Methadone Pharmacokinetics Are Independent of Cytochrome P4503A (CYP3A) Activity and Gastrointestinal Drug Transport: Insights from Methadone Interactions with Ritonavir/Indinavir

The investigation was approved by the University of Washington Institutional Review Board (Seattle, Washington), and subjects provided written informed consent. Eligibility criteria were (1) normal healthy volunteers aged 18–40 yr and (2) within 25% of ideal body weight (body mass index <30). Exclusion criteria were (1) major medical problems, (2) history of hepatic or renal disease, (3) family history of type 2 diabetes, (4) use of medications or nonprescription preparations known to alter CYP3A activity, (5) a known history of addiction to drugs or alcohol, and (6) access to and routine handling of addicting drugs in the regular course of employment. Females taking hormonal contraceptives were excluded. Both smokers and nonsmokers were enrolled. Subjects underwent a screening visit at which fasting blood glucose concentration and HIV serologic status were determined. Subjects were excluded if their glucose exceeded 110 mg/dl (because protease inhibitors can cause glucose intolerance) or if they were HIV seropositive (because monotherapy can cause HIV resistance). The final study population was 12 subjects (6 men, 6 women; age, 23 ± 5 yr [range, 18–34 yr]; weight, 68 ± 13 kg [range, 50–95 kg]).

The protocol was a three-period sequential crossover (control first, for logistical considerations) with each subject as their own control (fig. 1). Subjects were instructed to consume no food or beverages that contain grapefruit, apples, or oranges for 7 days before any study day, no alcohol or caffeine for 1 day before each study session, and no food or water after midnight before each study session. For each session, a catheter was placed in an arm vein for blood sampling; if needed, a second catheter was placed for drug administration. Subjects (supine) were monitored with a pulse oximeter and automated blood pressure cuff, and they received supplemental oxygen for saturations less than 94%.

First-pass CYP3A activity and intestinal P-gp (and other transporters) activity were evaluated on day 1 using oral alfentanil and fexofenadine as in vivo  probes.28,29,33,34Subjects received ondansetron (4 mg IV) for antinausea prophylaxis followed 30 min later by oral alfentanil (43 and 23 μg/kg at control and ritonavir/indinavir sessions, respectively) with 100 ml of water. Fexofenadine (60 mg) was administered with 100 ml of water 1 h after alfentanil. Subjects received a standard breakfast and lunch 3 and 6 h, respectively, after alfentanil. Venous blood was sampled for 48 h after alfentanil dosing, and plasma was stored at –20°C for later analysis. Coincident with blood sampling, dark-adapted pupil diameter was measured using a Pupilscan Model 12A infrared pupillometer with 0.1 mm of resolution (Keeler Instruments, Broomall, PA).28,29Each recorded value was the mean of triplicate measurements, which typically agreed to within 0.1–0.3 mm.

Hepatic CYP3A activity was evaluated on day 2 using intravenous alfentanil as an in vivo  probe.28,29,34Subjects received ondansetron followed 30 min later by alfentanil bolus (15 and 10 μg/kg at control and ritonavir/indinavir sessions, respectively). Subjects received a standard breakfast 4 h after alfentanil and had free access to food and water thereafter. Venous blood was sampled for 24 h after alfentanil dosing, and dark-adapted pupil diameter was measured coincident with blood sampling.

Methadone metabolism and clearance were assessed on day 3 (with follow-up on days 4–7) by simultaneously administering IV and oral methadone.24,32Subjects received IV ondansetron followed 30 min later by oral deuterated racemic (d5)-methadone HCl (11.0 mg, equivalent to 9.86 mg free base) with 100 ml water, and IV racemic unlabeled (d0)-methadone HCl (6.0 mg, equivalent to 5.4 mg free base; Roxane Laboratories, Columbus, OH). Deuterated methadone hydrochloride was synthesized and used under Investigational New Drug approval.32Venous blood samples were obtained for 96 h after methadone, plasma was separated and stored at –20°C for later analysis, and dark-adapted pupil diameter was measured coincident with blood sampling. Subjects were fed a standard breakfast 4 h after methadone and had free access to food and water thereafter. Continuous urine samples were collected at 24, 48, 72, and 96 h. Nausea and/or vomiting were treated with ondansetron (4 mg IV or 8 mg orally) as needed.

Hepatic and first-pass CYP3A, intestinal transporters activity, and methadone disposition were assessed at baseline. After 1–2 months, subjects then began taking ritonavir/indinavir 100/800 mg twice daily (approximately 7 am and 7 pm). Dosing was adjusted on study days (morning dose at lunch; evening dose at 11 pm) to preclude an acute inhibitor effect from the morning dose. First-pass CYP3A and transporter activities were determined on day 2 of ritonavir/indinavir, and methadone disposition was determined on day 3 (with follow-up on days 4–7). At steady-state ritonavir/indinavir, first-pass CYP3A and intestinal transporters activity were assessed on day 15, hepatic CYP3A assessed on day 17, and methadone disposition on day 17 (with follow-up on days 18–21). Because the duration of ritonavir administration may affect the degree of CYP3A alteration,23,24,34,35both acute and steady-state ritonavir/indinavir were evaluated. A potential for CYP3A induction by steady-state ritonavir was reported.35The time course of potential CYP3A induction was unknown, and time to evaluate hepatic CYP3A, first-pass CYP3A, and P-gp activities were potentially not sufficient before commencement of induction; therefore, only first-pass CYP3A and P-gp activities were evaluated in the acute ritonavir/indinavir phase. First-pass is a greater determinant than hepatic metabolism of oral drug disposition, and hence the more important parameter to evaluate.

Sample size was determined using a simplified analysis (paired t  test) for comparing the outcome variable methadone systemic clearance. A previous study found 22 and 33% interday/intrasubject variability in IV and oral methadone clearances, respectively.32To detect a 30% change in clearance, using a paired t  test, with 33% variability, 1-ß= 0.8, α= 0.05, would require 12 subjects.

Plasma alfentanil and fexofenadine concentrations were simultaneously quantified using solid-phase extraction and electrospray liquid chromatography-mass spectrometry as described previously.33Plasma and urine methadone and EDDP enantiomer concentrations were quantified using automated on-line extraction, stereoselective liquid chromatography, and electrospray mass spectrometry as described previously, except that standard curves for both d0- and d5-methadone were used.24,36 

Plasma methadone and EDDP data were analyzed using noncompartmental methods, assuming complete absorption, as described previously (WinNonlin 5.2; Pharsight Corp, Mountain View, CA).28Systemic clearance of IV alfentanil and d0-methadone was (CL^IV) = dose^IV/AUC^IV; apparent oral clearance of alfentanil and d5-methadone was (CL/F) = dose^oral/AUC^oral; bioavailability was (F^oral) = (AUC^oral/dose^oral) × (dose^IV/AUC^IV); volume of distribution based on the terminal phase was (Vz) = Dose/(AUC ×λ), where λ is the terminal elimination rate constant; steady-state volume of distribution was (V^ss) = CL × mean residence time. Hepatic extraction (E^H) was CL^H/Q^p, where hepatic plasma flow (Q^p) was estimated as 15.2 ml · kg^–1· min^–1. Renal clearance (Cl^rand CL^r/F) was determined as amount excreted in urine/AUC^0-∞. EDDP formation clearance was determined from urine data as Cl^f= fraction of dose recovered in urine × CL^IVand Cl^f/F = fraction of dose recovered in urine × CL/F^l, for IV and oral dosing, respectively. Hepatic clearance (CL^H) was CL^IV– CL^r. Alfentanil plasma data were similarly analyzed, as described previously.28,29Gastrointestinal extraction was E^G= 1–F^G, where F^Gwas F^oral/(F^abs(1–E^H)); the oral dose was assumed to be entirely absorbed, and F^abswas thus considered to be unity. Alfentanil and methadone effect (miosis) versus  time curves were treated analogously to conventional plasma concentration curves, to determine the area under the effect curve (AUEC).28,29 

Differences between treatment groups for pharmacokinetic and effect parameters were analyzed using one-way repeated measures analysis of variance followed by the Student-Newman-Keuls test for multiple comparisons or paired t  tests, as appropriate (SigmaStat 3.5; Systat Corp, Point Richmond, CA). Nonnormal data were log transformed for analysis but reported as the nontransformed results. Statistical significance was assigned at P < 0.05. Results are reported as the arithmetic mean ± SD (SD). Plasma AUC and urine data were also assessed as ratios (treated/control) and the geometric mean and 90% confidence interval of the geometric mean. Confidence intervals excluding 1.0 were considered statistically significant. Relationships between methadone and alfentanil clearances were evaluated by linear regression analysis.

Ritonavir/indinavir profoundly inhibited both hepatic and intestinal CYP3A activities. Effects of ritonavir/indinavir on alfentanil plasma concentrations are shown in figure 2, and pharmacokinetic parameters are provided in table 1. The AUC^0-∞/dose ratio (ritonavir-indinavir/control) for IV alfentanil was increased nearly 13-fold, half-life increased 11-fold, and systemic clearance and hepatic extraction were both decreased 92% by steady-state ritonavir/indinavir, indicating more than 90% inhibition of hepatic CYP3A activity. Acute and steady-state ritonavir/indinavir increased oral alfentanil AUC^0-∞/dose ratio (ritonavir-indinavir/control) 40- and 30-fold and decreased apparent oral clearance 98% and 97%, respectively. Both protease inhibitor regimens increased the C^max/dose ratio 4-fold and the half-life 12- to 13-fold. These results indicate >97% inhibition of first-pass CYP3A activity. Ritonavir/indinavir increased alfentanil bioavailability from 34 to 81% and reduced alfentanil intestinal extraction to 38% of control (from 48 to 18%), indicating significant inhibition of intestinal CYP3A activity.

Pupil data were available before plasma alfentanil concentrations and were used as an early surrogate for alfentanil clearance to assess ritonavir-indinavir effects on CYP3A (fig. 3and table 2). Ritonavir-indinavir increased and prolonged alfentanil miosis. The AUEC^0-∞/dose ratio for IV alfentanil was significantly increased by steady-state ritonavir-indinavir, and the ratio for oral alfentanil was significantly increased by both acute and steady-state ritonavir/indinavir.

Disposition of oral fexofenadine was used to probe the activity of the intestinal efflux pump P-gp and any other intestinal transporters for which fexofenadine is a substrate. Both acute and steady-state ritonavir/indinavir significantly increased fexofenadine peak plasma concentrations and AUC (fig. 4and table 3). There were no differences between acute and steady-state ritonavir/indinavir effects on intestinal transporters.

Disposition of methadone, after both IV and oral administration, was minimally affected by either acute or steady-state ritonavir/indinavir. Plasma R- and S-methadone concentrations in ritonavir/indinavir–treated subjects were not different from those of untreated subjects after IV methadone (fig. 5) and oral methadone (fig. 6). Acute and steady-state ritonavir/indinavir had no effect on the plasma AUC^0-∞ratio, systemic clearance, hepatic clearance, or hepatic extraction of R- and S- methadone after IV administration (table 4). Ritonavir/indinavir had no effect on the plasma AUC^0-∞ratio, apparent oral clearance, or bioavailability of R- and S- methadone after oral administration (table 5). Neither acute nor steady-state ritonavir/indinavir altered IV or oral methadone enantiomers renal clearance, which accounted for approximately 25% of both methadone enantiomers systemic clearance (table 6). Methadone N -demethylation was somewhat induced by ritonavir/indinavir, with greater effects on S- than R-methadone. Plasma R-EDDP but not S-EDDP concentrations were greater than controls in steady-state but not acute ritonavir/indinavir subjects for 10 h after IV and oral methadone. S-EDDP/ S-methadone AUC ratios were greater in acute and steady-state ritonavir/indinavir subjects than in controls for both IV and oral methadone. R-EDDP/R-methadone AUC ratios were not different from controls. EDDP enantiomer formation clearances after both IV and oral methadone were essentially unchanged.

The relationship between methadone clearance and CYP3A activity measured as alfentanil clearance is shown in figure 7. For IV R- and S-methadone, there was no significant correlation between systemic methadone clearance and hepatic CYP3A activity (r²≤ 0.01 for both; P > 0.05). For oral R- and S-methadone, there was similarly no significant correlation between methadone oral clearance and first-pass CYP3A activity (r²= 0.04 and 0.01; both P > 0.05).

Methadone effects were assessed using changes in dark-adapted pupil diameter (miosis) (fig. 8). Plasma concentrations of total (sum of IV d0- and oral d5) R-methadone (the pharmacologically active enantiomer) are shown in figure 8A. There was a second plasma concentration peak at 3 h after the initial IV peak, reflecting the slow absorption of oral methadone. R-methadone concentrations were not different between treatments. Miosis was also not different in ritonavir-indinavir–treated subjects and controls (fig. 8B), as was R-methadone pharmacodynamics, assessed by maximum miosis and AUEC (table 2) and the plasma concentration-effect curve (fig. 8C).

One major finding of this investigation was that ritonavir/indinavir profoundly inhibited CYP3A activity. Steady-state ritonavir/indinavir inhibited more than 90% of hepatic CYP3A activity, as assessed using alfentanil clearance. Both acute and steady-state ritonavir/indinavir inhibited more than 97% of first-pass CYP3A activity, attributable to inhibition of both intestinal and hepatic CYP3A, as alfentanil hepatic extraction was reduced from 0.35 to 0.03 and intestinal extraction from 0.48 to 0.18. Changes in alfentanil miosis were qualitatively similar to the alterations in alfentanil plasma AUC ratios and hepatic and first-pass CYP3A activities, indicating that alfentanil miosis qualitatively reflected ritonavir/indinavir effects on CYP3A activity. There are considered to be relatively few well-characterized protease inhibitor interactions with model CYP3A substrates,37and ritonavir/indinavir drug interactions in general are among the least well-studied. This is the first investigation to evaluate ritonavir/indinavir effects on both hepatic and first-pass CYP3A and to show that this protease inhibitor combination diminishes both hepatic and intestinal CYP3A activities, and thus both liver and intestine are the site(s) of ritonavir/indinavir-CYP3A drug interactions.

CYP3A inhibition by ritonavir/indinavir was not unexpected. Indinavir and ritonavir are each potent inhibitors of hepatic and/or intestinal CYP3A4 and CYP3A5,20,21,37both of which metabolize alfentanil.30Whereas indinavir alone only increased the AUC of IV and oral alfentanil 2- and 3-fold, respectively (Kharasch, unpublished results, 2004), and ritonavir alone only increased the AUC of IV and oral alfentanil 4- and 10-fold,34ritonavir/indinavir in the current investigation increased the AUC of IV and oral alfentanil 13- and 30-fold. Thus, ritonavir/indinavir effects likely reflect CYP3A inhibition by both protease inhibitors.

A second major finding of this investigation, which is the first to evaluate ritonavir/indinavir effects on apparent gastrointestinal P-gp activity, was that this protease inhibitor combination inhibited P-gp activity. Both short-term and steady-state ritonavir/indinavir significantly increased fexofenadine maximum plasma concentrations and AUC, which was not due to diminished elimination, suggesting impaired intestinal efflux. Fexofenadine is a well-characterized human P-gp substrate, is frequently used to assess P-gp pharmacogenetics and drug interactions,38–40and well-known P-gp inhibitors such as ketoconazole, verapamil, and erythromycin40–42increase plasma fexofenadine concentrations. It is unlikely that ritonavir/indinavir effects on fexofenadine AUC are attributable to changes in CYP3A or renal P-gp activity; these comprise less than 1% and only 5–10%, respectively, of fexofenadine elimination.38,40Rather, they suggest ritonavir/indinavir inhibition of intestinal and/or hepatic P-gp. A previous investigation showed that acute and steady-state ritonavir alone increased fexofenadine AUC.34Similarly, single-dose ritonavir and ritonavir/lopinavir and steady-state ritonavir/lopinavir also increased fexofenadine AUC, similarly ascribed to hepatic P-gp inhibition.43Increases in fexofenadine AUC caused by acute and steady-state ritonavir/indinavir were greater than in these previous reports. An explanation is not apparent because indinavir is a comparatively weak inhibitor of P-gp44,45and other transporters.46Fexofenadine is not transported exclusively by P-gp, but it is a substrate for organic anion uptake transporters; therefore, exact mechanisms by which ritonavir/indinavir alters drug transport remain unknown.

The primary purpose of this investigation was to determine the effect of ritonavir/indinavir on methadone disposition, metabolism, and clearance. This is the first investigation to evaluate ritonavir/indinavir effects on IV or oral methadone concentrations, concurrent effects on both IV and oral methadone, and effects on methadone metabolism and on renal excretion, and to compare short-term and steady-state ritonavir/indinavir effects on methadone disposition. There was some evidence for very minor induction of methadone N -demethylation. Ritonavir/indinavir (steady-state more than acute) mildly increased (S- more then R-) EDDP/methadone plasma AUC ratios, although EDDP enantiomer formation clearances were essentially unchanged. Any minor induction of methadone metabolism was not at all reflected by changes in methadone plasma concentrations, systemic or apparent oral clearance, hepatic clearance, hepatic extraction, or bioavailability. Ritonavir/indinavir had no effect on methadone renal clearance.

Another purpose of this investigation was to test the hypothesis that CYP3A mediates methadone disposition. For years, clinical methadone metabolism, clearance and disposition in general have been attributed to CYP3A4,4,9,10,15,19,47and variability in CYP3A4 activity was considered the major factor responsible for interindividual differences in methadone bioavailability,6based on extrapolation of early in vitro  studies that identified methadone N -demethylation by human recombinant and hepatic and intestinal CYP3A4.11,32,48–50Methadone interactions with highly active antiretroviral therapy and other drugs have been attributed to CYP3A4-mediated alterations in methadone disposition.12,19,26Nevertheless, the current results provide robust and unequivocal evidence against a major role for CYP3A in IV and oral methadone clearance. Specifically, ritonavir/indinavir caused minor induction of methadone metabolism and no changes in hepatic, systemic, or oral clearance, despite more than 90% inhibition of hepatic CYP3A activity and more than 97% inhibition of first-pass CYP3A activity. There was no correlation between IV methadone clearance and hepatic CYP3A activity, and none between oral methadone apparent clearance and first-pass CYP3A activity. The hypothesis that CYP3A mediates methadone disposition is rejected. Although the present investigation offers no insights into the CYP isoform, which is responsible for clinical methadone metabolism and clearance, investigations stimulated in part by rejection of this hypothesis have identified the prominent role of CYP2B6 in methadone metabolism in vitro ,32,50–52and clinical studies are consistent with this hypothesis.15,24,32,34,47,53 

A fifth major finding of this investigation, which is the first to evaluate the relationship between ritonavir/indinavir effects on gastrointestinal P-gp activity and methadone bioavailability, was that bioavailability was unchanged despite significant inhibition of P-gp. Although methadone appears to be a P-gp substrate,54and an animal experiment suggested a role for intestinal P-gp in methadone absorption,55the present investigation does not support a hypothesis that intestinal P-gp is significantly involved in methadone absorption and first-pass extraction. Clinically, P-gp polymorphisms had no effect on methadone enantiomer peak concentrations after oral administration.15Thus, although methadone may be a P-gp substrate, passive methadone permeability may be sufficiently high to enable high absorption uninfluenced by P-gp activity.

The sixth major purpose of this investigation was to evaluate ritonavir/indinavir effects on methadone pharmacodynamics. Pupil diameter changes were used to assess plasma concentration-effect relationships. Methadone miosis and concentration-effect relationships were not significantly changed by acute or steady-state ritonavir/indinavir. This is in contrast to the effects of acute and steady-state ritonavir and less so steady-state nelfinavir, which appeared to cause leftward and upward shifts of the R-methadone concentration-response curve, consistent with an increase in apparent potency and efficacy.24,53Such effects were postulated to suggest methadone brain access by an active process, influenced by one or more blood brain barrier drug efflux and/or influx transporters, susceptible to a transport-mediated interaction with protease inhibitors. Consistent with this interpretation is the finding that indinavir was the least potent of any protease inhibitor in inhibiting activity of primary cultured bovine brain microvessel endothelial cell P-gp and other efflux transporters.56 

The results of this investigation have significant clinical implications. Methadone is a cornerstone therapy for opiate dependence, but it is being increasingly used to treat acute, chronic, perioperative, neuropathic, and cancer pain. Greater methadone use for pain therapy is primarily responsible for a 1,300% increase in methadone prescriptions between 1997 and 2006.§||Unfortunately, however, an exponential decade-long rise in methadone toxicity has accompanied this growing clinical use. Methadone-related adverse events increased 1,800% between 1997 and 2004, and fatalities increased 390% from 1999 to 2004, both of which have been attributed primarily to increased methadone use for pain treatment not addiction.3#**The long elimination half-life and extreme variability in clearance (pharmacogenetic and/or drug interaction-related) are considered the greatest impediment to predictable methadone dosing and clinical effect.9Concerted research efforts for over a decade have endeavored to identify the enzymes responsible for methadone elimination and pertinent drug interactions and to provide practitioner guidance. On the basis of extrapolation of in vitro  studies, methadone metabolism and clearance in vivo  have been ubiquitously attributed to CYP3A4,4,9,10,15,47and methadone dosing guidelines warn about the potential for CYP3A4-mediated methadone drug interactions and the need to adjust dosing accordingly.4,6,9–11,13–15The Food and Drug Administration–approved methadone label states: “Since the metabolism of methadone is mediated primarily by CYP3A4, coadministration of drugs that inhibit CYP3A4 may cause decreased clearance of methadone. The expected clinical results would be increased or prolonged opioid effects.”††Nevertheless, neither the role of CYP3A4 in methadone clearance nor effects of CYP3A4 inhibition were ever clinically tested. We performed the first ever assessment of CYP3A activity and methadone clearance, and found no influence of CYP3A inhibition (60%) on methadone plasma concentrations or clearances and no correlation between methadone clearance and hepatic CYP3A activity.32Subsequently, ritonavir (>90%),24nelfinavir (>75%),53and in this investigation ritonavir/indinavir (>90%) were found to inhibit hepatic and first-pass CYP3A profoundly without reducing methadone clearance, and lack of a correlation between methadone clearance and CYP3A activity has been a consistent finding.15,24,32,53These results strengthen the original findings and provide substantive evidence against a role for CYP3A in methadone drug interactions and clinical methadone N -demethylation and clearance in general. Practitioner guidelines identifying methadone as a CYP3A substrate and warning of CYP3A-mediated drug interactions require thorough and thoughtful reevaluation. Indeed, the important message is that CYP3A-based guidelines used by clinicians to direct methadone therapy may be incorrect.

There are potential limitations with this investigation. First, a single methadone dose was evaluated. The risk of causing addiction make it neither possible nor ethical to study healthy volunteers at steady-state. Nonetheless, the kinetics of a single low methadone dose appear similar to those of higher, steady-state methadone doses. Second, ritonavir/indinavir effects were evaluated in healthy volunteers, not HIV-infected patients. This was deliberate, because antiretroviral therapy involves several drugs, thereby precluding a mechanistic evaluation and attribution of results to any one specific agent (i.e. , ritonavir-boosted indinavir). The present investigation was conducted when the standard ritonavir-boosted indinavir regimen was 100/800 mg twice daily, but that has now been changed to 100/400 mg twice daily.57Nonetheless, a similar absent effect of lower dose ritonavir-indinavir would be expected, and the mechanistic results of the present investigation remain valid.

In summary, ritonavir/indinavir had no significant effects on methadone enantiomers systemic or apparent oral clearance, renal clearance, hepatic extraction or clearance, bioavailability, or plasma concentrations, despite significant (>90%) inhibition of hepatic and first-pass CYP3A activity. This suggests little or no role for CYP3A in clinical methadone metabolism and clearance.