentanil target concentration pair that resulted in the fastest return of responsiveness. The bottom panel (Figure 2.5b) shows the time in minutes to the return of responsiveness after a 1 hour procedure in which sevoflurane and remifentanil were administered to target concentration pairs on the EC 95 isobole for tetanic stimulation. The highlighted trace on the panel on the left is shown topographically. The minimum time to regain responsiveness represents the target concentration pairs for a 1 hour procedure.

48 Figure a) b)

49 39 Figure 2.6: The optimal combinations of remifentanil and sevoflurane to maintain adequate anesthesia and promote rapid emergence. The combinations that produced the quickest time to regain responsiveness (Observer s Assessment of Alertness/Sedation score 4) at various durations (in hrs) are shown. For example: In a 1 hour procedure target concentrations of 1.05 vol % of sevoflurane and 4.3 ng ml -1 of remifentanil result in the fastest return of responsiveness. The simulations show that optimal combination changes as a function of length of procedure. Although a target concentration pair with higher remifentanil concentrations provides a faster recovery in longer cases, remifentanil-sevoflurane mixtures in which sevoflurane is less than 0.75 vol % show no significant advantage.

50 Table 2.5: Simulation Results for Anesthetics Minutes in Length * 40 Length of Shortest Remifentanil Remifentanil Sevoflurane Anesthetic Recovery Time C e Infusion Rate Alveolar [hr] [min] [ng ml -1 ] [mcg kg -1 min -1 ] vol % * The effect site concentration (C e ) and infusion rate for remifentanil and the alveolar end tidal concentration of sevoflurane that produced the shortest recovery times are reported for anesthetics lasting hours. All simulations were performed for a hypothetical 30 year old male who weighed 80 kg and was 183 cm tall.

51 41 probability of no response to painful stimulation) while allowing the quickest time to awakening (e.g., 5% probability of OAAS 4) for surgical procedures of increasing duration. These simulations demonstrated that there was a plateau in the utility of remifentanil to decrease the amount of sevoflurane necessary to produce clinically adequate anesthesia (sedation and nonresponsiveness to noxious stimulation) Response Surface Models Response surface methods have been utilized to model the interactions between a variety of combinations of anesthetics, the most common being that of propofol and remifentanil. 8,13,14,29-31 Our results are similar to the findings with propofol and remifentanil, in that our data demonstrate that the addition of remifentanil to sevoflurane results in a synergistic effect for both analgesia and sedation. Our results do not agree with the study by Dahan who found that alfentanil produced no synergistic effect on sevoflurane induced sedation. 32 Dahan used Bispectral Index rather than OAA/S to measure sedation and used a relatively lower concentration of alfentanil. Our data evaluated the contribution of higher levels of opioid effect (remifentanil) relative to the alfentanil concentration range studied by these investigators. Furthermore, we specifically evaluated the effects of combinations of sevoflurane and remifentanil on clinical sedation, as measured by the OAA/S, as opposed to the surrogate marker of the Bispectral Index. Perhaps the limitations of the Bispectral Index algorithm, specifically its insensitivity to the effect of an opioid on sedation, 33 may explain differences in our results. Alternatively, the fact that we utilized the Logit model for our response surface data whereas Dahan utilized the Minto response surface models, may have resulted in a

52 42 forced fit of our data to the relatively constrained model. However, the response surface generally predicted the observed data extremely well (Figure 2.2a and 2.2b and Table 2.3), and therefore is most likely not a forced fit. Over the past few years, several investigators have utilized response surface models to determine the interactions between propofol and remifentanil, 8,11,13,30 propofol and alfentanil, 34,35 and sevoflurane and alfentanil. 32 Each of these authors utilized a single type of pharmacodynamic model to develop their response surface models. The pharmacodynamic model described by Greco, 12 and utilized by Kern, 13 differs from the pharmacodynamic model developed by Minto, 15 and utilized by Dahan, 32 in that it requires the exponent of the response to be fixed, therefore limiting the flexibility of the model to fit optimally the response data. However, the Greco form of this model provides a specific parameter that examines the interaction between the two drugs. The models proposed by Bouillon, 11 Bol, 30,36 and the Logit model also differ in their mathematical complexity and physiologic plausibility. Choosing the right model to describe the data is an empirical process in which the error statistics of each model are used to determine if increasing the level of complexity allows a better fit of the measured response data. 23 However, if a model that has many degrees of freedom is chosen, it is possible to fit a surface to data from poorly designed trials or studies with inadequate response sampling. 15 For the analysis of our data, we chose the Logit model because it easily allowed the analysis of data from volunteers with different baseline and maximal responses to the surrogate pain stimuli and the clinical assessment of sedation. Given the diversity of different response surfaces models published in the anesthesia literature, the fact that we

53 43 were able to characterize adequately our data set with the Logit model, which is a moderately constrained model compared to those proposed by Greco, 12 Minto 14 or Bol, 36 may indicate that the synergism observed by these surfaces is accurate. Minto, et al., have proposed that there are several criteria necessary for an Ideal Pharmacodynamic Interaction Model. 14 The Logit model is able to predict additive, synergistic, and antagonistic interactions. Simulations of the isoboles that result with changes in the Logit model s ß 3 coefficient-the coefficient that controls the interaction between the two drugsproduce isoboles consistent with those of Berenbaum 37 (Figure 2.7). The response surfaces derived from the Logit model were easily derived from a relatively small number of volunteers from predicted effect-site remifentanil concentrations and measured alveolar end-tidal sevoflurane concentrations covering the entire clinical range of concentration pairs. In addition, the response surface reduces to single drug concentration-response curves that are similar to those that would be derived by single drug analysis 17,38 as shown in Figures 2.4a and 2.4b. However, the mathematics of logarithms dictates that when there is no drug present (i.e., sevoflurane-remifentanil target concentration pair of 0 vol % and 0 ng ml -1 ) there is still a slight effect (approximately probability of no response). Therefore, the Logit model that we have chosen as the basis of our response surface analysis meets all but one of the criteria proposed by Minto, et al., 15 albeit that the predictions made when there are no drugs present is close to no drug effect.

54 44 Figure 2.7: The isoboles derived from simulated Logit model of the sedation response surface (Observer s Assessment of Alertness/Sedation score 1) to demonstrate additive, synergistic, and antagonistic interactions, by only modifying the ß 3 coefficient. In the Logit model, the ß 3 coefficient controls the interaction between the two drugs- ß 3 = 0, ß 3 > 0, and ß 3 < 0, producing additive, synergistic, and antagonistic interactions. The dotted line represents the isobole predicted by the Logit model when the drug interaction is simply additive (ß 3 = 0), while the solid line and the dotted line represent the predicted isoboles when there is a synergistic (ß 3 = 3.94) or antagonistic (ß 3 = -0.22) drug interaction.

55 Combined Pharmacokinetic and Pharmacodynamic Simulations The simulations utilizing pharmacokinetic models and our pharmacodynamic response surfaces to determine the combination of sevoflurane and remifentanil that would produce the fastest return of responsiveness for anesthetics of varying durations provided interesting insight into the role of pharmacokinetics and pharmacodynamics in optimizing clinical anesthetics. As shown in Figure 2.5a, for a 1 hour duration anesthetic, the optimum combination of sevoflurane and remifentanil is at the point in the center of the edge of the plateau of maximum response-on the isobole that defines 95% probability of not responding to electrical tetanic stimulation. As the duration of the anesthetic increases, the optimal combinations shifted toward higher remifentanil concentrations due to the rapid elimination of remifentanil. Despite the synergistic interactions between remifentanil and sevoflurane in providing analgesia and sedation, there was a discrete plateau in the sevofluraneremifentanil combinations for the longest of procedures (Figure 2.6). This plateau occurs at a sevoflurane concentration of 0.75 vol % which correlates with an approximately 66% reduction in the Mean Alveolar Concentration (MAC) of sevoflurane (2.2 vol % for adult males and females between the ages of years). 38 The 66% reduction in sevoflurane requirements coincidentally is between the amount of reduction of MAC (61%) and MAC BAR (Blocks Autonomic Responses, 83%) expected when high doses of opioids are combined with the modern, potent volatile anesthetics. 3,7,38,39 Furthermore, this value is similar to the MAC AWAKE of sevoflurane (0.35 MAC, approximately 0.75 vol %), 40 thereby demonstrating that these response-surface models may account for the fact that opioids themselves cannot provide complete anesthesia The major factor

56 preventing a further decrease in the sevoflurane requirement may be the limited reduction 46 of the MAC AWAKE observed with opioids. 44 That these sevoflurane-remifentanil response surface pharmacodynamic models predict interactions that are consistent with clinical practice further demonstrates that these response surfaces may be useful tools for understanding anesthetic interactions in the clinical realm Clinical Implications These response surface models allow the creation of two-dimensional concentration-effect curves that demonstrate an approximately 6-fold decrease in the EC 95 for sedation and an approximately 10-fold decrease in the EC 95 for tolerance of tetanic stimulation with the addition of 7.5 ng ml -1 remifentanil (0.27 µg kg -1 min -1 infusion) to a sevoflurane anesthetic (Figure 2.4a and 2.4b and Table 2.4). Based on the synergistic interaction between sevoflurane and remifentanil in preventing a response to the surrogate surgical stimuli and in producing sedation, the response surfaces from this study confirm the utility of administrating balanced anesthetics with a combination of a volatile anesthetic and an opioid. The pharmacokinetic-pharmacodynamic simulations illustrate the benefit of minimizing the administered dose of even a low solubility volatile anesthetic to near 0.5 MAC in the presence of remifentanil, an opioid with rapid elimination. This is especially true for anesthetics with duration of over 5 hours. Whether this results in a pharmacoeconomic advantage of combining a low dose of sevoflurane with a higher dose of remifentanil will require prospective studies, because the pharmacoeconomic advantages of a drug are certainly not limited to just minimizing the time until awakening or the drug acquisition costs. 46

57 2.5.4 Limitations 47 One of the limitations of our study design is that the response surface model for sedation was determined in unstimulated volunteers. Because the level of stimulation can change the depth of sedation, it is possible that our unstimulated volunteer response surface analysis for sedation may not accurately predict the sedation response of patients undergoing surgical procedures. In particular, the lack of an endotracheal tube in the volunteers may have resulted in our measuring deeper levels of sedation than would be apparent if the endotracheal tube was stimulating a patient or volunteer receiving the same target concentration pairs of sevoflurane and remifentanil. However, the difficulty in measuring the level of sedation during stimulation in a volunteer setting (e.g., confounding sedation score by stimulation response, etc.) prevented us from collecting the data that would be needed to estimate a surface with continual stimulation. A further limitation of our study design was that the surrogate pain stimuli used to measure the analgesic response in volunteers is only a surrogate of intraoperative surgical pain. By including a range of experimental pain stimuli to cover the range expected during a surgical procedure, it is probable that the most stimulating intraoperative eventssurgical incision and laryngoscopy-have been recreated in the volunteer laboratory. However, since key surgical stimuli can only be applied once (e.g., skin incision, etc.), and since surgical patients cannot ethically be provided with subtherapeutic combinations of anesthetics or serve as their own pharmacologic control, volunteer studies are essential to allow the collection of the high resolution data needed to achieve the goal of mapping the interaction surface between two agents over the entire concentration spectrum.

58 48 Another limitation in this study is that we used a pharmacokinetic model to predict remifentanil effect site concentrations rather than drawing blood samples during pseudo-steady state to measure remifentanil plasma concentrations. This limitation may explain the variability found in the single drug dose-response data for remifentanil. 47 Mertens, et al., determined that remifentanil can be delivered accurately by target controlled infusions. 48 However, they found that the most accurate and least biased delivery was achieved when the pharmacokinetic set(s) determined by Egan, et al were utilized. Given the fact that the pharmacokinetic set utilized (by Minto, et al. 17 ) was determined in a population very similar to that being studied here, the accuracy and bias of the target controlled infusion should be at least as accurate as employing the pharmacokinetic sets of Egan, et al. 48 Although we had an unequal number of males and females in our groups, it is unlikely that this accounted for the pharmacodynamic variability given that sex has little influence on the pharmacokinetics or pharmacodynamics of remifentanil 17 or sevoflurane. 52 Other sources of pharmacokinetic variability (e.g., age, body weight, cardiac output, etc.) most likely did not contribute much to the pharmacodynamic variability, given the similarities between groups in the important covariates. For the analgesic response measurements we were forced to both limit the maximum stimulus applied and discard those responses that were below the respective baseline values. We limited the maximum stimulus applied in order to prevent irreversible tissue damage in the volunteers. In a previous investigation in our laboratory, 13 we found levels of the pressure, temperature, and electrical current that could be tolerated without any evidence of long lasting damage. However, this approach

59 49 may result in censored data that can result in pharmacodynamic response curves that predict potency lower than the true values. Therefore, extending the application of these response surfaces beyond the range of concentrations examined by these response surfaces may result in erroneous conclusions. Just as difficult of a statistical problem is how to deal with those analgesic responses that were below the baseline values. This hyperalgesic response has been observed when low doses of volatile anesthetics are administered to animals and humans. 21 Unfortunately, the models utilized to construct response surfaces require a monotonic function, and therefore are unable to characterize this phenomenon. Other investigators often do not observe this hyperalgesic response because the step change in inhaled anesthetic concentration is either so large that the hyperalgesic concentrations are jumped over or the variability in the analgesic response measurement is so large that this small hyperalgesic effect is unidentifiable. The hyperalgesia associated with the presence of low concentrations of volatile anesthetics 21 is different from the hyperalgesia phenomenon occasionally observed after the administration of remifentanil The hyperalgesia observed by some investigators after remifentanil administration is associated with a rightward shift in the subsequent analgesic concentration-response curves. Although we did not design this study to specifically address the presence or absence of remifentanil induced hyperalgesia, we did not find any difference between the baseline levels of tolerated stimuli (e.g., prior to remifentanil administration) and the levels of stimuli tolerated at the lowest level of remifentanil with the first doses of sevoflurane (Study Period II, Remifentanil Group, One-sided paired t-test, P > 0.05 for all three stimuli). This is consistent with the

60 50 observations of Lotsch and Angst where hyperalgesia to pressure and electrical stimulation was not induced by remifentanil. 55 The Logit model offered the advantage of being able to easily compensate for data from volunteers with different baseline and maximal responses to the surrogate pain stimuli and the clinical assessment of sedation. However, the mathematics of logarithms dictates that when there is no drug present (i.e., sevoflurane-remifentanil target concentration pair of 0 vol % and 0 ng ml -1 ) there is still a very slight effect (approximately probability of no response). Furthermore, the Logit model requires a dichotomous response- response vs. no response to a single stimulus intensity. For the surrogates for surgical stimulus, this was the equivalent of having no movement and no hemodynamic change when a volunteer received the maximum possible intensity of the pain surrogate. However, the OAA/S is an ordinal scale that consists of five different scores (Table 2.1). The Logit model mandated that we choose which OAA/S scores defined patients who were awake and those who were asleep. In order to represent the state most consistent with adequate sedation for surgery, the response surface model for general anesthesia was based on an OAA/S 1 ( no response to shake and shout ). On the other hand, to most accurately represent the emergence from general anesthesia (i.e., suitable for extubation), we chose an OAA/S 4 ( responds to normal voice ) as the basis of the response surface for awakening from anesthesia. Although this dichotomous view of general anesthesia is not reflected by the OAA/S score, it is more consistent with adequate general anesthesia-for any given stimulus at any given time point, anesthesia can be either considered adequate or not. 20 The models described by Greco, 12 Minto, 14 and Bouillon 11 would have avoided this

61 51 complexity because all of these models easily handled continuous response variables. However, each of these alternative model architectures would have had difficulty resolving the intersubject variability that naturally exists in the baseline and maximal tolerated stimulus Future Work Our response surface models for sevoflurane and remifentanil interactions were developed in volunteers exposed to a variety of surrogate pain stimuli. These models will need to be validated in a variety of surgical patients receiving these two drugs as the only anesthetic agents. Further work will need to be done to determine if the surrogate pain stimuli accurately predict the responses to different surgical stimuli (e.g., skin incision, abdominal insufflation, placement of Mayfield head fixation, etc.). In addition, there are conceivably 15 different sedative-opioid combinations that could be generated when one considers the pharmacodynamic and pharmacokinetic differences between the clinically available volatile anesthetics (desflurane, sevoflurane, and isoflurane) and commonly utilized opioids (morphine, fentanyl, alfentanil, sufentanil, and remifentanil), not to mention the alternative of a propofol based anesthetic. Response surface models of these combinations would be necessary to develop a comprehensive library of models for use in everyday anesthesia practice that would not constrain the clinician to a single pair of anesthetics (i.e., sevoflurane and remifentanil only) Conclusion In summary, the sevoflurane-remifentanil response surfaces estimated in this study demonstrate clear and profound synergism for both analgesia and sedation. Furthermore, combined with pharmacokinetic models, the response surfaces provide the

62 52 scientific foundation to identify the optimal combinations of sevoflurane and remifentanil required to produce the fastest return to alertness (OAA/S 4) after anesthetics varying in duration from minutes. The reduction in sevoflurane requirements predicted by these simulations plateaus at a value (0.75 vol %, 0.34 MAC) comparable to that of MAC AWAKE (0.35 MAC) of sevoflurane and in the range of the maximum reduction in MAC (61%) and MAC BAR (85%) that results from coadministration of high doses of remifentanil with sevoflurane, acting as indirect validation of the response surfaces. These response surfaces may potentially be used to clinical advantage, such as their incorporation into real-time, pharmacokineticpharmacodynamic display systems. 45, Appendix A: The Logit Model For Pharmacodynamics The pharmacodynamic response to a single drug can be described by the logistic regression model. In the logistic regression model, the natural logarithm of the odds ratio of drug effect (the Logit) is described as a function of drug concentration (C): Logit P ln( oddsratio) = ln = 1 P = β C β (1) where P is the probability of the desired effect, and ß 0 and ß 1 are estimated parameters. The Logit model can be generalized to multiple drugs, using the linear function of the concentrations of the two drugs sevoflurane (C s ) and remifentanil (C r ) 22 : Logit P 1 P = ln( oddsratio) = ln = β 0 + β1 Cs + β2 Cr + β3 Cs Cr (2) where P is the probability of the desired effect, and ß 0, ß 1, ß 2, ß 3 are estimated coefficients of the linear function.

63 equation (3): Rearranging equation (2) to solve for the probability of effect, P, results in 53 P = β β Cs β Cr β Cs Cr ) (3) 1+ e 1 ( Equation (3) can be rearranged to compute the 50% (equation (4a)) and 95% (equation (4b)) probability isoboles for sevoflurane: β0 β2 Cr EC50, S = β + β Cr 1 3 (4a) EC95, S = 1 ln( ) + β β Cr β + β Cr (4b) The Logit model reduces to a simpler form that allows calculation of the concentration-effect relationship for sevoflurane or remifentanil when administered alone. By substituting into equation (3) a value of 0 for remifentanil or sevoflurane, respectively, the concentration of each drug needed to produce 50% probability of effect (EC 50 ) when each of the drugs is used individually, can be calculated by equations (5a) and (5b). EC 50, S = β0 β 1 (5a)

64 EC 50, R = β0 β 2 (5b) Appendix B: Pharmacokinetic-Pharmacodynamic Simulations Pharmacodynamic End Points Examining the response surface models generated for adequate sedation (95% probability of OAA/S 1) and adequate analgesia (95% probability of having no movement or hemodynamic response to a 50 ma electrical stimulus), it is clear that there are many target concentration pairs of sevoflurane and remifentanil that would provide adequate surgical anesthesia. The concentration pairs on the EC 95% isobole for noresponse to a 50 ma electrical stimulation (Figure 2.3b) is consistently greater than the concentration pairs on the EC 95% isobole for adequate sedation (Figure 2.2b). Therefore, providing combinations of sevoflurane and remifentanil that are on the electrical stimulation EC 95% isobole will provide adequate surgical anesthesia. Clinical recovery from surgical anesthesia is characterized by the ability to follow simple commands (e.g., eye opening, squeezing hands, etc.) upon discontinuing drug administration. The state of clinical recovery from anesthesia corresponds to an OAA/S 4 (Table 1). Therefore, in order to model the response surface for clinical recovery from administration of combinations of sevoflurane and remifentanil, a Logit model can be constructed with OAA/S 4 defined as adequate recovery and an OAA/S < 4 defined as asleep. This model has a correlation coefficient of 0.83 and the model coefficients, ß 0, ß 1, ß 2, ß 3 are estimated as 2.97, 4.98, 0.33, and 3.15, respectively. Because the Logit model has the limitation that a small effect remains when there is no drug administered, the EC 80% isobole for OAA/S 4 was used to determine the sevoflurane-remifentanil concentration

65 pairs that resulted in clinical recovery after discontinuing administration of sevoflurane and remifentanil Pharmacokinetic Models As detailed above, the time until clinical recovery after the discontinuation of the administration of sevoflurane and remifentanil can be defined as the time that it takes for the sevoflurane and remifentanil concentrations to reach a combination on the EC 80% isobole for OAA/S 4. In order to simulate the elimination of sevoflurane and remifentanil, it is necessary to know the concentrations in all of the pharmacokinetic compartments prior to the cessation of drug administration. Administration and elimination of sevoflurane was simulated utilizing the 14 compartment physiologic model described by Lerou, et al., 26 with the volumes and blood flows reported by Lowe and Ernst, 57 and partition coefficients reported by Kennedy, et al. 58 Simulation of the administration of propofol required the use of the target controlled infusion algorithm described by Van Puocke, et al., 59 employing the remifentanil pharmacokinetic model described by Minto, et al., 17 to maintain a remifentanil effect site concentration on the EC 95% isobole for no-response to 50 ma electrical stimulus Determination of the Shortest Time to Awakening The EC 95% isobole for no-response to a 50 ma electrical stimulus provides a large number of concentration pairs of sevoflurane and remifentanil. An initial concentration pair was randomly picked from those concentration pairs located on the EC 95% isobole for tetanic stimulation. The alveolar concentration of sevoflurane and the effect site concentration of remifentanil were maintained constant for the predetermined duration ( minutes). For example, to simulate the administration of 1.05 vol %

66 56 sevoflurane and 4.53 ng ml -1 of remifentanil, the uptake and distribution of sevoflurane throughout the body was simulated to maintain an alveolar concentration of 1.05% and the uptake and distribution of remifentanil was simulated for utilizing the target controlled infusion algorithm to maintain a constant value of 4.53 ng ml -1 at the effect site. At the end of the predetermined length of drug administration, the decay of the effect site concentration of remifentanil and alveolar concentration of sevoflurane were observed. The time at which these combinations fell below levels on the EC 80% isobole for OAA/S 4 were noted. For this example, the recovery time was 5 minutes (see Figure 2.5b). This procedure was repeated with a binary search algorithm to determine the combination of sevoflurane and remifentanil that started on the EC 95% isobole for tetanic stimulation and had the fastest recovery time for the predetermined duration of drug administration. Using the same methods the ratio that had the fastest recovery time was determined for each procedure length (0.5, 1, 2, 4, 7, 10, 15, 20 and 24 hrs). 2.8 References 1. Zbinden AM, Petersen-Felix S, Thomson DA: Anesthetic depth defined using multiple noxious stimuli during isoflurane/oxygen anesthesia. II. Hemodynamic responses. Anesthesiology 1994; 80: Eger EI, 2nd, Shafer SL: Tutorial: context-sensitive decrement times for inhaled anesthetics. Anesth Analg 2005; 101: Katoh T, Kobayashi S, Suzuki A, Iwamoto T, Bito H, Ikeda K: The effect of fentanyl on sevoflurane requirements for somatic and sympathetic responses to surgical incision. Anesthesiology 1999; 90: Kazama T, Ikeda K, Morita K: Reduction by fentanyl of the Cp50 values of propofol and hemodynamic responses to various noxious stimuli. Anesthesiology 1997; 87:

67 5. Smith C, McEwan AI, Jhaveri R, Wilkinson M, Goodman D, Smith LR, Canada AT, Glass PS: The interaction of fentanyl on the Cp50 of propofol for loss of consciousness and skin incision. Anesthesiology 1994; 81: 820-8; discussion 26A 6. Vuyk J, Lim T, Engbers FH, Burm AG, Vletter AA, Bovill JG: Pharmacodynamics of alfentanil as a supplement to propofol or nitrous oxide for lower abdominal surgery in female patients. Anesthesiology 1993; 78: ; discussion 23A 7. Hall RI, Szlam F, Hug CC, Jr.: The enflurane-sparing effect of alfentanil in dogs. Anesth Analg 1987; 66: Vuyk J, Mertens MJ, Olofsen E, Burm AG, Bovill JG: 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: Vuyk J, Lim T, Engbers FH, Burm AG, Vletter AA, Bovill JG: The pharmacodynamic interaction of propofol and alfentanil during lower abdominal surgery in women. Anesthesiology 1995; 83: Glass PS, Gan TJ, Howell S, Ginsberg B: Drug interactions: volatile anesthetics and opioids. J Clin Anesth 1997; 9: 18S-22S 11. Bouillon TW, Bruhn J, Radulescu L, Andresen C, Shafer TJ, Cohane C, Shafer SL: Pharmacodynamic interaction between propofol and remifentanil regarding hypnosis, tolerance of laryngoscopy, bispectral index, and electroencephalographic approximate entropy. Anesthesiology 2004; 100: Greco WR, Bravo G, Parsons JC: The search for synergy: a critical review from a response surface perspective. Pharmacol Rev 1995; 47: Kern SE, Xie G, White JL, Egan TD: A response surface analysis of propofol-remifentanil pharmacodynamic interaction in volunteers. Anesthesiology 2004; 100: Minto C, Vuyk J: Response surface modelling of drug interactions. Adv Exp Med Biol 2003; 523: Minto CF, Schnider TW, Short TG, Gregg KM, Gentilini A, Shafer SL: Response surface model for anesthetic drug interactions. Anesthesiology 2000; 92: Short TG, Ho TY, Minto CF, Schnider TW, Shafer SL: Efficient trial design for eliciting a pharmacokinetic-pharmacodynamic model-based response surface describing the interaction between two intravenous anesthetic drugs. Anesthesiology 2002; 96:

68 17. Minto CF, Schnider TW, Egan TD, Youngs E, Lemmens HJ, Gambus PL, Billard V, Hoke JF, Moore KH, Hermann DJ, Muir KT, Mandema JW, Shafer SL: Influence of age and gender on the pharmacokinetics and pharmacodynamics of remifentanil. I. Model development. Anesthesiology 1997; 86: Chernik DA, Gillings D, Laine H, Hendler J, Silver JM, Davidson AB, Schwam EM, Siegel JL: Validity and reliability of the Observer's Assessment of Alertness/Sedation Scale: study with intravenous midazolam. J Clin Psychopharmacol 1990; 10: Cormack RS, Lehane J: Difficult tracheal intubation in obstetrics. Anaesthesia 1984; 39: Prys-Roberts C: Anaesthesia: a practical or impractical construct? Br J Anaesth 1987; 59: Zhang Y, Eger EI, 2nd, Dutton RC, Sonner JM: Inhaled anesthetics have hyperalgesic effects at 0.1 minimum alveolar anesthetic concentration. Anesth Analg 2000; 91: Egan TD, Minto C: Common Pharmacodynamic Drug Interactions in Anesthetic Practice, Anesthetic Pharmacology: Physiologic Principles & Clinical Practice. Edited by Evers AX, Maze M. London, Churchill Livingstone, 2004, pp Jacquez JA, Perry T: Parameter estimation: local identifiability of parameters. Am J Physiol 1990; 258: E Glantz SA, Slinker KK: Primer of Applied Regression and Analysis of Variance, 2 Edition. New York, McGraw-Hill, Inc., 2001, pp Shafer SL, Varvel JR: Pharmacokinetics, pharmacodynamics, and rational opioid selection. Anesthesiology 1991; 74: Lerou JG, Booij LH: Model-based administration of inhalation anaesthesia. 1. Developing a system model. Br J Anaesth 2001; 86: Zbinden AM, Maggiorini M, Petersen-Felix S, Lauber R, Thomson DA, Minder CE: Anesthetic depth defined using multiple noxious stimuli during isoflurane/oxygen anesthesia. I. Motor reactions. Anesthesiology 1994; 80: Knuth D: Sorting and Searching, The Art of Computer Programming, 3 Edition. Reading, Massachusetts, Addison-Wesley, 1997, pp Short TG, Plummer JL, Chui PT: Hypnotic and anaesthetic interactions between midazolam, propofol and alfentanil. Br J Anaesth 1992; 69:

69 30. Mertens MJ, Olofsen E, Engbers FH, Burm AG, Bovill JG, Vuyk J: Propofol reduces perioperative remifentanil requirements in a synergistic manner: response surface modeling of perioperative remifentanil-propofol interactions. Anesthesiology 2003; 99: Nieuwenhuijs DJ, Olofsen E, Romberg RR, Sarton E, Ward D, Engbers F, Vuyk J, Mooren R, Teppema LJ, Dahan A: Response surface modeling of remifentanilpropofol interaction on cardiorespiratory control and bispectral index. Anesthesiology 2003; 98: Dahan A, Nieuwenhuijs D, Olofsen E, Sarton E, Romberg R, Teppema L: Response surface modeling of alfentanil-sevoflurane interaction on cardiorespiratory control and bispectral index. Anesthesiology 2001; 94: Lysakowski C, Dumont L, Pellegrini M, Clergue F, Tassonyi E: Effects of fentanyl, alfentanil, remifentanil and sufentanil on loss of consciousness and bispectral index during propofol induction of anaesthesia. Br J Anaesth 2001; 86: Vuyk J, Hennis PJ, Burm AG, de Voogt JW, Spierdijk J: Comparison of midazolam and propofol in combination with alfentanil for total intravenous anesthesia. Anesth Analg 1990; 71: Mertens MJ, Olofsen E, Burm AG, Bovill JG, Vuyk J: Mixed-effects modeling of the influence of alfentanil on propofol pharmacokinetics. Anesthesiology 2004; 100: Bol CJ, Vogelaar JP, Tang JP, Mandema JW: Quantification of pharmacodynamic interactions between dexmedetomidine and midazolam in the rat. J Pharmacol Exp Ther 2000; 294: Berenbaum MC: What is synergy? Pharmacol Rev 1989; 41: Katoh T, Ikeda K: The minimum alveolar concentration (MAC) of sevoflurane in humans. Anesthesiology 1987; 66: Hall RI, Murphy MR, Hug CC, Jr.: The enflurane sparing effect of sufentanil in dogs. Anesthesiology 1987; 67: Katoh T, Suguro Y, Ikeda T, Kazama T, Ikeda K: Influence of age on awakening concentrations of sevoflurane and isoflurane. Anesth Analg 1993; 76: Hug CC, Jr.: Does opioid "anesthesia" exist? Anesthesiology 1990; 73: Philbin DM, Rosow CE, Schneider RC, Koski G, D'Ambra MN: Fentanyl and sufentanil anesthesia revisited: how much is enough? Anesthesiology 1990; 73: 5-11

70 43. Wong KC: Narcotics are not expected to produce unconsciousness and amnesia. Anesth Analg 1983; 62: Katoh T, Ikeda K: The effects of fentanyl on sevoflurane requirements for loss of consciousness and skin incision. Anesthesiology 1998; 88: Syroid ND, Agutter J, Drews FA, Westenskow DR, Albert RW, Bermudez JC, Strayer DL, Prenzel H, Loeb RG, Weinger MB: Development and evaluation of a graphical anesthesia drug display. Anesthesiology 2002; 96: Miller DR, Tierney M: Observational studies and "real world" anesthesia pharmacoeconomics/les etudes par observation et la realite pharmacoeconomique de l'anesthesie. Can J Anesth 2002; 49: Avram MJ, Krejcie TC: Using front-end kinetics to optimize targetcontrolled drug infusions. Anesthesiology 2003; 99: Mertens MJ, Engbers FH, Burm AG, Vuyk J: Predictive performance of computer-controlled infusion of remifentanil during propofol/remifentanil anaesthesia. Br J Anaesth 2003; 90: Egan TD, Huizinga B, Gupta SK, Jaarsma RL, Sperry RJ, Yee JB, Muir KT: Remifentanil pharmacokinetics in obese versus lean patients. Anesthesiology 1998; 89: Egan TD, Lemmens HJ, Fiset P, Hermann DJ, Muir KT, Stanski DR, Shafer SL: The pharmacokinetics of the new short-acting opioid remifentanil (GI87084B) in healthy adult male volunteers. Anesthesiology 1993; 79: Egan TD, Minto CF, Hermann DJ, Barr J, Muir KT, Shafer SL: Remifentanil versus alfentanil: comparative pharmacokinetics and pharmacodynamics in healthy adult male volunteers. Anesthesiology 1996; 84: Eger EI, 2nd, Laster MJ, Gregory GA, Katoh T, Sonner JM: Women appear to have the same minimum alveolar concentration as men: a retrospective study. Anesthesiology 2003; 99: Angst MS, Koppert W, Pahl I, Clark DJ, Schmelz M: Short-term infusion of the mu-opioid agonist remifentanil in humans causes hyperalgesia during withdrawal. Pain 2003; 106: Guignard B, Bossard AE, Coste C, Sessler DI, Lebrault C, Alfonsi P, Fletcher D, Chauvin M: Acute opioid tolerance: intraoperative remifentanil increases postoperative pain and morphine requirement. Anesthesiology 2000; 93:

71 55. Lotsch J, Angst MS: The mu-opioid agonist remifentanil attenuates hyperalgesia evoked by blunt and punctuated stimuli with different potency: a pharmacological evaluation of the freeze lesion in humans. Pain 2003; 102: Schumacher PM, Bouillon TW, Leibundgut D, Sartori V, Zbinden AM: Anesthesia Advisory Display (AAD): Real Time Guidance through the Pharmacokinetic and Interaction Pharmacodynamic Relationship during Simultaneous Administration of Multiple Drugs. Anesthesiology 2004; 101: A Lowe HJ, Ernst EA: The Quantitative Practice of Anesthesia-- Use of Closed Circuit. Baltimore, Williams & Wilkens, 1981, pp Kennedy RR, French RA, Spencer C: Predictive accuracy of a model of volatile anesthetic uptake. Anesth Analg 2002; 95: , table of contents 59. Van Poucke GE, Bravo LJ, Shafer SL: Target controlled infusions: targeting the effect site while limiting peak plasma concentration. IEEE Trans Biomed Eng 2004; 51:

72 CHAPTER 3 CONTEXT SENSITIVE TARGETS FOR OPIOIDS AND INTRAVENOUS ANESTHETICS 3.1 Abstract Background Anesthesia is most often achieved by a combination of a hypnotic and an opioid. Utilizing pharmacokinetic models and pharmacodynamic response surface models, it should be possible to determine the combination of propofol and remifentanil that would result in the shortest time to awakening for anesthetics with different durations Methods Response surface models that described the interaction between propofol and remifentanil in providing adequate sedation and surgical analgesia generated from volunteer data. Pharmacokinetic models were used to simulate dosing regimens that maintained constant effect site target concentration pairs on the 95% isobole for adequate anesthesia and the opioid/sedative mixture that yielded the fastest recovery (Observer s Alertness and Assessment Scale, OAA/S 4) from anesthetics with durations varying from Submitted for review in Anesthesiology, June Original article titled: Does the ideal combination of remifentanil and propfol change with the duration of surgery? The text of Chapter 3 of this dissertation is primarily authored by Dhanesh K. Gupta M.D., Assistant Professor, Department of Anesthesiology, University of Utah. Sandeep C Manyam conducted research, performed data analysis, and generated figures and tables that form the basis of this manuscript.

73 0.5 to 24 hours were calculated Results Logit response surface models were able to characterize all the pharmacodynamic endpoints well. The pharmacokinetic-pharmacodynamic simulations revealed that as the length of the procedure increased, faster recovery was produced by mixtures containing higher amounts of remifentanil. This trend plateaued for anesthetics lasting two or more hours at effect site concentrations of 1 µg ml -1 propofol and 15 ng ml -1 remifentanil Conclusions For longer duration anesthetics, the pharmacokinetic advantage of remifentanil becomes more apparent. Therefore, it appears that the optimal target concentration pairs of propofol-remifentanil anesthetics only changes during the first two hours of anesthesia, before the optimal concentration pairs plateau at their final values Acknowledgements Supported in part by a research grant from Alaris Medical Systems, Inc., San Diego, CA, (TDE) and by the National Institute of Biomedical Imaging and Bioengineering of the National Institutes of Health 8 RO1 EB00294 (SCM and DRW). Portions of this work have been presented at the 79 th Annual Clinical and Scientific Congress of the International Anesthesia Research Society in Honolulu, HI, March 15, 2005, (Poster S-405) and the 80 th Annual Clinical and Scientific Congress of the International Society of Anesthesia Research in San Francisco, CA, March 27, 2006.

74 3.2 Introduction 64 The time until a patient regains responsiveness from a single drug anesthetic is determined by the pharmacokinetics of the individual drug, the concentration-effect relationship, and the duration of administration of the drug. 1,2 For two-drug anesthetics, the time to awakening is not only dependent on the individual drug pharmacokinetics and the duration of administration of the anesthetics, but it is also dependent on the target concentrations achieved for each of the drugs administered. 3 Attempting to run a lean anesthetic can result in an increased chance of intraoperative awareness, 4 while attempting to run a deep anesthetic can result in intraoperative hemodynamic instability 5 and possibly even an increase in one year mortality. 6 To provide a clinically useful context for applying the response surface models to everyday anesthesia practice, pharmacodynamic response surface models can be combined with pharmacokinetic models 7,8 using computer simulation as described by Vuyk, et al., 3 to identify target concentration pairs of the sedative/hypnotic and the opioid that provide a high probability of clinical sedation and nonresponsiveness to noxious stimulation and the most rapid emergence after cessation of anesthetic administration. 9 Several authors have developed response surface pharmacologic interaction models of the prototypic intravenous sedative/hypnotic, propofol and the prototypic synthetic opioid, remifentanil However, only the work by Mertens, et al., 10 applied these models to predict possible optimum target concentration pairs of propofol and remifentanil that would result in the fastest return of consciousness. Surprisingly, simulations based on these response surface models, and those based of isobologram data extrapolated form propofol-alfentanil interaction data 3 have both determined that with

75 65 increasing duration of anesthesia there was no change in the optimal propofolremifentanil target concentrations. This is in direct contrast to what would be predicted based on the complex pharmacokinetics of propofol and remifentanil 2 and the synergistic interactions that occur for a variety of pharmacologic end points. 3,11-14 In addition, increasing the duration of anesthesia changes the optimal sevoflurane-remifentanil target concentration pairs to contain higher concentrations of remifentanil. 9 The aim of this study was to utilize previously collected pharmacodynamic data and apply the generated propofol-remifentanil response surface models to determine if the optimal propofol-remifentanil concentrations changes as the anesthetic duration increased. We hypothesized that by combining Logit response surface models developed from volunteer data with published pharmacokinetic models, we could predict the target concentrations of propofol and remifentanil that resulted in the fastest time to awakening from anesthesia. We also hypothesized that the pharmacokinetic advantages of remifentanil over propofol would result in higher remifentanil concentrations being targeted as the duration of the anesthetic increased-the optimal propofol-remifentanil concentration would increase as the duration of anesthesia increased. 3.3 Materials and Methods Data from 32 of the 40 subjects included in this manuscript were acquired from the datasets reported in two manuscripts from our laboratory that examined the synergistic interaction between sedative/hypnotics and remifentanil in producing clinical sedation and analgesia to experimental painful stimuli that are surrogates for intraoperative painful stimuli. 9,11 The data from all 24 subjects reported by Kern, et al., 11 were included in the current analyses, while only the data acquired from the eight

76 66 subjects who received remifentanil alone during the initial phase of the study reported by Manyam, et al., 9 were included in the current analyses. The data from an additional eight subjects who received propofol as the initial anesthetic drug followed by two fixed doses of remifentanil were included in these analyses; these subjects were collected as part of the volunteer study conducted by Manyam, et al., 9 but have not been reported elsewhere. A written informed consent document that was approved by the Human Institutional Review Board at the University of Utah Health Sciences Center (Salt Lake City, Utah) was obtained from each of 40 volunteers in this open-label, randomized, parallel group crisscross designed study to asses drug interactions (Figure 3.1). 15 Each volunteer was randomized to receive a target controlled infusion of propofol (predicted effect site concentrations of µg ml -1 ) or remifentanil (predicted effect site concentrations of ng ml -1 ) as the primary agent with the other drug acting as the secondary agent (Figure 3.1). The reader is referred to the previous manuscripts by Kern, et al., 11 and Manyam, et al., 9 for complete details regarding the methods of volunteer preparation, drug administration, data collection, and data analysis). Only those portions of the data analysis that have substantial differences from the previous manuscripts are provided in detail below. All of the effect measurements utilized by Manyam, et al., 9 had maximum intensities that were decreased from those utilized by Kern, et al., 11 because intensity levels of 60 ma and 60 PSI were found to be well above the supra-maximal stimulus intensity. To adjust for the different supra-maximal stimulus levels applied between the two studies as well as the intersubject variation in baseline tolerance of noxious stimulation, the level of stimulus tolerated was normalized against each volunteer s

77 67 Figure 3.1: A schematic summary of the infusion scheme. During each of the three study periods the primary drug is administered in a stepwise fashion (solid black line), while in the second and third study periods, the second drug (grey filled area) is held at a constant predicted effect site concentration or measured alveolar concentration. In between each study period there is a washout phase, during which the primary and secondary drugs are allowed to decay to predicted concentrations below that of the subsequent target concentration pair.

78 68 baseline value, such that 0 represented baseline and 1 represented the maximal stimulus tolerated by the volunteer. This produced a quantal pharmacodynamic response of whether the volunteer could tolerate the maximal stimulus level (e.g., no withdrawal, no increase in heart rate or blood pressure). For each pharmacodynamic response, the data were combined and used to fit the three-dimensional response surface based on the Logit model. 9 Simulated two-dimensional concentration-effect relationship curves for propofol at a variety of remifentanil concentrations were utilized to determine the type of pharmacologic interaction produced by the addition of remifentanil to a propofol anesthetic. 13 The pharmacodynamic response surface models from this study were combined with previously published pharmacokinetic models 7,16 using computer simulation as described by Vuyk, et al., 3 to identify target concentration pairs of propofol and remifentanil that provided a high probability of nonresponsiveness to noxious stimulation and the most rapid emergence after cessation of anesthetic administration. 9 Because of the overlap between the propofol-remifentanil clinical sedation isobole (95% probability of achieving an OAA/S 1) and surgical analgesia isobole (95% probability of no movement response and no hemodynamic response to a 50 ma tetanic stimulation), the composite isobole predicting adequate surgical anesthesia-adequate clinical sedation and adequate surgical analgesia-was chosen to be the higher of the two isoboles at any given concentration pair (Figure 3.2). The propofol model described by Tackely, et al., 17 and the remifentanil model reported by Minto, et al., 7 were utilized with the target controlled infusion algorithm employed by STANPUMP 18 to simulate a range of propofol and remifentanil effect site concentrations that produced a 95% probability of

79 69 Figure 3.2: A representation of the isoboles predicting a 95% probability of clinical sedation (OAA/S 1, dotted line), a 95% probability of surgical analgesia (no movement and no hemodynamic response to a 50 ma tetanic stimulus, solid line), and an 80% probability of clinical awakening form anesthesia (OAA/S 4, dashed line). Because the isoboles for adequate sedation and adequate analgesia intersect and cross, the targets for adequate clinical anesthesia (sedation and analgesia) is determined by the isobole that is at the higher target concentration pairs (boundary of the hatched area).

80 70 surgical anesthesia, as determined by the composite clinical anesthesia isobole. These effect site concentrations were maintained at these levels for one hour, after which time the drugs were discontinued and the washout of the anesthetics was simulated. The shortest time during the washout until the drug interaction model predicted a 95% probability that OAA/S was 4 was found through iterative simulation utilizing a binary search algorithm. 19 The combination of propofol and remifentanil that resulted in the quickest recovery (OAA/S 4) was determined for anesthetics of minutes in duration. 3.4 Results All forty volunteers completed one of the two study protocols. The demographics of the four groups of patients are shown in Table 3.1. There was no difference between the groups except that the remifentanil patients from Manyam s study were predominately male, whereas the remainder of the groups contained equal numbers of male and female volunteers Response Surface Models and Determination of Synergy The parameters for all the response surface models were identifiable. The Logit model parameters estimated through the nonlinear regression are shown in Table 3.2. The models described the pharmacodynamic data reasonably well (R 2 > 0.5), with the models for clinical sedation score and pressure algometry performing best. Figure 3.3a shows the response surface for sedation (OAA/S 1) of the unstimulated volunteers and Figure 3.3b shows the data overlaid on the simulated isoboles that predict a 50% and 95% probability of having an OAA/S 1. Figure 3.4a and 3.4b demonstrate the tetanic stimulation response surface and the simulated isoboles that predict a 50% and 95%

81 Table 3.1: Demographics of Study Volunteers* 71 Propofol Propofol Remifentanil Remifentanil Kern Manyam Kern Manyam (n = 12) (n = 8) (n = 12) (n = 8) Age [years] 29.0 ± ± ± ± 3.0 Weight [kg] 69.0 ± ± ± ± 9.0 Height [cm] ± ± ± ± 8.0 Sex [M:F] 5 : 7 8 : 3 7 : 5 7: 1 * All values are given as mean ± standard deviation, except for the ratio of males to females.

82 Table 3.2: Mean Model Parameters for the Logit Response Surface* ß 0 ß 1 ß 2 ß 3 Log Likelihood Correlation Coefficient 72 Pressure algometry Tetanic Stimulation Laryngoscopy OAA/S * Model parameters are listed for all values. Standard errors for all parameters were < 0.01, as determined by the bootstrap method. OAA/S = Observer assessment of Alertness and sedation score.

83 73 Figure 3.3: The propofol-remifentanil interaction for sedation. The Logit response surface model prediction for sedation for unstimulated volunteers is presented in the top panel (Figure 3.3a). An Observer s Assessment of Alertness/Sedation (OAA/S) score 1 represents a sedated volunteer. A 0 indicates an OAA/S 2 and a 1 indicates an OAA/S 1. The symbols show measured responses and the surface predicted by the model is represented by the grid-lined surface. The raw data used to create this model are shaded based on the residual error. A topographic view of the 50% and 95% effect isoboles for probability of being sedated is presented in the bottom panel (Figure 3.3b). The OAA/S score at each target concentration pair is overlaid.

84 Figure a) b)

85 75 Figure 3.4: The remifentanil-sevoflurane interaction for electrical tetanic stimulation. The top panel (Figure 3.4a) shows the Logit response surface model prediction for tetanic stimulation of 50 ma. A 0 indicates a response (movement or a 10% increase in blood pressure or heart rate) to a 50 ma stimulus current and a 1 indicates no response to 50 ma stimulus current. The symbols show measured volunteer responses to 50 ma of stimulus current and the surface predicted by the model is represented by the grid-lined surface. The raw data used to create this model is shaded based on the residual error. The bottom panel (Figure 3.4b) shows a topographic view of the 50% and 95% effect isoboles for probability of tolerating a 50 ma stimulus current. The percentage of tolerated stimulus current at each target concentration pair is overlaid.

86 Figure a) b)

87 77 probability of not having a movement response or hemodynamic response to a 50 ma tetanic stimulation. All of the other pain stimuli surfaces (not shown) were of similar shape. The residual errors for both clinical sedation and surgical anesthesia were less than 10% throughout most of the clinically relevant range of concentrations (propofol 0-10 µg ml -1 and remifentanil 0-15 ng ml -1 ). Simulated concentration-response curves for propofol at a variety of remifentanil concentrations that are based on the response surface models for clinical anesthesia and surrogate surgical anesthesia are shown in Figures 3.5a and 3.5b, respectively Combined Pharmacokinetic and Pharmacodynamic Simulations The addition of a moderate dose of remifentanil (C e 9.03 ng ml -1 ) to a thirty minute anesthetic decreased the propofol effect site concentration 8 fold (C e µg ml -1 ) compared to the propofol effect site concentration required to produce surgical analgesia without any remifentanil (Tables 3.3 and 3.4). The concentration of propofol and remifentanil that resulted in the fastest emergence from surgical anesthesia plateaued at 1 µg ml -1 and 15 ng ml -1, respectively, for anesthetics lasting as short as two hours (Figure 3.6 and Table 3.4). By administering combinations of anesthetics containing higher amounts of remifentanil than propofol, it is possible to take advantage of the more favorable pharmacokinetic properties of remifentanil and exploit the synergistic pharmacodynamic effects.

88 78 Figure 3.5: The effect of adding remifentanil on the concentration-effect relationships of propofol for sedation (Figure 3.5a) and analgesia (Figure 3.5b). Each curve represents the concentration-effect relationship for propofol with a fixed effect site concentration of remifentanil simulated from the corresponding response surface model. The shift in the curves toward the left indicates that much less propofol is needed when remifentanil is added, demonstrating the significant pharmacodynamic synergy between0 the sedative and the opioid. Note that the magnitude of the leftward shift decreases as the remifentanil concentration increases (i.e., there is a ceiling effect). The addition of small to moderate amounts of remifentanil to a propofol anesthetic result in a large decrease in the amount of propofol required to produce clinically adequate sedation and surgical anesthesia (Figures 3.5a and 3.5b, and Table 3.3).

89 Figure a) b)

90 Table 3.3: Reduction in Propofol Requirements by Remifentanil * 80 Remifentanil Remifentanil Propofol Propofol C e Infusion Rate C 95% OAA/S 1 C 95% Tetanic [ng ml -1 ] [µ kg -1 min -1 ] [µ ml -1 ] Stimulation [µ ml -1 ] * The reduction in the effect site concentration (C e ) of propofol that produces a 95% probability (C 95% ) of an OAA/S score 1 or no movement or hemodynamic response to a 50 ma tetanic stimulation by the addition of remifentanil in doses ranging from mcg kg -1 min -1 are reported. All infusion rates were calculated for a hypothetical 30 year old male who weighed 80 kg and was 183 cm tall utilizing Stanpump.

91 Table 3.4: Simulation Results for Anesthetics Minutes in Length * 81 Length of Shortest Remifentanil Remifentanil Propofol Anesthetic Recovery Time C e Infusion Rate C e [hr] [min] [ng ml -1 ] [µ kg -1 min -1 ] [µ ml -1 ] * The effect site concentration (C e ) and infusion rate for remifentanil and effect site concentration (C e ) for propofol that produced the shortest recovery times are reported for anesthetics lasting hours. All simulations were performed for a hypothetical 30 year old male who weighed 80 kg and was 183 cm tall.

92 82 Figure 3.6: The results of computer simulations designed to identify optimal target concentration pairs of remifentanil- and propofol that minimize the time to responsiveness. The top panel (Figure 3.6a) shows the predicted decline in effect site concentrations for remifentanil and propofol after stopping drug administration regimens targeted to reach the EC 95 for adequate clinical anesthesia isobole for one hour. The EC 95 isobole is on the floor of the cube; the vertical axis represents time elapsed since stopping the administration of the drugs. The isobole representing a 95% probability of the return of responsiveness (Observer s Assessment of Alertness/Sedation score 4) is shown by a dotted line that is superimposed on the concentration decay curves. The highlighted curve is the sevoflurane and remifentanil target concentration pair that resulted in the fastest return of responsiveness. The bottom panel (Figure 2.5b) shows the time in minutes to the return of responsiveness after a 1hr procedure in which propofol and remifentanil were administered to target concentration pairs on the EC 95 isobole for adequate clinical anesthesia isobole. The highlighted trace on the panel on the left is shown topographically. The minimum time to regain responsiveness represents the target concentration pairs for a 1 hour procedure.

93 Figure a) b)

94 3.5 Discussion 84 Our simulations revealed that for short duration anesthetics, the pharmacokinetic advantage of remifentanil becomes more apparent. Between 0.5 hours and 2 hours, the propofol target effect site concentration decreased by 33% while the remifentanil target effect site concentration increased by 66% (Table 3.5). However, with further increases in the anesthetic duration, both the remifentanil and the propofol effect site concentrations rapidly reached their plateau values-for all anesthetics lasting two or more hours, the optimal target effect site concentration of propofol reached a nadir at 1.0 µg ml -1 while the target effect site concentration of remifentanil plateaued at 15.0 ng ml -1. Therefore, our results are similar to those predicted by the Mertens, et al., 10 and Vuyk, et al., 3 and different from our previous observations with sevoflurane-remifentanil anesthesia Response Surface Models and Determination of Synergy We chose to utilize a Logit model as the basis of our response surface analyses because the Logit model is able to characterize quantal pharmacologic responses. Transformation of the data from different data collection periods that have different baseline and maximal pharmacologic response into quantal responses allows the analysis of a larger data set derived from a variety of sources. With the Logit model we were able to confirm the synergistic interaction of propofol and remifentanil in producing clinical sedation and analgesia proposed by Minto, et al., 20 As before, the Logit model fulfills all but one of the criteria and the single unfulfilled criterion is that the Logit based response surface model dictates that there is a slight effect (< 0.1% probability of a response) when no drug is administered. 9

95 85 Our predictions for the EC 50, PROP for sedation and tetanic stimulation are very close to those predicted by the previous analysis of portion of this dataset utilizing the Greco form of the response surface model (2.2 vs 1.8 µg ml -1 and 6.7 vs 4.6 ng ml -1, sedation and tetanic stimulation, respectively). 11 The estimates for the EC 50, PROP for sedation are also in agreement with those reported for surgical patients by other investigators. 10,21 The relative agreement between the EC 50, PROP for tetanic stimulation and the reported EC 50, PROP for laryngoscopy in this manuscript ( 6.7 vs. 6.6 µg ml -1, tetanic stimulation vs. laryngoscopy) and by others 10,12 suggests that electric stimulation may provide an stimulus of an intensity comparable to surgical incision or laryngoscopy. Therefore, the volunteer paradigm utilized in this an other studies, 9,11 is able to predict results that are consistent with similar pharmacologic end points in surgical patients. The volunteer paradigm offers several advantages over the surgical patient for studying pharmacodynamic interactions between two anesthetic drugs. The main two advantages are the ability to study subtherapeutic combinations of drugs without concern of providing inadequate clinical effect and the ability to perform repeated measurement of responses thereby allowing characterization of the entire spectrum of concentration pairs. However, one of the remaining challenges in pharmacodynamics research is the validation of the surrogates of surgical stimulation (e.g., electrical stimulation) by other means than comparing the predicted concentration effect relationship to those reported for the same drug in surgical patients Combined Pharmacokinetic and Pharmacodynamic Simulations There appears to be a very limited range of anesthetic lengths over which the optimal propofol-remifentanil effect site concentrations change before reaching plateau

96 values. The minimum sedative concentration of propofol plateaus at 1.0 µg ml which correlates with approximately a 71% reduction in the EC 95, PROP for sedation. Coincidently, various opioid and sedative/hypnotic combinations have revealed that even high concentrations of opioid are unable to produce more than a 60-70% reduction in pharmacologic requirements for a sedative/hypnotic. 10,13,22-25 Examined another way, it appears that the fundamental processes of anesthesia (amnesia) require a modest amount of sedative/hypnotic even in the presence of extremely high opioid concentrations-the ceiling effect. 10,13,22-25 Accordingly, the EC 95, REMI for sedation and EC 95, REMI for surgical analgesia were two to three times higher than the maximum remifentanil concentrations simulated (32.7 and 41.4 ng ml -1, sedation and tetanic stimulation, respectively), which is consistent with the findings of other investigations. 3,9, Clinical Implications The concentration-effect curves generated by these response surface models demonstrate an approximately 2-fold reduction in the EC 95, PROP for sedation and an approximately 8-fold decrease in the EC 95, PROP for surgical analgesia with the addition of 7.5 ng ml -1 of remifentanil (0.27 µg kg -1 min -1 infusion) to a propofol anesthetic (Figure 3.5a and 3.5b and Table 3.3). Comparing this to the 6-fold reduction in the EC 95, SEVO for sedation and an approximately 10-fold decrease in the EC 95, SEVO for surgical analgesia under similar conditions, 9 one can see that propofol is a more potent sedative/hypnotic than sevoflurane and about equally as good of an immobilizer or analgesic. The propofol-remifentanil synergy in producing clinical sedation and surgical analgesia supports the utility of administrating balanced anesthetics with a combination

97 87 of sedative/hypnotic and opioid. The pharmacokinetic-pharmacodynamic simulations demonstrate the benefit of administrating a very low target effect site concentration of propofol (1.0 µg ml -1 ) in the presence of remifentanil, an opioid with a very rapid elimination clearance. This is especially true for anesthetics 2 hours or longer. Because the pharmacoeconomic advantages of the drug are not limited to just minimizing the time until awakening or the drug acquisition costs, it is unclear whether these high dose remifentanil-low dose propofol anesthetics will be of a pharmacoeconomic advantage Limitations Although the volunteer study design affords the advantages of ethically allowing the investigation of multiple concentration pairs spanning from subtherapeutic to supratherapeutic combinations, it is limited in that the sedation end point is determined in unstimulated volunteers because of the lack of an endotracheal tube. Since every surgical patient has at least the mild to moderate stimulation provided by the constant presence of an endotracheal tube, there is a possibility that our predictions for the drug concentration pairs that provide adequate sedation may underestimate the sedative requirements of an intubated surgical patient. However, the robustness of the OAA/S criteria utilized to define sedation (OAA/S 1-does not respond to shaking or shouting) may compensate by providing intermittent stimulation. The similarities between the EC 50, PROP for sedation determined from this model and that determined in unintubated patients during placement of an intracerebral stimulating electrode 21 or intubated surgical patients undergoing gynecological operations 10 is reassuring that this limitation is not too large. The application of surgical stimulus surrogates also has advantages and disadvantages. Although it is possible to apply these surgical surrogates multiple times in

98 88 order to determine the pharmacologic response spanning the entire concentration range, in order to assure that the volunteers were not injured, we had to limit the maximum stimulus applied. This limitation may have resulted in censored data that could result in pharmacodynamic response curves that predict a falsely low potency. Therefore, care must be used if the data are extrapolated above the concentration range examined. The other disadvantage of the surrogates to surgical stimulus is that it is unclear what the true intraoperative correlates are for the tested stimuli just as it is unclear what the appropriate surrogate stimulus is for many intraoperative stimuli (e.g., pneumoperitoneum, subcutaneous tunneling, etc.). Based on our growing experience with these surgical surrogate stimuli and observations of the pharmacodynamic and physiologic responses, 9,11 it is probably safe to state that the most stimulating intraoperative eventssurgical incision, placement of a Mayfield head fixation device, sternotomy, and laryngoscopy-are mimicked by the most intense surrogate stimuli-tetanic stimulation and laryngoscopy. The use of intravenous anesthetics brought about two possible limitations. First, the use of pharmacokinetic models to predict the propofol and remifentanil effect site concentrations in lieu of measuring the actual blood drug concentration may compound some of the variability in the opioid only, single drug data. 27 However, as in our previous study, 9 there is convincing evidence to demonstrate that this may not be a major source of pharmacokinetic variability. Second, continuous infusions of remifentanil has been shown to induce hyperalgesia in patients 28 and volunteers. 29 As detailed in our prior manuscript, 9 we did not design the studies to detect the existence of or the presence of remifentanil induced hyperalgesia. However, we did not find any differences between the

99 89 baseline levels of tolerated stimuli and the levels of stimuli tolerated at the lowest doses of propofol. In addition, one could conjecture that any opioid hyperalgesia that developed would not effect the clinical sedation scores (OAA/S) that were determined during quiet periods prior to the determination of the analgesic response of each of the targeted concentration pairs Future Work Our response surface models for propofol and remifentanil interactions were developed in volunteers exposed to a variety of surrogate pain stimuli. These models will need to be validated in a variety of surgical patients receiving these two drugs as the only anesthetic agents. Further work will need to be done to determine if the surrogate pain stimuli accurately predict the responses to different surgical stimuli (e.g., skin incision, abdominal insufflation, placement of Mayfield head fixation, etc.). In addition, there are conceivably 20 different sedative-opioid combinations that could be generated when one considers the pharmacodynamic and pharmacokinetic differences between the clinically available sedative/hypnotics (propofol, desflurane, sevoflurane, and isoflurane) and commonly utilized opioids (morphine, fentanyl, alfentanil, sufentanil, and remifentanil). Simply utilizing the potency ratio of one drug compared to another member of it s drug class and performing computer simulations of the results based on previous developed models may not yield correct results if the concentration-effect relationships are not parallel and the appropriate dose ranges are not selected. But in order to develop a comprehensive library of models for use in everyday anesthesia practice that would not constrain the clinician to a single pair of anesthetics (i.e., sevoflurane and remifentanil only) response surface models of these combinations would be necessary.

100 3.5.6 Conclusion 90 Several authors have used a variety of response surface models to characterize the pharmacodynamic interactions of propofol and remifentanil for a variety of pharmacologic end points in volunteers and surgical patients. We demonstrated that by combining Logit response surface models developed from volunteer data with pharmacokinetic models, we could identify target concentrations of propofol and remifentanil that resulted in the fastest time to awakening from anesthesia. The pharmacokinetic advantages of remifentanil over propofol resulted in higher remifentanil concentrations being targeted as the duration of the anesthetic increased. 3.6 References 1. Eger EI, 2nd, Shafer SL: Tutorial: context-sensitive decrement times for inhaled anesthetics. Anesth Analg 2005; 101: Shafer SL, Varvel JR: Pharmacokinetics, pharmacodynamics, and rational opioid selection. Anesthesiology 1991; 74: Vuyk J, Mertens MJ, Olofsen E, Burm AG, Bovill JG: 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: Practice Advisory for Intraoperative Awareness and Brain Function Monitoring: A Report by the American Society of Anesthesiologists Task Force on Intraoperative Awareness. Anesthesiology 2006; 104: Zbinden AM, Petersen-Felix S, Thomson DA: Anesthetic depth defined using multiple noxious stimuli during isoflurane/oxygen anesthesia. II. Hemodynamic responses. Anesthesiology 1994; 80: Monk TG, Saini V, Weldon BC, Sigl JC: Anesthetic management and one-year mortality after noncardiac surgery. Anesth Analg 2005; 100: Minto CF, Schnider TW, Egan TD, Youngs E, Lemmens HJ, Gambus PL, Billard V, Hoke JF, Moore KH, Hermann DJ, Muir KT, Mandema JW, Shafer SL: Influence of age and gender on the pharmacokinetics and pharmacodynamics of remifentanil. I. Model development. Anesthesiology 1997; 86: 10-23

101 8. Schnider TW, Minto CF, Shafer SL, Gambus PL, Andresen C, Goodale DB, Youngs EJ: The influence of age on propofol pharmacodynamics. Anesthesiology 1999; 90: Manyam SC, Gupta DK, Johnson KB, White JL, Pace NL, Westenskow DR, Egan TD: Opiod-Volatile Anesthetic Synergy: A Response Surface Model with Remifentanil and Sevoflurane as Prototypes. Anesthesiology 2006: in press 10. Mertens MJ, Olofsen E, Engbers FH, Burm AG, Bovill JG, Vuyk J: Propofol reduces perioperative remifentanil requirements in a synergistic manner: response surface modeling of perioperative remifentanil-propofol interactions. Anesthesiology 2003; 99: Kern SE, Xie G, White JL, Egan TD: A response surface analysis of propofol-remifentanil pharmacodynamic interaction in volunteers. Anesthesiology 2004; 100: Bouillon TW, Bruhn J, Radulescu L, Andresen C, Shafer TJ, Cohane C, Shafer SL: Pharmacodynamic interaction between propofol and remifentanil regarding hypnosis, tolerance of laryngoscopy, bispectral index, and electroencephalographic approximate entropy. Anesthesiology 2004; 100: Vuyk J, Lim T, Engbers FH, Burm AG, Vletter AA, Bovill JG: The pharmacodynamic interaction of propofol and alfentanil during lower abdominal surgery in women. Anesthesiology 1995; 83: Schraag S, Mohl U, Bothner U, Georgieff M: Interaction modeling of propofol and sufentanil on loss of consciousness. J Clin Anesth 1999; 11: Short TG, Ho TY, Minto CF, Schnider TW, Shafer SL: Efficient trial design for eliciting a pharmacokinetic-pharmacodynamic model-based response surface describing the interaction between two intravenous anesthetic drugs. Anesthesiology 2002; 96: Lerou JG, Booij LH: Model-based administration of inhalation anaesthesia. 1. Developing a system model. Br J Anaesth 2001; 86: Tackley RM, Lewis GT, Prys-Roberts C, Boaden RW, Dixon J, Harvey JT: Computer controlled infusion of propofol. Br J Anaesth 1989; 62: Shafer SL, Gregg KM: Algorithms to rapidly achieve and maintain stable drug concentrations at the site of drug effect with a computer-controlled infusion pump. J Pharmacokinet Biopharm 1992; 20: Knuth D: Sorting and Searching, The Art of Computer Programming, 3 Edition. Reading, Massachusetts, Addison-Wesley, 1997, pp

102 20. Minto CF, Schnider TW, Short TG, Gregg KM, Gentilini A, Shafer SL: Response surface model for anesthetic drug interactions. Anesthesiology 2000; 92: Fabregas N, Rapado J, Gambus PL, Valero R, Carrero E, Salvador L, Nalda-Felipe MA, Troconiz IF: Modeling of the sedative and airway obstruction effects of propofol in patients with Parkinson disease undergoing stereotactic surgery. Anesthesiology 2002; 97: Smith C, McEwan AI, Jhaveri R, Wilkinson M, Goodman D, Smith LR, Canada AT, Glass PS: The interaction of fentanyl on the Cp50 of propofol for loss of consciousness and skin incision. Anesthesiology 1994; 81: Sebel PS, Glass PS, Fletcher JE, Murphy MR, Gallagher C, Quill T: Reduction of the MAC of desflurane with fentanyl. Anesthesiology 1992; 76: Hall RI, Szlam F, Hug CC, Jr.: The enflurane-sparing effect of alfentanil in dogs. Anesth Analg 1987; 66: Brunner MD, Braithwaite P, Jhaveri R, McEwan AI, Goodman DK, Smith LR, Glass PS: MAC reduction of isoflurane by sufentanil. Br J Anaesth 1994; 72: Miller DR, Tierney M: Observational studies and "real world" anesthesia pharmacoeconomics. Can J Anaesth 2002; 49: Avram MJ, Krejcie TC: Using front-end kinetics to optimize targetcontrolled drug infusions. Anesthesiology 2003; 99: Crawford MW, Hickey C, Zaarour C, Howard A, Naser B: Development of Acute Opioid Tolerance During Infusion of Remifentanil for Pediatric Scoliosis Surgery. Anesth Analg 2006; 102: Angst MS, Koppert W, Pahl I, Clark DJ, Schmelz M: Short-term infusion of the mu-opioid agonist remifentanil in humans causes hyperalgesia during withdrawal. Pain 2003; 106: Prys-Roberts C: Anaesthesia: a practical or impractical construct? Br J Anaesth 1987; 59: Greco WR, Bravo G, Parsons JC: The search for synergy: a critical review from a response surface perspective. Pharmacol Rev 1995; 47: Minto C, Vuyk J: Response surface modelling of drug interactions. Adv Exp Med Biol 2003; 523:

103 CHAPTER 4 PROCESSED EEG TARGETS REQUIRED FOR ADEQUATE ANESTHESIA 4.1 Abstract Background Opioids are commonly used in conjunction with sedative drugs to provide anesthesia. Previous studies have shown that opioids reduce the clinical requirements of sedative drug needed to provide adequate anesthesia. Processed EEG parameters, such as the Bispectral Index (BIS, Aspect Medical Systems, Newton, MA) and Auditory Evoked Potential Index (AAI, Alaris Medical Systems, ), can be used intra-operatively to assess the depth of sedation. The aim of this study was to characterize how the addition of opioids sufficient to change the clinical level of sedation, influenced the BIS and AAI Methods Twenty four adult volunteers received a target controlled infusion of remifentanil (0-15 ng ml -1 ) and inhaled sevoflurane (0-6 vol %) at various target concentration pairs. After reaching pseudo-steady-state drug levels, the Observer's Assessment of Alertness/Sedation (OAAS) score, BIS, and AAI were measured at each Accepted for publication in Anesthesiology, August Copyright 2006, American Society of Anesthesiologists. Original article titled: When is a bispectral index of 60 too low? Rational processed EEG targets are dependent on the sedative-opioid ratio.

104 target concentration pair. Response surface pharmacodynamic interaction models were built using the pooled data for each pharmacodynamic end point Results Response surface models adequately characterized all pharmacodynamic end points. Despite the fact that sevoflurane-remifentanil interactions were strongly synergistic for clinical sedation, BIS and AAI were minimally affected by the addition of remifentanil to sevoflurane anesthetics Conclusion Although clinical sedation increases significantly with the addition of a small to moderate dose of remifentanil to a sevoflurane anesthetic, the BIS and AAI are insensitive to this change in clinical state. Therefore, during sevoflurane-remifentanil anesthesia, targeting a BIS < 60 or an AAI <30 may result in an unnecessarily deep anesthetic state Acknowledgements Supported in part by a research grant from Alaris Medical Systems, Inc., San Diego, CA, (TDE) and by the National Institute of Biomedical Imaging and Bioengineering of the National Institutes of Health 8 RO1 EB00294 (SCM and DRW). 4.1 Introduction Explicit recall of intraoperative events (intraoperative awareness) is a major concern of both patients undergoing anesthetics and health-care providers administering anesthetics. 1 With an incidence of 0.13% in the general population, 2 the topic of intraoperative awareness has come under scrutiny by both the lay press and the scientific

105 community. This has intensified the search for the Holy Grail of intraoperative 95 anesthesia-a reliable, continuous monitor of the depth of anesthesia. 3 However, adequate depth of anesthesia is a vague term that spans from a state of sedation and amnesia that prevents explicit recall 4 to a state where there is no movement 5 or no hemodynamic response to surgical stimuli. 6 Furthermore, delivery of a single anesthetic drug class (e.g., volatile anesthetic or propofol) results in a different anesthetic profile than when a balanced anesthetic is delivered. 7 Therefore, complete monitors of the depth of anesthesia must characterize these clinical endpoints during the administration of a variety of combinations of anesthetics. 8 Processed EEG parameters are gaining popularity as intra-operative monitors of depth of anesthesia. 3 One depth of anesthesia monitor, the Bispectral Index (BIS, Aspect Medical Systems, Newton, MA), is based on Bispectral analysis of the EEG. 9 The propriety BIS algorithm was a unique step forward in the use of EEG parameters to determine anesthetic depth because it combined multiple distinct EEG parameters and a large volume of prospectively collected clinical observations into a single descriptive variable which was then prospectively tested and validated. 3 The BIS is the only processed EEG that has been found to decrease the incidence of intraoperative awareness in a randomized controlled trial of patients with a large number of risk factors for intraoperative awareness. 10 In addition, titrating anesthetics to specific BIS target values has been found to effect clinical outcomes-a BIS of results in faster emergence from anesthesia, 11 whereas avoiding deep anesthesia (BIS < 40) may improve one year survival of patients. 12

106 96 During general anesthesia, the brainstem and the midbrain auditory function is preserved, although meaningful interpretation of the auditory stimulus is inhibited. 13,14 These brainstem responses to an auditory stimulus correlate with motor signs of wakefulness and intraoperative awareness. 15 The preservation of brainstem responses that correlate with inadequate anesthesia (movement or awareness) suggests that the auditory evoked potential (AEP) might be more robust in detecting inadequate anesthesia as opposed to the EEG which solely monitors the cortical activity. 16,17 The A-Line AEP Index (AAI, Danmeter, Odense, Denmark) is the first commercially available monitor that utilizes changes within the AEP to measure the depth of anesthesia. 18 Like the BIS, the AAI correlates well with the clinical level of sedation produced by increasing doses of sevoflurane 13,19 or propofol. 9,20 Although adequate surgical anesthesia can be produced utilizing a volatile anesthetic alone, 5,6 hemodynamic depression 21 and prolonged time to awakening 22 limit the practicality of utilizing a volatile anesthetic as the sole anesthetic agent. Therefore, an opioid analgesic is commonly coadministered with smaller doses of a volatile anesthetic to provided adequate analgesia and maintain a state of nonresponsiveness to surgical stimulation. 23 The addition of opioids is known to synergistically increase the clinical level of sedation produced by propofol 24,25 and volatile anesthetics. 26,27 However, the effects of the addition of an opioid on the processed EEG parameters is controversialsome reports show that the processed EEG is insensitive to opioids, 9,28,29 whereas others suggest that opioids do alter processed EEG parameters Therefore, the true effects of the addition of opioids to hypnotic drugs on the BIS (and AAI) are unclear.

107 97 The principle aim of this study was to characterize how the addition of opioids sufficient to change the clinical level of sedation influenced processed EEG parameters such as BIS and AAI. Data acquired from volunteers receiving various target concentration pairs of sevoflurane and remifentanil were utilized to construct response surfaces models of the observed level of sedation and the measured EEG parameters. We hypothesized that the processed EEG parameters (BIS and AAI) do not accurately reflect the level of clinical sedation observed with the addition of remifentanil to a sevoflurane anesthetic. In addition, we hypothesized that with the co-administration of remifentanil and sevoflurane, attempting to maintain a target BIS of or a target AAI of would result in overdosing the anesthetic-sevoflurane-remifentanil target concentration pairs well above those that provide clinically adequate anesthesia (e.g., no awareness, no movement, and no hemodynamic response in response to stimulation). 4.3 Materials and Methods A portion of the data from this data set were published previously in a manuscript examining the synergistic interaction between remifentanil and sevoflurane in producing clinical sedation and analgesia to experimental painful stimuli that are surrogates for intraoperative painful stimuli. 33 Because of the minor overlap between the hypotheses of the previous and the current manuscript and the large amount of data reported in each manuscript, each analysis is reported in a separate manuscript. A written informed consent document that was approved by the Human Institutional Review Board at the University of Utah Health Sciences Center (Salt Lake City, Utah) was obtained from each of 24 volunteers in this open-label, randomized, parallel group crisscross designed study to asses drug interactions (Figure 4.1). 34 Each

108 98 Figure 4.1: A schematic summary of the infusion scheme. During each of the three study periods the primary drug is administered in a stepwise fashion (solid black line), while in the second and third study periods, the second drug (grey filled area) is held at a constant predicted effect site concentration or measured alveolar concentration. In between each study period there is a washout phase, during which the primary and secondary drugs are allowed to decay to predicted concentrations below that of the subsequent target concentration pair.

109 99 volunteer was randomized to receive a target controlled infusion of remifentanil (predicted effect site concentrations of ng ml -1 ) or sevoflurane (0.3-6 vol % end tidal alveolar concentration) as the primary agent with the other drug acting as the secondary agent (Figure 4.1). The reader is referred to the previous manuscript by Manyam, et al., 33 for complete details regarding the methods of volunteer preparation, drug administration, and data collection. Because the methods of data analysis and statistical analysis have substantial differences from the previous manuscript, they are provided in detail BIS and AAI Measurements To avoid variability arising from hysteresis between plasma concentration and effect site, BIS and AAI were measured at each assessment point 5 minutes after the targeted effect-site concentration (or stable end-tidal concentration) for a primary drug step, was reached. The EEG parameters were averaged in a 40 second interval that preceded the assessment of the Observer s Assessment of Alertness/Sedation score (OAA/S, Table 4.1). 35 This interval was also considered a quiet time where no other changes or assessments were made in the volunteers. Data resulting from faulty sensors or monitor malfunction were not included in the subsequent analyses Demographic Data Analysis Demographic data for the volunteers in each group were compared utilizing an unpaired, two-sided t-test using StatView version (SAS Institute, Inc., Cary, NC) with P < 0.05 considered significant. All demographic data were reported as means with standard deviations.

110 Table 4.1: Observer s Assessment of Alertness/Sedation (OAA/S) Score* 100 Responsiveness Score Responds readily to name spoken in normal tone 5 Lethargic response to name spoken in normal tone 4 Responds only after name is called loudly and/or repeatedly 3 Responds only after mild prodding or shaking 2 Does not respond to mild prodding or shaking 1 Does not respond to noxious stimulus 0 * For the purposes of this study, an OAA/S 1 was considered nonresponsive, whereas an OAA/S 4 was considered awake.

111 Measurement of Association The performance of each of the processed EEG parameters was assessed by comparison against the sedation score (OAA/S). Because a direct correlation can not be calculated between an ordinal variable (OAA/S score) and either of the continuous variables (processed EEG parameters), we calculated the prediction probability (P k ) as described by Smith and Dutton 36 for the association between the clinical sedation scale (OAA/S) and BIS and AAI using SPSS Version 14 (SPSS Inc., Chicago, IL). The P k values were also calculated for BIS and AAI to test their ability to detect the anesthetic state that corresponds with loss of shake and shout responses (OAA/S <=1, Table 4.1) Response Surface Models of the Processed EEG Parameters Response surface models were constructed for each processed EEG parameter using the Greco-Berenbaum model as shown below: 37 E E C. EC 50A 50B C + α. EC C + α. EC 50 A C. EC C. EC A B A B max 50 A EC50B 50 A 50B = n C EC A + C + EC B C where C A, C B are the concentrations of the two drugs, EC 50A, EC 50B are drug concentrations causing 50% of the maximal drug effect, E MAX is the maximal drug effect, α characterizes the extent of interaction between both drugs, n is a measure of response steepness For each processed EEG parameter, the data were pooled and used to fit the three-dimensional response surface using a naïve pooled technique. Model coefficients and standard errors were estimated using MATLAB (MathWorks Inc., Natick, MA). A B 50B n + 1

112 102 Models were built by an iterative process in which the log likelihood between the observations and the model predictions was maximized. The contribution of each coefficient was evaluated by excluding it from the model and determining whether the model deteriorated significantly using the likelihood ratio test ( Likelihood Ratio 30%). The standard error of the model parameters was estimated using the bootstrap method for 5000 iterations. 38 Model performance was evaluated by assessment of Error Prediction (observed vs. predicted probability of effect for each dose combination) and the correlation coefficient. The Error Prediction is defined as the following: Error ediction = 100X Observed Pr edicted / Observed Pr The correlation coefficient of the regression parameter estimates was used to evaluate how well the nonlinear regression models described the observed data. A large value of the correlation coefficient ( 0.7) indicates that the responses predicted from the surface described the observed data well Results All 24 volunteers completed the study. The demographics of the two groups are shown in Table 4.2. There were no differences between the groups except that the sevoflurane group contained equal numbers of male and female volunteers, whereas the remifentanil group was predominately male volunteers. For individual drugs, the relationship between the processed EEG parameters, the measured drug concentrations, and the OAA/S score at each assessment point is shown in Figure 4.2 and summarized in Table 4.3. We observed that most volunteers were sedated (OAA/S 1) at sevoflurane concentrations greater than 1.5 vol. %. Adequate sedation

113 Table 4.2: Demographics of Study Volunteers* 103 Group 1 Sevoflurane Group 2 Remifentanil Age [years] 25.0 ± ± 2.7 Weight [kg] 70.8 ± ± 9.3 Height [cm] ± ± 8.4 Sex [M:F] 4:4 7:1 * All values are given as mean ± standard deviation, except for the ratio of males to females.

114 104 Figure 4.2: A scatter plot showing the relationship among processed EEG parameters, individual anesthetic drug concentrations, and the clinical sedation scores. Each point represents an assessment after target concentrations of the drug were achieved. Open circles represent observations classified as conscious (volunteers responded to verbal command, OAA/S 3), whereas filled circles are considered unconscious.

115 Table 4.3: Prediction Probability (P k ) - OAA/S Score* 105 BIS AAI Sevoflurane End Tidal [vol %] Remifentanil C e [ng ml -1 ] SEVO 0.97 (0.01) 0.87 (0.03) 0.99 (0.01) N/A SEVO-REMI 0.87 (0.01) 0.75 (0.02) 0.87 (0.01) 0.56 (0.03) REMI 0.76 (0.04) 0.52 (0.05) N/A 0.93 (0.02) * Standard Errors are given in parentheses

116 106 could not be achieved at remifentanil concentrations in the clinical range (5-10 ng ml -1 ). Sedation using remifentanil could be achieved at concentrations higher than 20 ng ml -1. Figures 4.3a and 4.3b show the distribution of BIS and AAI at clinically relevant sedation states-loss of responsiveness to shouting (OAA/S = 2), loss of responsiveness to shaking and shouting (OAA/S = 1), and loss of responsiveness to noxious stimulus (OAA/S = 0). The data are presented in a group where only sevoflurane was administered and in a group in which the volunteers received a combination of sevoflurane and remifentanil Response Surface Models The parameters for all the response surface models were identifiable. The Greco model parameters estimated through nonlinear regression are shown in Table 4.4. The estimates of goodness of fit (e.g., Log Likelihood, Standard Errors, and Correlation Coefficient) suggest that the models described the BIS data better than the AAI data. The response surfaces that describe BIS and AAI at various target concentrations of sevoflurane and remifentanil are shown in Figures 4.4a and 4.4b, respectively. Throughout most of the clinically relevant range of concentrations (sevoflurane 0-3 vol % and remifentanil ng ml -1 ) the residual error is below 10%. Isoboles from Logit response surface models for clinical sedation (95% probability of OAA/S score 1) and tolerance of surgical incision (the 95% probability of no movement or hemodynamic response to a 50 ma electric tetanic stimulation) previously reported by Manyam, et al., 33 are shown in Figures 4.5 and 4.6. In addition, the raw data for each of the processed EEG parameters and the predicted processed EEG parameter values for the concentration target pairs on the previously described isoboles

117 107 Figure 4.3: A box plot showing the distribution of BIS (top panel, Figure 4.3a) and AAI (bottom panel, Figure 4.3b) at clinically relevant sedation states (OAA/S 2). The data are presented in two groups-the first group (open boxes) show the distribution in the processed EEG parameters where volunteers received only sevoflurane. The second group (shaded boxes) show the distribution in the processed EEG parameters when the volunteers received a combination of sevoflurane and remifentanil.

118 Figure

119 Table 4.4: Mean Model Parameters for the Greco Response Surface for Sevoflurane and Remifentanil * 109 EC 50,Sevoflurane EC 50,Remifentanil Synergy Gamma Log Correlation [vol %] [ng ml -1 ] (α) (γ) Likeli Coefficient hood BIS 2.37 (0.06) (2.57) 0.52(0.39) 1.12(0.02) AAI 0.62 (0.06) (17.40) 1.15(1.33) 1.12(0.08) * Standard Errors are given in parentheses

120 110 Figure 4.4: The Greco response surface model predictions of the sevofluraneremifentanil interaction for BIS (top panel, Figure 4.4a) and AAI (bottom panel, Figure 4.4b) for unstimulated volunteers are presented. The symbols show measured responses and the surface predicted by the model is represented by the grid-lined surface. The raw data used to create this model is shaded based on the residual error.

121 111 Figure 4.5: The panel on the top (Figure 4.5a) shows a topographic view of the raw data (BIS) overlaid upon isoboles for adequate clinical sedation (95% probability of achieving an OAA/S score 1) and adequate surgical analgesia (95% probability of no movement response or hemodynamic response to a 50 ma tetanic electrical stimulation). The panel on the bottom (Figure 4.5b) demonstrates the predictions of the BIS response surface model (mean and standard deviation) at different concentration pairs along the isoboles for adequate clinical sedation and surgical analgesia.

122 Figure Sevoflurane %V/V EC 95 P(OAAS 1) EC 95 P(Tol. 50 ma) BIS < Remifentanil ng/ml 3 43(0.8) 45(0.8) EC 95 P(OAAS 1) EC 95 P(Tol. 50 ma) 49(0.8) Sevoflurane %V/V (0.9) 57(0.9) 61(1) 68(0.5) 71(0.5) 65(1) 74(0.5) 77(0.6) 69(1) 80(0.7) 81(0.9) 79(1) 69(1) 68(2) 77(2) 72(2) Remifentanil ng/ml

123 113 Figure 4.6: The panel on the top (Figure 4.6a) shows a topographic view of the raw data (AAI) overlaid upon isoboles for adequate clinical sedation (95% probability of achieving an OAA/S score 1) and adequate surgical analgesia (95% probability of no movement response or hemodynamic response to a 50 ma tetanic electrical stimulation). The panel on the bottom (Figure 4.6b) demonstrates the predictions of the AAI response surface model (mean and standard deviation) at different concentration pairs along the isoboles for adequate clinical sedation and surgical analgesia.

124 Figure Sevoflurane %V/V EC 95 P(OAAS 1) EC 95 P(Tol. 50 ma) AAI < Remifentanil ng/ml 3 28(10) 29(10) EC 95 P(OAAS 1) EC 95 P(Tol. 50 ma) 30(10) Sevoflurane %V/V (10) 34(10) 37(9) 39(9) 42(9) 40(9) 44(9) 47(9) 51(9) 55(10) 45(9) 58(10) 49(9) 52(9) 60(11) 62(11) Remifentanil ng/ml

125 115 are overlaid onto the isoboles. These figures clearly demonstrate that the addition of small amounts of remifentanil (2.5 ng ml -1 ) results in an increase in the target BIS and AAI necessary to produce clinically adequate sedation or anesthesia (Figures 4.5b and 4.6b). 4.5 Discussion In this study, we utilized the volunteer paradigm previously employed by our laboratory 25,33 and others 27,40,41 to generate response surface models for two anatomically distinct processed EEG parameters (BIS and AAI) during the concomitant administration of a wide range of target concentration pairs of a prototypic potent volatile anesthetic, sevoflurane, and a prototypic potent synthetic opioid, remifentanil. Although we had previously demonstrated that remifentanil synergistically potentiates the sedative effects of sevoflurane, 33 we did not observe more than a mild, additive increase in BIS and AAI with the addition of remifentanil to a sevoflurane anesthetic. The fact that the BIS and AAI are both insensitive to the observed changes in the clinical sedation state produced by the addition of a small to moderate dose of remifentanil to a sevoflurane anesthetics suggests that sevoflurane-remifentanil anesthetics titrated to traditional BIS or AAI targets would result in a deeper than predicted anesthetic state. With an estimated effect site concentration of 5 ng ml -1 of remifentanil (an infusion of approximately 0.2 µg kg - 1 min -1 ), no more than 1% sevoflurane is required to produce clinically adequate anesthesia without any concern of explicit recall, and yet the BIS would > 65 and the AAI would be > 40. Therefore, during sevoflurane-remifentanil anesthesia, targeting a BIS < 60 or an AAI <30 may result in too deep of an anesthetic state. This work identifies an important limitation of the currently available algorithms of two distinct

126 processed EEG parameters and should serve as the basis for future development and validation of any depth of anesthesia monitor Concentration-Effect Relationship When examining the effects of prototypic anesthetic agents from a single drug class on the processed EEG parameters, the administration of sedatives-hypnotic agents (e.g., sevoflurane or propofol) results in a clear dose dependent increase in anesthetic depth. In contrast, the administration of an opioid in isolation does very little to decrease the processed EEG parameter (increase anesthetic depth) until extremely high concentrations of the opioid are achieved. Our results were similar-we observed that the BIS and AAI correlate well with sevoflurane concentrations (P k s 0.97 and 0.87, respectively) and more poorly with remifentanil (P k s 0.76 and 0.52, respectively). In addition, the BIS had a wider dynamic range in response to increasing drug concentration than the AAI, consistent with previous reported response of the BIS and AAI The wider dynamic range available with the BIS could potentially translate into easier titration of sevoflurane than with the small dynamic range of the AAI. However, the ability of a monitor to track the concentration changes of a drug does not necessarily improve its performance in predicting the depth of anesthesia. Therefore, when developing algorithms to measure clinical depth of anesthesia, it is more important to focus on capturing the clinical anesthetic state rather than the change in anesthetic drug concentration(s). We determined the concentration-cns effect relationship of opioids using remifentanil as a prototype opioid. Although, a remifentanil effect site concentration above 15 ng ml -1 (an infusion of approximately 0.6 µg kg -1 min -1 ) is rarely used in

127 clinical practice we sampled remifentanil concentrations up to 60 ng ml in an attempt to rigorously capture the sedative effects of remifentanil. Within the clinical range, we did not observe a clinically significant level of sedation with remifentanil. The variability in BIS and AAI within this range was similar to that observed when volunteers did not reach a clinical level of sedation with sevoflurane. At supra-clinical remifentanil concentrations, remifentanil produced a clinically significant level of sedation; however, this opioid induced sedation rarely approached an OAA/S score of 1. In addition, increasing the level of clinical sedation with remifentanil did not alter the AAI although the BIS decreased modestly. Our results are similar to previous reports that showed that the processed EEG parameters are insensitive to opioids 9,28,29 within the clinical range Prediction Probability Several previous reports have demonstrated that the BIS and the AAI are useful surrogates of depth of anesthesia. 3 The BIS showed less variation at each level of clinical sedation than did the AAI (Figures 4.3a and 4.4b). This may be an intrinsic characteristic of the arbitrary scaling of the AAI to have its operating range for general anesthesia between 0-30, therefore, small changes in clinical state might result in a large (erroneous) increase in AAI. An alternative explanation might be the fact that the brainstem auditory pathways are well preserved during moderate levels of anesthetics resulting in an increased sensitivity to ascending (sensory) signals. 13,14 Finally, the increased variability may simply be the result of the more primitive (and poorer performing) electromyogram filtering algorithms available on the early model AAI compared to the more developed BIS.

128 118 Our results are in agreement with previous reports that have demonstrated that the BIS outperforms the AAI when evaluating the performances of the processed EEG parameters utilizing Prediction Probabilities (P k ). 18,45 However, prediction probabilities are limited in that they report only the direction and the goodness of correlation between the clinical sedation score and the processed EEG parameter- they do not give any feel to whether the change in the parameter is large or small. Therefore, even though the addition of remifentanil to a sevoflurane anesthetic resulted in a minor change in processed EEG parameters that was underwhelming compared to the large change in clinical sedation, the modest decrease in the prediction probabilities does not reflect the inability of the BIS or the AAI to capture the observed clinical change Response Surface Models As a complement to prediction probability analyses, response surfaces analysis was used to study the pharmacodynamic effects of adding remifentanil to a sevoflurane anesthetic. Response surface methods have been utilized to model the interactions between varieties of combinations of anesthetics. Using the Greco form of the response surfaces models, we were able to characterize the relationship between the effect site concentrations of remifentanil, the end tidal concentrations, and the BIS with a low amount of error (R 2 > 0.8). The response surface model for AAI had moderately good correlation (R 2 > 0.8), with the poorer fit most likely related to the larger variability in the response and the smaller operating range. The pharmacodynamic response surface revealed that the addition of remifentanil decreased the BIS in a minor and additive fashion, whereas the AAI response surface showed that AAI is not significantly affected by the addition remifentanil. A possible explanation for this difference is that the

129 119 brainstem responses are relatively resistant to opioid effects 46 while the cortical responses are decreased with the inhibition of ascending sensory signals. 47 In order to give clinical meaning to the predictions made by the response surface models for processed EEG parameters, we utilized the Logit response surface models for adequate sedation for general anesthesia (probability of providing an OAA/S score 1) and for adequate analgesia for general anesthesia (probability of no movement of hemodynamic response to a 50 ma electrical tetanic current) that were previously described by our laboratory 33 to generate 95% tile isoboles. Then the predicted values for the BIS (Figure 4.5b) and the AAI (Figure 4.6b) for a variety of target concentration pairs of sevoflurane and remifentanil that lay on the two isoboles were calculated from the response surface models generated in this manuscript. These figures demonstrate that with the addition of a modest remifentanil effect site concentration of 5 ng ml -1 (an infusion of approximately 0.2 µg kg -1 min -1 ), adequate sedation would be provided with a BIS of 81 and an AAI of 57 and adequate general anesthesia would be provided with a BIS of 65 and an AAI of 41- all values considerably higher than the usual target range of either of the processed EEG parameters (BIS and AAI 15-30). Therefore, the inability of the two anatomically distinct processed EEG parameters to characterize the increase in clinical sedation and the increase in clinical anesthetic depth brought about by the addition of even modest doses of remifentanil to a sevoflurane anesthetic would result in an overdose in the amount of sevoflurane administered and too deep of a clinical anesthetic level being targeted (Figure 4.7).

130 120 Figure 4.7: The isoboles that produce target BIS values of 40, 50, 60, and 70 are overlaid upon isoboles for adequate clinical sedation (95% probability of achieving an OAA/S score 1) and adequate surgical analgesia (95% probability of no movement response or hemodynamic response to a 50 ma tetanic electrical stimulation).

131 4.5.4 Clinical Implications 121 The processed electroencephalogram (EEG) has emerged as an important surrogate measure of the depth of anesthesia. 9,48 Surrogate measures are employed when the clinical drug effect of interest is difficult or impossible to measure. The processed EEG has many characteristics of the ideal surrogate. In contrast to more clinically oriented measures of drug effect, it is can be an objective, continuous, reproducible, noninvasive, high resolution signal. 3 It can also be used as an effect measure when an experimental subject is unconscious, whereas many of the more clinically oriented measurements require awake, cooperative subjects. 49 The ability of the addition of even a small amount of synthetic opioid to decrease the amount of potent volatile anesthetic required to produce clinically adequate anesthesia has been reported in surgical patients using isobologram or dose reduction analyses. Furthermore, previous work from our laboratory has demonstrated that the addition of remifentanil to sevoflurane 33 or propofol 25 anesthetics results in a synergistic increase in depth of anesthesia. In contrast, the lack of ability of the two processed EEG parameters studied here to detect the increase in anesthetic depth produced by the addition of even modest amounts of a synthetic opioid has been demonstrated with isobologram analysis of surgical patients. Similar to our results here, response surface analysis performed by Dahan, et al., 27 investigating the interaction of moderate levels of alfentanil and sevoflurane anesthesia has shown that there is no increase in anesthetic depth as measured by the BIS. Therefore, it would appear that despite the clinically significant increase in the clinical sedation level and the anesthetic depth produced by the addition of modest amounts of remifentanil to a sevoflurane anesthetic, there is minimal

132 122 effect of even supra-therapeutic doses of opioid on the depth of anesthesia measured by the BIS and the AAI. Our response surface models demonstrate that the targeting the familiar operating range for the BIS of would result in a % higher end tidal sevoflurane concentration being administered than would be needed to provided clinically adequate anesthesia if a modest dose of remifentanil (effect site concentration of 5 ng ml -1 ) was administered (Figure 4.7). Besides the anticipated hemodynamic side effects expected from this anesthetic overdose, 21 if delivering too deep of an anesthetic (BIS < 40) results in a reproducible increase in one year mortality, 12 the resulting deep anesthesia could have significant implications long after the perioperative period has ended. Therefore, either new context sensitive operating ranges for the processed EEG parameters must be derived to account for the unmeasured effects of the addition of varying doses of opioids; a suitable easily usable fudge factor should be derived for adjusting the measured processed EEG parameter for the opioid contribution; any anesthesiologist who wanted to utilize a processed EEG parameter to titrate the administered anesthetic should limit the administration of opioid to the emergence period as to avoid needing to calculate the corrected BIS or AAI; or a monitor sensitive to the actual clinical conditions, with or without opioids, needs to be developed. It is possible that the combination of real-time pharmacokinetic-pharmacodynamic displays 50 with the addition of the response surfaces described here would be able to numerically and graphically provide anesthesiologists with real time feedback as to the actual (predicted) clinical depth of anesthesia during a balanced anesthetic. However, the lack of a ready solution suggests that the delivery of a balanced anesthetic utilizing a closed loop controlled based on any of the conventional

133 123 processed EEG parameters could possibly result in clinically deeper anesthetics than desired, especially if the algorithm attempts to utilize the unique pharmacokinetic and pharmacodynamic characteristics of remifentanil to improve responsiveness and pharmacologic control. 51 Previously, we had identified optimum target combinations of sevoflurane and remifentanil that provided adequate surgical anesthesia and minimized the time to awakening. 33 For anesthetics ranging in length between hours, the target sevoflurane concentration varied from % and the target remifentanil concentration ranged from ng ml -1 (infusion rates of µg kg -1 min -1 ). Targeting these optimum combinations would produce clinically adequate surgical anesthesia with BIS (65-69) and AAI (41-46) higher than the normal operating ranges suggested by the manufactures Limitations The fact that our response surface models were determined in unstimulated volunteers is a major constraint that may limit the applicability of our results. In particular, the lack of constant stimulation from an endotracheal tube or the continuous pain form a surgical incision may result in our volunteer data underestimating the anesthetic requirements of surgical patients. However, the advantages of the volunteer study paradigm to develop response surface models-key surgical stimulation can be applied multiple times, repeated measurements can be made on the same subject, and the entire dynamic range of anesthetic combinations can be examined, all without ethical concerns of providing inadequate anesthesia during a surgical procedure, continues to make the volunteer study paradigm popular.

134 124 The fact that we utilize pharmacokinetic models to predict the remifentanil effect site concentration in lieu of measuring the actual blood drug concentration may compound some of the variability in the opioid only, single drug data. 52 However, as in our previous study, 33 there is convincing evidence to demonstrate that this may not be a major source of pharmacokinetic variability. Another source of pharmacokinetic variability may be the targeting of an end tidal alveolar pseudo-steady state of volatile anesthetic instead of targeting the effect site concentration. The steady-state partial pressure of the volatile anesthetic at the effect site correlates with the measured end tidal alveolar partial pressure at steady state. However, achieving pseudo-steady state at the alveoli results in an effect site concentration that would most likely not reach its own pseudo-steady state. We did not choose to target a pseudo-steady state at the effect site because we would have to assume a priori knowledge of which anatomic compartment contained the pharmacologic effect site for sedation and for clinical anesthesia. Given the fact that volatile anesthetics produce sedation through a supra-spinal site of action while immobility is produced at the spinal cord level, 53 the choice of effect site to target in the pharmacokinetic simulations to determine when pseudo-steady state at the effect site is achieved is one of many difficult assumptions that would be needed to construct an accurate pharmacokinetic-pharmacodynamic model for sevoflurane. In addition, the time involved in achieving a steady state alveolar concentration or a pseudo-steady state effect site concentration would be prohibitively longer than that required to achieve alveolar pseudo-steady state. Although remifentanil induced hyperalgesia has been observed in the patients 54 and volunteers 55 receiving infusions of various durations, as detailed in our prior

135 125 manuscript, 33 we did not find any differences between the baseline levels of tolerated stimuli and the levels of stimuli tolerated at the lowest doses of sevoflurane. In addition, one could conjecture that any opioid hyperalgesia that developed would not effect the clinical sedation score (OAA/S) or the processed EEG parameters that were determined during quiet periods prior to the determination of the analgesic response of each of the targeted concentration pairs. The Greco response surface model used to describe the response surface models generated here is different than the Logit model utilized in the previous manuscripts from our laboratory. 33,56 Although the Logit model proved advantageous for the modelling of stimuli whose responses can be dichotomized, the Greco model, 37 along with the models described by Minto 57 and Bouillon, 58 all handle continuous response variables (e.g., processed EEG parameters) extremely well. The main advantage of the Greco model is that it assumes a sigmoidal E max structure that is readily familiar to most readers of pharmacodynamic modelling. The biggest limitation of the Greco model is that it cannot account for a partial agonist-it presumes that remifentanil at sizeable concentrations will produce a BIS or AAI of 0. This assumption causes a bias in the determination of the response surface, however, because no model that accounts for partial agonists currently exists, there is no way to overcome this limitation. Even with the assumption that Greco model does not account for a partial agonist, by setting the C MAX, REMI at a high enough value (i.e., 400 ng ml -1 ), the error in the response surface is not significantly large to cause a change in model predictions.

136 4.5.6 Conclusions 126 Although clinical sedation increases significantly with the addition of a small to moderate dose of remifentanil to a sevoflurane anesthetic, the BIS and AAI are insensitive to this change in clinical state. Therefore, during sevoflurane-remifentanil anesthesia, targeting a BIS < 60 or an AAI <30 may result in too deep of an anesthetic state. If providing too deep of an anesthetic state produces undesirable side effects, such as intraoperative hemodynamic instability or an increase in one year mortality, correcting the measured processed EEG parameter to account for the actual measured clinical anesthetic depth would be required to prevent these undesirable side effects. As a first step, by superimposing the isobolograms for adequate surgical anesthesia and adequate sedation on top of the isobolograms for various targets values for BIS or AEP, a figure is developed that can be utilized to make crude clinical adjustments to either the combination of sevoflurane and remifentanil administered or the targeted BIS or AEP value necessary to produce the desired clinical depth of anesthesia (Figure 4.7). Incorporation of these response surfaces into a real-time, pharmacokineticpharmacodynamic display system 50 may allow more precise concentration pairs or target adjustments. 4.6 References 1. Practice Advisory for Intraoperative Awareness and Brain Function Monitoring: A Report by the American Society of Anesthesiologists Task Force on Intraoperative Awareness. Anesthesiology 2006; 104: Sebel PS, Bowdle TA, Ghoneim MM, Rampil IJ, Padilla RE, Gan TJ, Domino KB: The incidence of awareness during anesthesia: a multicenter United States study. Anesth Analg 2004; 99: 833-9, table of contents 3. Rampil IJ: A primer for EEG signal processing in anesthesia. Anesthesiology 1998; 89:

137 Stoelting RK, Longnecker DE, Eger EI, 2nd: Minimum alveolar concentrations in man on awakening from methoxyflurane, halothane, ether and fluroxene anesthesia: MAC awake. Anesthesiology 1970; 33: Eger EI, 2nd, Saidman LJ, Brandstater B: Minimum alveolar anesthetic concentration: a standard of anesthetic potency. Anesthesiology 1965; 26: Roizen MF, Horrigan RW, Frazer BM: Anesthetic doses blocking adrenergic (stress) and cardiovascular responses to incision--mac BAR. Anesthesiology 1981; 54: Glass PS, Gan TJ, Howell S, Ginsberg B: Drug interactions: volatile anesthetics and opioids. J Clin Anesth 1997; 9: 18S-22S 8. Kalkman CJ, Drummond JC: Monitors of depth of anesthesia, quo vadis? Anesthesiology 2002; 96: Glass PS, Bloom M, Kearse L, Rosow C, Sebel P, Manberg P: Bispectral analysis measures sedation and memory effects of propofol, midazolam, isoflurane, and alfentanil in healthy volunteers. Anesthesiology 1997; 86: Myles PS, Leslie K, McNeil J, Forbes A, Chan MT: Bispectral index monitoring to prevent awareness during anaesthesia: the B-Aware randomised controlled trial. Lancet 2004; 363: Song D, Joshi GP, White PF: Titration of volatile anesthetics using bispectral index facilitates recovery after ambulatory anesthesia. Anesthesiology 1997; 87: Monk TG, Saini V, Weldon BC, Sigl JC: Anesthetic management and one-year mortality after noncardiac surgery. Anesth Analg 2005; 100: Schwender D, Conzen P, Klasing S, Finsterer U, Poppel E, Peter K: The effects of anesthesia with increasing end-expiratory concentrations of sevoflurane on midlatency auditory evoked potentials. Anesth Analg 1995; 81: Plourde G, Belin P, Chartrand D, Fiset P, Backman SB, Xie G, Zatorre RJ: Cortical processing of complex auditory stimuli during alterations of consciousness with the general anesthetic propofol. Anesthesiology 2006; 104: Schwender D, Klasing S, Madler C, Poppel E, Peter K: Depth of anesthesia. Midlatency auditory evoked potentials and cognitive function during general anesthesia. Int Anesthesiol Clin 1993; 31:

138 Schwender D, Daunderer M, Mulzer S, Klasing S, Finsterer U, Peter K: Midlatency auditory evoked potentials predict movements during anesthesia with isoflurane or propofol. Anesth Analg 1997; 85: Schwender D, Golling W, Klasing S, Faber-Zullig E, Poppel E, Peter K: Effects of surgical stimulation on midlatency auditory evoked potentials during general anaesthesia with propofol/fentanyl, isoflurane/fentanyl and flunitrazepam/fentanyl. Anaesthesia 1994; 49: Struys MM, Jensen EW, Smith W, Smith NT, Rampil I, Dumortier FJ, Mestach C, Mortier EP: Performance of the ARX-derived auditory evoked potential index as an indicator of anesthetic depth: a comparison with bispectral index and hemodynamic measures during propofol administration. Anesthesiology 2002; 96: Katoh T, Suzuki A, Ikeda K: Electroencephalographic derivatives as a tool for predicting the depth of sedation and anesthesia induced by sevoflurane. Anesthesiology 1998; 88: Schwender D, Faber-Zullig E, Klasing S, Poppel E, Peter K: Motor signs of wakefulness during general anaesthesia with propofol, isoflurane and flunitrazepam/fentanyl and midlatency auditory evoked potentials. Anaesthesia 1994; 49: Zbinden AM, Petersen-Felix S, Thomson DA: Anesthetic depth defined using multiple noxious stimuli during isoflurane/oxygen anesthesia. II. Hemodynamic responses. Anesthesiology 1994; 80: Eger EI, 2nd, Shafer SL: Tutorial: context-sensitive decrement times for inhaled anesthetics. Anesth Analg 2005; 101: , table of contents 23. Kissin I: General anesthetic action: an obsolete notion? Anesth Analg 1993; 76: Vuyk J: Pharmacokinetic and pharmacodynamic interactions between opioids and propofol. J Clin Anesth 1997; 9: 23S-26S 25. Kern SE, Xie G, White JL, Egan TD: A response surface analysis of propofol-remifentanil pharmacodynamic interaction in volunteers. Anesthesiology 2004; 100: Olofsen E, Sleigh JW, Dahan A: The influence of remifentanil on the dynamic relationship between sevoflurane and surrogate anesthetic effect measures derived from the EEG. Anesthesiology 2002; 96:

139 Dahan A, Nieuwenhuijs D, Olofsen E, Sarton E, Romberg R, Teppema L: Response surface modeling of alfentanil-sevoflurane interaction on cardiorespiratory control and bispectral index. Anesthesiology 2001; 94: Guignard B, Menigaux C, Dupont X, Fletcher D, Chauvin M: The effect of remifentanil on the bispectral index change and hemodynamic responses after orotracheal intubation. Anesth Analg 2000; 90: Iselin-Chaves IA, Flaishon R, Sebel PS, Howell S, Gan TJ, Sigl J, Ginsberg B, Glass PS: The effect of the interaction of propofol and alfentanil on recall, loss of consciousness, and the Bispectral Index. Anesth Analg 1998; 87: Heck M, Kumle B, Boldt J, Lang J, Lehmann A, Saggau W: Electroencephalogram bispectral index predicts hemodynamic and arousal reactions during induction of anesthesia in patients undergoing cardiac surgery. J Cardiothorac Vasc Anesth 2000; 14: Koitabashi T, Johansen JW, Sebel PS: Remifentanil dose/electroencephalogram bispectral response during combined propofol/regional anesthesia. Anesth Analg 2002; 94: , table of contents 32. Lysakowski C, Dumont L, Pellegrini M, Clergue F, Tassonyi E: Effects of fentanyl, alfentanil, remifentanil and sufentanil on loss of consciousness and bispectral index during propofol induction of anaesthesia. Br J Anaesth 2001; 86: Manyam SC, Gupta DK, Johnson KB, White JL, Pace NL, Westenskow DR, Egan TD: Opiod-Volatile Anesthetic Synergy: A Response Surface Model with Remifentanil and Sevoflurane as Prototypes. Anesthesiology 2006: in press 34. Short TG, Ho TY, Minto CF, Schnider TW, Shafer SL: Efficient trial design for eliciting a pharmacokinetic-pharmacodynamic model-based response surface describing the interaction between two intravenous anesthetic drugs. Anesthesiology 2002; 96: Chernik DA, Gillings D, Laine H, Hendler J, Silver JM, Davidson AB, Schwam EM, Siegel JL: Validity and reliability of the Observer's Assessment of Alertness/Sedation Scale: study with intravenous midazolam. J Clin Psychopharmacol 1990; 10: Smith WD, Dutton RC, Smith NT: Measuring the performance of anesthetic depth indicators. Anesthesiology 1996; 84: Greco WR, Bravo G, Parsons JC: The search for synergy: a critical review from a response surface perspective. Pharmacol Rev 1995; 47:

140 38. Jacquez JA, Perry T: Parameter estimation: local identifiability of parameters. Am J Physiol 1990; 258: E Glantz SA, Slinker KK: 2nd Edition. Primer of Applied Regression and Analysis of Variance Vuyk J: Clinical interpretation of pharmacokinetic and pharmacodynamic propofol-opioid interactions. Acta Anaesthesiol Belg 2001; 52: Minto CF, Schnider TW, Short TG, Gregg KM, Gentilini A, Shafer SL: Response surface model for anesthetic drug interactions. Anesthesiology 2000; 92: Ekman A, Brudin L, Sandin R: A comparison of bispectral index and rapidly extracted auditory evoked potentials index responses to noxious stimulation during sevoflurane anesthesia. Anesth Analg 2004; 99: , table of contents 43. Alpiger S, Helbo-Hansen HS, Vach W, Ording H: Efficacy of A-line AEP Monitor as a tool for predicting acceptable tracheal intubation conditions during sevoflurane anaesthesia. Br J Anaesth 2005; 94: Alpiger S, Helbo-Hansen HS, Vach W, Ording H: Efficacy of the A-line AEP monitor as a tool for predicting successful insertion of a laryngeal mask during sevoflurane anesthesia. Acta Anaesthesiol Scand 2004; 48: Kreuer S, Bruhn J, Larsen R, Buchinger H, Wilhelm W: A-line, bispectral index, and estimated effect-site concentrations: a prediction of clinical end-points of anesthesia. Anesth Analg 2006; 102: Schwender D, Rimkus T, Haessler R, Klasing S, Poppel E, Peter K: Effects of increasing doses of alfentanil, fentanyl and morphine on mid-latency auditory evoked potentials. Br. J. Anaesth. 1993; 71: Morley AP, Derrick J, Seed PT, Tan PE, Chung DC, Short TG: Isoflurane dosage for equivalent intraoperative electroencephalographic suppression in patients with and without epidural blockade. Anesth Analg 2002; 95: , table of contents 48. Gan TJ, Glass PS, Windsor A, Payne F, Rosow C, Sebel P, Manberg P: Bispectral index monitoring allows faster emergence and improved recovery from propofol, alfentanil, and nitrous oxide anesthesia. BIS Utility Study Group. Anesthesiology 1997; 87: Leslie K, Sessler DI, Smith WD, Larson MD, Ozaki M, Blanchard D, Crankshaw DP: Prediction of movement during propofol/nitrous oxide anesthesia. Performance of concentration, electroencephalographic, pupillary, and hemodynamic indicators. Anesthesiology 1996; 84: 52-63

141 Syroid ND, Agutter J, Drews FA, Westenskow DR, Albert RW, Bermudez JC, Strayer DL, Prenzel H, Loeb RG, Weinger MB: Development and evaluation of a graphical anesthesia drug display. Anesthesiology 2002; 96: Struys MM, Mortier EP, De Smet T: Closed loops in anaesthesia. Best Pract Res Clin Anaesthesiol 2006; 20: Avram MJ, Krejcie TC: Using front-end kinetics to optimize targetcontrolled drug infusions. Anesthesiology 2003; 99: Rampil IJ, Mason P, Singh H: Anesthetic potency (MAC) is independent of forebrain structures in the rat. Anesthesiology 1993; 78: Crawford MW, Hickey C, Zaarour C, Howard A, Naser B: Development of Acute Opioid Tolerance During Infusion of Remifentanil for Pediatric Scoliosis Surgery. Anesth Analg 2006; 102: Angst MS, Koppert W, Pahl I, Clark DJ, Schmelz M: Short-term infusion of the mu-opioid agonist remifentanil in humans causes hyperalgesia during withdrawal. Pain 2003; 106: Gupta DK, Manyam SC, Johnson KB, White JL, Pace NL, Westenskow DR, Egan TD: Does the Ideal Combination of Remifentanil and Propfol Change with the Duration of Surgery? Anesthesiology 2006: submitted 57. Minto C, Vuyk J: Response surface modelling of drug interactions. Adv Exp Med Biol 2003; 523: Bouillon TW, Bruhn J, Radulescu L, Andresen C, Shafer TJ, Cohane C, Shafer SL: Pharmacodynamic interaction between propofol and remifentanil regarding hypnosis, tolerance of laryngoscopy, bispectral index, and electroencephalographic approximate entropy. Anesthesiology 2004; 100:

142 CHAPTER 5 PROCESSED EEG SIGNALS AS INDICATORS OF INADEQUATE ANESTHESIA 5.1 Abstract Background The processed auditory evoked potential (AAI, Danmeter, Odense, Denmark) and the Bispectral Index (BIS, Aspect Medical Systems, Newton, MA) of the electroencephalogram are two mechanistically different technologies used to assess the functional depression of the central nervous system during general anesthesia. The aim of this study was to compare how the AAI and BIS perform in response to noxious stimulation in volunteers who are profoundly sedated. This study examines the possibility of using the AAI and BIS monitors intraoperatively to detect patient responses to stimulation under inadequate anesthesia Methods After obtaining institutional approval and informed consent, twenty two healthy adult male and female volunteers were enrolled. Volunteers received a combination of opioid (remifentanil, REMI) and hypnotic drug (sevoflurane, SEVO or propofol, PROP) Will be submitted for review in Anesthesia & Analgesia, July Will be published in Anesthesia & Analgesia pending review. Original article titled: The auditory evoked potential and bispectral index: A comparison of signal performance during clinically inadequate anesthesia.

143 133 at various target concentration pairs spanning the entire clinical spectrum (i.e. REMI 0-80 ng ml -1 (computer controlled infusion), PROP mcg ml -1 (computer controlled infusion) and end-tidal SEVO ranging from 0-7% atm). AAI, BIS, and heart rate were digitally acquired throughout the experiment. Baseline AAI and BIS values were recorded after volunteers reached steady-state drug levels. A series of randomly applied experimental pain stimuli (pressure algometry on the leg to 50 psi, electrical tetany on the leg to 50 mamps, and thermal stimuli on the forearm to 50 C) were used to assess the level of anesthesia. Response to stimulation was defined as withdrawal movement or a heart rate increase of 20%. The magnitude and time course of AAI and BIS changes in the first minute after volunteer response were considered the outcome of interest. Artifactual corrupted AAI and BIS signals (movement artifact, etc.) were not analyzed. For volunteers with OAAS <= 1 (i.e., subjects sedated as during general anesthesia), the magnitude of the change in the AAI and the BIS values were plotted versus time to examine the signal response in patients with and without adequate anesthesia, as assessed by heart rate change and withdrawal movement in response to stimulation Results All 22 subjects completed the experiment. The temporal profiles of AAI and BIS values showed responses at a latency of 40 and 50 seconds respectively. For volunteers with sedation scores equivalent to loss of conscious response(oaas >= 1) both the AAI and the BIS values showed robust responses when there was a heart rate or withdrawal movement response to experimental pain stimuli. In those subjects in whom there were no responses to pain, the AAI and BIS values showed no change compared to prestimulation values.

144 5.1.4 Discussion 134 In this observational study, application of experimental pain measures to volunteers receiving various combinations of remifentanil and sevoflurane producing sedation scores equivalent to adequate anesthesia (OASS <= 1) resulted in robust increases in the AAI and BIS values only in those volunteers who showed other signs of inadequate anesthesia-withdrawal movement or increase in heart rate. 5.2 Introduction Clinicians often depend on unreliable, nonspecific measures of anesthetic effect 1 such as hemodynamics, reflexes to stimuli, spontaneous respiration rate, etc. to determine the level of anesthetic effect. To use these methods the clinician is dependent on a number of factors such as training, experience and availability of intraoperative monitoring methods. Some measures such as blood pressure are rarely available on a continuous basis intraoperatively. Hemodynamic responses are often affected by the presence of vasoactive and ionotropic drugs. 2 A practical, more reliable method is needed to determine patient responses to inadequate anesthesia. Such methodology would improve intraoperative monitoring, enable more accurate drug administration, and may eventually lead to closed loop computer controlled drug delivery. 3,4 Processed EEG parameters are gaining popularity as intraoperative monitors of depth of anesthesia. 5 One such example, the Bispectral Index (BIS, Aspect Medical Systems, Newton, MA), is based on Bispectral analysis of the EEG. 6 The propriety BIS algorithm was a unique step forward in the use of EEG parameters to determine anesthetic depth because it combined multiple distinct EEG parameters and a large

145 135 volume of prospectively collected clinical observations into a single descriptive variable which was then prospectively tested and validated. 5 During general anesthesia, the brainstem and the midbrain auditory function is preserved, although meaningful interpretation of the auditory stimulus is inhibited. 7,8 These brainstem responses to an auditory stimulus correlate with motor signs of wakefulness and intraoperative awareness. 9 The preservation of brainstem responses that correlate with inadequate anesthesia (movement or awareness) suggests that the auditory evoked potential (AEP) might be more robust in detecting inadequate anesthesia as opposed to the EEG which solely monitors the cortical activity. 10,11 The A-Line AEP Index (AAI, Danmeter, Odense, Denmark) is the first commercially available monitor that utilizes changes within the AEP to measure the depth of anesthesia. 12 Like the BIS, the AAI correlates well with the clinical level of sedation produced by increasing doses of sevoflurane 7,13 or propofol. 6,14 The principle aim of this study was to measure AAI and BIS responses to stimulation in volunteers who were clinically sedated. We use multimodal experimental pain measures to elicit movement or heart rate responses in volunteers anesthetized using a combination of sevoflurane and remifentanil or propofol or remifentanil. The magnitude and latency of BIS and AAI responses were estimated off line. We hypothesized that the responses shown by processed EEG parameters (BIS and AAI) are comparable to traditional markers of inadequate anesthesia such as increased heart rate or movement. In addition, we hypothesized that the modality of stimulus, i.e., thermal, electrical and mechanical, did bias the responses shown by processed EEG parameters.

146 Materials and Methods A portion of the data from this data set were published previously in a manuscript examining the synergistic interaction between remifentanil and sevoflurane in producing clinical sedation and analgesia to experimental painful stimuli that are surrogates for intraoperative painful stimuli. 15 Because of the minor overlap between the hypotheses of the previous and the current manuscript and the large amount of data reported in each manuscript, each analysis is reported in a separate manuscript. A written informed consent document that was approved by the Human Institutional Review Board at the University of Utah Health Sciences Center (Salt Lake City, Utah) was obtained from each of 24 volunteers in this open-label, randomized, parallel group crisscross designed study to asses drug interactions (Figure 5.1). 16 Each volunteer was randomized to receive a target controlled infusion of remifentanil (predicted effect site concentrations of ng ml -1 ) or target controlled infusion of propofol (predicted effect site concentrations of mcg ml -1 ) or sevoflurane (0.3-6 vol % end tidal alveolar concentration) as the primary agent with the other drug acting as the secondary agent (Figure 5.1). Five minutes after achieving the targeted effect-site concentration (or stable end-tidal concentration) for a primary drug step, a battery of pharmacodynamic assessments were made. Effect measures included the Observer s Assessment of Alertness/Sedation score (OAA/S) 17 and three surrogates for surgical stimulus- pressure algometry and tetanic electrical stimulation, as previously described by Kern, 18 and thermal stimulation. The reader is referred to the previous manuscript by Manyam, et al., 15 for complete details regarding the methods of volunteer preparation, drug administration, and

147 137 Figure 5.1: A schematic summary of the data collection and analysis. At each target concentration pair, baseline measurements of AAI and BIS were determined by averaging monitor indices in a 40 second time window (upper panel). Responses were elicited by gradually increasing stimulus level in until the volunteers showed signs of discomfort (20% increase in heart rate or a movement response). A safety limit of stimulation was defined to prevent long term pain that could confound successive measurements (middle panel). The AAI and BIS signals in the response time window (bottom panel) were used for data analysis. Time zero in the response time window corresponds to the time at which the volunteers responded in case of responders or the time at which the safety limit was reached in the case of non responders.

148 Figure

149 data collection. 139 Because the methods of data analysis and statistical analysis have substantial differences from the previous manuscript, they are provided in complete detail Baseline BIS and AAI Measurements To avoid variability arising from hysteresis between plasma concentration and effect site, BIS and AAI were measured at each assessment point five minutes after the targeted effect-site concentration (or stable end-tidal concentration) for a primary drug step, was reached. The processed EEG parameters were averaged in a 40 second interval that preceded the assessment of the Observer s Assessment of Alertness/Sedation score (OAA/S). 17 This interval was also considered a quiet time where no other changes or assessments were made in the volunteers. Data resulting from faulty sensors or monitor malfunction were not included in the subsequent analyses Demographic Data Analysis Demographic data for the volunteers in each group were compared utilizing an unpaired, two-sided t-test using StatView version (SAS Institute, Inc., Cary, NC) with P < 0.05 considered significant. All demographic data were reported as means with standard deviations Definition of Volunteer Responses Volunteer responses to stimulation were defined as a movement and/ or a 20% increase in heart rate.

150 5.3.4 Time Series Analysis 140 The protocol for determining responses in processed EEG parameters is outlined in Figure 5.1. AAI and BIS signals stored were time aligned with the patient responses. Time zero represents the time at which the volunteers responded to stimulation or the maximal permissible stimulus was reached. Signal analysis was performed using MATLAB (MathWorks Inc., Natick, MA). The magnitude of the response was defined as the percentage change from the baseline assessment. The Percent change is defined as the following: ( StimBISorAEP BaselineBISorAEP ) BaselineBISorAEP PercentCha nge = 100X / The latency of responses was identified by time-averaging all the responses. The window in which the percentage change of the processed EEG parameters exceeded baseline variation was defined as the Time-Window. The average signal within this time window was used in comparing responses across stimuli and comparing responses in volunteers who were awake from those who were sedated. 5.4 Results All 22 volunteers completed the study. The demographics of volunteers are shown in Table 5.1. The time course of AAI and BIS signal changes is shown in Figure 5.2 and Figure 5.3. The percent change at 30 seconds prior to volunteer response, the time of response (0 sec.), 30 and 60 seconds after the response are represented by the box plots. The central line indicates the median value and the whiskers indicate 10 and 90% intervals. The average response, computed by averaging all signals within the response time window is shown in as a gray trace in Figure 5.2 and Figure 5.3. An unpaired t-test indicated that

151 Table 5.1: Demographics of Study Volunteers* 141 Group 1 Sevoflurane Group 2 Propofol Group 2 Remifentanil Age [years] 25.0 ± ± ± 3.0 Weight [kg] 70.8 ± ± ± 9.0 Height [cm] ± ± ± 8.0 Sex [M:F] 4:4 8 : 3 7: 1 * All values are given as mean ± standard deviation, except for the ratio of males to females.

152 142 Figure 5.2: A box plot showing the time course of AAI response to stimulation. The average signal change is shown by the gray trace. The upper panel shows the AAI signal change in volunteers who showed signs of discomfort when stimulated. The bottom panel shows percent change AAI signal change in volunteers who showed no signs of discomfort. Filled circles indicate outlier data.

153 143 Figure 5.3: A box plot showing the time course of BIS response to stimulation. The average signal change is shown by the gray trace. The upper panel shows the BIS signal change in volunteers who showed signs of discomfort when stimulated. The bottom panel shows percent change BIS signal change in volunteers who showed no signs of discomfort. Filled circles indicate outlier data.

154 144 responders and non responders differed with a significance values of < 0.01 for AAI and <0.001 for BIS. The latency of responses was defined as the average time at which the signal increased more that two standard deviations from its mean value Sedated vs. Awake volunteers Signal responses in volunteers who did not have a clinical level of sedation, i.e., OAA/S >=2 volunteers were not oblivious to shaking, were compared against those who were sedated to level at which they did not respond to shaking prior to stimulation. The average percent change indicated in Figure 5.4 indicates larger changes in the signal in volunteers who were awake than sedated Modality of Stimulus Signal responses were compared during stimulation with multiple experimental pain measures, attempted laryngoscopy and OAA/S assessment (Figure 5.5). The monitors showed no preference to a particular stimulus modality although laryngoscopy, considered a much more intense form of stimulation than other experimental showed the largest change. 5.5 Discussion In this study, we utilized the volunteer paradigm previously employed by our laboratory 15,18 and others to elicit responses to stimulation at varying levels of anesthesia. Processed EEG parameters (BIS and AAI) were recorded during stimulation using a variety of experimental pain measures and attempted laryngoscopy. The average change in processed EEG parameter differed in those volunteers that exhibited movement or heart rate increases in response to stimulation from those volunteers that did not respond to stimulation. The changes in processed EEG signals were observed with in a

155 145 Figure 5.4: A box plot comparing signal responses in sedated (OAA/S <=1, or loss of responsiveness to shaking and shouting) and awake (OAA/S >=2, or responsive to shouting) volunteers. The upper and lower panels show AAI and BIS responses respectively. Data is only shown for volunteers who responded to stimulation. Filled circles indicate outlier data.

156 146 Figure 5.5: A box plot comparing signal responses among different stimuli. OAA/S assessment, a predominantly auditory stimulus was considered to see if it produced any change in signals. Data is only shown for volunteers who responded to stimulation. Filled circles indicate outlier data.