Lack of effect of ondansetron on the pharmacokinetics and analgesic effects of morphine and metabolites after single-dose morphine administration in healthy volunteers

Abstract

Aims

The purpose of this investigation was to study the influence of ondansetron on the single-dose pharmacokinetics and the analgesic effects elicited by morphine and the 3- and 6-glucuronide metabolites of morphine in healthy volunteers.

Methods

This was a randomized, double-blind, placebo-controlled, two-way crossover study in which six male and six female subjects were administered a single 10 mg intravenous dose of morphine sulphate, followed 30 min later by a single 16 mg intravenous dose of ondansetron hydrochloride or placebo. Serum and urine concentrations of morphine, morphine-3-glucuronide (M3G) and morphine-6-glucuronide (M6G) samples were quantified over 48 h using high performance liquid chromatography with detection by mass spectrometry. Analgesia was assessed in the volunteers with a contact thermode device to provide a thermal pain stimulus. Four analgesic response variables were measured including thermal pain threshold, thermal pain tolerance, temporal summation of pain and mood state.

Results

The two treatments appeared to be equivalent based on the 90% confidence intervals (0.6, 1.67) of the least squares means ratio. All least squares means ratio confidence intervals for each parameter, for each analyte fell within the specified range, demonstrating a lack of an interaction.

Conclusions

The results of this study suggest that administration of ondansetron (16 mg i.v.) does not alter the pharmacokinetics of morphine and its 3- or 6-glucuronide metabolites to a clinically significant extent, nor does it affect the overall analgesic response to morphine as measured by the contact thermode system.

Introduction

Ondansetron (OND) is a potent and selective 5-HT 3 receptor antagonist used as an antiemetic. The metabolism of OND is complex, involving both Phase I and Phase II pathways. Several CYP450 isoforms contribute to the metabolism of OND including CYP1A1, CYP1A2, CYP2D6, and the CYP3A family [1]. In terms of overall metabolism, no single enzyme dominates. Phase I reactions result in the N-demethylation and hydroxylation of the indole nucleus. Both glucuronide and sulphate conjugates of the oxidative metabolites are also formed [1].

Like OND, morphine undergoes glucuronidation resulting in the formation of two metabolites, morphine-3-glucuronide (M3G) and morphine-6-glucuronide (M6G). Approximately 50% of an orally administered dose is biotransformed to the primary metabolite, M3G, while formation of M6G accounts for approximately 5% of the dose [2]. Studies have shown that M3G does not demonstrate analgesic activity [3]. In contrast, M6G produces analgesia and is as potent or more potent than morphine. Therefore, alterations in M6G disposition may be of clinical significance [1, 3, 4].

In the treatment of postoperative nausea and vomiting, OND is frequently given after morphine is administered for pain relief. Because glucuronidation is a major elimination pathway for both drugs, competition for glucuronidation may occur between these agents when administered concomitantly [5, 6]. Alterations in the pharmacokinetics of morphine may lead to changes in the analgesic response elicited by morphine. To date, morphine pharmacokinetics and pharmacodynamics have not been evaluated when morphine administration is followed by OND.

The objectives of the present study were to determine whether the pharmacokinetics of morphine and its glucuronide conjugates were altered following a single, intravenous dose of OND, and to evaluate the analgesic effects of morphine in response to experimentally induced pain, in the presence and absence of ondansetron. Specifically, a contact thermode device, previously employed to measure the analgesic effects of morphine [7], provided the thermal stimuli. Four analgesic response measures were examined including thermal threshold, thermal tolerance, temporal summation, and mood.

Part of this study was presented at the American College of Clinical Pharmacy Spring Research Forum, Palm Springs, California, USA, April 1998.

Methods

The study protocol was approved by the Committee on the Protection of the Rights of Human Subjects at University of North Carolina Hospitals, and written informed consent was obtained from each subject.

Subjects

Twelve normal, healthy volunteers participated in this study. Subjects were eligible for inclusion if they were healthy males or nonpregnant, nonlactating females between 18 and 45 years of age, inclusive, weighed between 55 and 85 kg, inclusive, and were free from significant physical or psychiatric disorders as determined by history, physical examination and laboratory screens. Subjects with any clinically significant haematological endocrine, cardiovascular, hepatic, renal, gastrointestinal and/or pulmonary disorders were excluded. Subjects were excluded if they had a history of hypersensitivity to opioids or OND, a history of alcohol or drug abuse, if they had used tobacco or marijuana within the previous year, or if they had donated blood within the previous 30 days. In addition, individuals chronically taking prescription or nonprescription medication were excluded. Females maintained on an oral contraceptive or a progesterone regimen were excluded.

Demographics

Six male and six female subjects were enrolled and completed this study. The mean age (± s.d.) was 25.8 ± 8.8 years, and the mean weight was 72.9 ± 8.1 kg. Two subjects were African-American, one was Hispanic and nine were Caucasian.

Study design

This was a single centre (General Clinical Research Center [GCRC], University of North Carolina at Chapel Hill, USA) randomized, double-blind, placebo-controlled, two-way crossover study. The following restrictions applied: no methylxanthine-containing foods or beverages for 24 h before and 48 h after drug administration; no alcohol for 24 h before and 48 h after drug administration. Subjects were admitted to the GCRC on the evening prior to study drug administration. On the study day, after an overnight fast, subjects received a single 10 mg intravenous dose of morphine sulphate (Wyeth-Ayerst Laboratories, Philadelphia, Pennsylvania, USA) over 5 min, followed 30 min later by either a single 16 mg intravenous dose of OND hydrochloride (Glaxo Wellcome, Inc., Research Triangle Park, North Carolina, USA) or 0.9% sodium chloride (NS; Abbott Laboratories, North Chicago, Illinois, USA) administered over 5 min. The two treatments were administered in a random order separated by a 3–21 day washout period.

Pharmacokinetic sampling

Blood samples were collected through a venous catheter placed in the forearm (kept patent with normal saline solution) or by direct venepuncture. In order to clear the dead space from the i.v. line, 2 ml of blood were withdrawn and discarded. A 5 ml blood sample was obtained before and at 5, 10, 20, 30, 45, 60 and 90 min and 2, 3, 4, 6, 8, 12, 24, 36 and 48 h after the start of the morphine administration. Urine collections were obtained before and from 0 to 2, 2–6, 6–12, 12–24, 24–36 and 36–48 h after the start of the morphine administration. Serum and urine samples were stored below −20 °C until analysed.

Pharmacodynamic measurements

Analgesia was assessed following application of a thermogenic stimulus delivered by a contact thermode system based upon previously reported methods [7]. All methods were performed on each subject at least once on the evening prior to the study.

Study day measurements were obtained before and at 20, 30, 60 and 90 min and 2, 3, 4, 6, and 8 h after the start of the morphine administration.

Thermal pain threshold and tolerance

Thermal pain threshold and tolerance were measured by an ascending method of limits using a 1 cm diameter copper contact thermode located at the thermode tip. An initial (adapting) temperature of 38 °C was held constant for 10 s. The temperature then increased to 41.5 °C, and subsequently increased at a controlled rate of 0.5 °C/5 s until an intolerable temperature was perceived or a maximum temperature of 48 °C was reached. Cooling was achieved by circulating Freon on the inside of the thermode.

The thermode was applied to the volar forearm on premarked spots approximately 2 cm apart, indicating placement. To avoid repeated application in a particular area of the forearm, a grid was drawn on the subject's forearm indicating 20 distinct testing spots. The probe was placed on one spot, located near the wrist for the first measurement, and was moved up the forearm in a linear fashion to collect three to five measurements. After all spots on the grid were used, they were re-used in the same order. If the skin surrounding a spot was erythematous after the measurements, the area was not retested in order to prevent injury.

To determine thermal pain threshold, subjects were instructed to say ‘painful’ when they first perceived the thermal stimulation as painful. Tolerance was determined by instructing the subjects to say ‘stop’ when they were no longer able to tolerate the thermal stimulation. Temperatures corresponding to thermal pain threshold and tolerance were recorded.

Temporal summation of pain

To examine the temporal summation of heat pain [7, 8], a total of 10 heat pulses at a temperature of 53 °C were applied to the ventral surface of the right hand at the base of the index finger. Each heat pulse was 1.5 s in duration and was delivered from a 40 °C base temperature with an intertrial interval of 1.5 s. In effect, this produced a transient 53 °C heat pulse with a peak-to-peak interpulse interval of 3 s. Subjects were instructed to verbally rate the intensity of each thermal pulse using a 0–100 numerical scale with ‘0’ representing ‘no sensation’, ‘20’ representing ‘just painful’ and ‘100’ representing the ‘most pain I can bear’. Subjects were informed that the procedure would be terminated when they reported a value of ‘100’ or when 10 trials had elapsed. For subjects who terminated the procedure prior to the completion of 10 trials, a value of 100 was assigned to the missing trials for statistical analysis.

The thermode was calibrated prior to use and after the 4 h time point. Baseline thermal threshold, thermal tolerance and temporal summation of pain determinations were performed at each subject's first visit prior to the predose determination.

Mood assessment

Subjects were asked to mark a visual analogue scale (VAS) depicting his or her mood state at 15–30 min predose and immediately prior to each pharmacodynamic evaluation time point. The VAS was 10 cm in length with the word ‘Unpleasant’ to the far left and ‘Pleasant’ to the far right. Subjects were instructed to mark the spot on the scale which best depicted their mood at that time.

Drug analysis

Morphine, M3G and M6G, were analysed from plasma and urine using a validated, proprietary assay developed at PPD-Development (Richmond, VA). Morphine, M3G and M6G along with their corresponding stable isotope-labelled internal standards, morphine-D₃, M3G-D₃ and M6G-D₃, were isolated from a 1 ml aliquot of heparinized human plasma (for urine samples, 0.25 ml of urine was brought to a volume of 1.0 ml with blank human plasma) by solid-phase extraction onto a SPEC SCX disc cartridge. The glucuronides were not retained on the disc and were collected as a separate fraction. The cartridge was washed with dilute acetic acid, followed by methanol; morphine and its internal standard were eluted with 1% acetic acid in methanol. Following evaporation under nitrogen, the morphine extract was reconstituted in 0.15 ml of a dilute acetic acid solution. The morphine extract was injected using a CTC A200SE autosampler onto a Hewlett-Packard 1100 series h.p.l.c. equipped with a Perkin Elmer R Activity C8 column. The retention time for morphine was approximately 1 min using an isocratic mobile phase composed of 15% acetonitrile/85% 20 mm ammonium acetate flowing at a rate of 1 ml min⁻¹. The column effluent was directed into a Finnigan MAT SSQ7000 LC/MS system and the analytes were detected using positive ion atmospheric pressure chemical ionization (heated nebulizer) mode with selected ion monitoring. The glucuronide fraction was further purified by solid-phase extraction onto a 200 mg C2 silica cartridge. The cartridge was washed with 20 mm ammonium acetate (pH 9) and the glucuronides and their internal standards were eluted with 20% acetonitrile/80% 10 mm ammonium acetate (pH 8). Following evaporation under nitrogen, the glucuronide extract was reconstituted in 0.15 ml of a dilute acetic acid. The glucuronide extract was injected using a CTC A200SE autosampler onto a Hewlett–Packard 1100 series h.p.l.c. equipped with a YMC ODS-AQ C18 column. The glucuronides were eluted using a step gradient mobile phase beginning with 1% acetonitrile/99% 20 mm ammonium acetate (pH 8) and ending with 40% acetonitrile/60% 20 mm ammonium acetate (pH 8) at a flow rate of 0.8 ml min⁻¹. The retention times for M3G and M6G were approximately 1.9 min and 2.0 min, respectively. The column effluent was directed into the LC/MS system described above for detection.

The lower limits of quantification for morphine, M3G and M6G in serum were 0.5, 4 and 2 ng ml⁻¹, respectively, and in urine were 5, 20 and 20 ng ml⁻¹, respectively. Inter-day precision coefficients of variation for morphine, M3G and M6G serum samples were < 4.7%, < 4.0% and < 7.8%, respectively, and for urine samples were < 5.8%, < 3.5% and < 7.7%, respectively. The accuracy ranges for morphine, M3G and M6G in serum were 96.4–103.5%, 96.1–100.5% and 90.5–106.9%, respectively. The accuracy ranges for morphine, M3G and M6G in urine were 100.0–107.8%, 104.2–110.7% and 99.1–103.3%, respectively.

Pharmacokinetic analysis

Noncompartmental pharmacokinetic analysis of concentration-time data for morphine, M3G, and M6G was performed using the software WinNonlin version 1.1 (Scientific Consulting, Inc., Cary, North Carolina, USA). Actual blood collection times were used for all calculations. Maximum observed plasma concentration (C_max) and the first time to C_max (t_max) were determined by visual inspection of individual concentration vs time curves. Standard methods were used to calculate terminal elimination rate constant (λ_z); total systemic clearance for morphine (CL); apparent half-life (t_½); renal clearance (CLr) over 48 h of each analyte; amount of each analyte excreted in the urine (Ae) over 48 h; and fraction of dose excreted in the urine (fe) as each analyte.

Area under the concentration vs time curve through C_last, the last measurable concentration (AUC_last) was calculated by the trapezoidal method. Area under the concentration vs time curve extrapolated to infinity (AUC_inf) was calculated according to the following equation:

The AUC ratio of metabolite to morphine was calculated using AUC_inf values, after conversion to molar units. The molecular weights used in the conversions were 285.34 for morphine and 478.5 for M3G and M6G.

Pharmacodynamic analysis

Four pharmacodynamic parameters for the morphine-associated analgesic response were analysed: thermal pain threshold, thermal pain tolerance, mood (as measured on a VAS), and temporal summation of pain. Three to five determinations for each time point were measured. If five determinations were taken, the highest and lowest determinations were excluded, and the remaining three determinations were used to calculate a mean. In the case of four determinations, the three closest together (or the three lowest in the case of a tie) were selected for averaging. Mood assessment was made on a 10-cm VAS ranging from −5 to 5. For temporal summation of pain, the number of trials to maximum pain was the analysis variable.

Statistical analysis

Power analysis

The power analysis was based on variability in M3G and M6G AUC and C_max from the results of the original pilot study (Data on File, GlaxoSmithKline). A total of 12 patients was determined to be sufficient to demonstrate equivalence between the two treatments with at least 80% power using two one-sided, 5% significance levels (90% confidence intervals).

Pharmacokinetic parameters

The primary statistical analysis was performed on log-transformed AUC_inf, C_max, CL, and t_½ for morphine and metabolites. Each parameter was analysed separately using analysis of variance (anova) with the software SAS® version 6.12 (SAS Institute, Cary, North Carolina, USA) General Linear Models procedure. Full error diagnostics were performed on the studentized residuals from the anova model to confirm that the assumptions underlying the model had not been violated. Geometric means and 95% confidence intervals were calculated for the primary pharmacokinetic parameter estimates. Median values were calculated for the secondary pharmacokinetic parameters.

The objective of the pharmacokinetic statistical analysis was to demonstrate equivalence between the derived pharmacokinetic parameters for both treatments, in order to establish a lack of an interaction. Estimates of treatment effect were based on the ratio of the least squares means for each pharmacokinetic parameter. The lack of an interaction was delineated when the 90% confidence interval for the ratio of the least squares means fell within the range of 0.6–1.67, corresponding to 40% of the least squares means ratio.

t_max for each of the metabolites was compared for each treatment on a pairwise basis using the standard Koch methodology [9] for a two-period crossover design. Estimates of the treatment effect were based on median differences and presented with the associated nonparametric 90% confidence interval [10].

Pharmacodynamic parameters

Weighted means and 95% confidence intervals were computed for each of the analgesic response variables (threshold, tolerance, temporal summation of pain, mood). Each was analysed separately using analysis of covariance (ancova). Estimates of treatment effect were based on the difference of the least squares means (morphine + OND – morphine + placebo). Full error diagnostics were performed on the studentized residuals from the ancova model to confirm that the assumptions underlying the model had not been violated.

Safety

Adverse events were recorded in the case report forms for each individual subject during each of the study sessions, using standardized case report forms. The date, time of onset after administration of study drug, frequency, outcome, action taken, and description of the event were recorded. The investigator's assessment of severity (mild, moderate, severe), the investigator's opinion of the potential relationship of each event to the subjects receipt of the study drug (unlikely, possible, and probably related), and any comments from the investigator also were recorded.

Results

Pharmacokinetic parameters

Serum concentration-time profiles for morphine, M3G and M6G in one representative subject are presented in Figure 1. Geometric mean pharmacokinetic parameters for morphine, M3G and M6G are presented in Table 1. Median values for the secondary pharmacokinetic parameters for morphine, M3G and M6G are presented in Table 2. The two treatments appeared to be equivalent based on the 90% confidence intervals (0.6, 1.67) for the ratios of the least square means for morphine, M3G and M6G (Table 1).

Figure 1.

Serum concentration-time profiles for a representative subject for a) morphine, b) M3G and c) M6G after administration of morphine with ondansetron (▪) or placebo (Δ).

Table 1.

Primary pharmacokinetic parameters for morphine, M3G and M6G.

Parameter	Morphine + OND*	Morphine Morphine + NS*	Ratio†	Morphine +.OND*	M3G Morphine + NS*	Ratio†	Morphine + OND	M6G Morphine + NS*	Ratio†
AUCinf	96.0	103.4	0.93	1326.0	1358.6	0.98	128.7	125.0	1.03
(ng ml−1 h)	(85.9–107.2)	(92.2,116.0)	(0.85,1.01)	(1193.8,1472.9)	(1202.1,1535.3)	(0.93,1.02)	(109.3,151.6)	(103.3,151.4)	(0.95,1.11)
CL	88.6	82.2	1.08	NA	NA	NA	NA	NA	NA
(l h−1)	(79.1–99.2)	(73.1,92.4)	(0.99,1.17)
t½ (h)	2.6	2.5	1.05	15.2	15.7	0.97	2.5	2.2	1.16
	(2.4–2.9)	(2.2,2.9)	(0.97,1.15)	(12.9,17.9)	(13.3,18.6)	(0.83,1.13)	(2.2,2.9)	(1.9,2.5)	(1.04,1.28)
Cmax	204.9	193.3	1.06	167.9	174.1	0.96	29.0	30.1	0.97
(ng ml−1)	(150.2–279.5)	(145.5,256.7)	(0.89,1.26)	(147.2,191.4)	(151.2,200.3)	(0.90,1.03)	(25.4,33.1)	(26.3,34.4)	(0.90,1.03)
tmax (h)	0.08	0.08	NA	0.42	0.50	0.00	1.00	0.75	0.08
	(0.08–0.08)	(0.08,0.08)		(0.25,1.00)	(0.33,0.75)	(−0.11,0.13)	(0.75,1.05)	(0.33,1.50)	(−0.13,0.28)

^*Geometric means (95% confidence intervals); median value (range) for t_max.

^†Ratio of the least squares means for the log transformed parameters (90% confidence intervals); median difference (90% confidence intervals) for t_max. OND, ondansetron; NS, normal saline solution; AUC_inf, area under the concentration vs time curve extrapolated to infinity; CL, total systemic clearance of morphine; t_½, apparent half-life based on the terminal elimination rate constant λ_z; C_max, maximum observed plasma concentration; t_max, first time to C_max. Note: the molecular weights for morphine, M3G and M6G are 285.34, 478.5, and 478.5, respectively.

Table 2.

Secondary pharmacokinetic parameters for morphine, M3G, and M6G represented as median values (range).

	Morphine		M3G		M6G
Parameter	Morphine+ OND	Morphine+ NS	Morphine+OND	Morphine+ NS	Morphine+OND	Morphine+NS
CL r (l h−1)	5.9 (1.1,8.1)	6.8 (3.8,9.9)	4.7 (2.1,6.2)	4.9 (1.6,7.3)	10.9 (4.2,15.2)	12.0 (2.8,21.3)
AUC ratio	NA	NA	8.5 (5.1,11.1)	7.9 (6.1,10.1)	0.8 (0.5,1.1)	0.7 (0.5,1.3)
(metabolite: morphine)
fe (%)	7.0 (1.2,11.4)	7.7 (5.7,13.1)	38.3 (17.8,58.8)	42.9 (19.3,51.0)	9.6 (3.7,12.5)	10.0 (4.0,11.7)

CL r, renal clearance; fe, percent of dose excreted in urine as each analyte.

Pharmacodynamic parameters

Table 3 lists weighted mean threshold, tolerance, number of temporal summation trials to maximum pain and mood scores with summary statistics. Median threshold and tolerance profiles are presented in Figure 2. Median number of trials to maximum pain for temporal summation are presented in Figure 3. No differences were detected between treatment groups for any of the analgesic response variables, based on the minimal differences of the least squares means (Table 3).

Table 3.

Mean pharmacodynamic parameters (95% confidence interval).

Parameter	Morphine+OND	Morphine+NS	Difference*
Threshold (°C)	44.7	44.8	0.09
	(43.7,45.7)	(43.8,45.8)	(−0.49,0.66)
Tolerance (°C)	46.5	46.7	−0.04
	(45.6,47.7)	(45.7,47.6)	(−0.48,0.41)
Temporal summation	7.2	7.7	−0.45
(# pulses to maximum pain)	(5.6,8.8)	(6.4,8.9)	(−1.44,0.53)
Mood score (cm)	2.1	2.0	−0.13
	(1.1,3.1)	(1.0,3.0)	(−1.23,0.96)

OND, ondansetron; NS, normal saline solution.

^*Difference of the least squares means [morphine+ OND – morphine+placebo] (95% confidence interval).

Figure 2.

a) Median threshold and b) median tolerance profiles after administration of morphine with ondansetron (▪) or placebo (▵).

Figure 3.

Median number of trials to maximum pain score for temporal summation after administration of morphine with ondansetron (▪) or placebo (▵).

Safety

A total of 23 adverse events were reported by nine subjects during the trial. Headache was the most frequently reported adverse event, with seven episodes reported by five subjects. Four headache episodes were reported during OND coadministration, and three episodes were reported during placebo coadministration. One episode of nausea and one episode of vomiting were reported in two subjects during OND coadministration. A total of four accounts of nausea and three accounts of vomiting occurred collectively in four subjects following morphine + placebo treatment. One episode of urinary retention occurred during each treatment in the same subject. One episode of pruritis occurred during each treatment in two separate subjects. One episode of nightmares occurred during OND coadministration. Nausea and vomiting occurred more frequently with placebo coadministration than with OND coadministration. There were no other obvious differences in adverse events between treatment groups.

Discussion

Results of this study indicate that the pharmacokinetic disposition of morphine is not altered when coadministered with OND. Renal clearance, AUC ratio and fe values for morphine, M3G, and M6G were consistent with values previously reported in the literature [11, 12]. All primary pharmacokinetic parameter estimates were consistent with those reported in the literature, except for the half-life of M3G [12]. The apparent half-life of M3G was approximately 15 h, which is greater than the literature reported value of 3.9 h [12]. Previous studies may have failed to adequately characterize the terminal elimination half-life of M3G owing to differences in assay sensitivity and in the number of time points in the terminal elimination phase. In addition, this study was designed to describe more accurately the elimination half-life of M3G by extending the collection period to 48 h, as opposed to previous studies that only sampled through 24 h.

This study was powered to detect equivalence between the treatment groups based on the pharmacokinetics of ondansetron, such that the 90% confidence intervals for the ratio of the least squares means for the derived pharmacokinetic parameters fell within the range of 0.6–1.67. Therefore, the study population was of sufficient size to demonstrate equivalence. Because the two treatments appear to be equivalent in terms of the pharmacokinetics, differences in the analgesic response variables between treatment groups would not be expected.

The contact thermode system has been used to elicit pain in young healthy individuals following morphine administration [7]. It consists of a copper probe that delivers a thermal stimulus to the skin. Price et al. examined the effects of morphine on sensory and affective visual analogue scale responses to suprathreshold levels of experimental pain [7]. The results of the study showed that morphine produced a dose-dependent reduction in both sensory and affective responses to the experimentally induced pain delivered by the contact thermode system. Based on these results, the contact thermode system was chosen as a suitable device to detect changes in the morphine-associated analgesic response if alterations were to occur during OND coadministration.

Despite the established use of the contact thermode system in pain studies [7], this method has not been validated for morphine-induced analgesia. Nevertheless, the effect of morphine on the analgesic response was demonstrated by the contact thermode system in the present study. Figure 2 presents both thermal threshold vs time and tolerance vs time profiles. By visual inspection, 10 mg of morphine clearly increased the subjects' threshold and tolerance for pain during the first 2 hours. The profiles reflect median values for each endpoint as a function of time and the values after morphine administration exceeded the baseline values. Therefore, the contact thermode system was able to depict morphine-induced analgesia during the first 2 hours following morphine administration.

Interestingly, the median values for both thermal threshold and tolerance, as shown in Figure 2, failed to return to baseline. This may be the result of a learning effect, meaning the subjects were conditioned to respond in a certain manner. The placebo response is a powerful phenomenon, which depends on the patient's expectation that the therapy is effective [13]. Expectancy theory maintains that conditioning trials produce placebo response expectancies, thereby causing the expectancies to elicit the response [14]. Montgomery et al. tested this theory and confirmed that conditioning trials significantly enhanced placebo responses [14]. In the present study, multiple applications with the heat probe served as the conditioning trials and it is possible that the subjects learned to expect a certain degree of pain and were able to anticipate their responses.

In place of serial pharmacodynamic sampling, single time point measurements would minimize the placebo response. In future studies, it may be more advantageous to collect single pharmacodynamic measurements, corresponding to peak concentrations of morphine and M6G. The deviation from baseline for each of the analgesic response variables would then be used to assess whether the contact thermode system detects significant changes in the overall analgesic response to morphine. Rather than comparing effect-time profiles between treatment groups, this analysis may better characterize alterations in the pharmacodynamics of morphine.

In conclusion, the pharmacokinetics of morphine and its metabolites, M3G and M6G, are not altered significantly when morphine is administered as a single intravenous dose followed by a single intravenous dose of OND in healthy volunteers. In terms of the pharmacokinetics, the two treatments appear to be equivalent. No significant treatment differences were detected for any of the pain response variables. Results of this study reveal that a pharmacokinetic interaction between OND and morphine does not exist, therefore alterations in the overall analgesic response to morphine would not be expected.