- Email: [email protected]
Effect of dexmedetomidine on propofol requirements in healthy subjects
Effect of Dexmedetomidine on Propofol Requirements in Healthy Subjects SANDEEP DUTTA,1 MICHAEL D. KAROL,1 THEODORA COHEN,2 RON M. JONES,3 TIMOTHY MANT4 1
Clinical Pharmacokinetics Department, Abbott Laboratories, 100 Abbott Park Road, Department 4PK, AP13A-3, Abbott Park, Illinois 60064-6104
2
Biometrics, IBAH, Lake Bluff, Illinois
3
St. Mary's Hospital Medical School, Imperial College, London, United Kingdom
4
Guys Drug Research Unit, London, United Kingdom
Received 30 September 1999; revised 25 April 2000; accepted 23 August 2000
ABSTRACT: Dexmedetomidine±propofol pharmacodynamic interaction was evaluated in nine healthy subjects in a crossover design. Dexmedetomidine/placebo was infused using a computer-controlled infusion pump (CCIP) to maintain a pseudo-steady-state plasma concentration of 0.66 0.080 or 0 ng/mL, respectively. Forty-®ve minutes after the dexmedetomidine/placebo infusion was started, propofol was infused using a second CCIP to achieve a stepwise logarithmically ascending propofol concentration (1.00 to 13.8 mg/mL) pro®le. Each propofol step lasted 10 min. Blood was sampled for plasma concentration determination, and pharmacodynamic endpoint assessments were made during the study. Propofol and dexmedetomidine/placebo infusions were terminated when three endpoints (subjects were too sedated to hold a syringe, followed by loss of eyelash re¯ex, followed by loss of motor response to electrical stimulation) were achieved sequentially. The concentration of propofol associated with 50% probability of achieving a pharmacodynamic endpoint in the absence of dexmedetomidine (EC50; placebo treatment) was 6.63 mg/mL for motor response to electrical stimulation and ranged from 1.14 to 1.98 mg/mL for the ability to hold a syringe, eyelash re¯ex, and sedation scores. The apparent EC50 values of propofol (EC50APP; concentration of propofol at which the probability of achieving a pharmacodynamic endpoint is 50% in the presence of dexmedetomidine concentrations observed in the current study; dexmedetomidine treatment) were 0.273, 0.544±0.643, and 3.89 mg/mL for the ability to hold a syringe, sedation scores, and motor response, respectively. Dexmedetomidine reduced propofol concentrations required for sedation and suppression of motor response. Therefore, the propofol dose required for sedation and induction of anesthesia may have to be reduced in the presence of dexmedetomidine. ß 2001 Wiley-Liss, Inc. and the American Pharmaceutical Association J Pharm Sci 90:172±181, 2001 Keywords: dexmedetomidine; Precedex; propofol; pharmacokinetics; pharmacody-
namics; modeling; EC50; sedation; motor response; anesthesia
INTRODUCTION Dexmedetomidine is the pharmacologically active d-isomer of medetomidine. It is a new generation Correspondence to: S. Dutta (Telephone: 847-937-8502; Fax: 847-939-5193, E-mail: [email protected]). Journal of Pharmaceutical Sciences, Vol. 90, 172±181 (2001) ß 2001 Wiley-Liss, Inc. and the American Pharmaceutical Association
172
a2-adrenoceptor agonist with high selectivity for the a2- versus a1-receptor.1,2 It has been shown to reduce the anesthetic requirements in experimental animals and human subjects, with effects similar to clonidine.3 Dexmedetomidine has greater a2- versus a1-selectivity (2000:1) than clonidine (300:1).1,2 Administration of dexmedetomidine during the perioperative period has been
JOURNAL OF PHARMACEUTICAL SCIENCE, VOL. 90, NO. 2, FEBRUARY 2001
DEXMEDETOMIDINE±PROPOFOL PHARMACODYNAMIC INTERACTION IN HEALTHY SUBJECTS
demonstrated to reduce the requirements for anesthetics (e.g., thiopental, iso¯urane), opioid analgesics (e.g., fentanyl, oxycodone), and other sedative and hypnotic drugs.3±6 This reduction in anesthetic and opioid requirements as a result of dexmedetomidine coadministration may potentially alleviate the ventilatory depression associated with these agents.3±6 Propofol is an intravenous (iv) sedative/hypnotic agent that is extensively used for the induction and maintenance of anesthesia and for sedation in the intensive care units (ICU).7±9 Because of its rapid induction and recovery characteristics and anti-emetic properties, propofol is commonly incorporated into the anesthetic regimen for outpatient surgeries.8 Because dexmedetomidine reduces anesthetic requirements, it is anticipated that dexmedetomidine may have pharmacodynamic interactions with the extensively used nonbarbiturate anesthetic agent propofol. This study evaluated, in healthy male subjects, the impact of a single, constant dexmedetomidine concentration on propofol dose requirements by assessments of the pharmacodynamic endpoints of the ability to hold a syringe, eyelash re¯ex, motor response to electrical stimulation, and sedation scores.
METHODS Study Design This was a single center, double-blind, randomized, placebo-controlled, two-period, crossover, Phase I study in 10 male subjects. The study was approved by the Guy's Hospital Research Ethics Committee. A 14-day washout period was allowed between treatment sessions. During a treatment session, the subjects received iv infusions of dexmedetomidine (Precedex, Abbott Laboratories, Abbott Park, IL) or placebo (0.9% sodium chloride solution packaged identical to dexmedetomidine and provided by Abbott Laboratories) using a computer-controlled infusion pump (CCIP). A constant dexmedetomidine concentration of 0.66 0.080 ng/mL was achieved and maintained for the duration of the infusion in the dexmedetomidine treatment group. This dexmedetomidine concentration is in the middle of the anticipated therapeutic range. Dexmedetomidine or placebo was administered through an iv catheter dedicated for drug delivery. The study site personnel operating the CCIP were blinded to the study
173
drug. The CCIP used was a Harvard Model 22 that was connected to an IBM portable computer via an RS232 serial connector. The computer program that controlled the dosing was STANPUMP,10 which drives the infusion pump to administer drugs according to prede®ned pharmacokinetic models and parameters, and desired drug plasma concentration pro®les.11 For dexmedetomidine, a two-compartment model, with drug input into and elimination from the central compartment and ®rst-order intercompartmental transfer, was used to drive the infusion pump. For propofol, a three-compartment model, with drug input into and elimination from the central compartment and ®rst-order intercompartmental transfer between the central and two peripheral com-partments, was used to drive the infusion pump. Forty-®ve minutes after the start of the dexmedetomidine or placebo infusion, propofol (Dipravan1 1%, Zeneca Pharmaceuticals, Wilmington, DE) infusion was initiated with a second CCIP in a manner designed to achieve a stepwise logarithmically ascending propofol concentration pro®le. Propofol was administered through a different iv catheter using a second CCIP and the program STANPUMP with the intent of achieving up to 11 pseudo-steady-state plasma concentrations ranging from 1.00 to 13.8 mg/mL. Each propofol concentration step lasted 10 min. During each propofol concentration step, blood samples for determination of drug concentrations were drawn, and the pharmacodynamic assessments were made. Propofol and dexmedetomidine/placebo infusions were terminated when all three of the following pharmacodynamic endpoints were achieved sequentially; (a) the subjects were too sedated to hold a syringe, followed by (b) loss of eyelash re¯ex, followed by (c) lack of motor response to a 10-s, 100-Hz electrical stimulation. Response to electrical stimulation was de®ned as gross purposeful movement of any part of the body except the stimulated limb. Blood samples (5 mL) were collected for determination of dexmedetomidine and propofol plasma concentrations. The blood samples were collected prior to the dexmedetomidine infusion at 0 h or baseline, prior to the propofol infusion, 10 min after the start of every propofol step, at emergence from anesthesia, and at 1, 2, 4, and 8 h after termination of dexmedetomidine/placebo and propofol infusions. All blood samples were assayed for propofol plasma concentrations. All blood samples, except for every odd-numbered propofol step, JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 90, NO. 2, FEBRUARY 2001
174
DUTTA ET AL.
were assayed for dexmedetomidine plasma concentrations. Prospective power calculations based on response to tetanic stimulation were used to calculate the number of subjects for the study. The use of 10 subjects in a two-way crossover study was expected to enable detection of a 25% decrease in the propofol concentration required for loss of response to tetanic stimulation at a 0.05 level of signi®cance with statistical power of > 80%. Subjects Subjects selected were healthy nonsmoking males between the ages of 18 and 45 years and within 15% of the upper and lower boundaries of weight set forth in the Metropolitan Life Insurance Table. In the screening period prior to dosing, candidates received a full explanation of the study, provided written informed consent, and underwent pre-study procedures. Subjects were judged to be in good health based on the results of medical history, physical examination, vital signs, electrocardiogram, and routine clinical laboratory evaluations. Subjects had not taken any prescription medication within 4 weeks or overthe-counter medication within 2 weeks prior to study drug administration. Analytical Methods Dexmedetomidine concentrations were determined by gas chromatography/mass spectroscopy (GC/MS). Brie¯y, the extraction procedure involved the addition of a 75-mL aliquot internal standard (4-[1-(2-methylphenyl)ethyl]-1H-imidazole) to 1.0 mL of sample. The samples were then basi®ed with 0.5 mL of a 1 M sodium carbonate solution and extracted into 7.0 mL of hexane while simultaneously undergoing derivatization by the addition of 0.5 mL of 0.5% penta¯uorobenzoyl chloride solution. The organic layer was transferred into a clean, dry tube and dried under nitrogen. The samples were then reconstituted in 0.2 mL of toluene, and 3 mL was injected onto the GC/MS system. The lower limit of quanti®cation was 10 pg/mL and the within-day and interday coef®cients of variation (CVs) were < 15%. Propofol concentrations were determined by high-performance liquid chromatography (HPLC) with ¯uorescence detection (HPLC assay was performed at Pheonix International Life Sciences Inc., St-Laurent, Canada). Brie¯y, propofol and JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 90, NO. 2, FEBRUARY 2001
its internal standard, thymol, were extracted with pentane/dichloromethane and evaporated to dryness. The residue was reconstituted with mobile phase and injected onto the HPLC system. Separation was obtained using a reversed-phase column under isocratic conditions, and the ef¯uent was continuously monitored by ¯uorescence detection. The lower limit of quanti®cation was 0.061 mg/mL and the within-day and interday CVs were < 5%. Pharmacodynamic Measurements The ability to hold a syringe, the presence of eyelash re¯ex, and the presence of motor response were measured on a binary scale (Yes/No). Only the ability to hold a syringe was measured immediately prior to starting the propofol infusion. During each propofol infusion step, the ability to hold a syringe, the presence of eyelash re¯ex, and the presence of motor response to electrical stimulation were measured 8 min after the start of each step. Sedation was measured on a six point Ramsay scale (1Ðpatient anxious, agitated, or restless; 2Ðpatient cooperative, oriented, and tranquil; 3Ðpatient responds to commands; 4Ðpatient asleep but with brisk response to light labellar tap or loud auditory stimulus; 5Ðpatient asleep, with sluggish response to light labellar tap or loud auditory stimulus; 6Ðpatient asleep, no response). Ramsay sedation scores were recorded prior to dexmedetomidine/placebo infusion, immediately prior to propofol infusion, 5 min after the start of each propofol infusion step, every 5 min following the end of propofol and dexmedetomidine/placebo infusions until emergence, and every 15 min for 4 h after the termination of both infusions. Pharmacokinetic and Pharmacodynamic Analyses Pharmacokinetic Analyses Noncompartmental analyses were performed to obtain the following pharmacokinetic parameters. The maximum plasma concentration (Cmax), clearance (CL), the terminal elimination half-life (t1/2), and the volume of distribution at steady state (VSS). Treatment effects (placebo versus dexmedetomidine) on propofol pharmacokinetic parameters were evaluated ( p < 0.05) using analysis of variance. The analysis of variance model had sequence, subject within sequence, treat-
DEXMEDETOMIDINE±PROPOFOL PHARMACODYNAMIC INTERACTION IN HEALTHY SUBJECTS
ment, and the treatment by sequence interaction as independent factors. Pharmacodynamic Analyses The results for the tests for the ability to hold a syringe, the presence of eyelash re¯ex, and the presence of motor response to electrical stimulation were expressed on a binary scale (Yes/No). The effect of propofol and dexmedetomidine on the probability of a Ramsay sedation score greater than a speci®ed value was modeled using logistic regression. Because the Ramsay sedation score may be one of six values (1, 2, 3, 4, 5, or 6), there are ®ve possible binary logistic regressions; they are the probabilities of Ramsay sedation score > 1, > 2, > 3, > 4, or > 5. The orientation of probability of alertness (1 ÿ probability of sedation) was selected, rather than probability of sedation, for ease of graphic interpretation. When probabilities versus concentration are plotted in this way, decreases in the plotted curves indicate decreased alertness. Logistic regression analyses with repeated measurements12,13 ( p< 0.05) were performed on these binary data to determine the effect of dexmedetomidine and propofol and their interaction on the probability of alertness for each pharmacodynamic endpoint. The mathematical theory,12 derivations, and equations used for logistic regression can be found in the Appendix. The logistic function12 (eq. A1, Appendix) gives two curves for probability of alertness; one for the placebo and one for the dexmedetomidine treatment for each pharmacodynamic endpoint. The true EC50 of propofol is the concentration of propofol associated with 50% probability of achieving a pharmacodynamic endpoint in the absence of dexmedetomidine (placebo treatment). The apparent EC50 of propofol (EC50APP) is the concentration of propofol at which the probability of achieving a pharmacodynamic endpoint is 50% in the presence of dexmedetomidine concentrations observed in the current study (dexmedetomidine treatment). Propofol EC50 and EC50APP were calculated from the coef®cients of the logistic regression function (Appendix). The 95% con®dence intervals for EC50 and EC50APP were calculated by Fieller's method.14 Treatment effects (placebo versus dexmedetomidine) on recovery time (time from termination of infusions to emergence from anesthesia) and the propofol concentration at the end of the dexmedetomidine and propofol infusions were
175
evaluated ( p < 0.05) by analysis of variance. The analysis of variance model had sequence, subject within sequence, treatment, and the treatment by sequence interaction as independent factors. Dexmedetomidine plasma concentrations at 10 min after the start of each odd-numbered step were estimated by linear interpolation if the next even-numbered propofol infusion step took place. If the odd-numbered step was the last propofol infusion step, the dexmedetomidine concentration was set equal to the previous measurement. Both dexmedetomidine and propofol concentrations at the time when Ramsay sedation scores were recorded were estimated by linear interpolation. Only the data during the dexmedetomidine/ placebo and propofol infusions were used for the logistic regression analysis. Time to recovery was de®ned as the time between the end of the dexmedetomidine/placebo and propofol infusions and the time of emergence. The logistic regression analyses were performed with Procedure GENMOD and analyses of variance were performed with Procedure GLM of the SAS software.13 A signi®cance level of 0.05 was used for all tests.
RESULTS AND DISCUSSION The mean age, weight, and height of the nine subjects who completed the study were 24 years (range, 18±28 years), 73.7 kg (range, 61.0±87.3 kg), and 175 cm (range, 168±183 cm), respectively. One subject did not complete the study for personal reasons. The partial data (only dexmedetomidine treatment) from this subject were not included in the pharmacokinetic, pharmacodynamic, and statistical analyses. Pharmacokinetics Figure 1 presents the mean dexmedetomidine and propofol plasma concentration±time pro®les. The mean standard deviation (SD) pseudosteady-state concentration of dexmedetomidine during the computer-controlled infusion was 0.66 0.080 ng/mL (Table 1). The mean dexmedetomidine concentration at emergence was 0.470 ng/mL. The mean propofol concentrations at emergence were 2.32 0.694 and 0.803 0.226 mg/mL for the placebo and dexmedetomidine treatments, respectively. For the placebo treatment, four subjects completed Step 9, with an average concentration of 7.45 1.38 mg/mL. JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 90, NO. 2, FEBRUARY 2001
176
DUTTA ET AL.
Figure 1. Mean dexmedetomidine (top panels) and propofol (bottom panels) plasma concentrations during and after termination of the infusions. Error bars indicate standard deviations. Propofol pseudo-steady-state concentrations were increased logarithmically every 10 min.
Two of these subjects continued on to Step 10, where they had an average propofol concentration of 6.38 1.01 mg/mL. For the dexmedetomidine treatment, the mean propofol concentrations achieved in the last two steps, Steps 7 and 8, were 4.05 0.220 (n 3) and 4.94 (n 1) mg/mL, respectively. Dexmedetomidine and propofol doses, infusion durations, and pharmacokinetic parameters (mean SD) are summarized in the Table 1. Dexmedetomidine pharmacokinetic parameters were consistent with values obtained from other studies.15 Propofol pharmacokinetic parameters were not signi®cantly in¯uenced by the presence of dexmedetomidine in this study, and these JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 90, NO. 2, FEBRUARY 2001
parameter values were consistent with values reported in the literature.7,9,16,17 Propofol VSS was probably underestimated because blood samples were collected for only 8 h after termination of infusion. Studies with prolonged infusions and extended sampling have shown larger VSS and longer t1/2 values associated with slowly equilibrating tissue compartments.7 Pharmacodynamics Propofol Dose and Concentration at the End of Infusion Dexmedetomidine decreased the mean propofol dose required to reach the termination of infusion
DEXMEDETOMIDINE±PROPOFOL PHARMACODYNAMIC INTERACTION IN HEALTHY SUBJECTS
Table 1.
177
Dexmedetomidine and Propofol Pharmacokinetic Parameters (Mean SD) Propofol
Parameter Dose Infusion Duration Cmax Ccss Tmax te1=2 MRT CL Vss Vss
Units
Dexmedetomidine
Placebo Treatment
Dexmedetomidine Treatment
mga mgb h ng/mLa , mg/mLb ng/mL h h h L/h L L/kg
106 14 1:69 0:29 0:72 0:060 0:66 0:080 1:10 0:47 2:27 0:456 3:14 0:682 41:3 8:15 126 18:1 1:71 0:193
994 404 1:34 0:23 6:00 2:00 Ðd 2:08 0:21 2:33 0:700 2:45 0:388 119 22:4 291 69:9 3:93 0:820
492 259 0:93 0:29 3:97 2:34 Ðd 1:70 0:30 2:27 0:815 2:18 0:744 124 46:8 264 113 3:53 1:36
a
Dexmedetomidine. Propofol. c Pseudo-steady-state concentration. d Plasma concentrations were increased every 10 min. e Presented as the harmonic mean and pseudo standard deviation; the terminal elimination rate constant was tested for treatment effect. b
criteria by approximately one-half (Table 1). The mean propofol infusion durations were also decreased in presence of dexmedetomidine (Table 1). Therefore, propofol doses for induction and maintenance of anesthesia may have to be reduced in presence of dexmedetomidine. The effect of dexmedetomidine on propofol plasma concentration at the end of the infusion was evaluated using analysis of variance. Mean propofol plasma concentration at the end of the infusion for the placebo and dexmedetomidine treatments were 5.97 2.00 and 3.97 2.34 mg/ mL, respectively. These concentrations were statistically signi®cantly different. Dexmedetomidine and propofol infusions were terminated when a motor response to electrical stimulation could not be elicited in a subject. Based on this termination of dosing criterion, it can be concluded that dexmedetomidine reduced the propofol concentration required to suppress motor response to electrical stimulation.
concentrations of 0.66 0.080 ng/mL, is unlikely to alter emergence from anesthesia when coadministered with propofol. Logistic Regression Analyses The ability to hold a syringe, presence of eyelash re¯ex, presence of motor response to electrical stimulation, and the binary variables based on the Ramsay sedation scores were analyzed using logistic regression models with dexmedetomidine treatment groups, propofol concentrations, and their interaction as independent variables. Figure 2 shows the rank order of steady-state propofol concentrations required to elicit the different pharmacodynamic endpoints in the absence (top panel) and presence of dexmedetomidine (bottom panel). Comparison of these two plots reveal that, in general, the propofol concentration required to achieve a given pharmacodynamic endpoint is reduced in the presence of dexmedetomidine.
Recovery Time
Sedation
The effect of dexmedetomidine on the recovery time was evaluated using analysis of variance. The mean recovery times were 16 10 and 15 7.7 min for the placebo and dexmedetomidine treatments, respectively. These recovery times were not statistically signi®cantly different. Therefore, the presence of dexmedetomidine at
Propofol EC50 and EC50APP values for each pharmacodynamic endpoint are reported in Table 2. Propofol EC50 (Table 2 and Figure 2, top panel) for the ability to hold a syringe, eyelash re¯ex, and Ramsay sedation scores > 2, 3, 4, and 5 ranged from 1.14 to 1.98 mg/mL. Propofol EC50APP (Table 2 and Figure 2, bottom panel) could not be JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 90, NO. 2, FEBRUARY 2001
178
DUTTA ET AL.
Figure 2. The probability of different pharmacodynamic endpoints evaluated in this study versus propofol concentration, for the placebo (top panel) and dexmedetomidine (bottom panel) treatments. These plots show the rank order of steady-state propofol concentrations eliciting the different pharmacodynamic endpoints in the absence (top panel) and presence of dexmedetomidine (bottom panel).
determined for eyelash re¯ex and Ramsay sedation score > 2. Propofol EC50APP for the ability to hold a syringe and Ramsay sedation scores > 3, 4, and 5 were 65±80% lower than EC50. The presence of dexmedetomidine had a statistically signi®cant effect (coef®cient a3 was signi®cantly different from zero, Appendix) on the propofol concentrations required to suppress these pharmacodynamic measures of sedation. JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 90, NO. 2, FEBRUARY 2001
Signi®cant Interactions. In these analyses, a value for the a4 coef®cient (Appendix) that is statistically signi®cantly different from zero represents a pharmacodynamic interaction. Speci®cally, a positive a4 value indicates antagonism and a negative a4 value indicates synergism. Statistical analyses indicated that there was a signi®cant interaction between dexmedetomidine and propofol for the pharmacodynamic endpoints
DEXMEDETOMIDINE±PROPOFOL PHARMACODYNAMIC INTERACTION IN HEALTHY SUBJECTS
179
Table 2. Dexmedetomidine±Propofol Pharmacodynamic Interaction Variable Ability to hold syringe Eyelash re¯ex Motor response Ramsay >2 Ramsay >3 Ramsay >4 Ramsay >5
Propofol EC50 (mg/mL)a 1.37 1.86 6.63 1.14 1.54 1.70 1.98
(1.18±1.60) (1.61±2.21) (5.50±9.32) (0.922±1.33) (1.35±1.80) (1.50±2.00) (1.70±2.31)
Propofol EC50APP (mg/mL)b 0.273 ( ÿ 0.128±0.625) Ðc 3.89 (3.17±6.69)d Ðc 0.544 (0.260±0.766) 0.610 (0.325±0.814) 0.643 (0.358±0.879)
All Ramsay scores were > 1; parameters reported as mean (95% con®dence interval). a Propofol concentration associated with 50% probability of achieving a pharmacodynamic endpoint. b Propofol concentration associated with 50% probability of achieving a pharmacodynamic endpoint with dexmedetomidine coadministration. c Probability of Ramsay sedation score > 2 and eyelash relfex were 0.4. d Dexmedetomidine did not have a statistically signi®cant effect on slope of motor response versus propofol concentration curve.
of eyelash re¯ex and Ramsay sedation score > 2. In the presence of pseudo-steady-state dexmedetomidine plasma concentration of 0.66 0.080 ng/ mL, the subjects were markedly sedated even when no propofol was present, with probabilities of 0.42 and 0.14 for eyelash re¯ex and Ramsay sedation score > 2, respectively. Therefore, an EC50APP for propofol could not be determined for eyelash re¯ex and Ramsay sedation score > 2 (Figure 2). For these pharmacodynamic endpoints, antagonism was observed between dexmedetomidine and propofol. For other sedation parameters (i.e., the ability to hold a syringe and Ramsay sedation scores > 3, 4, and 5), no antagonism or synergism was observed between dexmedetomidine and propofol. It should be noted that large variability of the interaction coef®cient (a4) for some of the metrics may have limited our ability to detect statistically signi®cant synergism or antagonism. Motor Response Propofol EC50 for motor response to electrical stimulation was 6.63 mg/mL. Propofol EC50APP for motor response to electrical stimulation was 3.89 mg/mL, 40% lower than EC50. However, logistic regression analysis failed to reveal that the effect of dexmedetomidine was statistically signi®cant (coef®cient a3 was not signi®cantly different from zero, Appendix). High variability and limited sample size may be responsible for lack of signi®cant effect of dexmedetomidine on the slope of the logistic regression curves for motor response. Interaction (i.e., antagonism or synergism; coef®cient a4, Appendix), between
dexmedetomidine and propofol for motor response was not statistically signi®cant. Although dexmedetomidine did not have a statistically signi®cant effect in the logistic regression analyses, there is evidence from the two analyses performed that indicate dexmedetomidine reduces propofol concentrations required to suppress motor response to electrical stimulation. First, propofol EC50APP was 40% lower than EC50, and second, based on the termination of infusion criterion of no motor response to electrical stimulation, statistically signi®cantly lower propofol concentrations were achieved at the end of the infusions for the dexmedetomidine treatment compared with the placebo treatment. The relationships between propofol plasma concentrations and different pharmacodynamic measures of sedation evaluated in the current study in the absence of dexmedetomidine are consistent with our understanding of propofol pharmacodynamics. Propofol plasma concentrations had a statistically signi®cant effect on all the sedation parameters examined. Propofol EC50 values for all the sedation parameters (except motor response) evaluated in the current study ranged from 1.14 to 1.98 mg/mL. These values are in excellent agreement with literature data that suggest that propofol concentrations of 1±2 mg/ mL are required for maintenance of sleep, 24 mg/ mL are needed for maintenance of anesthesia, and 7±10 mg/mL are necessary for surgical anesthesia in humans. Others have found that propofol concentrations from 1 to 6 mg/mL covers a range of states from conscious sedation to full anesthesia.16,18 JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 90, NO. 2, FEBRUARY 2001
180
DUTTA ET AL.
The propofol concentration required for suppression of motor response determined in the current study also agrees with concentrations required for achieving similar pharmacodynamic endpoints in humans and rats. Propofol blood concentrations of 6±10 mg/mL are associated with onset of unconsciousness, electroencephalographic (EEG) burst suppression, and surgical anesthesia in humans.16,18 In rats, plasma concentrations of > 7 mg/mL are required to produce EEG burst suppression.19 These values are similar to propofol EC50 for motor response (6.63 mg/mL) determined in the current study.
antagonism or additivity). Using the orientation of probability of alertness, when a4 is negative, there is synergism; when it is positive, there is antagonism; and when a4 is zero, there is simple additivity. A statistically signi®cant interaction (i.e., synergism or antagonism) is said to occur when a4 is signi®cantly different ( p< 0.05) from zero. For simplicity let us de®ne P P(1|Pr,D) and x a1 a2Pr a3D a4PrD. Then eq. A1 becomes: P1ÿ x ln
CONCLUSIONS Dexmedetomidine did not have a signi®cant in¯uence on propofol pharmacokinetics in this study. Dexmedetomidine (0.66 0.080 ng/mL) reduced propofol concentrations required for sedation by 65±80% and for suppression of motor response to electrical stimulation by 40%. Therefore, the propofol dose administered for sedation and induction of anesthesia may have to be reduced in the presence of dexmedetomidine.
APPENDIX Mathematical Theory and Equations for Logistic Regression Equation A1 shows a logistic function for modeling the pharmacodynamic interaction between two drugs: P 1jPr; D 1 ÿ
Pr
ln
ÿ
P 1ÿP
1 e a1 a2 Pra3 Da4 PrD 1 e a1 a2 Pra3 Da4 PrD
A1
where P(1|Pr,D) is the probability of response for a given pharmacodynamic endpoint, (e.g., ability to hold syringe, eyelash re¯ex, etc.) at a given propofol concentration, Pr and a given dexmedetomidine concentration D, and where both Pr and D are continuous variables. The terms a1, a2, a3, and a4 are coef®cients obtained by logistic regression. Coef®cient a1 is referred to as the intercept coef®cient. Terms a2 and a3 are the concentration or slope coef®cients for propofol and dexmedetomidine, respectively. The term a4 describes the pharmacodynamic interaction (synergism or JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 90, NO. 2, FEBRUARY 2001
A2
P 1ÿP
a1 a2 Pr a3 D a4 Pr D ln
A3 P A4 1ÿP
ÿ a1 a3 D a2 a4 D
A5
The intersection of a horizontal plane with the three-dimensional response pro®le de®ned by the logistic function (eq. A1) represents a two-dimensional isobologram. Equation A5 (a two-dimensional isobologram) gives the solution of eq. A1 for one drug concentration given a selected probability, a concentration of the second drug, and the function parameters. In the case of the 50% isobologram, where P 0.5, eq. A5 simpli®es to: EC50APP-Pr Pr
1 e a1 a2 Pra3 Da4 PrD
1 1 ex
ÿ a1 a3 D a2 a4 D
A6
Equation A6 permits computation of the apparent EC50 (i.e., propofol concentration required to elicit a 50% probability of response, given any particular concentration of dexmedetomidine). Conversely, this expression may be rearranged to permit the computation of the dexmedetomidine concentration required for the same endpoint given a concentration of propofol. That form of the expression appears below. EC50APP-D D
ÿ a1 a2 Pr a3 a4 Pr
A7
The negative of the ratio of the intercept coef®cient to the appropriate concentration coef®cient
DEXMEDETOMIDINE±PROPOFOL PHARMACODYNAMIC INTERACTION IN HEALTHY SUBJECTS
gives the concentration at which a 50% probability of response occurs when only one drug is present (i.e., Pr or D is equal to zero). Thus, the true or intrinsic EC50 values may be computed by:
4.
ÿa1 a2 ÿa1 EC50-D a3
5.
EC50-Pr
A8 A9
where EC50-Pr and EC50-D are the true EC50 values of propofol and dexmedetomidine, respectively. In the current study, the pharmacodynamic responses to different propofol plasma concentrations were measured in the absence of dexmedetomidine (placebo) and presence of dexmedetomidine. Because the pharmacodynamic responses were measured only at one pseudosteady-state dexmedetomidine concentration, the variable D was expressed as a categorical variable that had a value of 0 (dexmedetomidine not present) or 1 (dexmedetomidine present). Therefore, P(1|Pr,D) represents the probability of being alert at a given propofol concentration (Pr) in the presence or absence of dexmedetomidine (D). Computation of the true EC50 of propofol under these conditions is unaffected by the reparameterization of the dexmedetomidine variable ``D''. Thus, eq. A8 can be used. With dexmedetomidine variable ``D'' reparameterized as either 0 or 1, the equation for calculation of propofol apparent EC50, eq. A6, becomes: EC50APP EC50APP-Pr
ÿ a1 a3 a2 a4
A10
In the current study, dexmedetomidine was presented only at one pseudo-steady-state concentration and was represented as a categorical variable (present or absent). Therefore, EC50 and EC50APP for dexmedetomidine were not calculated.
6.
7. 8. 9.
10.
11.
12. 13. 14. 15. 16. 17.
REFERENCES 1. Virtanen R. 1989. Pharmacological pro®les of medetomidine and its antagonist, atipamezole. Acta Vet Scand Suppl. 85:29±37. 2. Tong C, Eisenach JC. 1994. Alpha-2 adrenergic agonists. New Drugs Anesth: Part II. 12:49±63. 3. Aantaa R, Kanto J, Scheinin M, Kallio A, Scheinin H. 1990. Dexmedetomidine, an alpha-2-adrenocep-
18.
19.
181
tor agonist, reduces anesthetic requirements for patients undergoing minor gynecologic surgery. Anesthesiology 73:230±235. Jaakola ML, Ali-Melkkila T, Kanto J, Kallio A, Scheinin H, Scheinin M. 1992. Dexmedetomidine reduces intraocular pressure, intubation responses and anaesthetic requirements in patients undergoing opthalmic surgery. Br J Anaesth 68:570±575. Jaakola ML. 1994. Dexmedetomidine premedication before intravenous regional anesthesia in minor outpatient hand surgery. J Clin Anesth 6: 204±211. Aho MS, Erkola OA, Scheinin H, Lehtinen AM, Korttila KT. 1991. Effect of intravenously administered dexmedetomidine on pain after laparoscopic tubal ligation. Anesth Analg 73:112±118. Physicians' Desk Reference; 52nd edition. 1998. Montvale, NJ: Medical Economical Company, Inc. Deegan RJ. 1992. Propofol: A review of the pharmacology and applications of an intravenous anesthetic agent. Am J Med Sci 304:45±49. Shafer A, Doze VA, Shafer SL, White PF. 1988. Pharmacokinetics and pharmacodynamics of propofol infusions during general anesthesia. Anesthesiology 69:348±356. The software, STANPUMP, is available at no charge from Dr. Shafer SL, VA Medical center, Palo Alto, CA, and at his web site (http://pkpd. icon.palo-alto.med.va.gov/default.htm). Bailey JM, Shafer SL. 1991. A simple analytical solution to the three-compartment pharmacokinetic model suitable for computer-controlled infusion pumps. IEEE Trans Biomed Eng 38:522± 525. Homer DW, Lemeshow S. 1989. Applied Logistic Regression. New York: John Wiley & Sons. SAS/STAT User's Guide, Version 6, Fourth Edition. 1990. Cary, NC: SAS Institue, Inc. Finney DJ. 1978. Parallel line Assays. In: Statistical method in biological assay. London: Charles Grif®n and Company Ltd., pp. 80±82. Dutta S, Lal R, Karol MD, Cohen T, Ebert T. 2000. In¯uence of cardiac output on dexmedetomidine pharmacokinetics. J Pharm Sci 89:519±527. Kanto J, Gepts E. 1989. Pharmacokinetic implications for the clinical use of propofol. Clin Pharmacokinet 17:308±326. Servin F, Desmonts JM, Haberer JP, Cockshott ID, Plummer GF, Farinotti R. 1988. Pharmacokinetics and protein binding of propofol in patients with cirrhosis. Anesthesiology 69:887±891. Forrest FC, Tooley MA, Saunders PR, Prys-Roberts C. 1994. Propofol infusion and the suppression of consciousness: The EEG and dose requirements. Br J Anaesth 72:35±41. Dutta S, Matsumoto Y, Gothgen NY, Ebling WF. 1997. Concentration-EEG effect relationship of propofol in rats. J Pharm Sci 86:37±43.
JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 90, NO. 2, FEBRUARY 2001
Clinical Pharmacokinetics Department, Abbott Laboratories, 100 Abbott Park Road, Department 4PK, AP13A-3, Abbott Park, Illinois 60064-6104
2
Biometrics, IBAH, Lake Bluff, Illinois
3
St. Mary's Hospital Medical School, Imperial College, London, United Kingdom
4
Guys Drug Research Unit, London, United Kingdom
Received 30 September 1999; revised 25 April 2000; accepted 23 August 2000
ABSTRACT: Dexmedetomidine±propofol pharmacodynamic interaction was evaluated in nine healthy subjects in a crossover design. Dexmedetomidine/placebo was infused using a computer-controlled infusion pump (CCIP) to maintain a pseudo-steady-state plasma concentration of 0.66 0.080 or 0 ng/mL, respectively. Forty-®ve minutes after the dexmedetomidine/placebo infusion was started, propofol was infused using a second CCIP to achieve a stepwise logarithmically ascending propofol concentration (1.00 to 13.8 mg/mL) pro®le. Each propofol step lasted 10 min. Blood was sampled for plasma concentration determination, and pharmacodynamic endpoint assessments were made during the study. Propofol and dexmedetomidine/placebo infusions were terminated when three endpoints (subjects were too sedated to hold a syringe, followed by loss of eyelash re¯ex, followed by loss of motor response to electrical stimulation) were achieved sequentially. The concentration of propofol associated with 50% probability of achieving a pharmacodynamic endpoint in the absence of dexmedetomidine (EC50; placebo treatment) was 6.63 mg/mL for motor response to electrical stimulation and ranged from 1.14 to 1.98 mg/mL for the ability to hold a syringe, eyelash re¯ex, and sedation scores. The apparent EC50 values of propofol (EC50APP; concentration of propofol at which the probability of achieving a pharmacodynamic endpoint is 50% in the presence of dexmedetomidine concentrations observed in the current study; dexmedetomidine treatment) were 0.273, 0.544±0.643, and 3.89 mg/mL for the ability to hold a syringe, sedation scores, and motor response, respectively. Dexmedetomidine reduced propofol concentrations required for sedation and suppression of motor response. Therefore, the propofol dose required for sedation and induction of anesthesia may have to be reduced in the presence of dexmedetomidine. ß 2001 Wiley-Liss, Inc. and the American Pharmaceutical Association J Pharm Sci 90:172±181, 2001 Keywords: dexmedetomidine; Precedex; propofol; pharmacokinetics; pharmacody-
namics; modeling; EC50; sedation; motor response; anesthesia
INTRODUCTION Dexmedetomidine is the pharmacologically active d-isomer of medetomidine. It is a new generation Correspondence to: S. Dutta (Telephone: 847-937-8502; Fax: 847-939-5193, E-mail: [email protected]). Journal of Pharmaceutical Sciences, Vol. 90, 172±181 (2001) ß 2001 Wiley-Liss, Inc. and the American Pharmaceutical Association
172
a2-adrenoceptor agonist with high selectivity for the a2- versus a1-receptor.1,2 It has been shown to reduce the anesthetic requirements in experimental animals and human subjects, with effects similar to clonidine.3 Dexmedetomidine has greater a2- versus a1-selectivity (2000:1) than clonidine (300:1).1,2 Administration of dexmedetomidine during the perioperative period has been
JOURNAL OF PHARMACEUTICAL SCIENCE, VOL. 90, NO. 2, FEBRUARY 2001
DEXMEDETOMIDINE±PROPOFOL PHARMACODYNAMIC INTERACTION IN HEALTHY SUBJECTS
demonstrated to reduce the requirements for anesthetics (e.g., thiopental, iso¯urane), opioid analgesics (e.g., fentanyl, oxycodone), and other sedative and hypnotic drugs.3±6 This reduction in anesthetic and opioid requirements as a result of dexmedetomidine coadministration may potentially alleviate the ventilatory depression associated with these agents.3±6 Propofol is an intravenous (iv) sedative/hypnotic agent that is extensively used for the induction and maintenance of anesthesia and for sedation in the intensive care units (ICU).7±9 Because of its rapid induction and recovery characteristics and anti-emetic properties, propofol is commonly incorporated into the anesthetic regimen for outpatient surgeries.8 Because dexmedetomidine reduces anesthetic requirements, it is anticipated that dexmedetomidine may have pharmacodynamic interactions with the extensively used nonbarbiturate anesthetic agent propofol. This study evaluated, in healthy male subjects, the impact of a single, constant dexmedetomidine concentration on propofol dose requirements by assessments of the pharmacodynamic endpoints of the ability to hold a syringe, eyelash re¯ex, motor response to electrical stimulation, and sedation scores.
METHODS Study Design This was a single center, double-blind, randomized, placebo-controlled, two-period, crossover, Phase I study in 10 male subjects. The study was approved by the Guy's Hospital Research Ethics Committee. A 14-day washout period was allowed between treatment sessions. During a treatment session, the subjects received iv infusions of dexmedetomidine (Precedex, Abbott Laboratories, Abbott Park, IL) or placebo (0.9% sodium chloride solution packaged identical to dexmedetomidine and provided by Abbott Laboratories) using a computer-controlled infusion pump (CCIP). A constant dexmedetomidine concentration of 0.66 0.080 ng/mL was achieved and maintained for the duration of the infusion in the dexmedetomidine treatment group. This dexmedetomidine concentration is in the middle of the anticipated therapeutic range. Dexmedetomidine or placebo was administered through an iv catheter dedicated for drug delivery. The study site personnel operating the CCIP were blinded to the study
173
drug. The CCIP used was a Harvard Model 22 that was connected to an IBM portable computer via an RS232 serial connector. The computer program that controlled the dosing was STANPUMP,10 which drives the infusion pump to administer drugs according to prede®ned pharmacokinetic models and parameters, and desired drug plasma concentration pro®les.11 For dexmedetomidine, a two-compartment model, with drug input into and elimination from the central compartment and ®rst-order intercompartmental transfer, was used to drive the infusion pump. For propofol, a three-compartment model, with drug input into and elimination from the central compartment and ®rst-order intercompartmental transfer between the central and two peripheral com-partments, was used to drive the infusion pump. Forty-®ve minutes after the start of the dexmedetomidine or placebo infusion, propofol (Dipravan1 1%, Zeneca Pharmaceuticals, Wilmington, DE) infusion was initiated with a second CCIP in a manner designed to achieve a stepwise logarithmically ascending propofol concentration pro®le. Propofol was administered through a different iv catheter using a second CCIP and the program STANPUMP with the intent of achieving up to 11 pseudo-steady-state plasma concentrations ranging from 1.00 to 13.8 mg/mL. Each propofol concentration step lasted 10 min. During each propofol concentration step, blood samples for determination of drug concentrations were drawn, and the pharmacodynamic assessments were made. Propofol and dexmedetomidine/placebo infusions were terminated when all three of the following pharmacodynamic endpoints were achieved sequentially; (a) the subjects were too sedated to hold a syringe, followed by (b) loss of eyelash re¯ex, followed by (c) lack of motor response to a 10-s, 100-Hz electrical stimulation. Response to electrical stimulation was de®ned as gross purposeful movement of any part of the body except the stimulated limb. Blood samples (5 mL) were collected for determination of dexmedetomidine and propofol plasma concentrations. The blood samples were collected prior to the dexmedetomidine infusion at 0 h or baseline, prior to the propofol infusion, 10 min after the start of every propofol step, at emergence from anesthesia, and at 1, 2, 4, and 8 h after termination of dexmedetomidine/placebo and propofol infusions. All blood samples were assayed for propofol plasma concentrations. All blood samples, except for every odd-numbered propofol step, JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 90, NO. 2, FEBRUARY 2001
174
DUTTA ET AL.
were assayed for dexmedetomidine plasma concentrations. Prospective power calculations based on response to tetanic stimulation were used to calculate the number of subjects for the study. The use of 10 subjects in a two-way crossover study was expected to enable detection of a 25% decrease in the propofol concentration required for loss of response to tetanic stimulation at a 0.05 level of signi®cance with statistical power of > 80%. Subjects Subjects selected were healthy nonsmoking males between the ages of 18 and 45 years and within 15% of the upper and lower boundaries of weight set forth in the Metropolitan Life Insurance Table. In the screening period prior to dosing, candidates received a full explanation of the study, provided written informed consent, and underwent pre-study procedures. Subjects were judged to be in good health based on the results of medical history, physical examination, vital signs, electrocardiogram, and routine clinical laboratory evaluations. Subjects had not taken any prescription medication within 4 weeks or overthe-counter medication within 2 weeks prior to study drug administration. Analytical Methods Dexmedetomidine concentrations were determined by gas chromatography/mass spectroscopy (GC/MS). Brie¯y, the extraction procedure involved the addition of a 75-mL aliquot internal standard (4-[1-(2-methylphenyl)ethyl]-1H-imidazole) to 1.0 mL of sample. The samples were then basi®ed with 0.5 mL of a 1 M sodium carbonate solution and extracted into 7.0 mL of hexane while simultaneously undergoing derivatization by the addition of 0.5 mL of 0.5% penta¯uorobenzoyl chloride solution. The organic layer was transferred into a clean, dry tube and dried under nitrogen. The samples were then reconstituted in 0.2 mL of toluene, and 3 mL was injected onto the GC/MS system. The lower limit of quanti®cation was 10 pg/mL and the within-day and interday coef®cients of variation (CVs) were < 15%. Propofol concentrations were determined by high-performance liquid chromatography (HPLC) with ¯uorescence detection (HPLC assay was performed at Pheonix International Life Sciences Inc., St-Laurent, Canada). Brie¯y, propofol and JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 90, NO. 2, FEBRUARY 2001
its internal standard, thymol, were extracted with pentane/dichloromethane and evaporated to dryness. The residue was reconstituted with mobile phase and injected onto the HPLC system. Separation was obtained using a reversed-phase column under isocratic conditions, and the ef¯uent was continuously monitored by ¯uorescence detection. The lower limit of quanti®cation was 0.061 mg/mL and the within-day and interday CVs were < 5%. Pharmacodynamic Measurements The ability to hold a syringe, the presence of eyelash re¯ex, and the presence of motor response were measured on a binary scale (Yes/No). Only the ability to hold a syringe was measured immediately prior to starting the propofol infusion. During each propofol infusion step, the ability to hold a syringe, the presence of eyelash re¯ex, and the presence of motor response to electrical stimulation were measured 8 min after the start of each step. Sedation was measured on a six point Ramsay scale (1Ðpatient anxious, agitated, or restless; 2Ðpatient cooperative, oriented, and tranquil; 3Ðpatient responds to commands; 4Ðpatient asleep but with brisk response to light labellar tap or loud auditory stimulus; 5Ðpatient asleep, with sluggish response to light labellar tap or loud auditory stimulus; 6Ðpatient asleep, no response). Ramsay sedation scores were recorded prior to dexmedetomidine/placebo infusion, immediately prior to propofol infusion, 5 min after the start of each propofol infusion step, every 5 min following the end of propofol and dexmedetomidine/placebo infusions until emergence, and every 15 min for 4 h after the termination of both infusions. Pharmacokinetic and Pharmacodynamic Analyses Pharmacokinetic Analyses Noncompartmental analyses were performed to obtain the following pharmacokinetic parameters. The maximum plasma concentration (Cmax), clearance (CL), the terminal elimination half-life (t1/2), and the volume of distribution at steady state (VSS). Treatment effects (placebo versus dexmedetomidine) on propofol pharmacokinetic parameters were evaluated ( p < 0.05) using analysis of variance. The analysis of variance model had sequence, subject within sequence, treat-
DEXMEDETOMIDINE±PROPOFOL PHARMACODYNAMIC INTERACTION IN HEALTHY SUBJECTS
ment, and the treatment by sequence interaction as independent factors. Pharmacodynamic Analyses The results for the tests for the ability to hold a syringe, the presence of eyelash re¯ex, and the presence of motor response to electrical stimulation were expressed on a binary scale (Yes/No). The effect of propofol and dexmedetomidine on the probability of a Ramsay sedation score greater than a speci®ed value was modeled using logistic regression. Because the Ramsay sedation score may be one of six values (1, 2, 3, 4, 5, or 6), there are ®ve possible binary logistic regressions; they are the probabilities of Ramsay sedation score > 1, > 2, > 3, > 4, or > 5. The orientation of probability of alertness (1 ÿ probability of sedation) was selected, rather than probability of sedation, for ease of graphic interpretation. When probabilities versus concentration are plotted in this way, decreases in the plotted curves indicate decreased alertness. Logistic regression analyses with repeated measurements12,13 ( p< 0.05) were performed on these binary data to determine the effect of dexmedetomidine and propofol and their interaction on the probability of alertness for each pharmacodynamic endpoint. The mathematical theory,12 derivations, and equations used for logistic regression can be found in the Appendix. The logistic function12 (eq. A1, Appendix) gives two curves for probability of alertness; one for the placebo and one for the dexmedetomidine treatment for each pharmacodynamic endpoint. The true EC50 of propofol is the concentration of propofol associated with 50% probability of achieving a pharmacodynamic endpoint in the absence of dexmedetomidine (placebo treatment). The apparent EC50 of propofol (EC50APP) is the concentration of propofol at which the probability of achieving a pharmacodynamic endpoint is 50% in the presence of dexmedetomidine concentrations observed in the current study (dexmedetomidine treatment). Propofol EC50 and EC50APP were calculated from the coef®cients of the logistic regression function (Appendix). The 95% con®dence intervals for EC50 and EC50APP were calculated by Fieller's method.14 Treatment effects (placebo versus dexmedetomidine) on recovery time (time from termination of infusions to emergence from anesthesia) and the propofol concentration at the end of the dexmedetomidine and propofol infusions were
175
evaluated ( p < 0.05) by analysis of variance. The analysis of variance model had sequence, subject within sequence, treatment, and the treatment by sequence interaction as independent factors. Dexmedetomidine plasma concentrations at 10 min after the start of each odd-numbered step were estimated by linear interpolation if the next even-numbered propofol infusion step took place. If the odd-numbered step was the last propofol infusion step, the dexmedetomidine concentration was set equal to the previous measurement. Both dexmedetomidine and propofol concentrations at the time when Ramsay sedation scores were recorded were estimated by linear interpolation. Only the data during the dexmedetomidine/ placebo and propofol infusions were used for the logistic regression analysis. Time to recovery was de®ned as the time between the end of the dexmedetomidine/placebo and propofol infusions and the time of emergence. The logistic regression analyses were performed with Procedure GENMOD and analyses of variance were performed with Procedure GLM of the SAS software.13 A signi®cance level of 0.05 was used for all tests.
RESULTS AND DISCUSSION The mean age, weight, and height of the nine subjects who completed the study were 24 years (range, 18±28 years), 73.7 kg (range, 61.0±87.3 kg), and 175 cm (range, 168±183 cm), respectively. One subject did not complete the study for personal reasons. The partial data (only dexmedetomidine treatment) from this subject were not included in the pharmacokinetic, pharmacodynamic, and statistical analyses. Pharmacokinetics Figure 1 presents the mean dexmedetomidine and propofol plasma concentration±time pro®les. The mean standard deviation (SD) pseudosteady-state concentration of dexmedetomidine during the computer-controlled infusion was 0.66 0.080 ng/mL (Table 1). The mean dexmedetomidine concentration at emergence was 0.470 ng/mL. The mean propofol concentrations at emergence were 2.32 0.694 and 0.803 0.226 mg/mL for the placebo and dexmedetomidine treatments, respectively. For the placebo treatment, four subjects completed Step 9, with an average concentration of 7.45 1.38 mg/mL. JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 90, NO. 2, FEBRUARY 2001
176
DUTTA ET AL.
Figure 1. Mean dexmedetomidine (top panels) and propofol (bottom panels) plasma concentrations during and after termination of the infusions. Error bars indicate standard deviations. Propofol pseudo-steady-state concentrations were increased logarithmically every 10 min.
Two of these subjects continued on to Step 10, where they had an average propofol concentration of 6.38 1.01 mg/mL. For the dexmedetomidine treatment, the mean propofol concentrations achieved in the last two steps, Steps 7 and 8, were 4.05 0.220 (n 3) and 4.94 (n 1) mg/mL, respectively. Dexmedetomidine and propofol doses, infusion durations, and pharmacokinetic parameters (mean SD) are summarized in the Table 1. Dexmedetomidine pharmacokinetic parameters were consistent with values obtained from other studies.15 Propofol pharmacokinetic parameters were not signi®cantly in¯uenced by the presence of dexmedetomidine in this study, and these JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 90, NO. 2, FEBRUARY 2001
parameter values were consistent with values reported in the literature.7,9,16,17 Propofol VSS was probably underestimated because blood samples were collected for only 8 h after termination of infusion. Studies with prolonged infusions and extended sampling have shown larger VSS and longer t1/2 values associated with slowly equilibrating tissue compartments.7 Pharmacodynamics Propofol Dose and Concentration at the End of Infusion Dexmedetomidine decreased the mean propofol dose required to reach the termination of infusion
DEXMEDETOMIDINE±PROPOFOL PHARMACODYNAMIC INTERACTION IN HEALTHY SUBJECTS
Table 1.
177
Dexmedetomidine and Propofol Pharmacokinetic Parameters (Mean SD) Propofol
Parameter Dose Infusion Duration Cmax Ccss Tmax te1=2 MRT CL Vss Vss
Units
Dexmedetomidine
Placebo Treatment
Dexmedetomidine Treatment
mga mgb h ng/mLa , mg/mLb ng/mL h h h L/h L L/kg
106 14 1:69 0:29 0:72 0:060 0:66 0:080 1:10 0:47 2:27 0:456 3:14 0:682 41:3 8:15 126 18:1 1:71 0:193
994 404 1:34 0:23 6:00 2:00 Ðd 2:08 0:21 2:33 0:700 2:45 0:388 119 22:4 291 69:9 3:93 0:820
492 259 0:93 0:29 3:97 2:34 Ðd 1:70 0:30 2:27 0:815 2:18 0:744 124 46:8 264 113 3:53 1:36
a
Dexmedetomidine. Propofol. c Pseudo-steady-state concentration. d Plasma concentrations were increased every 10 min. e Presented as the harmonic mean and pseudo standard deviation; the terminal elimination rate constant was tested for treatment effect. b
criteria by approximately one-half (Table 1). The mean propofol infusion durations were also decreased in presence of dexmedetomidine (Table 1). Therefore, propofol doses for induction and maintenance of anesthesia may have to be reduced in presence of dexmedetomidine. The effect of dexmedetomidine on propofol plasma concentration at the end of the infusion was evaluated using analysis of variance. Mean propofol plasma concentration at the end of the infusion for the placebo and dexmedetomidine treatments were 5.97 2.00 and 3.97 2.34 mg/ mL, respectively. These concentrations were statistically signi®cantly different. Dexmedetomidine and propofol infusions were terminated when a motor response to electrical stimulation could not be elicited in a subject. Based on this termination of dosing criterion, it can be concluded that dexmedetomidine reduced the propofol concentration required to suppress motor response to electrical stimulation.
concentrations of 0.66 0.080 ng/mL, is unlikely to alter emergence from anesthesia when coadministered with propofol. Logistic Regression Analyses The ability to hold a syringe, presence of eyelash re¯ex, presence of motor response to electrical stimulation, and the binary variables based on the Ramsay sedation scores were analyzed using logistic regression models with dexmedetomidine treatment groups, propofol concentrations, and their interaction as independent variables. Figure 2 shows the rank order of steady-state propofol concentrations required to elicit the different pharmacodynamic endpoints in the absence (top panel) and presence of dexmedetomidine (bottom panel). Comparison of these two plots reveal that, in general, the propofol concentration required to achieve a given pharmacodynamic endpoint is reduced in the presence of dexmedetomidine.
Recovery Time
Sedation
The effect of dexmedetomidine on the recovery time was evaluated using analysis of variance. The mean recovery times were 16 10 and 15 7.7 min for the placebo and dexmedetomidine treatments, respectively. These recovery times were not statistically signi®cantly different. Therefore, the presence of dexmedetomidine at
Propofol EC50 and EC50APP values for each pharmacodynamic endpoint are reported in Table 2. Propofol EC50 (Table 2 and Figure 2, top panel) for the ability to hold a syringe, eyelash re¯ex, and Ramsay sedation scores > 2, 3, 4, and 5 ranged from 1.14 to 1.98 mg/mL. Propofol EC50APP (Table 2 and Figure 2, bottom panel) could not be JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 90, NO. 2, FEBRUARY 2001
178
DUTTA ET AL.
Figure 2. The probability of different pharmacodynamic endpoints evaluated in this study versus propofol concentration, for the placebo (top panel) and dexmedetomidine (bottom panel) treatments. These plots show the rank order of steady-state propofol concentrations eliciting the different pharmacodynamic endpoints in the absence (top panel) and presence of dexmedetomidine (bottom panel).
determined for eyelash re¯ex and Ramsay sedation score > 2. Propofol EC50APP for the ability to hold a syringe and Ramsay sedation scores > 3, 4, and 5 were 65±80% lower than EC50. The presence of dexmedetomidine had a statistically signi®cant effect (coef®cient a3 was signi®cantly different from zero, Appendix) on the propofol concentrations required to suppress these pharmacodynamic measures of sedation. JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 90, NO. 2, FEBRUARY 2001
Signi®cant Interactions. In these analyses, a value for the a4 coef®cient (Appendix) that is statistically signi®cantly different from zero represents a pharmacodynamic interaction. Speci®cally, a positive a4 value indicates antagonism and a negative a4 value indicates synergism. Statistical analyses indicated that there was a signi®cant interaction between dexmedetomidine and propofol for the pharmacodynamic endpoints
DEXMEDETOMIDINE±PROPOFOL PHARMACODYNAMIC INTERACTION IN HEALTHY SUBJECTS
179
Table 2. Dexmedetomidine±Propofol Pharmacodynamic Interaction Variable Ability to hold syringe Eyelash re¯ex Motor response Ramsay >2 Ramsay >3 Ramsay >4 Ramsay >5
Propofol EC50 (mg/mL)a 1.37 1.86 6.63 1.14 1.54 1.70 1.98
(1.18±1.60) (1.61±2.21) (5.50±9.32) (0.922±1.33) (1.35±1.80) (1.50±2.00) (1.70±2.31)
Propofol EC50APP (mg/mL)b 0.273 ( ÿ 0.128±0.625) Ðc 3.89 (3.17±6.69)d Ðc 0.544 (0.260±0.766) 0.610 (0.325±0.814) 0.643 (0.358±0.879)
All Ramsay scores were > 1; parameters reported as mean (95% con®dence interval). a Propofol concentration associated with 50% probability of achieving a pharmacodynamic endpoint. b Propofol concentration associated with 50% probability of achieving a pharmacodynamic endpoint with dexmedetomidine coadministration. c Probability of Ramsay sedation score > 2 and eyelash relfex were 0.4. d Dexmedetomidine did not have a statistically signi®cant effect on slope of motor response versus propofol concentration curve.
of eyelash re¯ex and Ramsay sedation score > 2. In the presence of pseudo-steady-state dexmedetomidine plasma concentration of 0.66 0.080 ng/ mL, the subjects were markedly sedated even when no propofol was present, with probabilities of 0.42 and 0.14 for eyelash re¯ex and Ramsay sedation score > 2, respectively. Therefore, an EC50APP for propofol could not be determined for eyelash re¯ex and Ramsay sedation score > 2 (Figure 2). For these pharmacodynamic endpoints, antagonism was observed between dexmedetomidine and propofol. For other sedation parameters (i.e., the ability to hold a syringe and Ramsay sedation scores > 3, 4, and 5), no antagonism or synergism was observed between dexmedetomidine and propofol. It should be noted that large variability of the interaction coef®cient (a4) for some of the metrics may have limited our ability to detect statistically signi®cant synergism or antagonism. Motor Response Propofol EC50 for motor response to electrical stimulation was 6.63 mg/mL. Propofol EC50APP for motor response to electrical stimulation was 3.89 mg/mL, 40% lower than EC50. However, logistic regression analysis failed to reveal that the effect of dexmedetomidine was statistically signi®cant (coef®cient a3 was not signi®cantly different from zero, Appendix). High variability and limited sample size may be responsible for lack of signi®cant effect of dexmedetomidine on the slope of the logistic regression curves for motor response. Interaction (i.e., antagonism or synergism; coef®cient a4, Appendix), between
dexmedetomidine and propofol for motor response was not statistically signi®cant. Although dexmedetomidine did not have a statistically signi®cant effect in the logistic regression analyses, there is evidence from the two analyses performed that indicate dexmedetomidine reduces propofol concentrations required to suppress motor response to electrical stimulation. First, propofol EC50APP was 40% lower than EC50, and second, based on the termination of infusion criterion of no motor response to electrical stimulation, statistically signi®cantly lower propofol concentrations were achieved at the end of the infusions for the dexmedetomidine treatment compared with the placebo treatment. The relationships between propofol plasma concentrations and different pharmacodynamic measures of sedation evaluated in the current study in the absence of dexmedetomidine are consistent with our understanding of propofol pharmacodynamics. Propofol plasma concentrations had a statistically signi®cant effect on all the sedation parameters examined. Propofol EC50 values for all the sedation parameters (except motor response) evaluated in the current study ranged from 1.14 to 1.98 mg/mL. These values are in excellent agreement with literature data that suggest that propofol concentrations of 1±2 mg/ mL are required for maintenance of sleep, 24 mg/ mL are needed for maintenance of anesthesia, and 7±10 mg/mL are necessary for surgical anesthesia in humans. Others have found that propofol concentrations from 1 to 6 mg/mL covers a range of states from conscious sedation to full anesthesia.16,18 JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 90, NO. 2, FEBRUARY 2001
180
DUTTA ET AL.
The propofol concentration required for suppression of motor response determined in the current study also agrees with concentrations required for achieving similar pharmacodynamic endpoints in humans and rats. Propofol blood concentrations of 6±10 mg/mL are associated with onset of unconsciousness, electroencephalographic (EEG) burst suppression, and surgical anesthesia in humans.16,18 In rats, plasma concentrations of > 7 mg/mL are required to produce EEG burst suppression.19 These values are similar to propofol EC50 for motor response (6.63 mg/mL) determined in the current study.
antagonism or additivity). Using the orientation of probability of alertness, when a4 is negative, there is synergism; when it is positive, there is antagonism; and when a4 is zero, there is simple additivity. A statistically signi®cant interaction (i.e., synergism or antagonism) is said to occur when a4 is signi®cantly different ( p< 0.05) from zero. For simplicity let us de®ne P P(1|Pr,D) and x a1 a2Pr a3D a4PrD. Then eq. A1 becomes: P1ÿ x ln
CONCLUSIONS Dexmedetomidine did not have a signi®cant in¯uence on propofol pharmacokinetics in this study. Dexmedetomidine (0.66 0.080 ng/mL) reduced propofol concentrations required for sedation by 65±80% and for suppression of motor response to electrical stimulation by 40%. Therefore, the propofol dose administered for sedation and induction of anesthesia may have to be reduced in the presence of dexmedetomidine.
APPENDIX Mathematical Theory and Equations for Logistic Regression Equation A1 shows a logistic function for modeling the pharmacodynamic interaction between two drugs: P 1jPr; D 1 ÿ
Pr
ln
ÿ
P 1ÿP
1 e a1 a2 Pra3 Da4 PrD 1 e a1 a2 Pra3 Da4 PrD
A1
where P(1|Pr,D) is the probability of response for a given pharmacodynamic endpoint, (e.g., ability to hold syringe, eyelash re¯ex, etc.) at a given propofol concentration, Pr and a given dexmedetomidine concentration D, and where both Pr and D are continuous variables. The terms a1, a2, a3, and a4 are coef®cients obtained by logistic regression. Coef®cient a1 is referred to as the intercept coef®cient. Terms a2 and a3 are the concentration or slope coef®cients for propofol and dexmedetomidine, respectively. The term a4 describes the pharmacodynamic interaction (synergism or JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 90, NO. 2, FEBRUARY 2001
A2
P 1ÿP
a1 a2 Pr a3 D a4 Pr D ln
A3 P A4 1ÿP
ÿ a1 a3 D a2 a4 D
A5
The intersection of a horizontal plane with the three-dimensional response pro®le de®ned by the logistic function (eq. A1) represents a two-dimensional isobologram. Equation A5 (a two-dimensional isobologram) gives the solution of eq. A1 for one drug concentration given a selected probability, a concentration of the second drug, and the function parameters. In the case of the 50% isobologram, where P 0.5, eq. A5 simpli®es to: EC50APP-Pr Pr
1 e a1 a2 Pra3 Da4 PrD
1 1 ex
ÿ a1 a3 D a2 a4 D
A6
Equation A6 permits computation of the apparent EC50 (i.e., propofol concentration required to elicit a 50% probability of response, given any particular concentration of dexmedetomidine). Conversely, this expression may be rearranged to permit the computation of the dexmedetomidine concentration required for the same endpoint given a concentration of propofol. That form of the expression appears below. EC50APP-D D
ÿ a1 a2 Pr a3 a4 Pr
A7
The negative of the ratio of the intercept coef®cient to the appropriate concentration coef®cient
DEXMEDETOMIDINE±PROPOFOL PHARMACODYNAMIC INTERACTION IN HEALTHY SUBJECTS
gives the concentration at which a 50% probability of response occurs when only one drug is present (i.e., Pr or D is equal to zero). Thus, the true or intrinsic EC50 values may be computed by:
4.
ÿa1 a2 ÿa1 EC50-D a3
5.
EC50-Pr
A8 A9
where EC50-Pr and EC50-D are the true EC50 values of propofol and dexmedetomidine, respectively. In the current study, the pharmacodynamic responses to different propofol plasma concentrations were measured in the absence of dexmedetomidine (placebo) and presence of dexmedetomidine. Because the pharmacodynamic responses were measured only at one pseudosteady-state dexmedetomidine concentration, the variable D was expressed as a categorical variable that had a value of 0 (dexmedetomidine not present) or 1 (dexmedetomidine present). Therefore, P(1|Pr,D) represents the probability of being alert at a given propofol concentration (Pr) in the presence or absence of dexmedetomidine (D). Computation of the true EC50 of propofol under these conditions is unaffected by the reparameterization of the dexmedetomidine variable ``D''. Thus, eq. A8 can be used. With dexmedetomidine variable ``D'' reparameterized as either 0 or 1, the equation for calculation of propofol apparent EC50, eq. A6, becomes: EC50APP EC50APP-Pr
ÿ a1 a3 a2 a4
A10
In the current study, dexmedetomidine was presented only at one pseudo-steady-state concentration and was represented as a categorical variable (present or absent). Therefore, EC50 and EC50APP for dexmedetomidine were not calculated.
6.
7. 8. 9.
10.
11.
12. 13. 14. 15. 16. 17.
REFERENCES 1. Virtanen R. 1989. Pharmacological pro®les of medetomidine and its antagonist, atipamezole. Acta Vet Scand Suppl. 85:29±37. 2. Tong C, Eisenach JC. 1994. Alpha-2 adrenergic agonists. New Drugs Anesth: Part II. 12:49±63. 3. Aantaa R, Kanto J, Scheinin M, Kallio A, Scheinin H. 1990. Dexmedetomidine, an alpha-2-adrenocep-
18.
19.
181
tor agonist, reduces anesthetic requirements for patients undergoing minor gynecologic surgery. Anesthesiology 73:230±235. Jaakola ML, Ali-Melkkila T, Kanto J, Kallio A, Scheinin H, Scheinin M. 1992. Dexmedetomidine reduces intraocular pressure, intubation responses and anaesthetic requirements in patients undergoing opthalmic surgery. Br J Anaesth 68:570±575. Jaakola ML. 1994. Dexmedetomidine premedication before intravenous regional anesthesia in minor outpatient hand surgery. J Clin Anesth 6: 204±211. Aho MS, Erkola OA, Scheinin H, Lehtinen AM, Korttila KT. 1991. Effect of intravenously administered dexmedetomidine on pain after laparoscopic tubal ligation. Anesth Analg 73:112±118. Physicians' Desk Reference; 52nd edition. 1998. Montvale, NJ: Medical Economical Company, Inc. Deegan RJ. 1992. Propofol: A review of the pharmacology and applications of an intravenous anesthetic agent. Am J Med Sci 304:45±49. Shafer A, Doze VA, Shafer SL, White PF. 1988. Pharmacokinetics and pharmacodynamics of propofol infusions during general anesthesia. Anesthesiology 69:348±356. The software, STANPUMP, is available at no charge from Dr. Shafer SL, VA Medical center, Palo Alto, CA, and at his web site (http://pkpd. icon.palo-alto.med.va.gov/default.htm). Bailey JM, Shafer SL. 1991. A simple analytical solution to the three-compartment pharmacokinetic model suitable for computer-controlled infusion pumps. IEEE Trans Biomed Eng 38:522± 525. Homer DW, Lemeshow S. 1989. Applied Logistic Regression. New York: John Wiley & Sons. SAS/STAT User's Guide, Version 6, Fourth Edition. 1990. Cary, NC: SAS Institue, Inc. Finney DJ. 1978. Parallel line Assays. In: Statistical method in biological assay. London: Charles Grif®n and Company Ltd., pp. 80±82. Dutta S, Lal R, Karol MD, Cohen T, Ebert T. 2000. In¯uence of cardiac output on dexmedetomidine pharmacokinetics. J Pharm Sci 89:519±527. Kanto J, Gepts E. 1989. Pharmacokinetic implications for the clinical use of propofol. Clin Pharmacokinet 17:308±326. Servin F, Desmonts JM, Haberer JP, Cockshott ID, Plummer GF, Farinotti R. 1988. Pharmacokinetics and protein binding of propofol in patients with cirrhosis. Anesthesiology 69:887±891. Forrest FC, Tooley MA, Saunders PR, Prys-Roberts C. 1994. Propofol infusion and the suppression of consciousness: The EEG and dose requirements. Br J Anaesth 72:35±41. Dutta S, Matsumoto Y, Gothgen NY, Ebling WF. 1997. Concentration-EEG effect relationship of propofol in rats. J Pharm Sci 86:37±43.
JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 90, NO. 2, FEBRUARY 2001