Rifampin Greatly Reduces the Plasma Concentrations of Intravenous and Oral Oxycodone

In view of previous studies2,5it was calculated that 10 subjects would be required to demonstrate a 30% difference in the area under the oxycodone plasma concentration–time curve (AUC) at level of significance P = 0.05 and a power of 80%. Because we wanted to be prepared for dropouts, 12 healthy volunteers (7 women and 5 men; age range, 20–31 yr; weight range, 53–110 kg), were enrolled in the study after giving written informed consent. All subjects were genotyped for CYP2D619; one was a poor CYP2D6 metabolizer, and the others were extensive metabolizers. The volunteers were ascertained to be in good health by medical history, clinical examination, and standard hematologic and blood chemistry test results. Urine toxicology and pregnancy test results were negative. None of the volunteers were receiving any continuous medication or were a smoker. The risk of participants to develop opioid abuse was considered low as assessed by the Finnish translation of the Abuse Questions.20 

The study was conducted according to the revised Declaration of Helsinki, and it was approved by the Ethics Committee of the Hospital District of Southwest Finland and the Finnish National Agency for Medicines. We used a four-session, paired, crossover study design (fig. 1). The volunteers were given in randomized order either 600 mg oral rifampin (Rimapen®; Orion, Espoo, Finland) or placebo once daily at 8 pm for 7 days. Oxycodone was administered on day 6, intravenously in the first part of the study and orally in the second part.

In the first part of the study, on two occasions at an interval of 4 weeks, all volunteers received 0.1 mg/kg intravenous oxycodone hydrochloride (Oxynorm®, 10 mg/ml; Mundipharma, Bard Pharmaceuticals Ltd., Cambridge, United Kingdom) on day 6 at 8 am in 2 min, 12 h after the fifth dose of rifampin or placebo. Two more doses of rifampin or placebo were given on days 6 and 7. In the second part, the volunteers were given correspondingly 15 mg oral oxycodone hydrochloride (Oxynorm®, 10-mg and 5-mg capsules; Mundipharma, Bard Pharmaceuticals, Ltd.) with 150 ml warm water. Adherence with the rifampin/placebo-dosing schedule was assessed by use of mobile telephone text messages. The volunteers fasted for 8 h before the administration of oxycodone. They were given standard meals at 4 and 8 h after oxycodone. They were not allowed to smoke or to drink grapefruit juice during the study. Alcohol, coffee, tea, and cola drinks were not permitted on test days.

On the test days, a forearm vein of each subject was cannulated with a plastic cannula, and timed blood samples were drawn into EDTA-containing tubes immediately before and 0.5, 1, 1.5, 2, 3, 4, 5, 6, 8, 10, 12, 24, and 48 h after administration of oxycodone. An additional blood sample was drawn 0.25 h after intravenous oxycodone administration. In the first part of the study, another venous cannula was inserted into the opposite forearm for the intravenous administration of oxycodone. Plasma was separated within 30 min and stored at −70°C until oxycodone, noroxycodone, oxymorphone, and noroxymorphone concentrations were analyzed with a validated liquid chromatography tandem mass spectrometric method as described previously.21The lower limits of quantification were 0.1 ng/ml for oxycodone and oxymorphone and 0.25 ng/ml for noroxycodone and noroxymorphone. The interday coefficients of variation were less than 10% for all analytes at relevant plasma concentrations.

The peak plasma concentrations (C^max) and corresponding C^maxtimes (t^max) of oxycodone and its metabolites were observed directly from the data. The areas under the oxycodone, noroxycodone, oxymorphone, and noroxymorphone concentration–time curves were estimated by means of the trapezoidal rule with extrapolation to infinity (AUC^0–∞). We used the linear trapezoidal rule when successive concentration values were increasing and the logarithmic trapezoidal rule when successive concentration values were decreasing after the peak concentration value. For each subject, the terminal log-linear phase of the oxycodone, noroxycodone, oxymorphone, and noroxymorphone plasma concentration–time curve was identified visually, and the elimination rate constant (k^e) was determined by regression analysis. The elimination half-life (t^½) was then calculated from the following equation: t^½= ln 2/k^e.

After intravenous administration of oxycodone, plasma clearance (Cl) and steady state volume of distribution (V^ss) of oxycodone were calculated by use of noncompartmental methods based on statistical moment theory. The oral bioavailability of oxycodone (F) was calculated as follows: F =[AUC^{0–∞ oral}· Dose^intravenous]/[AUC^{0–∞ intravenous}· Dose^oral]. After oral administration of oxycodone, we calculated the apparent clearance (Cl/F) and the apparent volume of distribution of oxycodone during elimination (V^z/F). The pharmacokinetic data were analyzed using the WinNonlin pharmacokinetic program (version 4.1; Pharsight, Mountain View, CA). The metabolite–to–parent drug AUC ratios were calculated as indexes of metabolism through different metabolic pathways.

Subjective effects of oxycodone were recorded with 100-mm visual analog scales from the following items: alert–drowsy, very good performance–very poor performance, no effects of the drug–very strong effects of the drug, unpleasant feeling–pleasant feeling.22The Maddox wing test was used to measure the effects of oxycodone on the coordination of extraocular muscles. The result was given in diopters.23The digit symbol substitution test was used to measure both cognitive and motor functions by recording the number of correct symbols substituted in 3 min.24Pupil diameter was measured using a Cogan pupillometer under stable lighting conditions.25These tests were performed before and 1, 2, 3, 4, 5, 6, 8, 10, and 12 h after the administration oxycodone. For each effect variable, the area under the response–time curve was determined by use of the trapezoidal rule for 12 h (AUC^0–12).

Sensitivity to tactile stimuli was assessed with von Frey monofilaments (18011 Semmes–Weinstein Aesthesiometer Kit, Stoelting Co., Wood Dale, IL) before the administration of oxycodone and 2 h after dosing. The subject lay comfortably in a supine position with the eyes blindfolded while monofilaments of five strengths (0.23, 0.27, 1.63, 4.00, and 11.8 mN) and a sham filament (no stimulus) were applied to the tip of the right index finger. Each stimulus was randomly applied on the skin eight times at an interval of 5 s. The subjects rated each stimulus on a scale of 0–2 (0 = no stimulus, 1 = a possible stimulus, 2 = a definite stimulus). Tactile threshold was determined before and 2 h after the administration of oxycodone. For the analysis, scores 1 and 2 were pooled together. Tactile threshold was depicted from psychometric curve as the 50% detection rate of stimuli.

Cutaneous heat pain threshold was determined before the administration of and 1 h after the dosing of oxycodone. The thermode site on the skin was the same during both measurement sessions. During each session, cutaneous heat pain threshold was defined as the mean of three measurements, which were performed at 30-s intervals on three adjacent sites on the right forearm to avoid sensitization and hyperalgesia. A feedback-controlled contact thermode (TSA-2; Medoc Inc., Rehovot, Israel) was used to deliver three heat pain stimuli of slowly increasing (1°C/s) temperature on the right volar forearm. The size of the thermode was 30 × 30 mm. Heat increased linearly from a baseline temperature of 35°C until the subject reported the first sensation of pain by pressing a button. This temperature was recorded, and after this, the temperature returned to baseline. The interval between the stimuli was approximately 30 s.

A cold pressor test was used to assess cold pain sensitivity. The subject immersed his or her left hand into ice water (temperature from 0.5° to 2°C) up to the wrist for 1 min. The latency from the immersion to the first sensation of pain (in seconds) was defined as the cold pain threshold. During the immersion, the subject reported the intensity and unpleasantness of cold pain at 30 and 60 s using a numerical rating scale (0 = no pain or unpleasantness, 100 = maximal pain or unpleasantness). If pain became intolerable, the subject was instructed to withdraw his or her hand from the water. In such case, the intensity and unpleasantness after withdrawal were reported as maximal (100). The cold pressor test was performed before and 1, 2, 3, 4, 5, 6, 8, 10, and 12 h after the dose of oxycodone. The AUC for cold pain was determined by the trapezoidal rule for 0–12 h.

The results are expressed as mean ± SD. The paired Student t  test was used, except for t^max, cutaneous heat pain threshold, and tactile sensitivity. The values for t^maxwere compared by the use of Wilcoxon signed rank test. The values for heat pain threshold and tactile sensitivity were compared by use of analysis of variance for repeated measures. The statistical significance level was P < 0.05. Because pharmacokinetic drug interactions can also be assessed statistically by the same methods that are standard for the investigation of bioequivalence, we calculated geometric mean ratios (rifampin/control) with 90% confidence intervals for the parameter AUC^0–∞of oxycodone. All data were analyzed by use of the statistical program SYSTAT for Windows (version 10.2; Systat Software, Richmond, CA).

Our pharmacokinetic results after placebo are well in line with the previous information on the pharmacokinetics of intravenous and oral oxycodone. Rifampin affected the pharmacokinetics of both intravenously and orally administered oxycodone (figs. 2 and 3, tables 1 and 2). The plasma concentrations of parent oxycodone were greatly decreased, particularly after its oral administration (fig. 2), and also the oxycodone-associated effects were reduced by rifampin.

Rifampin decreased the AUC^0–∞of intravenous oxycodone by 53%, increased the mean plasma Cl by 2.2-fold, and decreased its t^½from 3.7 to 2.4 h (fig. 2and table 1). Rifampin decreased the mean AUC^0–∞and C^maxof oral oxycodone by 86% and 63%, respectively (fig. 2and table 2). The geometric mean ratios with 90% confidence intervals for AUC^0–∞(rifampin/control) were 0.47 (0.42–0.52) and 0.14 (0.12–0.16) for intravenous and oral oxycodone, respectively. The mean oral bioavailability of oxycodone was decreased from 69% to 21% by rifampin. The intervention sequence (placebo–rifampin or rifampin–placebo) had no effect on the observed rifampin-associated changes in the pharmacokinetics of oxycodone.

Rifampin increased the C^maxof noroxycodone by 2.7-fold after intravenous oxycodone and by 2.0-fold after oral oxycodone compared with the control values (figs. 2 and 3, tables 1 and 2). The corresponding metabolite–to–parent drug AUC ratios (AUC^m/AUC^p) were increased 2.4-fold and 7.6-fold, respectively.

Rifampin reduced the AUC^0–∞of oxymorphone to approximately 5–10% of the corresponding control value after intravenous and oral administration of oxycodone. Intravenous oxycodone produced no detectable oxymorphone concentrations in two subjects after placebo and in eight subjects after rifampin. After rifampin and oral oxycodone, five of the subjects had every measured oxymorphone concentration below the lower limit of quantification. Rifampin increased the C^maxof noroxymorphone more than twofold in both the intravenous and the oral part of the study. The corresponding metabolite–to–parent drug AUC ratios increased 2.4-fold after intravenous and 9.6-fold after oral oxycodone.

After the oral administration of oxycodone, self-reported drowsiness, drug effect, and deterioration of performance were less pronounced after rifampin than after placebo as judged by AUC^0–12. Rifampin attenuated the miotic effect of intravenous oxycodone and decreased the oxycodone-associated heterotropia after oral oxycodone (fig. 4and table 3).

Oxycodone decreased cold pain sensitivity after intravenous and oral administration (fig. 4and table 3). During the control phase, oxycodone increased the heat pain threshold from 44.9°± 2.1° to 46.0°± 2.6°C after oral dosing of oxycodone, and from 43.9°± 1.9° to 45.1°± 2.3°C after intravenous dosing. Rifampin attenuated the increase in AUC^0–12of cold pain threshold, decrease in AUC^0–12of cold pain intensity, and unpleasantness after oral but not after intravenous administration of oxycodone.

Rifampin greatly increased the formation of both CYP3A-dependent metabolites of oxycodone as evidenced by the increased values for metabolite–to–parent drug AUC ratios of noroxycodone and noroxymorphone. The changes were more pronounced after oral administration of oxycodone because rifampin increased the first-pass metabolism in the intestinal wall and in the liver and enhanced the elimination of oxycodone. The results are in good agreement with our previous studies on the effect of rifampin on the pharmacokinetics of midazolam,26triazolam,14and ropivacaine27considering the relative contribution of CYP3A in their metabolism.

Another explanation for the observed pharmacokinetic changes especially after oral oxycodone could be the induction of the P-glycoprotein. Rifampin is known to induce P-glycoprotein, which is a membrane efflux transporter found in the intestinal wall and brain capillary endothelium.28Several opioids, such as loperamide,29alfentanil,30fentanyl,31morphine,32and methadone,33are substrates for P-glycoprotein. If oxycodone were a P-glycoprotein substrate, induction of intestinal P-glycoprotein would decrease its oral bioavailability and concentrations. Similarly, induction of brain capillary endothelium P-glycoprotein would decrease oxycodone penetration to the effect site in the brain and decrease its pharmacologic effects. On the basis of animal data, the role of P-glycoprotein in the pharmacokinetics of oxycodone seems to be of minor importance,34though this issue is controversial.35In humans, quinidine, which is a P-glycoprotein and CYP2D6 inhibitor, did not change the effects of oxycodone.36 

Inhibition of CYP2D6 by quinidine blocks the production of the active metabolite oxymorphone from oxycodone.36Interestingly, in the current study also rifampin seemed to reduce the formation of oxymorphone, as judged by means of the metabolite–to–parent drug AUC ratio. This is logical because rifampin shifted the metabolism of oxycodone toward the CYP3A-mediated route to noroxycodone, and the metabolism of oxymorphone via  CYP3A was also accelerated. Although some studies have suggested that rifampin may somewhat increase the activity of CYP2D6,37,38this enzyme is generally considered not to be inducible.

Oxymorphone has been used in the treatment of postoperative pain,39low back pain,40and cancer-related pain.41Its binding affinity for μ-opioid receptors is higher than that of oxycodone.4The role of oxymorphone in the analgesic effects of oxycodone, however, has not systematically been investigated.4,36In the current study, plasma concentrations of oxymorphone were generally very low, which is in agreement with the findings of previous studies,4,36and five subjects had plasma oxymorphone concentrations below the lower limit of quantification after rifampin and oral oxycodone. Considering these findings and the fact that we assessed the analgesic effect of oxycodone on experimental but not clinical pain, any conclusions about the role of oxymorphone in the analgesic effects of oxycodone would be highly speculative.

Oxycodone increased heat pain threshold and decreased cold pain sensitivity, supporting the findings of previous studies.42–44These effects seemed to be pain specific because oxycodone did not affect tactile sensitivity. Pharmacologic effects were not measured until 60 min after oxycodone dosing, which is a limitation of our study because we may have missed differences in the effects of oxycodone especially after intravenous administration. The changes in pupil diameter and behavioral test results were consistent with the well-known clinical effects of opioids. Induction of CYP3A with rifampin decreased oxycodone-induced drowsiness, drug effect, and deterioration of performance, and attenuated oxycodone-induced pain relief in the cold pressor test. These findings are in line with previous clinical observation of poor analgesic effect of oxycodone when used concomitantly with rifampin.7 

Rifampin is an effective inducer of many CYP enzymes both in vitro  and in humans.8Rifampin induced methadone withdrawal symptoms in 70% of opioid-addicted patients given maintenance therapy with methadone, a substrate of CYP3A and CYP2B6.9,45The concomitant administration of rifampin together with oral midazolam26and triazolam14has resulted in an approximately 95% reduction in the exposure to these hypnotics and a total loss of the hypnotic action. In the current study, the influence of rifampin on the pharmacologic effects of a single dose of oxycodone was rather modest. It is possible that the reduction of opioid effects after rifampin might have been of greater magnitude had we used a protocol with multiple doses of oxycodone.

The interaction of orally administered oxycodone with rifampin is clearly of clinical significance. It is reasonable to assume that other strong inducers of CYP3A, such as carbamazepine and phenobarbital, may also substantially reduce the therapeutic effect of oxycodone. Because the concomitant use of rifampin with oxycodone may attenuate its analgesic effect, dose adjustments or a switch to another method of analgesia may be necessary. Discontinuation of rifampin may result in excessive unwanted effects of opioids. Further studies are needed to investigate the possible effect of other CYP3A inducers on the pharmacology of oxycodone in patients.

The authors thank Elina Kahra (Medical Laboratory Technologist, Clinical Pharmacology, TYKSLAB, Hospital District of Southwest Finland, Turku, Finland) for her skillful technical assistance.