Identification of optimal dosing regimens for procedures requiring esophageal instrumentation through multiobjective optimization

The use of propofol and propofol in combination with remifentanil by nonanesthesiologists is a controversial topic. Much of the concern centers on adverse respiratory effects: loss of responsiveness, respiratory depression, and airway obstruction. The aim of this study was to investigate these adverse drug effects at propofol-remifentanil combinations commonly used in procedures requiring esophageal instrumentation and build response surface models of drug effects. A second aim was to investigate published dosing regimens through simulation with these models. A third aim was to develop an optimization algorithm to identify an ideal propofol-remifentanil dosing regimen for upper endoscopy procedures. Twenty-four volunteers received escalating target controlled remifentanil and propofol infusions. Responses to insertion of a bougie (40 cm), responsiveness, respiratory rate, and tidal volume were recorded at 384 targeted concentration pairs. Four published dosing regimens of propofol alone or in combination with opioids were simulated for a 10-min procedure. An optimization algorithm was developed to identify an optimal propofol-remifentanil dosing regimen from a set of possibilities. Models for loss of response to esophageal instrumentation, intolerable ventilatory depression, and respiratory compromise were built. Simulations of published dosing regimens showed that once drug administration ended, loss of responsiveness, and respiratory depression effects dissipated quickly. Respiratory compromise dissipated more quickly in propofol only techniques compared to propofol-opioid techniques. An optimal dosing recommendation was identified for a simulated 55 year-old, 75 kg, 175 cm male undergoing an anticipated 10-min upper endoscopy and consisted of a propofol bolus of 0.8 mg/kg and infusion rate of 40 mcg/kg/min and a remifentanil bolus of 0.2 mcg/kg and an infusion rate of 0.05 mcg/kg/min. High propofol-low remifentanil concentration pairs can block the response to esophageal instrumentation while avoiding intolerable ventilatory depression in spontaneously breathing volunteers. Propofol combined with remifentanil or fentanyl improved conditions for esophageal instrumentation and had a rapid return to responsiveness. Optimization techniques identified a remifentanil propofol dosing regimen that minimizes the duration of loss of responsiveness, respiratory depression, and airway obstruction and, according to expert opinion and models of drug effect, provides conditions that will permit upper endoscopy procedures. This dosing regimen merits clinical validation in patients undergoing brief endoscopic procedures.

IDENTIFICATION OF OPTIMAL DOSING REGIMENS FOR PROCEDURES REQUIRING ESOPHAGEAL INSTRUMENTATION THROUGH MULTIOBJECTIVE OPTIMIZATION by Cristen Doyle LaPierre 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 May 2012 Copyright © Cristen Doyle LaPierre 2012 All Rights Reserved The Uni ve r si ty of Utah Gr adua t e School STATEMENT OF DISSERTATION APPROVAL The dissertation of Cristen Doyle LaPierre has been approved by the following supervisory committee members: Dwayne R. Westenskow , Chair 01/06/12 Date Approved Kenward B. Johnson , Member 01/06/12 Date Approved Rob S. MacLeod , Member 01/06/12 Date Approved Edward W. Hsu , Member 01/06/12 Date Approved Talmage D. Egan , Member 01/06/12 Date Approved and by Patrick A. Tresco , Chair of the Department of Bioengineering and by Charles A. Wight, Dean of The Graduate School. ABSTRACT The use of propofol and propofol in combination with remifentanil by nonanesthesiologists is a controversial topic. Much of the concern centers on adverse respiratory effects: loss of responsiveness, respiratory depression, and airway obstruction. The aim of this study was to investigate these adverse drug effects at propofol-remifentanil combinations commonly used in procedures requiring esophageal instrumentation and build response surface models of drug effects. A second aim was to investigate published dosing regimens through simulation with these models. A third aim was to develop an optimization algorithm to identify an ideal propofol-remifentanil dosing regimen for upper endoscopy procedures. Twenty-four volunteers received escalating target controlled remifentanil and propofol infusions. Responses to insertion of a bougie (40 cm), responsiveness, respiratory rate, and tidal volume were recorded at 384 targeted concentration pairs. Four published dosing regimens of propofol alone or in combination with opioids were simulated for a 10-min procedure. An optimization algorithm was developed to identify an optimal propofol-remifentanil dosing regimen from a set of possibilities. Models for loss of response to esophageal instrumentation, intolerable ventilatory depression, and respiratory compromise were built. Simulations of published dosing regimens showed that once drug administration ended, loss of responsiveness, and respiratory depression effects dissipated quickly. Respiratory compromise dissipated more quickly in propofol only techniques compared to propofol-opioid techniques. An optimal dosing recommendation was identified for a simulated 55 year-old, 75 kg, 175 iv cm male undergoing an anticipated 10-min upper endoscopy and consisted of a propofol bolus of 0.8 mg/kg and infusion rate of 40 mcg/kg/min and a remifentanil bolus of 0.2 mcg/kg and an infusion rate of 0.05 mcg/kg/min. High propofol-low remifentanil concentration pairs can block the response to esophageal instrumentation while avoiding intolerable ventilatory depression in spontaneously breathing volunteers. Propofol combined with remifentanil or fentanyl improved conditions for esophageal instrumentation and had a rapid return to responsiveness. Optimization techniques identified a remifentanil propofol dosing regimen that minimizes the duration of loss of responsiveness, respiratory depression, and airway obstruction and, according to expert opinion and models of drug effect, provides conditions that will permit upper endoscopy procedures. This dosing regimen merits clinical validation in patients undergoing brief endoscopic procedures. CONTENTS ABSTRACT ................................................................................................................... iii LIST OF FIGURES........................................................................................................ vii LIST OF TABLES ........................................................................................................ viii PREFACE ...................................................................................................................... ix 1. INTRODUCTION ......................................................................................................... 1 1.1 Background ................................................................................................... 1 1.2 Goals ............................................................................................................. 2 1.3 References .................................................................................................... 4 2. REMIFENTANIL-PROPOFOL PHARMACODYNAMIC MODELS FOR INTOLERABLE VENTILATORY DEPRESSION, LOSS OF RESPONSIVENESS, AND LOSS OF RESPONSE TO ESOPHAGEAL INSTRUMENTATION ................... 5 2.1 Abstract ......................................................................................................... 5 2.2 Introduction .................................................................................................... 7 2.3 Methods ......................................................................................................... 7 2.4 Results ........................................................................................................ 15 2.5 Discussion ................................................................................................... 28 2.6 Appendix A: Target Effect-site Concentration Sets and Observed Responses. .......................................................................................................................... 36 2.7 Appendix B: Target Effect-site Concentrations and Observed Responses for Intolerable Ventilatory Depression from Prior Work ............................................ 38 2.8 References .................................................................................................. 39 3. A SIMULATION STUDY OF COMMON PROPOFOL AND PROPOFOL-OPIOID DOSING REGIMENS FOR UPPER ENDOSCOPY: IMPLICATIONS ON THE TIME COURSE OF RECOVERY ...................................................................................... 42 3.1 Abstract ....................................................................................................... 42 3.2 Introduction .................................................................................................. 43 3.3 Materials and Methods ................................................................................. 45 3.4 Results ........................................................................................................ 50 3.5 Discussion ................................................................................................... 66 3.6 Appendix: Target Effect-site Concentrations and Respiratory and Esophageal Instrumentation Outcomes ................................................................................. 74 3.7 References .................................................................................................. 76 v i 4. DEVELOPMENT OF PROPOFOL-REMIFENTANIL DOSING MULTIOBJECTIVE OPTIMIZATION ALGORITHM FOR MODERATELY PAINFUL PROCEDURES REQUIRING ESOPHAGEAL INSTRUMENTATION ............................................... 80 4.1 Abstract ....................................................................................................... 80 4.2 Introduction .................................................................................................. 81 4.3 Methods ....................................................................................................... 83 4.4 Results ........................................................................................................ 95 4.5 Discussion ................................................................................................. 102 4.6 Appendix: Questionnaire ............................................................................ 108 4.7 References ................................................................................................ 109 5. DISCUSSION .......................................................................................................... 111 5.1 Summary ................................................................................................... 111 5.2 Future Work ............................................................................................... 114 LIST OF FIGURES FIGURE 2.1: Presentation of raw data. ....................................................................................... 16 2.2: Observed and response surface model predictions for loss of response to esophageal instrumentation and intolerable ventilatory depression. .............................. 23 2.3: Superimposed topographical plots for probability of loss of responsiveness, loss of response to esophageal instrumentation, and intolerable ventilatory depression response surface models .............................................................................................. 29 3.1: Observed responses and model predictions for respiratory compromise (RC) and loss of response to esophageal instrumentation ............................................................ 52 3.2: Predicted propofol, fentanyl (in remifentanil equivalents), and remifentanil effect-site concentrations for selected published dosing regimens for endoscopy. ........................ 60 3.3: Simulations of loss of response to esophageal instrumentation over time for selected published dosing regimens for upper endoscopy............................................. 65 4.1: An evaluation of the loss of response to esophageal instrumentation model in 110 ultrasonographic endoscopy patients ............................................................................ 84 4.2: Graphical representation of objective times and transformed objective scores ....... 89 4.3: Of the 3,840 dosing combinations simulated, only 17 had every objective score within the minimum and maximum ideal range. ........................................................... 100 4.4: Dosing scheme recommended for a 55 year-old, 75 kg, 175 cm male presenting for an anticipated 10-min upper endoscopy ...................................................................... 101 LIST OF TABLES TABLE 2.1: The Observers Assessment of Alertness/Sedation (OAA/S) score ......................... 11 2.2: Interaction model parameters, coefficients of variation, and goodness-of-fit parameters .................................................................................................................... 22 2.3: Target effect-site concentration sets and observed responses ............................... 36 2.4: Target effect-site concentrations and observed responses for intolerable ventilatory depression from prior work (data unpublished) .............................................................. 38 3.1: New, revised, and published propofol-remifentanil pharmacodynamic interaction model parameters for selected drug effects. .................................................................. 49 3.2: Selected published propofol and propofol - opioid dosing regimens for upper endoscopy for a 55 year old, 75 kg, 175 cm male. ........................................................ 58 3.3: Target effect-site concentrations and respiratory and esophageal instrumentation outcomes ...................................................................................................................... 74 4.1: Dosing bolus and infusion rate ranges simulated for each patient demographic ..... 96 4.2: Ideal objective times, ranges, and priorities obtained from experts in anesthesiology for the six optimization objectives following four rounds of questioning.......................... 97 4.3: Dosing regimens that produce objective scores within the ideal ranges for all six objective scores when run on a 55 year-old, 75 kg, 175 cm male demographic assuming a 10-min procedure length............................................................................................. 99 PREFACE The work presented in this thesis represents several years of careful study into the analgesic and sedative effects resulting from propofol-remifentanil dosing combinations. There has been some debate on whether it is safe for nonanesthesiologists to administer these drugs. This work will address this question in the realm of gastrointestinal procedures requiring esophageal instrumentation. In Chapter 2, we present probabilistic models of drug effect for loss of response to esophageal instrumentation, loss of responsiveness, and intolerable ventilatory depression and conclude that there is an area where a high percentage of volunteers tolerated esophageal instrumentation and avoided concentrations that would lead to involuntary respiratory depression. However, there was no concentration that achieved these conditions and avoided loss of responsiveness in a majority of volunteers. This chapter was published in Anesthesia & Analgesia in September 2011 with an accompanying editorial. Chapter 3 improves upon our models for loss of response to esophageal instrumentation and intolerable ventilatory depression. A revised model is introduced that more accurately reflects clinically acceptable conditions for esophageal instrumentation. A new model for both intolerable ventilatory depression and airway obstruction is presented and called respiratory compromise. In addition, simulations of published upper endoscopy dosing protocols are performed. Because models for respiratory compromise and loss of responsiveness were built from data collected in unstimulated patients, it was decided they only are accurate in unstimulated patients x following termination of the procedure. An abstract presented on this work was awarded a Best of: Clinical Science award at the American Society of Anesthesiologists 2011 Conference. Anesthesiology, the official journal of the American Society of Anesthesiologist, extended an invitation to submit a manuscript from this work for publication in their April 2012 issue and it is currently under final review. Reviewers from both manuscripts have encouraged us to use our expert position to provide a dosing recommendation. However, the complexity involved in addressing this issue required this be addressed in its own manuscript. Chapter 4 presents our work into providing an a priori dosing recommendation using multiobjective optimization techniques and our previously published propofol-remifentanil interaction models. This work has not yet been submitted to a journal for publication. 1 CHAPTER 1 INTRODUCTION Each day, thousands of patients undergo gastrointestinal endoscopy, with the number continually increasing. The advent of new, fast acting drugs such as propofol and remifentanil has helped decrease procedure and recovery times but also introduces the risk of cardiopulmonary complications.1 Of particular concern is that for the procedures, the anesthetics are commonly administered by nonanesthesiologists. This is worrisome because of the rapid onset of drug effects and lack of reversal agents, placing patients in potentially harmful situations very rapidly and leaving the clinician with a narrow window in which to react. The American Society of Anesthesiologists has issued a statement that the use of propofol be limited to those properly trained.2 1.1 Background Propofol or propofol in combination with an opioid is commonly used to provide sedation and analgesia for gastrointestinal procedures such as upper endoscopy.3 Propofol is a sedative that also provides amnesia but offers only minimal analgesia. It has a time to peak effect of 1.6 min.4 While it is possible to perform upper endoscopy without any anesthesia, sedatives and analgesics are commonly administered to improve patient comfort and procedure quality. While mild to moderate sedation is the ideal target, it is often not possible to place a scope in the esophagus and avoid deep sedation. Also, if propofol is dosed alone, there is also a tendency to oversedate in an attempt to compensate for its lack of 2 analgesic properties. Elevated levels of propofol can lead to apnea, ventilatory depression, desaturation, and hypotension. Opiates commonly used in combination with propofol include midazolam, fentanyl, alfentanil, and remifentanil. This work will focus on remifentanil, which has a time to peak effect of around 1 min.4-6 As an opiate, remifentanil has analgesic properties but little sedative or amnesic properties. High doses of remifentanil can lead to respiratory depression. Administering propofol in combination with an opioid is common, allowing the patient to receive the benefits of both drugs. In addition, the interaction between these propofol and opioids is synergistic for most effects, meaning that when both are administered, less of each drug is needed to reach the same effect as if either drug were given alone. However, this is not limited to just the desired effects - the patient is also exposed to the adverse effects of both drugs and the interaction for these effects may also be synergistic. 1.2 Goals The question this dissertation seeks to answer is "does a dosing combination exist that provides adequate sedation and analgesia for esophageal instrumentation while minimizing the risk of adverse effects?" Because remifentanil is a relatively new drug, its effects have not been thoroughly characterized. Studies do exist that report a propofol C50 for endoscopy procedures with propofol alone and in the presence of an opioid, but to our knowledge no interaction model exists. In order to determine if propofol and remifentanil can be safely administered for upper endoscopy, interaction models between propofol and remifentanil needed to be built for several drug effects. Research needs to be conducted to identify current dosing strategies, and an algorithm needs to be developed to identify the optimal dosing combination. 3 1.2.1 Pharmacodynamic models of drug effect The first aim of this study was to characterize the interaction between propofol and remifentanil for loss of response to esophageal instrumentation, intolerable ventilatory depression, and respiratory compromise (intolerable ventilatory depression and airway obstruction). Response surface models would be built that could predict the probability of effect for any drug combination. 1.2.2 Evaluation of common dosing strategies A second aim was to explore through simulation the behavior of common dosing regimens for loss of response to esophageal instrumentation, loss of responsiveness, respiratory depression, and respiratory compromise. Focus will be on evaluating the adverse effects encountered by these protocols following the end of the procedure, a time when patients are unstimulated and therefore at greatest risk. This aim would also partly serve as a validation of the models developed in the first aim. 1.2.3 Identification of optimal drug combination and dosing Once drug effect models are created, we will have a "view of the landscape", meaning we will know where the various effects occur and how they interact. This will help identify what if any propofol-remifentanil combination will provide a high probability of loss of response to esophageal instrumentation yet avoid loss of responsiveness, respiratory depression, and respiratory compromise. In addition, simulation of common dosing protocols would serve to validate these models as well as comment on which strategy may be best. Ultimately, these steps would contribute to our making a final dosing recommendation. With the experience obtained in the first two aims, objective functions will be constructed that will define the properties of the ideal dose. It will include time until the procedure can begin, time needed to perform the procedure, and total recovery time. 4 Ideal times for each objective will be obtained from experts in the field. Finally, an optimization algorithm will be developed to identify the dosing combination that comes closest to these ideal times. The algorithm will evaluate the tradeoff between the various objectives and select the best compromise solution. Algorithm performance will be evaluated by comparing recommendations from the optimization routine to actual dosings administered to patients. Objective scores for the actual and recommended dosings will be computed and a final recommendation made. 1.3 References 1. Cohen LB, Delegge MH, Aisenberg J, Brill JV, Inadomi JM, Kochman ML, Piorkowski JD, Jr.: AGA Institute review of endoscopic sedation. Gastroenterology 2007; 133: 675-701 2. AANA-ASA: AANA-ASA Joint Statement Regarding Propofol Administration, 2005 3. Rex DK, Deenadayalu VP, Eid E, Imperiale TF, Walker JA, Sandhu K, Clarke AC, Hillman LC, Horiuchi A, Cohen LB, Heuss LT, Peter S, Beglinger C, Sinnott JA, Welton T, Rofail M, Subei I, Sleven R, Jordan P, Goff J, Gerstenberger PD, Munnings H, Tagle M, Sipe BW, Wehrmann T, Di Palma JA, Occhipinti KE, Barbi E, Riphaus A, Amann ST, Tohda G, McClellan T, Thueson C, Morse J, Meah N: Endoscopist-directed administration of propofol: a worldwide safety experience. Gastroenterology 2009; 137: 1229-37; quiz 1518-9 4. Schnider TW, Minto CF, Shafer SL, Gambus PL, Andresen C, Goodale DB, Youngs EJ: The influence of age on propofol pharmacodynamics. Anesthesiology 1999; 90: 1502-16 5. Burkle H, Dunbar S, Van Aken H: Remifentanil: a novel, short-acting, mu-opioid. Anesth Analg 1996; 83: 646-51 6. 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: 821-33 2 CHAPTER 2 REMIFENTANIL-PROPOFOL PHARMACODYNAMIC MODELS FOR INTOLERABLE VENTILATORY DEPRESSION, LOSS OF RESPONSIVENESS, AND LOSS OF RESPONSE TO ESOPHAGEAL INSTRUMENTATION* 2.1 Abstract 2.1.1 Introduction Remifentanil and propofol are increasingly used for short duration procedures in spontaneously breathing patients. In this setting, it is preferable to block the response to moderate stimuli while avoiding loss of responsiveness (LOR) and intolerable ventilatory depression (IVD). The aim of this study was to explore selected effects of combinations of remifentanil-propofol effect-site concentrations (Ces) that lead to a loss of response to esophageal instrumentation (EI), a LOR, and/or onset of IVD. A secondary aim was to use these observations to create response surface models for each effect measure. We hypothesized that (1) in a high percentage of volunteers, selected remifentanil and propofol Ces would allow EI yet avoid LOR and IVD and (2) the drug interaction for these effects would be synergistic. * Reprinted with permission from Wolters Kluwer Health: LaPierre CD, Johnson KB, Randall BR, White JL, Egan TD: An Exploration of Remifentanil-Propofol Combinations That Lead to a Loss of Response to Esophageal Instrumentation, a Loss of Responsiveness, and/or Onset of Intolerable Ventilatory Depression. Anesth Analg 2011; 113: 490-9 ©Wolters 2011 6 2.1.2 Methods Twenty-four volunteers received escalating target controlled remifentanil and propofol infusions over ranges of 0-6.4 ng∙mL-1 and 0-4.3 mcg∙mL-1, respectively. At each set of target concentrations, responses to insertion of a blunt end bougie into the mid-esophagus (40 cm), level of responsiveness, and respiratory rate were recorded. From these data, response surface models of loss of response to EI and IVD were built and characterized as synergistic, additive, or antagonistic. A previously published model of LOR was used. 2.1.3 Results Of the possible 384 assessments, volunteers were unresponsive to EI at 105 predicted R-P Ces; in 30 of these, volunteers had no IVD; in 30 of these, volunteers had no LOR; and in 9 of these, volunteers had no IVD or LOR. Many other assessments over the same concentration ranges, however, did have LOR and/or IVD. The combinations that allowed EI and avoided IVD and/or LOR primarily clustered around remifentanil propofol Ces ranging from 0.8 to 1.6 ng∙mL-1 and 1.5 to 2.7 mcg∙mL-1, respectively, and to a lesser extent around 3.0 to 4.0 ng∙mL-1 and 0.0 to 1.1 mcg∙mL-1, respectively. Models of loss of response to EI and IVD both demonstrated a synergistic interaction between remifentanil and propofol. 2.1.4 Discussion Selected remifentanil-propofol concentration pairs, especially higher propofol-lower remifentanil concentration pairs, can block the response to EI while avoiding IVD in spontaneously breathing volunteers. It is, however, difficult to block the response to EI and avoid both LOR and IVD. It may be necessary to accept some discomfort and blunt rather than block the response to EI in order to consistently avoid LOR and IVD. 7 2.2 Introduction Propofol in combination with remifentanil is useful for medical procedures that require moderate sedation and analgesia. Both drugs are rapid acting and quickly dissipate once administration is terminated. They interact synergistically with one another,1 requiring less of each drug to achieve a desired effect when used in combination. For example, a synergistic interaction is present for loss of response to laryngoscopy1-3 and moderately painful stimuli,4 and to a lesser extent for loss of responsiveness.1-3,5 Propofol and combinations of propofol with an opioid have been used to block the response to noxious stimuli during procedures in spontaneously breathing patients in the context of moderate sedation.6-11 In this setting, it is preferable to block the response to moderate noxious stimuli while avoiding intolerable ventilatory depression and minimizing loss of responsiveness. The aim of this study was to explore the effects of selected combinations of remifentanil and propofol. Effects of interest included a loss of response to esophageal instrumentation, a loss of responsiveness, and intolerable ventilatory depression. A secondary aim was to use these observations to create response surface models for each effect measure. We hypothesized that in a high percentage of volunteers, selected remifentanil-propofol effect-site concentrations would allow esophageal instrumentation yet avoid intolerable ventilatory depression and that the drug interaction for these effects would be synergistic. 2.3 Methods 2.3.1 Volunteer recruitment and instrumentation After approval by the Institutional Review Board at the University of Utah, informed written consent was obtained from 12 male and 12 female 8 (nonpregnant/nonlactating) volunteers. Eligible volunteers had an American Society of Anesthesiologists' Physical Status of I or II, were nonsmokers 18 years of age or older, and had a body mass index between 18 and 28. Volunteers were not eligible if they had a history of significant alcohol or drug abuse, allergy to opioids or propofol, sleep apnea, or chronic drug requirements or medical illness that are known to alter the pharmacokinetics or pharmacodynamics of opioids or intravenous anesthetics. 2.3.2 Monitoring Following overnight fasting, volunteers had a 20-gauge intravenous catheter placed for fluid and drug administration. A maintenance infusion of 0.9% sodium chloride was administered at 1 mL∙kg-1∙hour-1 throughout the study period. In addition, a 20- gauge arterial catheter was placed in a radial artery for continuous blood pressure monitoring and intermittent arterial blood gas analyses. Volunteers were monitored with an electrocardiogram, pulse oximeter, noninvasive blood pressure, and expired carbon dioxide and inspired oxygen monitor. Inspired and expired airway flow and volumes were measured using a pneumotachometer (Novametrix, Louisville, KY) attached to a tight fitting mask. All volunteers received oxygen by face mask at 2 L∙min-1. A Mapleson E circuit was used to provide manual ventilation if required to maintain adequate oxygenation and ventilation. Before administration of the study drugs, volunteers were treated with 0.2 mg glycopyrrolate to prevent bradycardia and 30 mL sodium citrate by mouth. 2.3.3 Experimental design The study was an open-label, randomized, parallel group study using a crisscross design as described by Short et al. to assess drug interactions.12 Each volunteer was randomly assigned to one of two groups: a basal infusion group of remifentanil or propofol. Each group was further randomized to receive three of six 9 possible sets of escalating predicted target effect-site concentrations (Ces) (Appendix A). For each set, one drug was stepped through five predetermined Ce targets (primary agent) while the second drug was held at a constant Ce (secondary agent). Following each set, the infusions were stopped until predicted Ces for both drugs returned to near 0, at which time the next set would begin. This design provided a total of 61 possible pairs: one at baseline prior to drug administration, 30 for the remifentanil basal infusion group, and 30 for the propofol basal infusion group. Based on prior work,1,5,13 8 to 9 volunteers were randomly assigned to eight of the twelve sets (sets 1-4 of the remifentanil and propofol groups) of concentration pairs in the anticipated transition zone (less than 5.0 ng∙mL-1 and 3.3 mcg∙mL-1 for remifentanil and propofol, respectively) and one to two volunteers to the remaining four sets (sets 5 and 6 of the remifentanil and propofol groups) anticipated to be near maximal effect. The predicted target effect-site concentrations ranged from (0.0-6.4 ng∙mL-1) for remifentanil and (0.0-4.3 mcg∙mL-1) for propofol. The study was designed so each experiment could be completed within 10 hours. 2.3.4 Drug delivery and effect measures Target controlled infusions were administered using computer controlled infusion pumps (Pump 22; Harvard Apparatus, Limited, Holliston, MA) and drug infusion software (STANPUMP, Available from Steven L. Shafer, M.D., at http://www.opentci.org/doku.php?id=code:code. Posted November 25, 2008. Last accessed June 3, 2010). Pharmacokinetic parameters published by Minto et al. were used for remifentanil14 and Schnider et al. for propofol.15 Effect measurements began 5 min after predicted Ces reached the targeted concentrations. At each target concentration pair, volunteers underwent an assessment period consisting of three measures. First, an assessment of responsiveness was made using 10 the Observers Assessment of Alertness and Sedation (OAA/S) scale (Table 2.1).16 A loss of response was defined as an OAA/S score = 1. Second, an assessment of respiratory rate was made using the capnography tracing. Intolerable ventilatory depression was defined as a respiratory rate of 4 or less breaths in a 1-min time window. During pilot studies, we arrived at this respiratory rate cutoff based on several observations: Below 4 breaths per minute, volunteers consistently began to (1) have a drop in their SpO2 levels (a rapid decline from 100 to low 90s), (2) the ETCO2 began to rise above 50 mmHg, and (3) without manual bag mask ventilation, the volunteers would become hypoxic. Third, an assessment of response to esophageal instrumentation was made. A 42 French (14 mm diameter, 215542, Teleflex Medical, RTP, NC) blunt end bougie was placed through the oropharynx and advanced 40 cm into the esophagus. Loss of response to esophageal instrumentation was defined as no gag reflex, no voluntary or involuntary movement, and no change in heart rate or blood pressure greater than 20% from baseline values recorded just prior to instrumentation. Each volunteer underwent a total of 16 assessment periods (one at baseline and five in each of three sets). Volunteers were verbally prompted to breathe if there were less than 2 breaths in 30 seconds. If SpO2 was below 95% on 2 liters per min of face mask oxygen or expired carbon dioxide levels were greater than 55 mmHg and they did not respond to prompts to breathe, mask ventilation was provided. If airway obstruction was present, the airway was opened using a head tilt and chin lift and/or placement of an oral pharyngeal airway. If volunteers developed a mean arterial blood pressure or heart rate less than 20% of baseline, drug administration was terminated and the washout period begun. Ephedrine 5-10 mg was administered intravenously to treat hypotension as needed. 11 Table 2.1: The Observers Assessment of Alertness/Sedation (OAA/S) score16 Value Description 5 Responds readily to name spoken in normal tone. 4 Lethargic response to name spoken in normal tone. 3 Responds only after name is called loudly and/or repeatedly 2 Responds only after moderate prodding or shaking. 1 Does not respond to moderate prodding or shaking An Observer's Assessment of Alertness/Sedation score of 1 was considered unresponsive. 12   1 1| , 50 50 50 50 50 50 50 50 max                                 P P R R P P R R P P R R P P R R R P C C C C C C C C C C C C C C C C E P LR C C           N i i i LL R P R P 1 2 2 ( ln( ) (1 ) (1 )) 2.3.5 Response surface models Using modeling software (MATLAB R2008b, The MathWorks, Inc., Natick, MA), binary data (presence or absence of a response) for loss of response to esophageal instrumentation and onset of intolerable ventilatory depression were fit to a Greco model17 adjusted for categorical data18 using equation 2.1. For loss of responsiveness, a previously reported model based on data collected in volunteers in a similar fashion was used.5 2.1 P(LR = 1|CR, CP) is the probability of loss of response at a given remifentanil (CR) and propofol (CP) concentration. Emax is the maximal effect (i.e. loss of response to esophageal instrumentation) and is 1 for categorical data. CR and CP are the predicted Ces of remifentanil and propofol (ng∙mL-1 and mcg∙mL-1) as predicted by Stanpump. C50R and C50P are the concentrations of remifentanil and propofol that alone achieve 50% probability of no response. The parameter γ (gamma) determines the slope along the sigmoid surface, and α (alpha) is the drug interaction term. Models were built using a naïve pooled technique.19 Effect ranged from 0 (100% probability of response) to 1 (100% probability of no response). Model parameters were determined using an iterative approach minimizing the -2 Log Likelihood (-2LL), presented in equation 2.2. 2.2 N is the number of observations made for all volunteers combined, Ri is the observed 13   CV  response, and P is the corresponding probability of loss of response. To characterize variability, coefficients of variation (CV) for each model parameter were estimated using a bootstrap technique. One thousand subsamples were randomly drawn (with replacement) from the raw data, with each subsample containing the same number of data points as the raw data set. Estimates of model parameters were generated from each subsample using the same techniques described previously. The mean (μ) and standard deviation () of the 1000 estimates were used to compute the CV for each model parameter (equation 2.3). 2.3 The CV was computed in this manner at least 10 times for each effect measure. It was continued until the percent change between the average of all iterations and the average from all previous iterations was less than 5%. The final averaged CV was reported. For each effect measure, model fits were evaluated using a Chi-square (2) goodness-of-fit test. Response/no response data were divided into probability bins with at least 5 no response data points in each bin. The expected frequency of no response for each bin (Pi) was calculated by multiplying the mean predicted probability by the total number of observations in the bin. Observed frequency of no response (Oi) was the number of observations where no response occurred. The 2 test statistic was computed using equation 2.4: 2.4 k is the number of bins. The null hypothesis was that the expected (based on the    k i i i i P O P 1 2 ( )  14 model's prediction of probability of no response) and observed frequencies were from the same distribution and was rejected if the 2 test statistic exceeded the 2 critical value at a significance level of 5% with k-5 degrees of freedom (four parameters used to compute expected frequency are estimated from the data). Two graphical approaches were used to assess model fits. The first plot presented the observed responses and a topographical rendering of model predictions. A graphical representation of the model was created by plotting the 5, 50, and 95% iso-effect lines (isoboles) representing predicted remifentanil-propofol Ces that produce an equivalent effect. This format was used to illustrate the number of volunteers that developed a loss of response alongside model predictions of the same effect measure. The second plot presented the observed responses and a three-dimensional rendering (response surface) of model predictions. This format was used to illustrate the differences between model predictions (ranging from 0 to 1 using equation 2.1) and observed responses (either 0 or 1). An assessment of how well the model predictions fit the observations was made by calculating the percentage of predictions that agreed with observations. Agreement was defined as an absolute difference less than 0.5. 2.3.6 Comparison of model profiles Topographical plots of models of loss of response to esophageal instrumentation, loss of responsiveness, and intolerable ventilatory depression were superimposed on one another. Each plot included the 5, 50, and 95% isoboles. Visual inspection of superimposed isoboles was used to identify potential concentration pairs with a high probability of loss of response to esophageal instrumentation, but avoid loss of responsiveness or intolerable ventilatory depression. 15 2.4 Results All twenty-four volunteers (12 male and 12 female) completed the study. The mean ± standard deviation of the height, weight, body mass index, and age were 174 ± 8 cm, 71 ± 12 kg, 23 ± 3 kg∙m-2, and 25 ± 4 years, respectively. Appendix A presents the observed responses for each effect measure over the 61 concentration pairs investigated. Seventeen assessment periods were completely or partially aborted at higher target concentrations because blood pressure and/or heart rate were less than 20% of baseline. Portions of three assessment periods were aborted due to inadequate oxygenation after maneuvers to correct it failed. Of the possible 384 evaluations and 61 possible concentration pairs, 367 were made for esophageal instrumentation at 56 concentration pairs, 373 were made for loss of responsiveness at 58 concentration pairs, and 376 were made for intolerable ventilatory depression at 59 concentration pairs (Appendix A). 2.4.1 Effect measures For esophageal instrumentation, some or all of the volunteers in 38 out of the 56 target concentration pairs exhibited no response (105 out of the 367 evaluations). Ten of the 38 concentration pairs consistently blocked the response to esophageal instrumentation (Figure 2.1). Responses at the remaining 28 concentration pairs were mixed (i.e. some volunteers responded, others did not). For example, with propofol at 2.7 mcg∙mL-1 and remifentanil at 0.8 ng∙mL-1, 4 volunteers tolerated esophageal instrumentation and 4 did not. Of the concentration pairs that blocked the response to esophageal instrumentation, 30 assessments at 19 concentration pairs had no intolerable ventilatory depression (Figure 2.1, Panel A). Of those, 4 assessments at 4 concentration pairs between 0.0 and 0.8 ng∙mL-1 for remifentanil and 3.3 and 4.3 mcg∙mL-1 for propofol 16 Figure 2.1: Presentation of raw data (observed responses) at 56 predicted remifentanil-propofol effect-site concentration pairs. Open circle size indicates the total number of esophageal instrumentation (EI) assessments made. Solid green circles represent a subset of those assessments where volunteers had no response to EI. Circles with two colors represent smaller subsets that had a combination of selected responses. In Panel A, red and green circles represent a loss of response to EI and no intolerable ventilatory depression (IVD). In Panel B, blue and green circles represent a loss of response to EI and no loss of responsiveness (LOR). In Panel C, blue and red circles represent a loss of response to EI, no LOR, and no IVD. Circle size represents the number of assessments (see legend) for each circle type (open, solid green, etc.). Ce indicates effect-site concentration. IVD was defined as a respiratory rate of 4 or less breaths per minute. 17 a) Panel A: Loss 01 Response to EI and no IVD .... ~ .".s. (.I ~ 0 "- .l.!. 5 • • • 0 0 4 • 0 3 (!X!)(j) @ 0 2 Cf!J@O~ 0 CD 0 CX)@@ 0 f @ CD10@ 1 0 0 ~ 0 0 0 0 0 0 0 2 3 4 5 6 7 R.em Ifenlanll Ce (ng/m L) o Number of Esophageallnstrumenlatlon evaluations • Number with loss of response 10 Esophageallnslrumentatlon • Number with loss of response 10 Eaophageallnslrumentatlon and no Intolerable Vantllatory Depre •• lon Sample SI.e ,,.me fo,.n symbol$ 0 1',1=1 0 N=2 0 1',1=3 0 N=4 0 1'1=6 0 N=6 0 1'1=7 0 1',1=8 0 N=9 0·=24 18 b) Figure 2.1 continued 19 c) Figure 2.1 continued 20 consistently had no intolerable ventilatory depression and tolerated esophageal instrumentation. All other pairs had a mixed response; some volunteers tolerated esophageal instrumentation but had intolerable ventilatory depression while others did not. For example, with remifentanil at 1.6 ng∙mL-1 and propofol at 2.0 mcg∙mL-1, 5 out of 7 volunteers tolerated esophageal instrumentation and 2 of those 5 (3 of the 7) had no intolerable ventilatory depression. Of the concentration pairs that blocked the response to esophageal instrumentation, 30 assessments at 19 concentrations pairs (not identical to the 30 above) had no loss of responsiveness (Figure 2.1, Panel B). At 8 of the concentration pairs, 9 volunteers also had no intolerable ventilatory depression and no loss of responsiveness (Figure 2.1, Panel C). For example, with propofol at 1.5 mcg∙mL-1 and remifentanil at 0.8 ng∙mL-1, 2 of 8 volunteers tolerated esophageal instrumentation with no intolerable ventilatory depression and no loss of responsiveness, but the other 6 did not tolerate esophageal instrumentation. 2.4.2 Response surface models With visual inspection of the raw data, it is clear that the development of intolerable ventilatory depression at high propofol, low remifentanil concentrations was beyond the range of target concentrations used in our study design. For model building purposes, 31 data points at higher concentrations taken from previous work in our laboratory as part of a study in similar volunteers conducted by Kern et al.1 were therefore included in our analysis. These additional data, presented in Appendix B, were collected using the same drugs, drug delivery technique, and approach to assessment of respiratory rate. Four hundred and seven data points were used to construct the model of intolerable ventilatory depression. Model parameters, coefficients of variation, and goodness-of-fit analysis for loss 21 of response to esophageal instrumentation and intolerable ventilatory depression are presented in Table 2.2. P values from the Chi squared goodness-of-fit test confirmed the null hypothesis that predicted and observed frequencies were from the same distribution, indicating a good fit for each model. Coefficients of variation ranged from 5 to 58%. More variability (i.e. larger coefficients of variation) was estimated about the alpha (interaction) model parameter. The positive alphas were consistent with a synergistic interaction for all models. The response surface models predicted transitions from responsive to unresponsive over a large range of the tested remifentanil and propofol concentrations (as indicated by the small gamma parameter values). Observed responses superimposed over response surface models for each effect measure are presented in Figure 2.2A and Figure 2.2C. In both models, predictions are consistent with observations; all volunteers above the 95% isobole are unresponsive, a large majority are unresponsive between the 50 and 95% isoboles, the responses are mixed between the 5 and 50% isoboles, and very few are unresponsive below the 5% isobole. Isoboles in both models bow toward the origin, indicating a synergistic interaction. The shape of the model of intolerable ventilatory depression and that of the model of esophageal instrumentation were different. Isoboles for intolerable ventilatory depression (Figure 2.2C) bow asymmetrically towards remifentanil, illustrating the large influence of opioids on this effect measure. By contrast, isoboles for esophageal instrumentation (Figure 2.2A) bow symmetrically between remifentanil and propofol. Agreement between model predictions and observations is presented graphically in Figure 2.2B and Figure 2.2D. For both models, agreement was high at concentration pairs below and above the slope of the response surface, but in the transition from 5 to 95%, the difference between predictions and observations were greater than 0.5 at several of the observations. Using an absolute difference less than 0.5 as a cutoff for 22 Table 2.2: Interaction model parameters, coefficients of variation, and goodness-of-fit parameters Stimulus C50 remi (CV) ng∙mL-1 C50 prop (CV) mcg∙mL-1 α (CV) (interaction) γ (CV) (slope) p,Χ2 LOR* 33.1 2.2 3.6 5.0 LREI 9.8 (25%) 3.8 (5%) 4.5 (58%) 3.7 (10%) 0.643 IVD 4.1 (24%) 7.0 (26%) 3.0 (38%) 3.2 (25%) 0.929 LOR = Loss of responsiveness (OAA/S = 1), LREI = Loss of response to esophageal instrumentation, and IVD = Intolerable ventilatory depression, CV = coefficients of variation, remi =remifentanil, prop = propofol, C50 = predicted concentration associated with a 50% probability of maximum effect. *Previously reported by Johnson et al.5 23 Figure 2.2: Observed and response surface model predictions for loss of response to esophageal instrumentation (EI) and intolerable ventilatory depression (IVD). Panels A and C present topographical views of raw data and model predictions. In Panel A, open circles represent assessments of a response to EI and solid green circles represent a subset of those assessments where there was a loss of response to EI. In Panel C, open circles represent assessments of IVD and solid red circles represent a subset of those assessments where there was IVD. The dotted, solid, and dashed lines represent the 5, 50, and 95% iso-effect lines (isoboles) for each model, respectively. Panels B and D present three-dimensional views of the raw data, model predictions, and an assessment of model error. The grid and colored lines represent response surface model predictions with their associated isoboles. Circles represent observed responses. Circles at the bottom of the response surface (0% probability) represent a response to EI (Panel B) or no IVD (Panel D). Circles at the top (100% probability) represent no response to EI (Panel B) or the presence of IVD (Panel D). Open circles represent assessments where the difference between predicted and observed response is less than 50% while solid circles represent assessments where the difference is greater than 50%. Circle size represents the number of assessments (see legend) for each circle type (open, solid, etc.). Ce indicates effect-site concentration. IVD was defined as a respiratory rate of 4 or less breaths per minute. 24 a) Panel A: Esophageal Instrumentation 5 ••••••• 5% Isobole Sample Size 50% Isobole same for all symbols --- 95% Isobole 0 N = 1 4 ,, • No response to EI 0 N=2 ~ , 0 Response to EI 0 N=3 ....I , E 0 , - , 0 N=4 Cl u 3 , E 0 N=5 - 0 ..... Q) ..... 0 N=6 U .E 2 0 Cb U' " 0 N=7 0 '- Q. 0 • -- 0 N=8 ... - a.. 0 N=9 1 O N=24 0 • 0 0 0 1 2 3 4 5 6 7 Remifentanil Ce (ng/mL) 25 b) Figure 2.2 continued Panel B: Esophageal Instrumentation w 0:: ...J '0 ~ :c .'c" e D.. 100 80 60 40 20 o o ••••••• 5% Isobole 50% Iso bole • o 0 95% Iso bole > 50% error between predicted and observed response ~ 50% error between predicted and observed response Number of Assessments same for all symbols 0 N = 1 0 N=2 t 0 N=3 0 N=4 0 N=5 0 N=6 0 N=7 0 N=B 0 N=9 0 N = 24 7 6 26 c) Figure 2.2 continued Panel C: Intolerable Ventilatory Depression 5 ••••••• 5% Isobole 50% Isobole Sample Size 95% Isobole • • ,\ -•-- IVD same0 f or alNl s =y m1 bols 4 \ 0 N=2 \ 0 NolVD .~..J \ 0 N= 3 E , - 0 • , 0 N =4 Cl .... uE 3 .... 0 N=5 - • • .... .... . Q) 0 N=6 () • .. •, ", . .2 2 0 N=7 0 , Q. , .0.. • ' .......... • 0 N=8 Il. • ...... 0 N=9 1 @ • ON=24 • 0 @ • 0 1 2 3 4 5 6 7 Remifentanil Ce (ng/mL) 27 d) Figure 2.2 continued Panel D: Intolerable Ventilatory Depression 100 ~ :c ~ 80 0 ~ .~ ~ ~ '0 ~ .-EQ ~. 60 -0 :;'i:- ~~ ,C_ 40 ra::: ,C ~ 0 ~ 0:> 20 0 ••••••• 5% Iso bole 50% Isobole • o 95% Isobole > 50% error between predicted and observed response ~ 50% error between predicted and observed response Number of Assessments same for al/ symbols 0 N=1 0 N=2 0 N=3 0 N=4 0 N= 5 0 N=6 0 N=7 0 N=8 0 N=9 0 N = 24 7 6 28 model goodness-of-fit, the percentage of model predictions consistent with observed responses was 79% and 81% for the EI (Figure 2.2B) and IVD (Figure 2.2D) models, respectively. Superimposed topographical plots of the loss of responsiveness, loss of response to esophageal instrumentation, and intolerable ventilatory depression models are presented in Figure 2.3. A comparison of isoboles between models revealed no regions of remifentanil-propofol concentration pairs that would have a high probability (>95%) of no response to EI and a low probability (<5%) of intolerable ventilatory depression and loss of responsiveness. Disregarding loss of responsiveness, there is a region of low remifentanil (0-1.5 ng∙mL-1) and high propofol (4-6 mcg∙mL-1) concentrations where there is a high probability (> 80-95%) of loss of response to esophageal instrumentation and a moderate probability (40-70%) of intolerable ventilatory depression. 2.5 Discussion We explored the effects of various combinations of remifentanil-propofol target concentrations on responsiveness, esophageal instrumentation, and ventilatory depression. We hypothesized that in a high percentage of volunteers, selected concentration pairs would allow esophageal instrumentation yet avoid intolerable ventilatory depression. Our results in part confirmed this hypothesis; we found that low remifentanil (0.8 ng∙mL-1) and high propofol (2 -3 mcg∙mL-1) concentration pairs blocked the response to esophageal instrumentation and avoided intolerable ventilatory depression in a majority of volunteers (Figure 2.1). At higher propofol concentrations, the response to esophageal instrumentation was blocked completely with no intolerable ventilatory depression, but the number of assessments was small, making it difficult to conclude that these concentration pairs would consistently lead to the desired response. 29 Figure 2.3: Superimposed topographical plots for probability of loss of responsiveness (blue), loss of response to esophageal instrumentation (green), and intolerable ventilatory depression (red) response surface models. Isobole probability is indicated by line style: dotted lines represent 5%, solid lines represent 50% and dashed lines represent 95%. The loss of responsiveness model was created using parameters previously reported by Johnson et al.5 30 2.5.1 Effect measures By comparison to studies by Kazama and Drover who also explored propofol requirements for esophageal instrumentation, our results are somewhat different; we had to use higher concentrations to achieve conditions that would allow esophageal instrumentation than what these authors have reported. The differences are most likely due to variations in study design. Kazama et al. studied the use of target controlled infusions in patients of various ages undergoing endoscopy 20. They reported a propofol C50 of 2.8 mcg∙mL-1 to blunt the response to esophageal instrumentation in 17-49 year old patients. Higher concentrations were required to blunt the gag reflex (C50 = 3.0 mcg∙mL-1). These are both lower than what we reported (C50 of 3.8 mcg∙mL-1). By design, they considered some movement and coughing NOT to be a response during endoscope placement. By comparison, our criteria to consider movement and heart rate change as responses are perhaps overly stringent and not reflective of clinical practice. Endoscopists may tolerate some level of patient movement or heart rate change to blunt rather than completely block the response to esophageal instrumentation. Drover et al. have studied the use of target controlled infusion in pediatric patients ages 3-10 years old undergoing endoscopy.21 Similar to Kazama et al., minimal movement was NOT considered a response to esophageal instrumentation. They reported a propofol C50 of 3.7 mcg∙mL-1. Drover also explored how a remifentanil infusion would alter propofol requirements for esophageal instrumentation. Using a continuous remifentanil infusion of 0.025 mcg∙kg-1∙min-1, the propofol requirement decreased to 2.8 mcg∙mL-1. For ease of comparison, we simulated this remifentanil infusion in a 55 year old, 75 kg, 175 cm male, which lead to a steady state predicted remifentanil Ce near 0.7 ng∙mL-1. This concentration pair is consistent with our findings and very close to the 50% isobole we reported in Figure 2.2. Drover et al. also explored 0.05 and 0.10 mcg∙kg- 1∙min-1 remifentanil infusion rates, which, when simulated in the same demographic, lead 31 to remifentanil Ces of 1.4 and 2.8 ng∙mL-1, but patients developed significant respiratory depression requiring positive pressure ventilation. They concluded that lower remifentanil infusion rates may be more appropriate for pediatric endoscopies. In addition to defining the loss of response to esophageal instrumentation, we also sought to characterize the extent of intolerable ventilatory depression and loss of responsiveness over the same set of target remifentanil and propofol concentrations. We found that many of the volunteers tolerated esophageal instrumentation and did not develop intolerable ventilatory depression, but this profile of responses was highly variable. At the same concentration pair, some volunteers would tolerate esophageal instrumentation, others would not; some would have significant ventilatory depression, others would not. The raw data revealed no pattern between volunteers who tolerated esophageal instrumentation and those that had intolerable ventilatory depression. A majority of the volunteers that tolerated esophageal instrumentation without significant ventilatory depression were at target concentration pairs consisting of high propofol, low remifentanil levels (Figure 2.1A). Similarly, we found that many of the volunteers tolerated esophageal instrumentation and did not lose responsiveness, but this profile was also quite variable (Figure 2.1B). By contrast, a majority of the volunteers that tolerated esophageal instrumentation and did not lose responsiveness were at target concentration pairs consisting of high remifentanil, low propofol levels. Finally, there were very few volunteers that tolerated esophageal instrumentation with no intolerable ventilatory depression and no loss of responsiveness. With regard to intolerable ventilatory depression, we made our evaluations in an un-stimulated state. This was done to facilitate data collection using the capnograph, mimicking the scenario where patients receive anesthetics to blunt the response to a brief, painful stimulus followed by a period of relatively little stimulus, and to explore the impact this dosing approach has on ventilatory function. It is conceivable that 32 observations of respiratory rate during stimuli such as calling out their name during the OAA/S assessment would increase their ventilatory rate and shift the observed onset of intolerable ventilatory depression to higher concentrations. We also chose respiratory rate as a measure of ventilatory function because of its familiarity among practitioners and its availability on many physiologic monitors. There are limitations to this measure. For example, we did not account for tidal volume; we acknowledge that minute volume may have been adequate to achieve both oxygenation and ventilation despite a slow respiratory rate. Many volunteers achieved tidal volumes greater than 1000 mL at slow respiratory rates. Furthermore, we did not account for changes in arterial CO2 on respiratory drive as many other authors have.22-25 Nevertheless, in the setting of moderate sedation, most clinicians would agree that a ventilatory rate of 4 or less per minute is concerning. 2.5.2 Response surface models We constructed a response surface model for loss of response to esophageal instrumentation and the presence of intolerable ventilatory depression. Both graphical and statistical approaches indicated that the models fit the observed data well. From a graphical perspective (Figure 2.2), the models appear to capture the transition from responsive to unresponsive well and this was confirmed by the 2 analysis and percentage of model predictions consistent with observed responses. We hypothesized that the interaction between these drugs would by synergistic for both effect measures. Our results confirmed this hypothesis as illustrated by the positive alpha values presented in Table 2.2. To our knowledge, no prior interaction model exists for esophageal instrumentation. Judged in terms of the concentrations required to blunt the response, the stimulus associated with esophageal instrumentation is much less than what we 33 previously reported for loss of response to laryngoscopy but similar to reports by Bouillon et al. (Table 2.2). For laryngoscopy, we reported remifentanil and propofol C50's for loss of response to laryngoscopy of 48.9 ng∙mL-1 and 5.6 mcg∙mL-1 respectively1 and Bouillon et al. reported 9.0 ng∙mL-1 and 5.6 mcg∙mL-1 respectively.2 With regard to intolerable ventilatory depression, prior work by Nieuwenhuijs et al. explored the onset of respiratory depression at remifentanil-propofol concentrations ranging from 0.0 to 2.0 ng∙mL-1 and 0.0 to 2.0 mcg∙mL-1 respectively.23 They used a 50% decrease from baseline minute ventilation as their effect measure (i.e. presence or absence of respiratory depression). They constructed a response surface model from their data using a nonlinear pharmacodynamic model structure. Although the effect measures and model constructs were different than ours, the C50's reported were similar considering the range of drugs they tested (4.2 versus 3.3 ng∙mL-1 for remifentanil and 6.8 versus 15.8 mcg∙mL-1 for propofol). To further explore the behavior of propofol in combination with remifentanil, we compared model predictions from three response surfaces: the two presented in this study and a previously reported response surface for loss of responsiveness.5 In attempting to orient oneself to the clinical meaning of response surfaces, a simple "take home" message is that target concentrations of approximately 2 ng∙mL-1 of remifentanil and 2 mcg∙mL-1 of propofol produce about a 50% probability of no response to esophageal instrumentation, no response to verbal and tactile stimuli, and intolerable ventilatory depression. Similarly, target Ces of 1 ng∙mL-1 of remifentanil and 1 mcg∙mL-1 of propofol have a low probability (i.e. 5%) and concentrations above 3 ng∙mL-1 for remifentanil and 3 mcg∙mL-1 have a high probability (i.e. 95%) (Figure 2.3) of producing those end points. As illustrated in Figure 2.3, model predictions from each model had considerable overlap. This was consistent with our observations; there was no set of concentration 34 pairs that consistently provided conditions for esophageal instrumentation yet avoided intolerable ventilatory depression and loss of responsiveness. In all models, the zone of transition from responsiveness to unresponsiveness (between the 5 and 95% isoboles) covered a wide range of remifentanil and propofol effect-site concentrations. In fact, some of the C50's are outside the range of predicted concentrations we used during data collection. This is a limitation of our study design. We designed our study with the intent of making assessments over a range of concentrations that were below, at, and above the concentrations necessary to produce a loss of response to esophageal instrumentation or intolerable ventilatory depression. In a majority of our observations, volunteers were either responsive or within the transition zone from responsive to unresponsive. Few of our observations were made where responses were completely blocked. With relatively little data at higher concentrations, our best fit models may have generated parameter sets that were skewed to higher concentrations due to the larger amount of response data at lower concentrations. With the Greco model structure, when data are well distributed about the C50, the fit is reasonable. When the C50 is outside the range of concentrations evaluated, it is extrapolated; in this scenario, small changes in the data can result in large changes in the C50, particularly when the interaction is synergistic. Model predictions will fit the data well at concentrations where observations were made, but can inflate to clinically unrealistic levels for just one drug (i.e. propofol in the intolerable ventilatory depression model). When this occurs, the alpha (interaction) term must also increase to ensure that the model characterizes the data rich portions of the response surface. Caution should be used when interpreting the magnitude of the alpha parameter when C50 estimates lie well outside the range of drugs tested. In summary, we explored the feasibility of blocking the response to esophageal instrumentation in volunteers at various target effect-site concentration pairs of 35 remifentanil and propofol. In general, our results suggest that although it is possible to identify target concentration pairs that produce significant sedation and analgesia while preserving responsiveness and adequate ventilation, rendering a patient completely unresponsive to esophageal instrumentation requires target concentration pairs that produce a clinical state beyond moderate sedation. In comparison to other similar work and typical clinical practice, the criteria we used to define a loss of response to esophageal instrumentation were perhaps too strict. Our results suggest that in order to stay within the boundaries of moderate sedation, it may be necessary to accept some discomfort and blunt rather than block the response to esophageal instrumentation in order to always avoid intolerable ventilatory depression. Alternatively, it may also be necessary to accept brief unresponsiveness while instrumenting the esophagus. An important clinical feature in this setting is the ability to prompt patients to breathe. Clinicians may tolerate a loss of responsiveness as long as patients continue to breathe; however, in the presence of intolerable ventilatory depression, clinicians are likely to find a prolonged loss of responsiveness and the inability to prompt a patient to breathe unacceptable. In conclusion, our results represent a preliminary finding in healthy volunteers. Further work is warranted to validate these models in patients undergoing moderate to deep sedation for procedures that require esophageal instrumentation. 36 2.6 Appendix A: Target Effect-site Concentration Sets and Observed Responses. Table 2.3: Target effect-site concentration sets and observed responses Remifentanil Group Propofol Group Primary Infusion Secondary Infusion Effect Measures Secondary Infusion Primary Infusion Effect Measures Set N Remi (ng∙mL-1) Prop (mcg∙mL-1) LREI LOR IVD Set N Remi (ng∙mL-1) Prop (mcg∙mL-1) LREI LOR IVD 0 12 0.0 0.0 0/12 0/12 0/12 0 12 0.0 0.0 0/12 0/12 0/12 1 9 0.0 0.8 0/9 0/9 0/9 1 8 1.2 0.0 0/8 0/8 1/8 1 9 0.4 0.8 0/9 0/9 0/9 1 8 1.2 0.3 0/8 0/8 0/8 1 9 0.8 0.8 0/9 0/9 2/9 1 8 1.2 0.6 0/8 0/8 0/8 1 9 1.6 0.8 1/9 0/9 3/9 1 8 1.2 1.1 1/8 0/8 2/8 1 9 3.3 0.8 3/9 0/9 6/9 1 8 1.2 2.2 5/8 6/8 5/8 2 8 0.0 1.5 0/8 0/8 0/8 2 8 2.2 0.0 0/9 0/9 0/9 2 8 0.4 1.5 0/8 2/8 0/8 2 8 2.2 0.3 0/9 0/9 1/9 2 8 0.8 1.5 2/8 0/8 0/8 2 8 2.2 0.6 0/9 0/9 2/9 2 8 1.6 1.5 1/7 3/8 2/8 2 8 2.2 1.1 3/9 2/9 6/9 2 8 3.3 1.5 5/7 5/7 7/7 2 8 2.2 2.2 6/7 6/8 9/9 3 9 0.0 2.0 0/9 1/9 0/9 3 8 3.0 0.0 2/8 0/8 5/8 3 9 0.4 2.0 2/9 5/9 0/9 3 8 3.0 0.3 1/8 0/8 3/8 3 9 0.8 2.0 4/9 7/9 1/9 3 8 3.0 0.6 2/8 0/8 5/8 3 9 1.6 2.0 5/7 7/8 3/7 3 8 3.0 1.1 6/8 2/8 6/8 3 9 3.3 2.0 5/6 7/7 6/6 3 8 3.0 2.2 7/8 7/8 8/8 4 8 0.0 2.7 1/8 5/8 0/8 4 8 4.0 0.0 1/8 0/8 4/8 4 8 0.4 2.7 2/8 8/8 0/8 4 8 4.0 0.3 1/8 0/8 1/8 4 8 0.8 2.7 4/8 8/8 1/8 4 8 4.0 0.6 1/8 0/8 4/8 4 8 1.6 2.7 7/8 8/8 5/8 4 8 4.0 1.1 2/8 1/8 6/8 4 8 3.3 2.7 8/8 8/8 8/8 4 8 4.0 2.2 6/7 5/7 8/8 5 1 0.0 3.3 1/1 0/1 0/1 5 2 5.0 0.0 0/2 0/2 1/2 5 1 0.8 3.3 0/1 1/1 0/1 5 2 5.0 0.6 0/2 0/2 1/2 5 1 1.6 3.3 - 1/1 1/1 5 2 5.0 1.1 2/2 2/2 2/2 5 1 3.3 3.3 - 1/1 - 5 2 5.0 2.2 - - 2/2 5 1 3.9 3.3 - 1/1 - 5 2 5.0 2.6 - - 2/2 37 Table 2.3 continued Remifentanil Group Propofol Group Primary Infusion Secondary Infusion Effect Measures Secondary Infusion Primary Infusion Effect Measures Set N Remi (ng∙mL-1) Prop (mcg∙mL-1) LREI LOR IVD Set N Remi (ng∙mL-1) Prop (mcg∙mL-1) LREI LOR IVD 6 1 0.0 4.3 1/1 1/1 0/1 6 2 6.4 0.0 0/1 0/1 1/1 6 1 0.4 4.3 1/1 1/1 0/1 6 2 6.4 0.3 0/1 0/1 1/1 6 1 0.8 4.3 1/1 1/1 0/1 6 2 6.4 0.6 1/1 0/1 1/1 6 1 1.6 4.3 1/1 1/1 1/1 6 2 6.4 1.1 1/1 1/1 1/1 6 1 2.4 4.3 1/1 1/1 1/1 6 2 6.4 1.6 1/1 - 1/1 total 192 56/182 83/188 47/184 192 49/185 32/185 89/192 Remi = Remifentanil, Prop = Propofol, N is the number of subjects assigned to each set based on the study design. Effect measures: LOR = Loss of responsiveness (OAA/S = 1), LREI = Loss of response to esophageal instrumentation, and IVD = Intolerable ventilatory depression defined as a respiratory rate of 4 breaths per minute or less. Dashes (-) = unable to complete evaluation of effect measure. The numerator represents the number of subjects at maximum effect and the denominator represents the total number of subjects assessed at that concentration pair. Subjects were randomly assigned to three sets in a two-step approach. Subjects were first randomized to either the remifentanil or the propofol group. Each subject was further randomized to receive three of the six possible sets of infusions within their group. In the propofol group, we incorrectly dosed one volunteer, which caused there to be nine subjects in set two instead of two subjects in set six. 38 2.7 Appendix B: Target Effect-site Concentrations and Observed Responses for Intolerable Ventilatory Depression from Prior Work Table 2.4: Target effect-site concentrations and observed responses for intolerable ventilatory depression from prior work (data unpublished).1 Secondary Infusion Primary Infusion Effect Measures Set Remi (ng∙mL-1) Prop (mcg∙mL-1) IVD 1 0.0 5.0 4/8 1 0.0 7.5 6/8 1 0.0 10.0 6/8 2 1.0 5.0 1/3 2 1.0 7.5 1/2 3 5.0 3.0 0/1 3 5.0 5.0 0/1 Remi = remifentanil, Prop = propofol, IVD = intolerable ventilatory depression defined as a respiratory rate of 4 or less breaths per minute. 39 2.8 References 1. Kern SE, Xie G, White JL, Egan TD: A response surface analysis of propofol-remifentanil pharmacodynamic interaction in volunteers. Anesthesiology 2004; 100: 1373-81 2. 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: 1353-72 3. 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: 347-59 4. Johnson KB, Syroid ND, Gupta DK, Manyam SC, Pace NL, LaPierre CD, Egan TD, White JL, Tyler D, Westenskow DR: An evaluation of remifentanil-sevoflurane response surface models in patients emerging from anesthesia: model improvement using effect-site sevoflurane concentrations. Anesth Analg 2010; 111: 387-94 5. Johnson KB, Syroid ND, Gupta DK, Manyam SC, Egan TD, Huntington J, White JL, Tyler D, Westenskow DR: An evaluation of remifentanil propofol response surfaces for loss of responsiveness, loss of response to surrogates of painful stimuli and laryngoscopy in patients undergoing elective surgery. Anesth Analg 2008; 106: 471-9 6. Tosun Z, Aksu R, Guler G, Esmaoglu A, Akin A, Aslan D, Boyaci A: Propofol-ketamine vs propofol-fentanyl for sedation during pediatric upper gastrointestinal endoscopy. Paediatr Anaesth 2007; 17: 983-8 7. Tosun Z, Esmaoglu A, Coruh A: Propofol-ketamine vs propofol-fentanyl combinations for deep sedation and analgesia in pediatric patients undergoing burn dressing changes. Paediatr Anaesth 2008; 18: 43-7 8. Fatima H, DeWitt J, LeBlanc J, Sherman S, McGreevy K, Imperiale TF: Nurse-administered propofol sedation for upper endoscopic ultrasonography. Am J Gastroenterol 2008; 103: 1649-56 9. VanNatta ME, Rex DK: Propofol alone titrated to deep sedation versus propofol in combination with opioids and/or benzodiazepines and titrated to moderate sedation for colonoscopy. Am J Gastroenterol 2006; 101: 2209-17 10. Cohen LB, Delegge MH, Aisenberg J, Brill JV, Inadomi JM, Kochman ML, Piorkowski JD, Jr.: AGA Institute review of endoscopic sedation. Gastroenterology 2007; 133: 675-701 11. Lichtenstein DR, Jagannath S, Baron TH, Anderson MA, Banerjee S, Dominitz JA, Fanelli RD, Gan SI, Harrison ME, Ikenberry SO, Shen B, Stewart L, Khan K, Vargo JJ: Sedation and anesthesia in GI endoscopy. Gastrointest Endosc 2008; 68: 815-26 40 12. 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: 400-8 13. Dahan A, Nieuwenhuijs DJ, Olofsen E: Influence of propofol on the control of breathing. Adv Exp Med Biol 2003; 523: 81-92 14. 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 15. Schnider TW, Minto CF, Shafer SL, Gambus PL, Andresen C, Goodale DB, Youngs EJ: The influence of age on propofol pharmacodynamics. Anesthesiology 1999; 90: 1502-16 16. Chernick 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 of Clin Psychopharmacology 1990; 10: 244-251 17. Greco WR, Bravo G, Parsons JC: The search for synergy: a critical review from a response surface perspective. Pharmacol Rev 1995; 47: 331-85 18. 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: 347-55 19. 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: 1430-43 20. Kazama T, Takeuchi K, Ikeda K, Ikeda T, Kikura M, Iida T, Suzuki S, Hanai H, Sato S: Optimal propofol plasma concentration during upper gastrointestinal endoscopy in young, middle-aged, and elderly patients. Anesthesiology 2000; 93: 662-9 21. Drover DR, Litalien C, Wellis V, Shafer SL, Hammer GB: Determination of the pharmacodynamic interaction of propofol and remifentanil during esophagogastroduodenoscopy in children. Anesthesiology 2004; 100: 1382-6 22. 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: 240-50 23. Nieuwenhuijs DJ, Olofsen E, Romberg RR, Sarton E, Ward D, Engbers F, Vuyk J, Mooren R, Teppema LJ, Dahan A: Response surface modeling of remifentanil-propofol interaction on cardiorespiratory control and bispectral index. Anesthesiology 2003; 98: 312-22 24. Romberg R, Olofsen E, Sarton E, Teppema L, Dahan A: 41 Pharmacodynamic effect of morphine-6-glucuronide versus morphine on hypoxic and hypercapnic breathing in healthy volunteers. Anesthesiology 2003; 99: 788-98 25. Romberg R, Sarton E, Teppema L, Matthes HW, Kieffer BL, Dahan A: Comparison of morphine-6-glucuronide and morphine on respiratory depressant and antinociceptive responses in wild type and mu-opioid receptor deficient mice. Br J Anaesth 2003; 91: 862-70 3 CHAPTER 3 A SIMULATION STUDY OF COMMON PROPOFOL AND PROPOFOL-OPIOID DOSING REGIMENS FOR UPPER ENDOSCOPY: IMPLICATIONS ON THE TIME COURSE OF RECOVERY 3.1 Abstract 3.1.1 Background Using models of respiratory compromise, loss of response to esophageal instrumentation and loss of responsiveness, we explored through simulation published dosing schemes for endoscopy using propofol alone and in combination with selected opioids. We hypothesized that models would predict adequate conditions for esophageal instrumentation and once drug administration is terminated, rapid return of responsiveness and minimal respiratory compromise. 3.1.2 Methods Four published dosing regimens of propofol alone or in combination with opioids were used to predict the probability of loss of response to esophageal instrumentation for a 10-min procedure and the probability of respiratory compromise and return of responsiveness once the procedure had ended. 3.1.3 Results Propofol alone provided a low probability (9-20%) and propofol-opioid techniques provided a moderate probability (15-58%) of loss of response to esophageal 43 instrumentation. Once the procedure ended, all techniques provided a high likelihood of rapid return of responsiveness (<3 min). Propofol-opioid techniques required more time than propofol alone to achieve a high probability of no respiratory compromise (7 versus 4 min). 3.1.4 Conclusions Propofol alone would likely lead to inadequate conditions for esophageal instrumentation but would provide a rapid return to responsiveness and low probability of respiratory compromise once the procedure ended. The addition of remifentanil or fentanyl improved conditions for esophageal instrumentation and had an equally rapid return to responsiveness. The time required to achieve a low probability of respiratory compromise was briefly prolonged; this is likely inconsequential given that patients are responsive and can be prompted to breathe. 3.2 Introduction Propofol alone and in combination with selected opioids are used by clinicians with no formal training in anesthesia to provide moderate or deep sedation for procedures associated with mild to moderately painful stimuli such as cardiac catheterizations,1 upper endoscopies,2-4 and colonoscopies.5 This is of particular clinical interest and controversy6,7† because doses used to blunt responses to moderately painful stimuli can be associated with loss of responsiveness,8-10 ventilatory depression9,11,12 and/or airway obstruction. Prior work in our laboratory on healthy unstimulated volunteers explored the presence or absence of intolerable ventilatory depression, defined as a respiratory rate † AANA-ASA Joint Statement Regarding Propofol Administration, April 14, 2004. Available at: http://www.aana.com/resources2/professionalpractice/Documents/PPM%20PS%20Joint%20AAN A-ASA%20Propofol.pdf. Accessed January 17, 2012 44 of ≤4 breaths per minute, over a wide range of propofol-remifentanil concentration pairs administered in a laboratory setting. From this data, a propofol-remifentanil interaction model of intolerable ventilatory depression was built. While conducting this study, it was clear that intolerable ventilatory depression was not the only adverse respiratory effect that developed. In many instances, volunteers developed partial to complete airway obstruction at higher drug concentrations. To build upon our interaction model for intolerable ventilatory depression, the first aim of this study was to construct a propofol-remifentanil interaction model that accounted for both airway obstruction and intolerable ventilatory depression. We named the combined effect respiratory compromise. We hypothesized that the interaction between propofol and remifentanil for respiratory compromise would be synergistic. Using the same volunteers, we also explored the loss of response to esophageal instrumentation, defined as no response to placing a surrogate of an endoscope (42F blunt end bougie) 40 cm into the esophagus. Non-responsiveness was defined as no gag, no change in heart rate or blood pressure greater than 20% from baseline, and no voluntary or involuntary movement. When comparing our model results with other similar modeling and dosing studies for endoscopy,8,9,12 the criteria we used to define loss of response to esophageal instrumentation were perhaps overly stringent and not reflective of clinical practice. Endoscopists may tolerate some level of patient movement, gag response, and heart rate or blood pressure change rather than expect to block completely the response to esophageal instrumentation in order to avoid intolerable ventilatory depression. Thus a second aim of our study was to revise our loss of response to esophageal instrumentation model by redefining the response criteria to better reflect clinical practice. We hypothesized that the revised model would predict adequate conditions at lower propofol-remifentanil target concentrations and that the interaction would be synergistic. 45 A third aim of our study was to explore through simulation the behavior of published dosing schemes for endoscopy in terms of the probability of loss of response to esophageal instrumentation during a brief (10-min) procedure and the probabilities of respiratory compromise and loss of responsiveness in an unstimulated state following the procedure. We hypothesized that simulations of these dosing regimens would predict a 50 to 95% probability of loss of response to esophageal instrumentation and a rapid decline in the probabilities of respiratory compromise and loss of responsiveness once drug administration ended. 3.3 Materials and Methods Previously collected data were used in this analysis; details regarding volunteer recruitment, study design, and physiologic monitoring have been previously reported.13 In brief, the University of Utah Internal Review Board (Salt Lake City, Utah, USA) approved the study. After receiving informed, written consent, twenty-four volunteers were enrolled and received escalating target controlled infusions of propofol and remifentanil covering a range of effect-site concentrations (Ces) for each drug (propofol 0 to 4.3 mcg∙mL-1 and remifentanil 0 to 6.4 ng∙mL-1). Volunteers were randomly assigned to receive three of 12 possible sets of target concentrations (360 evaluations at 60 unique target concentration pairs plus 24 baseline). Each set consisted of five target concentration pairs (Appendix). Measures of inspired and expired airway flow and tidal volumes were recorded using a pneumotachometer (Novametrix, Louisville, KY) and chest and abdominal wall excursion were recorded using inductive plethysmography (Respitrace, Ambulatory Monitoring Inc., Ardsley, NY) at each target concentration pair. 3.3.1 Effect measures Assessments of intolerable ventilatory depression and airway obstruction were made in the fourth minute after reaching predicted target Ces. We previously reported 46 the presence or absence of intolerable ventilatory depression (respiratory rate of ≤4 breaths per minute) at each target concentration pair.13 The presence of airway obstruction was defined as partial or complete. Partial airway obstruction was defined as a 30 second average inspired tidal volume <3 ml∙kg-1 AND >2 breaths in the same time period. Complete airway obstruction was defined as the absence of airway flow detected by the pneumotachometer in the presence of a respiratory effort detected by the plethysmograph. Respiratory compromise was defined as the presence of intolerable ventilatory depression and/or airway obstruction. Revised assessments of esophageal instrumentation were made at the same set of Ces as described for respiratory compromise. No response was defined as no voluntary movement when placing the bougie and no request by the volunteer (by raising their hand) that placement of the bougie stop. Involuntary movement, gag response, and changes in heart rate or blood pressure were not considered responses. 3.3.2 Response surface models Response surface models for respiratory compromise and loss of response to esophageal instrumentation were constructed by fitting binary effect data (presence or absence of effect) to a Greco model construct14 adjusted for categorical data15 using a naïve pooled technique16 and modeling software (MATLAB R2008b, The MathWorks, Inc., Natick, MA). Model parameters and their coefficients of variation were estimated as previously described.13 There were insufficient data points collected from individual subjects to construct post-hoc individual models. Model fits were evaluated using a Chi-square (2) goodness-of-fit test. Response/no response data were divided into probability bins with at least five no response data points in each bin. The expected frequency of no response for each bin (Pi) was calculated by multiplying the mean predicted probability by the total number of 47     k i i i i P O P 1 2 ( )  observations in the bin. Observed frequency of no response (Oi) was the number of observations where no response occurred. The 2 test statistic was computed using equation 3.1: 3.1 k is the number of bins. The null hypothesis was that the expected (based on the model's prediction of probability of no response) and observed frequencies were from the same distribution and was rejected if the 2 test statistic exceeded the 2 critical value at a significance level of 5% with k-5 degrees of freedom (four parameters used to compute expected frequency are estimated from the data). Two graphical approaches were used to assess model fits. The first plot presented the observed responses and a topographical rendering of model predictions created by plotting the 5%, 50%, and 95% iso-effect lines (isoboles). Isoboles represent all predicted propofol-remifentanil Ce combinations that produce the same probability of observing a modeled effect. This format was used to illustrate the number of volunteers that developed a loss of response alongside model predictions of the same effect measure. The second plot presented the observed responses on a three-dimensional rendering (response surface) of model predictions. This format was used to illustrate the differences between model predictions (ranging from 0 to 1) and observed responses (either 0 or 1). An assessment of how well model predictions fit the observations was made by calculating the percentage of predictions that agreed with observations. Agreement was defined as an absolute difference ≤0.5. 48 3.3.3 Identification of published endoscopy dosing regimens Keyword searches were performed in PubMed to identify published dosing regimens for upper endoscopy. Only those dosing schemes that administered propofol, remifentanil and/or fentanyl were considered. Any studies using additional local or topical agents were excluded. All searches included the keyword propofol in combination with one or more of the following: dosing, endoscopic retrograde cholangiopancreatography, endoscopic ultrasound, endoscopist-directed propofol sedation, endoscopy, esophagogastroduodenoscopy, nurse administered propofol sedation, protocol, and sedation. 3.3.4 Simulations of published dosing regimens for endoscopy A series of simulations were conducted to explore the duration of drug effects using published dosing regimens for endoscopy. Of particular interest was the ability of the dosing regimens to provide analgesia for esophageal instrumentation and the time to recovery (respiratory compromise and loss of responsiveness in an unstimulated state) once the procedure ended. Simulations consisted of an induction period and a 10-min maintenance period followed by a 10-min washout. Ces were estimated for remifentanil, propofol and fentanyl using published pharmacokinetic models.17-19 For purposes of using propofol-remifentanil models of drug effects, fentanyl was converted to remifentanil equivalents using a remifentanil:fentanyl equivalency ratio of 1:1.2.20,21 Simulated drug Ces from each dosing regimen were then used to predict the probability of drug effects over time using the response surface models described above for respiratory compromise and loss of response to esophageal instrumentation and a previously reported response surface model for loss of responsiveness22 (Table 3.1). Low, moderate, and high probabilities of drug effect were defined as <25%, 25-75% and 49 Table 3.1: New, revised, and published propofol-remifentanil pharmacodynamic interaction model parameters for selected drug effects. Effect C50 remi (CV) ng∙mL-1 C50 prop (CV) mcg∙mL-1 α (CV) (interaction) γ(CV) (slope) p,2 RC 6.7 (22%) 4.3 (26%) 9.7 (49%) 2.0 (14%) 0.724 LREI (revised) 9.6 (25%) 4.1 (8%) 7.7 (49%) 2.7 (11%) 0.708 LOR 22 33.1 2.2 3.6 5.0 - CV = coefficient of variation; remi =remifentanil, prop = propofol; C50 = predicted concentration associated with a 50% probability of effect; 2=Chi-square goodness-of-fit, RC = respiratory compromise; LREI = loss of response to esophageal instrumentation; LOR = loss of responsiveness. 50 >75% respectively. Once the simulated 10-min procedure ended, the time required for drug effects to dissipate were estimated using the time to reach a high probability of no respiratory compromise and no loss of responsiveness (<5% probability). 3.4 Results Data were obtained from all 24 subjects. The Appendix presents the observed responses for each effect measure. Of the possible 384 assessments at 61 possible concentration pairs, 376 assessments for intolerable ventilatory depression, 247 assessments for airway obstruction and 370 assessments for esophageal instrumentation were made at 59, 48 and 59 concentration pairs respectively. Twenty assessment periods were completely or partially aborted at higher target concentrations; seventeen because blood pressure and/or heart rate changed more than 20% from baseline and three due to inadequate oxygenation. This included eight intolerable ventilatory depression, eleven airway obstruction, three respiratory compromise and 14 esophageal instrumentation assessments. Results from an additional eight assessments were not used because of recording difficulties with the pneumotachometer. 118 assessments of airway obstruction could not be made because volunteers were experiencing intolerable ventilatory depression. 3.4.1 Effect measures Airway obstruction was observed in 27 of the 61 target concentration pairs (59 of 247 assessments) and consistently in ten (11 of 11 assessments). Airway obstruction occurred more often at high propofol Ces. Intolerable ventilatory depression was observed in 41 of the 61 target concentration pairs (137 of 376 assessments) and consistently in 17 (59 of 59 assessments). Intolerable ventilatory depression occurred more often at high remifentanil Ces. Combining airway obstruction and intolerable ventilatory depression, respiratory compromise was present in 54 of the target 51 concentration pairs (189 of 377 assessments). Volunteers in 25 of the 54 concentration pairs (86 of 86 assessments) consistently developed respiratory compromise (Figure 3.1A). Responses in the remaining 29 concentration pairs were mixed (i.e. some volunteers developed respiratory compromise others did not). For example, with propofol at 2.0 mcg∙mL-1 and remifentanil at 0.8 ng∙mL-1, 7 volunteers developed respiratory compromise and two did not. Loss of response to esophageal instrumentation was observed in 48 of the 61 target concentration pairs (135 of 370 assessments). Volunteers in 19 of the 48 concentration pairs (51 of 51 assessments) consistently had a loss of response to esophageal instrumentation (Figure 3.1B). Responses at the remaining 29 concentration pairs were mixed (i.e. some volunteers responded, others did not). For example, with propofol at 2.7 mcg∙mL-1 and remifentanil at 0.8 ng∙mL-1, 5 volunteers tolerated esophageal instrumentation and 3 did not. 3.4.2 Response surface models Model parameters, coefficients of variation, and the p-value from the Chi-square goodness-of-fit test are presented in Table 3.1. The positive alpha (interaction term) values indicate a synergistic relationship between remifentanil and propofol for respiratory compromise and loss of response to esophageal instrumentation. The small gamma value indicates a large range of concentrations covering the transition from responsive to unresponsive. Coefficients of variation indicated low parameter variability (<30%) except for the alpha parameters (49% for both the respiratory compromise and loss of response to esophageal instrumentation models). The Chi-square goodness-of-fit tests indicate good model fits to the raw data. Observed responses and topographical representation of model predictions are presented in Figure 3.1A for respiratory compromise and Figure 3.1B for loss of 52 Figure 3.1: Observed responses and model predictions for respiratory compromise (RC) and loss of response to esophageal instrumentation (EI). Panels A and B: Topographical plot of raw data and model predictions. Open circle size indicates the number of RC and loss of response to EI assessments made at the corresponding drug effect-site concentration (Ce) pairs respectively. Filled circle size indicates the number of subjects with RC and loss of response to EI. RC data is further characterized using pie charts to indicate the source of RC: either intolerable ventilatory depression (IVD, red) or airway obstruction (AO, black) or both (green). Panels C and D: Response surface plot of model prediction and model error. Model predictions are presented as a mesh surface. Dotted, solid, and dashed lines represent drug concentration pairs resulting in a 5%, 50% and 95% probability of effect (RC in orange and loss of response to EI in green). Model error is presented as open (error ≤0.5) and filled (error >0.5) circles. Circle size indicates the number of observations and corresponding effect at each concentration pair (0 = no RC or no loss of response to EI, 1 = RC or loss of response to EI). 53 a) Panel A: Respiratory Compromise 5 • • 4 ~ ...J -E • Cl 0 3 -E Q) U .E 2 0 Co .0.. Il. 1 0 0 1 \ \ \ • \ . \ , , • , , 2 , • •••••••• 5% Isobole 50% Isobole --- 95% Isobole • IVD • AO • IVD and AO 0 Total RC Assessments • .... : ', .... • ........ --•• • •• 3 4 5 6 Remifentanil Ce (ng/mL) Sample Size same for all symbols 0 N = 1 0 N = 2 0 N=3 0 N=4 0 N=5 0 N=6 0 N=7 0 N=8 -- 0 N=9 ON=24 7 54 b) Figure 3.1 continued Panel B: Revised Esophageal Instrumentation 5 4 .~.. -E 8' 3 E ~ til CJ ~ 2 o Il. E Q. o • • • • • o 2 3 --- • o • 4 5% Isobole 50% Isotlole 95% Isotlole Loss or Response to EI Total EI Assessments • • • o 5 6 Remlfentanll Ce (ng/mL) 7 55 c) Figure 3.1 continued Panel C: Respiratory Compromise 1.0 Q) .!!! -E.0. . 0.8 00. >E ~o 0.6 =.c u ."ec't:o-:- OA a.!!! .is. fA 0.2 Q) 0:: 0.0 ••••••• • o 5% Isobole 50% Isobole 95% Isobole > 50% error between predicted and observed response < 50% error between predicted and observed response 6 Sample Size same for all symbols 0 N = 1 0 N = 2 0 N = 3 0 N =4 0 N = 5 0 N = 6 0 N = 7 0 N = 8 0 N = 9 0 N = 24 7 56 d) Figure 3.1 continued 57 response to esophageal instrumentation. Model predictions were consistent with observations. The observed frequency of respiratory compromise and loss of response to esophageal instrumentation below the 5% isobole was 2.5% and 6.7% respectively and 100% for both above the 90% isobole. Along the 50% isobole, approximately half the assessments at each target concentration pair developed respiratory compromise or loss of response to esophageal instrumentation. Most assessments between the 50% and 95% isoboles had respiratory compromise and loss of response to esophageal instrumentation while most between 5% and 50% did not. Observed responses and prediction errors are presented in Figure 3.1C and Figure 3.1D. For respiratory compromise, 79% of the model predictions and for loss of response to esophageal instrumentation, 81% of the model predictions agreed with observed responses using an absolute difference of ≤0.5. One previously published propofol-remifentanil interaction model for loss of responsiveness is also presented in Table 3.1.22 Loss of responsiveness was defined as an Observer's Assessment of Alertness/Sedation score of 1.23 Volunteers experienced verbal and tactile stimuli during these assessments. 3.4.3 Identification of published endoscopy/colonoscopy dosing regimens Ten published manuscripts were identified using search criteria for endoscopy and propofol alone or in combination with an opioid. They were characterized according to drugs used: four describing techniques with propofol alone, three for propofol in combination with fentanyl using various bolus and infusion strategies for propofol, and one using target controlled infusion of propofol and remifentanil. Four dosing schemes were selected for simulation purposes and are presented in Table 3.2: (1) intermittent boluses of propofol alone,24 (2) loading bolus of fentanyl with intermittent boluses of propofol,25 (3) a loading bolus of fentanyl followed by a propofol bolus and infusion 58 Table 3.2: Selected published propofol and propofol - opioid dosing regimens for upper endoscopy for a 55 year old, 75 kg, 175 cm male. Author Technique Published Recommendation Simulated Dosing Regimen Technique #1: Cohen et al., 200724* Propofol Boluses Initial bolus of 10-60 mg. Additional 10- 20 mg boluses as needed with a minimum of 20 to 30 seconds between doses Initial bolus of 35 mg followed by 15 mg boluses 0.5, 3.5, 5.5, 8 and 10.5 minutes later Technique #2: Cohen et al., 200325 Propofol Boluses & Initial bolus of 5-10 mg. Additional 5-15 mg boluses as needed with a minimum of 30 seconds between doses Initial bolus of 7.5 mg followed by 10 mg boluses 0.5, 2, 4.5, 7, 9.5 and 12 minutes later Fentanyl Bolus Initial bolus of 75 mcg Initial bolus of 75 mcg Technique #3: Pambianco et al., 20084,26 Propofol Bolus and Infusion & Loading dose of 0.5 mg∙kg- 1∙(maintenance infusion rate)∙75-1 started 3 minutes after fentanyl bolus and administered over 3 minutes followed by a maintenance infusion of 25-75 mcg∙kg-1∙min-1 that is titrated to effect Three minutes after fentanyl bolus, a loading dose of 8.3 mg∙min-1 for 3 minutes followed by a 10 minute infusion at 50 mcg∙kg-1∙min-1 Fentanyl Bolus Initial bolus of 50-100 mcg 3 minutes prior to administration of propofol Initial bolus of 75 mcg Technique #4: Gambus et al., 201127 Propofol TCI & 2.8 to 1.8 mcg∙mL-1 Ce target of 1.8 mcg∙mL-1 Remifentanil TCI 0 to 1.5 ng∙mL-1 Ce target 1.5 ng∙mL-1 *Dosing recommendation reported by the American Gastroenterological Association Institute and cited by the American Society for Gastrointestinal Endoscopy. TCI=Target Controlled Infusion. Ce=effect-site concentration. 59 administered by SEDASYS (Ethicon Endo-Surgery, Inc., Cincinnati, OH),4,26 and target controlled infusions of propofol and remifentanil.27 3.4.4 Simulations of published dosing regimens for endoscopy Published dosing recommendations were used to simulate a 10-min upper endoscopy procedure. Published recommendations were converted to dosing regimens (Table 3.2) assuming a 75 kg, 175 cm, 55-year-old male patient. Predicted Ces for propofol, fentanyl (in remifentanil equivalents), and remifentanil for each dosing scheme are presented in Figure 3.2. During the 10-min procedure, estimated propofol concentrations ranged from 1.2 to 3.0 mcg·mL-1 and remifentanil concentrations ranged from 0.7 to 1.5 ng·mL-1. Predictions of time to recovery for respiratory compromise and loss of responsiveness are presented in Figure 3.2D. and predictions of loss of response to esophageal instrumentation throughout the 10-min procedure are presented in Figure 3.3. 3.4.4.1 Technique #1: For the intermittent propofol boluses, the resultant propofol concentrations during the 10 min procedure ranged from 2 to 3 mcg∙mL-1 and then dissipated to near 0.5 mcg∙mL-1 over the next 10 min. This led to a low probability of respiratory compromise and a moderate probability of loss of responsiveness at the end of the procedure that both quickly dissipated. This technique led to a low probability of loss of response to esophageal instrumentation during the 10-min procedure that dissipated within 3 min from the end of the procedure. 3.4.4.2 Technique #2: For the fentanyl bolus followed by intermittent propofol boluses, fentanyl reached a peak of about 1.2 ng∙mL-1 (in remifentanil equivalents) within 5 min of starting induction and then slowly dissipated to near 0.8 ng∙mL-1 at 10 min. The accompanying propofol concentrations ranged between 1 and 2 mcg∙mL-1 and dissipated to less than 0.5 mcg∙mL-1 over the next 10 min. This led to a moderate 60 Figure 3.2: Predicted propofol, fentanyl (in remifentanil equivalents), and remifentanil effect-site concentrations (Ce) for selected published dosing regimens for endoscopy (Panels A and B). Time 0 corresponds to the peak propofol Ce for techniques #1 and #2, the start of the propofol infusion for technique #3, and achievement of the propofol target for technique #4. Panel C presents a topographical plot of propofol versus remifentanil concentrations for each dosing regimen. Arrows indicate the time course of the dosing; dotted and solid orange lines represent drug concentration pairs that produce 5% and 50% probabilities of respiratory compromise. Panel D shows the time to recovery using a topographical plot of propofol versus remifentanil concentrations during the 10-min washout period for each dosing regimen. Arrows indicate the time course of the dosing; closed circles represent the Ces at the end of the procedure, and dotted blue and orange lines represent drug concentration pairs that produce 5% probabilities of loss of responsiveness and respiratory compromise. Numbers represent time (in minutes) to recovery, defined as a probability of effect <5%, and are placed next to the corresponding washout curve and isobole. 61 a) Panel A: Simulated Propofol Effect-Site Concentrations 3.0 2.5 .~... -E 2.0 CJ u -E- CD 1.5 0 ~ 0 a. 1.0 f Il.. 0.5 0.0 -6 -3 - Propofol Boluses ~-- Procedure --)~ I - Propofol Boluses and Fentanyl Bolus o 5 Time (mlns) I - Propofol Bolus and Infusion and Fentanyl Bolus I - Propofol and Remifentanil Target Controlled Infusion 10 15 20 62 b) Figure 3.2 continued Panel B: Simulated Remifentanil Effect-Site Concentrations 1.6 Procedure ..- ...J -E 1;1) ..s.:.::. . 1.2 CD 0 s: -III s: .! 0.8 E CD -a: I:: .! III 0.4 > ::I t7' ILl 0.0 -6 -3 0 ,') 10 15 20 Time (mlns) 63 c) Figure 3.2 continued Panel C: Simulated Propofol and Remifentanil Effect-Site Concentrations 3.0 2.5 -...J -E- 2.0 C) (,,) -E CI) 1.5 U .E 0 .0Q.. . 1.0 a.. t 0.5 0.0 00 0.4 •••• 5% Probability of Respiratory Compromise - 50% Probability of Respiratory Compromise ............ , •••• ••• f ••••• ••••••••••• 0.8 1.2 1.6 Equivalent Remifentanil Ce (ng/mL) 64 d) Figure 3.2 continued Panel 0 : Simulated Time to Recovery 2.5 2.0 -....I E --Cl E 1.5 -CI) U ~ 1.0 Q. ~ c.. 0.5 0.0 3min. •••• 5% Probability of Respiratory Compromise •••• 5% Probability of Loss of Responsiveness ••• ••••••••• •••••• 1 min. ~ ~~ min. • •••••• j. ••••••• 3 · •••••••••••••• min. 00 •• f', •• •• •• •• 7.5 mi n ~ · 0.4 min. 8• • •••••••••• mi• n. ••••••• • •••••••••• 0.8 1.2 Equivalent Remifentanil Ce (ng/mL) •••••• 1.6 65 Figure 3.3: Simulations of loss of response to esophageal instrumentation over time for selected published dosing regimens for upper endoscopy (solid lines). Simulations were designed to provide sedation and analgesia for a 10-min procedure (gray vertical lines). Horizontal dashed lines represent the boundary between low and moderate (25%) and moderate and high (75%) probabilities of effect. The horizontal dotted line represents the boundary for high probability of recovery (<5%). 66 probability of respiratory compromise and a low probability of loss of responsiveness at the end of the procedure. Respiratory compromise dissipated within 8 min while loss of responsiveness dissipated in less than 2. This technique led to a moderate probability of loss of response to esophageal instrumentation during the 10-min procedure that dissipated within 4 min. 3.4.4.3 Technique #3: For the fentanyl bolus 3 min prior to the start of a propofol bolus followed by infusion, fentanyl had a concentration profile similar to that of Technique #2 with the difference that it reached its peak near the start of the propofol bolus. Propofol concentrations ranged between 1.4 and 2 mcg∙mL-1 and then dissipated to less than 0.5 mcg∙mL-1 within 5 min following the procedure. This led to a moderate probability of respiratory compromise and a low probability of loss of responsiveness at the end of the 10-min procedure. Respiratory compromise dissipated within 8 min while loss of responsiveness dissipated in less than 2. This technique led to a moderate probability of loss of response to esophageal instrumentation for 8 min followed by a low probability for the rest of the procedure and dissipated within 4 min. 3.4.4.4 Technique #4: For the target controlled infusions, propofol was maintained at 1.8 mcg∙mL-1 and remifentanil at 1.5 ng∙mL-1 for 10 min. This led to a moderate probability of respiratory compromise and loss of responsiveness at the end of the procedure that required 8 and 3 min to dissipate, respectively. This technique also led to a moderate probability of loss of response to esophageal instrumentation during the procedure that dissipated within 5 min of terminating the infusions. 3.5 Discussion Predicting the likelihood, magnitude and duration of adverse effects such as ventilatory depression, airway obstruction and/or loss of responsiveness is important in formulating rational dosing regimens for procedural sedation. In a prior study, we 67 explored the feasibility of completely blocking the response to esophageal instrumentation in volunteers at various target remifentanil and propofol Ce pairs. Similar to what other authors have reported, we found that rendering a volunteer completely unresponsive to esophageal instrumentation often required doses that were associated with loss of responsiveness, intolerable ventilatory depression, or both.9,10,25 In clinical practice, patient movement and/or discomfort for a brief duration rather than completely blocking the response to esophageal instrumentation may be acceptable in order to avoid unwanted side effects from these drugs. In this present study, we modified our previously reported interaction model of intolerable ventilatory depression to include a measure of airway obstruction and called the combined effect respiratory compromise. We also modified our interaction model of loss of response to esophageal instrumentation by changing the criteria used to define a "response" to esophageal instrumentation. In this revised model, we categorized heart rate or blood pressure changes, non-purposeful movement, and gag response to esophageal instrumentation as "unresponsive" to be more consistent with other published work8,9,12 and better reflect clinical practice during endoscopy. 3.5.1 Effect measures By combining measures of partial or complete airway obstruction with the intolerable ventilatory depression data, volunteers were found to have respiratory compromise at more of the concentration pairs studied. As expected, airway obstruction primarily occurred at high propofol concentrations and intolerable ventilatory depression primarily occurred at high remifentanil concentrations. When interpreting this data, some important limitations merit discussion. First, all measures of loss of responsiveness, airway obstruction and intolerable ventilatory depression were made with volunteers in an unstimulated state. It is well known that 68 stimulation shifts the concentration-effect relationship of anesthetics to the right (i.e. higher concentrations are needed to achieve the same effect).28 Thus, in the presence of procedural stimulation, the number of volunteers that we observed with either loss of responsiveness and/or respiratory compromise would likely decrease. Second, it is likely that once an endoscope is in place, much of the partial or complete airway obstruction would resolve because the endoscope would stent the airway open.12 Furthermore, an increase in body habitus may lead to more prevalent airway obstruction than what we observed. Third, our measures of partial airway obstruction were rather simplistic. More sophisticated techniques exist.29-35 It is possible that our criteria for partial airway obstruction (tidal volume <3 mL·kg-1) did not accurately capture clinically significant partial airway obstruction. Fourth, the time course of airway obstruction or intolerable ventilatory depression necessary to produce clinically significant hypoxia or hypercarbia is not established; nevertheless, we believe that a respiratory rate ≤4 breaths per minute or a 30-second average tidal volume <3 mL·kg-1 would potentially lead to worrisome hypoxia and/or hypercarbia. Fifth, debilitated patients will likely require less propofol and remifentanil to achieve the same airway and respiratory effects. By changing our criteria for loss of response to esophageal instrumentation, more assessments were considered "unresponsive" than in our original model. In our prior work,13 volunteers were unresponsive in 105 out of 367 assessments,. With our revised criteria, 135 were unresponsive. For example, with propofol at 2.7 mcg∙mL-1 and remifentanil at 0.8 ng∙mL-1 and using the original response criteria, 4 volunteers tolerated esophageal instrumentation and 4 did not. With the revised criteria, 5 volunteers tolerated esophageal instrumentation and 3 did not. One potentially important nuance to consider when interpreting these results is the difference in anesthetic requirements between placing an endoscope versus tolerating one already in place. Our anecdotal experience was that during placement of 69 the bougie, some volunteers exhibited a gag response or involuntary movement that resolved once it was in place. Hence, less propofol or propofol with an opioid may be required to keep patients analgesic and sedated during endoscopy once the scope is in place. This concept is confirmed by other authors who observed that endoscope insertion is the most stimulating portion of the procedure.9,12 3.5.2 Response surface models We constructed a response surface model for respiratory compromise and loss of response to esophageal instrumentation. Graphical and statistical approaches indicated that the models fit the observed data well. From a graphical perspective (Figure 3.1C and D), the models captured the transition from no effect to effect well. This was confirmed by the 2 analysis and percentage of model predictions that agreed with observed responses. Our results confirmed our hypothesis that the interaction between propofol and remifentanil would be synergistic for both effect measures as illustrated by the positive alpha values presented in Table 3.1. The respiratory compromise model had a propofol C50 of 4.3 mcg·mL-1 compared to our previously reported 7.0 mcg·mL-1 for intolerable ventilatory depression and is due to the additional airway obstruction data along the propofol axis. In the region of low remifentanil Ces (i.e. <1 ng·mL-1), as propofol Ces increase from 0, any worrisome ventilation is preliminarily likely due to airway obstruction and can be resolved with a head tilt/chin lift and/or insertion of an oral airway. However, as propofol Ces approach 7.0 mcg·mL-1, intolerable ventilatory depression is increasingly present, requiring prompting to breathe or manual ventilations to maintain adequate ventilation. In contrast, the respiratory compromise model had a remifentanil C50 of 6.7 ng·mL-1 compared to our previously reported 4.1 ng·mL-1 for intolerable ventilatory depression. The increase in remifentanil C50 is likely a function of a few more volunteers 70 developing respiratory compromise due to airway obstruction at higher remifentanil concentrations (i.e. near 3 ng·mL-1) and the mathematical limitations of the Greco model structure.14,36 Specifically, the Greco model is an adaptation of the model proposed by Berenbaum for two non-interacting drugs37 and assumes each drug can be independently modeled using the Hill equation (sigmoid-Emax model).14 This assumption imposes mathematical constraints on what type of behavior can be modeled, a limitation that has been described by other authors as insufficiently flexible.36,38 Specifically, the interaction (alpha) and slope (gamma) are held constant for all drug combination ratios. In reality, each drug ratio can itself be considered a unique drug and could potentially have different alpha and gamma values from its neighbors. Additionally, assuming a sigmoid shape imposes an inflection point on the fit, which could lead to poor model fit in some data sets. Various models and techniques have been introduced by other authors to correct for these limitations.36,38,39 The revised loss of response to esophageal instrumentation model is somewhat similar to our previously reported model. The C50's and gamma terms are similar (propofol C50: original 9.8 versus revised 9.6 mcg·mL-1, remifentanil C50: original 3.8 versus revised 4.1 ng·mL-1, and gamma: original 3.7 versus revised 2.7) but the alpha term is larger in the revised model (original 4.5 versus revised 7.7). Although the C50's are similar, the larger alpha in the revised model indicates a more significant drug synergy, meaning less of either drug is required to achieve the same effect. In graphical terms, the iso-effect lines (isoboles) have more of a bow towards the origin with the larger alpha. For both models, the gamma value ranges from 2 to 3. These relatively small values indicate that the range between the 5% and 95% probability isoboles will be large. A wider range indicates more uncertainty of the concentration at which a given subject will transition from no effect to effect, with the typical patient transitioning near 71 the 50% isobole. 3.5.3 Simulations of published dosing regimens for endoscopy With regard to a rapid recovery, the propofol only technique had the fastest recovery (i.e. both no loss of responsiveness and no respiratory compromise within 3-4 min) once the procedure was completed. Nevertheless, it only achieved a low probability of loss of response to esophageal instrumentation during all but the very beginning of the procedure (Figure 3.3). An important clinical implication of these simulations is that should patients require prompting to breathe to avoid ventilatory depression, techniques that minimize loss of responsiveness may be more desirable because patients can respond to the prompting. This may be especially important when dosing with propofol alone; given that propofol has minimal analgesic effect, clinicians may be tempted to administer more propofol and over sedate patients to compensate for the lack of analgesia.24 Simulations of the propofol-opioid techniques did lead to moderate probabilities of respiratory compromise and loss of responsiveness at the end of the procedure, but they also provided a moderate probability of loss of response to esophageal instrumentation. Once drug administration was terminated, the time to return of responsiveness was faster for some of these techniques than it was for propofol alone (Figure 3.2D). More time was required for the respiratory compromise effect to dissipate with the propofol-opioid techniques (7-9 min), but a majority of this time would be with a patient in a responsive state and likely be receptive to prompting to breathe or open their airway. By way of comparison, authors have published observations using propofol in combination with opioids for endoscopy and colonoscopy. For example, in a trial where 496 patients received a fentanyl bolus followed by a computer administered feedback- 72 controlled propofol infusion (SEDASYS), Pambianco et al.26 found that over 95% of the patients experienced mild to moderate sedation during brief procedures (on average <4 min for upper gastrointestinal endoscopy and <14 min for colonoscopy) and a rapid recovery. There was a very low incidence of deeper than intended sedation and adverse respiratory events. They reported an area under the desaturation curve as a surrogate measure of the risk of hypoxic injury. Propofol combined with fentanyl led to a lower on average area under the curve than conventional dosing with midazolam, meperidine, and fentanyl (on average, 23 %∙seconds versus 88 %∙seconds). Although these results are not directly comparable, simulations presented in Figure 3.2C predicted only brief periods of a moderate probability of respiratory compromise. Some additional limitations deserve special emphasis. First, our models assume steady state conditions. This assumption is violated whenever drug concentrations are rapidly changing (e.g., such as after a bolus is injected). The respiratory depression associated with bolus doses of ventilatory depressants is greater than when the same drugs are administered by infusion to similar target concentrations.40,41 Thus, the simulations involving bolus drug administration are likely to be associated with more respiratory compromise than our models predict. Second, simulation predictions are based on population pharmacokinetic models associated with substantial variability. For example, using target controlled infusions, median absolute performance errors for propofol only and remifentanil only of 25%42 and 22%43 have been reported, respectively. The median performance error of propofol in the presence of remifentanil has been reported as 49%.43 Third, some of the published dosing regimens did not provide weight adjusted dosing. When conducting our simulations, we assumed a patient weight of 75 kg. Predictions would be different for simulations using a patient weight of 45 or 100 kg. Fourth, when implementing dosing recommendations, some degree of interpretation was required to formulate a dosing regimen; time intervals 73 between doses and infusion rates were published as ranges. We chose dosing intervals based on a published regimen44 but as with any simulation, we may have inappropriately interpreted the dosing regimens. Fifth, although the models we used to predict propofol and remifentanil concentrations do account for age, our pharmacodynamic models do not. As reported by Kazama et al. and Hammer et al., age is an important covariate when considering doses of propofol for endoscopy.8,9 Finally, although there are obvious limitations to the volunteer setting, particularly the problem of lack of stimulation, the volunteer paradigm is a necessary first step towards building models that can then be perfected and eventually validated in patients. In patients, it is not practically feasible to target the numerous concentration pairs necessary to build a response surface; these volunteer studies typically require an entire day for each subject. Furthermore, it is unethical to intentionally anesthetize or sedate a patient inadequately. In the volunteer setting, using noninvasive stimulation techniques, it is acceptable to produce inadequate anesthesia or sedation intentionally. In summary, we modified two previously reported propofol-remifentanil interaction models of loss of response to esophageal instrumentation and intolerable ventilatory depression. Revised models fit observed responses well. We used them and an additional model to make predictions regarding the temporal profile of recovery for sedation and respiratory endpoints using published dosing regimens for propofol alone and in combination with an opioid for upper endoscopy. Simulations of propofol-opioid techniques led to a moderate probability of conditions that allow esophageal instrumentation whereas propofol only techniques led to a low probability. Once the procedure was terminated, techniques that used a fentanyl bolus just prior to the procedure and propofol throughout the procedure provided the highest likelihood of rapid return of responsiveness. 74 3.6 Appendix: Target Effect-site Concentrations and Respiratory and Esophageal Instrumentation Outcomes Table 3.3: Target effect-site concentrations and respiratory and esophageal instrumentation outcomes Remifentanil Group Propofol Group Effect Measures Effect Measures Set n Remi (ng∙mL-1) Prop (mcg∙mL-1) IVD AO RC LREI (revised) Set n Remi (ng∙mL-1) Prop (mcg∙mL-1) IVD AO RC LREI (revised) 0 12 0.0 0.0 0/12 0/12 0/12 0/12 0 12 0.0 0.0 0/12 0/12 0/12 0/12 1 9 0.0 0.8 0/9 0/8 0/8 0/9 1 8 1.2 0.0 1/8 0/7 1/8 1/8 1 9 0.4 0.8 0/9 0/8 0/8 0/9 1 8 1.2 0.3 0/8 0/8 0/8 1/8 1 9 0.8 0.8 2/9 2/7 4/9 0/9 1 8 1.2 0.6 0/8 0/8 0/8 2/8 1 9 1.6 0.8 3/9 2/6 5/9 2/9 1 8 1.2 1.1 2/8 0/6 2/8 2/8 1 9 3.3 0.8 6/9 1/3 7/9 3/9 1 8 1.2 2.2 5/8 0/3 5/8 6/8 2 8 0.0 1.5 0/8 1/8 1/8 0/8 2 8 2.2 0.0 0/9 0/9 0/9 1/9 2 8 0.4 1.5 0/8 1/8 1/8 0/8 2 8 2.2 0.3 1/9 0/8 1/9 1/9 2 8 0.8 1.5 0/8 2/7 2/7 2/8 2 8 2.2 0.6 2/9 0/7 2/9 1/9 2 8 1.6 1.5 2/8 2/5 4/7 2/7 2 8 2.2 1.1 6/9 0/3 6/9 3/9 2 8 3.3 1.5 7/7 - 7/7 6/7 2 8 2.2 2.2 9/9 - 9/9 7/7 3 9 0.0 2.0 0/9 3/9 3/9 0/9 3 8 3.0 0.0 5/8 1/3 6/8 2/8 3 9 0.4 2.0 0/9 3/9 3/9 3/9 3 8 3.0 0.3 3/8 0/5 3/8 2/8 3 9 0.8 2.0 1/9 7/9 7/9 5/9 3 8 3.0 0.6 5/8 0/3 5/8 3/8 3 9 1.6 2.0 3/7 5/5 8/8 6/7 3 8 3.0 1.1 6/8 0/2 6/8 6/8 3 9 3.3 2.0 6/6 2/2 8/8 6/6 3 8 3.0 2.2 8/8 1/1 8/8 7/8 4 8 0.0 2.7 0/8 3/8 3/8 1/8 4 8 4.0 0.0 4/8 0/4 4/8 1/8 4 8 0.4 2.7 0/8 4/8 4/8 4/8 4 8 4.0 0.3 1/8 0/7 1/8 2/8 4 8 0.8 2.7 1/8 7/8 7/8 5/8 4 8 4.0 0.6 4/8 1/5 4/8 1/8 4 8 1.6 2.7 5/8 3/3 8/8 8/8 4 8 4.0 1.1 6/8 0/2 6/8 3/8 4 8 3.3 2.7 8/8 - 8/8 8/8 4 8 4.0 2.2 8/8 1/1 8/8 7/7 5 1 0.0 3.3 0/1 0/1 0/1 1/1 5 2 5.0 0.0 1/2 0/1 1/2 0/2 5 1 0.8 3.3 0/1 1/1 1/1 1/1 5 2 5.0 0.6 1/2 0/1 1/2 0/2 5 1 1.6 3.3 1/1 1/1 1/1 1/1 5 2 5.0 1.1 2/2 - 2/2 2/2 5 1 3.3 3.3 - 1/1 1/1 1/1 5 2 5.0 2.2 2/2 - 2/2 - 5 1 3.9 3.3 - 1/1 1/1 1/1 5 2 5.0 2.6 2/2 - 2/2 - 75 Table 3.3 continued Remifentanil Group Propofol Group Effect Measures Effect Measures Set n Remi (ng∙mL-1) Prop (mcg∙mL-1) IVD AO RC LREI (revised) Set n Remi (ng∙mL-1) Prop (mcg∙mL-1) IVD AO RC LREI (revised) 6 1 0.0 4.3 0/1 1/1 1/1 1/1 6 2 6.4 0.0 1/1 - 1/1 0/1 6 1 0.4 4.3 0/1 1/1 1/1 1/1 6 2 6.4 0.3 1/1 - 1/1 0/1 6 1 0.8 4.3 0/1 1/1 1/1 1/1 6 2 6.4 0.6 1/1 - 1/1 1/1 6 1 1.6 4.3 1/1 - 1/1 1/1 6 2 6.4 1.1 1/1 - 1/1 1/1 6 1 2.4 4.3 1/1 - 1/1 1/1 6 2 6.4 1.6 1/1 - 1/1 1/1 total 192 47/184 55/141 99/185 71/185 192 89/192 4/106 90/192 64/185 Remi = Remifentanil, Prop = Propofol, N is the number of subjects assigned to each set based on the study design. Effect measures: IVD = Intolerable ventilatory depression defined as a respiratory of ≤4 breaths per minute, AO = Airway obstruction defined as a 30 second average tidal volume <3 ml/kg AND respiratory rate >2 breaths in the same time period OR absence of airway flow in the presence of respiratory effort, RC = Respiratory compromise defined as the presence of IVD and/or AO, LREI = Loss of response to esophageal instrumentation. Dashes (-) = unable to complete evaluation of effect measure. The denominator is the total number of subjects assessed at that concentration pair for the corresponding effect. The numerator is the number of subjects at maximum effect. Totals for each effect are provided at the bottom. After being randomized to either the remifentanil or the propofol group, each subject was further randomized to receive three of the six possible sets of infusion targets within their group. One subject was incorrectly dosed in the propofol group, which caused there to be nine subjects in set two instead of two subjects in set six. 76 3.7 References 1. Lena P, Mariottini CJ, Balarac N, Arnulf JJ, Mihoubi A, Martin R: Remifentanil versus propofol for radio frequency treatment of atrial flutter. Can J Anaesth 2006; 53: 357-62 2. Fatima H, DeWitt J, LeBlanc J, Sherman S, McGreevy K, Imperiale TF: Nurse-administered propofol sedation for upper endoscopic ultrasonography. Am J Gastroenterol 2008; 103: 1649-56 3. Cohen LB: Nurse-administered propofol sedation for upper endoscopic ultrasonography: Not yet ready for prime time. Nat Clin Pract Gastroenterol Hepatol 2009; 6: 76-7 4. Pambianco DJ, Whitten CJ, Moerman A, Struys MM, Martin JF: An assessment of computer-assisted personalized sedation: A sedation delivery system to administer propofol for gastrointestinal endoscopy. Gastrointest Endosc 2008; 68: 542-7 5. Mandel JE, Tanner JW, Lichtenstein GR, Metz DC, Katzka DA, Ginsberg GG, Kochman ML: A randomized, controlled, double-blind trial of patient-controlled sedation with propofol/remifentanil versus midazolam/fentanyl for colonoscopy. Anesth Analg 2008; 106: 434-9 6. Dumonceau JM, Riphaus A, Aparicio JR, Beilenhoff U, Knape JT, Ortmann M, Paspatis G, Ponsioen CY, Racz I, Schreiber F, Vilmann P, Wehrmann T, Wientjes C, Walder B: European Society of Gastrointestinal Endoscopy, European Society of Gastroenterology and Endoscopy Nurses and Associates, and the European Society of Anaesthesiology Guideline: Non-anaesthesiologist administration of propofol for GI endoscopy. Eur J Anaesthesiol 2010; 27: 1016-30 7. Perel A: Non-anaesthesiologists should not be allowed to administer propofol for procedural sedation: A consensus statement of 21 European National Societies of Anaesthesia. Eur J Anaesthesiol 2011; 28: 580-4 8. Hammer GB, Litalien C, Wellis V, Drover DR: Determination of the median effective concentration (EC50) of propofol during oesophagogastroduodenoscopy in children. Paediatr Anaesth 2001; 11: 549-53 9. Kazama T, Takeuchi K, Ikeda K, Ikeda T, Kikura M, Iida T, Suzuki S, Hanai H, Sato S: Optimal propofol plasma concentration during upper gastrointestinal endoscopy in young, middle-aged, and elderly patients. Anesthesiology 2000; 93: 662-9 10. Nelson DB, Barkun AN, Block KP, Burdick JS, Ginsberg GG, Greenwald DA, Kelsey PB, Nakao NL, Slivka A, Smith P, Vakil N: Propofol use during gastrointestinal endoscopy. Gastrointest Endosc 2001; 53: 876-9 11. Nieuwenhuijs DJ, Olofsen E, Romberg RR, Sarton E, Ward D, Engbers F, Vuyk J, Mooren R, Teppema LJ, Dahan A: Response surface modeling of remifentanil-propofol interaction on cardiorespiratory control and bispectral index. Anesthesiology 2003; 98: 312-22 77 12. Drover DR, Litalien C, Wellis V, Shafer SL, Hammer GB: Determination of the pharmacodynamic interaction of propofol and remifentanil during esophagogastroduodenoscopy in children. Anesthesiology 2004; 100: 1382-6 13. LaPierre CD, Johnson KB, Randall BR, White JL, Egan TD: An exploration of remifentanil-propofol combinations that lead to a loss of response to esophageal instrumentation, a loss of responsiveness, and/or onset of intolerable ventilatory depression. Anesth Analg 2011; 113: 490-9 14. Greco WR, Bravo G, Parsons JC: The search for synergy: A critical review from a response surface perspective. Pharmacol Rev 1995; 47: 331-85 15. 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: 347-55 16. 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: 1430-43 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: 10-23 18. Schnider TW, Minto CF, Gambus PL, Andresen C, Goodale DB, Shafer SL, Youngs EJ: The influence