The use of cannabinoids, smoked marijuana and dronabinol by HIV-1-infected individuals has become an increasingly common therapy for the management of wasting, appetite stimulation, and antiretroviral side-effects . The Food and Drug Administration previously approved the use of dronabinol, an oral dosage form of delta-9-tetrahydrocannabinol (THC), for the treatment of anorexia in AIDS patients. However, smoking marijuana allows for easier titration and has been reported to be more effective than oral THC . California passed the Compassionate Use Act in 1996, which allows the use of smoked marijuana for a variety of medical conditions including HIV disease and its complications. The widespread use of protease inhibitors (PI), which are commonly associated with a multitude of drug interactions, creates concern regarding the combined use of cannabinoids and PI. The potential for interaction between the two warrants pharmacological investigation. Currently, there are no published data evaluating the potential drug–drug interaction between cannabinoids and antiretroviral agents.
The evolution of HIV pharmacotherapy has introduced a multitude of new medications, which are subject to multiple drug interactions occurring mainly through involvement of the cytochrome P-450 enzyme system . Indinavir (IDV) and nelfinavir (NFV) are metabolized by the cytochrome P-450 enzyme system, primarily by the 3A4 (CYP3A4) isoform in the liver and gastrointestinal tract. The liver similarly metabolizes cannabinoids via the cytochrome P-450 system and there exists a potential for metabolic interaction with the PI. The large number of active and inactive metabolites derived from cannabinoids, the psychoactive components of marijuana, complicates the understanding of cannabinoid metabolic pathways. Cannabinoids appear to have diverse pathways of cytochrome P-450 metabolism involving various families and subfamilies of hepatic microsomes, including 2C (CYP2C) and 3A (CYP3A) [4,5]. Cannabidiol, a primary cannabinoid in marijuana, was shown to inactivate CYP3A in mouse hepatic microsomes . In addition, constituents of marijuana have been shown to interact with CYP3A4 in human liver microsomes . Delta-9-THC, the primary psychoactive cannabinoid in marijuana, has been shown to interfere with the metabolism of antipyrine, phenobarbital, and hexobarbital .
The growing therapeutic use of marijuana in HIV-infected people creates the need to identify potential interactions with the antiretroviral agents. This is the first controlled study to assess the clinical consequences of short-term cannabinoid use in the context of antiretroviral therapy.
Materials and methods
Subjects were required to be at least 18 years old, have documented HIV infection, and be on a stable antiretroviral treatment regimen that included either IDV or NFV as the sole PI in addition to nucleoside and/or non-nucleoside reverse transcriptase inhibitors, for at least 8 weeks prior to enrollment. Subjects were also required to have a stable viral load, defined as less than a threefold change in HIV RNA level for the 16 weeks prior to enrollment. All subjects were required to have prior experience smoking marijuana (defined as six or more times) to ensure that they knew how to inhale and what neuropsychiatric effects to expect.
Exclusion criteria included the following: any active opportunistic infection or malignancy requiring acute treatment; unintentional loss of ≥ 10 percent of body weight during the past 6 months; current substance dependence; methadone maintenance; use of tobacco or cannabinoids (smoked or oral) within 30 days of enrollment; history of serious pulmonary disease; pregnancy; and Stage II or higher AIDS dementia complex. Laboratory exclusion criteria were hematocrit < 25% and hepatic transaminase elevations to greater than five times the upper limit of normal.
The University of California San Francisco Committee on Human Research approved the protocol and informed consent was obtained for all subjects. The trial consisted of a 4-day lead in (days –4 to –1) followed by a 21-day, randomized, placebo-controlled inpatient study conducted in 28 subjects receiving IDV and 34 subjects receiving NFV. Subjects were admitted to the General Clinical Research Center (GCRC) at the San Francisco General Hospital, remained within the GCRC for the entire study period, and were not permitted visitors.
Subjects received either IDV 800 mg every 8 h or NFV 750 mg three times daily throughout the study duration. A single subject on IDV had NVP included in their regimen and consequently was receiving an adjusted dose of 1000 mg every 8 h throughout the duration of the study. Dosage adjustments were not warranted for subjects on NFV who received either nevirapine or efavirenz. Subjects functioned as their own controls for pharmacokinetic evaluation and were maintained on stable therapy without regimen or dosage adjustments for 14 days prior to study entry and throughout the duration of the study. There were no restrictions on concurrent antiretroviral medications, other than additional PI. Concomitant drugs not permitted included megestrol acetate, testosterone, or other anabolic steroids, human growth hormone, thalidomide, pentoxyfylline or agents that might alter immune system parameters such as prednisone and interleukin-2. Subjects were stratified by PI and randomized to one of three treatment arms: 3.95% THC marijuana cigarettes to be inhaled; dronabinol 2.5 mg capsules; or placebo capsules. The GCRC staff administered all antiretroviral and treatment arm medications.
The marijuana cigarettes, weighing on average 0.9 g and containing 3.95% delta-9-THC, were provided by The National Institute on Drug Abuse. Additional details regarding handling and consumption of marijuana have been reported elsewhere . Subjects smoked up to three complete marijuana cigarettes daily, as tolerated, 1 h prior to meals.
Roxane Laboratories (Columbus, Ohio, USA) supplied dronabinol and matching placebo capsules. Subjects who were blinded and randomized to the oral regimens received dronabinol 2.5 mg or placebo, three times daily 1 hour prior to breakfast, lunch, and dinner.
Pharmacokinetic sample collection
Intensive blood sampling for IDV and NFV was done at baseline (day −1) in the absence of randomized treatment and day 14 of randomized treatment at the following times: pre-dose, 0.5, 1, 1.5, 2, 3, 4, 5, 6, and 8 h post-dose. IDV was administered on an empty stomach, 1 h prior to meals, while standard meals were supplied with NFV administration.
The selection of marijuana and dronabinol sampling times was based on previously studied cannabinoid pharmacokinetics . Trough levels of plasma delta-9-THC were obtained just prior to the second marijuana cigarette smoked on day 14. Additional levels were drawn 2 min, 60 min, and 6 h after smoking. For the oral dronabinol and placebo subjects, a trough level was also obtained just prior to the second dose on day 14, with subsequent levels obtained 2, 4, and 6 h after the dose.
Blood samples collected via indwelling peripheral catheters into heparinized Vacutainer tubes were kept at room temperature for no more than 2 h after collection. Plasma specimens were separated by centrifugation and frozen at −70°C until quantification.
PI pharmacokinetic analysis
The objectives of the study were to determine whether cannabinoids altered the steady-state pharmacokinetics of IDV, NFV, and M8 (hydroxy-t-butylamide metabolite of NFV). The pharmacokinetics of IDV, NFV, and M8 were computed using standard non-compartmental methods following multiple dose administration using WinNonlin software version 2.1 (Pharsight, Mountain View, California, USA). The primary endpoints of the study were the steady-state pharmacokinetic parameters of IDV, NFV, and M8 including the area under the plasma concentration-time curve to 8 h (AUC8), maximum concentration (Cmax), and minimum concentration (Cmin). The extent of metabolism of NFV to M8 is expressed as the ratio of M8/NFV AUC. AUC8 was derived by linear/logarithmic trapezoidal rule over the dosing interval. Maximum and minimum concentrations were appropriately identified from the plasma concentration versus time data for each subject.
Cannabinoid pharmacokinetic analysis
Cannabinoid pharmacokinetics were examined to characterize differences in plasma concentrations between the two forms of administration. The pharmacokinetics of delta-9-THC were computed using standard non-compartmental methods following multiple dose administration using WinNonlin software version 2.1 (Pharsight). Area under the concentration-time curve to 6 h (AUC6) was the primary measurement for delta-9-THC; secondary measurements were Cmax and time of maximum concentration (Tmax). AUC6 was derived by the linear/logarithmic trapezoidal rule over the dosing interval.
Plasma samples for IDV and NFV were analyzed by validated liquid chromatography/tandem mass spectrometry assays within the Drug Research Unit of the Department of Clinical Pharmacy, University of California San Francisco at the San Francisco General Hospital . Samples (200 μl of IDV or 50 μl NFV + 150 μl ammonium formate buffer pH 4.1) were processed with 50 μl of methyl-IDV as internal standard and vortexed. Acetonitrile (400 μl) was then added to denature proteins and release bound drug. After centrifugation, supernatants were injected for quantification. The IDV standard curve ranged from 3 ng/ml to 12000 ng/ml with a limit of quantitation of 3 ng/ml. NFV and M8 standard curves were 5 ng/ml to 4000 ng/ml with a limit of quantitation of 5 ng/ml. Intra-day analytical variability for IDV was 6.1–7.4% for mean concentrations of 31–8093 ng/ml, and for NDV it was 8.0–12.5% for mean concentrations of 32–3196 ng/ml.
The cannabinoid assays (lower limit of quantitation of 0.5 ng/ml) were performed by gas chromatography/MS at the Center for Human Toxicology at the University of Utah .
The effects of cannabinoids on the steady-state pharmacokinetics of IDV, NFV, and M8 were analyzed by comparing baseline parameters (day −1) with values obtained after concomitant cannabinoid treatment (day 14). The percent change in Cmax, Cmin and AUC8 from baseline to day 14 was calculated for each patient. The median and range of these changes are reported for IDV, NFV, and M8 within each study arm. Non-parametric statistical tests were performed as the distributions of these changes were positively skewed and thus did not meet the assumption of normality. Wilcoxon signed-rank tests were used to identify statistically significant changes for subjects on either PI within each treatment arm. Mann–Whitney rank-sum tests were performed to compare cannabinoid pharmacokinetic parameters between the marijuana and dronabinol treatment arms.
Sixty-two patients completed pharmacokinetic analyses at baseline and week 14 (Table 1). The majority of patients were male (89%) and over the age of 40 years (71%). Fifty percent of the patients were white. Twenty-eight patients on an IDV-containing regimen and 34 on a NFV-containing regimen were evaluated (Table 1). Baseline viral load and CD4 T-lymphocyte count characteristics have been described elsewhere .
The pharmacokinetics of IDV for each treatment arm are shown in Table 2. The most evident changes occurred in the marijuana arm (Fig. 1). A significant decrease of −14.1 % (range, −58 to +7;P = 0.039) in the Cmax was observed at day 14, after initiation of cannabinoid treatment. A similar decrease of −14.5% (range, −66 to +44;P = 0.074) in the AUC8 approached statistical significance. Interestingly, there was a modest reduction in Cmin of −33.7% (range, −64 to +98;P = 0.65), represented as a decline in median plasma concentration from 116 to 97 ng/ml. Because of large inter-patient variability the decrease was not significant.
There were no significant changes in the two oral treatment arms for subjects on IDV and no median percentage changes > 10% were observed in any of the pharmacokinetic parameters.
Similar decreases in Cmax, Cmin, and AUC for NFV were also observed in the marijuana arm (Table 3 and Fig. 1). Although no statistically significant changes were observed, there was a trend towards diminished plasma concentrations for all evaluated pharmacokinetic parameters in the NFV–marijuana arm. The largest median percentage change of −17.4% (range, −43 to +64;P = 0.46) was observed for Cmax. Minor reductions of −10.2% (range, −46 to +92;P = 0.15) and −12.2% (range, −79 to +104;P = 0.28) were observed for AUC8 and Cmin, respectively.
No significant changes were observed in either of the two oral treatment arms for the subjects on NFV and all median percentage changes from baseline were < 10%.
Changes in M8 metabolism are represented by the median percentage difference in M8/NFV AUC8 ratio at baseline and day 14 (results not shown as a table). The largest median percentage change of −18% (range, −124 to +19;P = 0.025) was observed in the marijuana arm. However, a decrease in the placebo arm of −16% (range, −75 to +14;P = 0.023) also reached significance. The −13% (range, −183 to 32;P = 0.35) change in the dronabinol arm did not approach significance due to extensive variability.
The pharmacokinetic comparisons of smoked marijuana versus dronabinol are shown in Table 4. Plasma levels of delta-9-THC in the marijuana arms were significantly higher than in the dronabinol arms. In the marijuana arm, the median Cmax and AUC6 values for subjects on NFV were 145 ng/ml (range, 43–261 ng/ml) and 84 ng/ml⋅h (range, 58–147 ng/ml⋅h) respectively. Three subjects had concentrations below the limit of quantification at hour 6, therefore an AUC6 could not be determined and these subjects were excluded from pharmacokinetic analysis. In the dronabinol arm on NFV, median Cmax and AUC6 were 1.2 ng/ml (range, 0.6–4.2 ng/ml) and 4.4 ng/ml⋅h (range, 0.6–11.5 ng/ml⋅h) respectively. In four subjects the delta-9-THC levels in plasma were below the limit of quantitation at all measured time points but were detectable in a single urine collection. Therefore it was not possible to calculate the AUC especially in light of the limited sampling schedule. Similar results were observed in the treatment arms on IDV. The marijuana arm achieved median Cmax and AUC6 values of 138 ng/ml (range, 10–190 ng/ml) and 70 ng/ml⋅h (range, 10–108 ng/ml⋅h) respectively. One subject had concentrations below the limit of quantification at hour 6, therefore an AUC6 could not be determined and the subject was excluded from pharmacokinetic analysis. The median Cmax in the dronabinol arm was 1.1 ng/ml (range, 0.6–3.2 ng/ml) and median the AUC was 3.7 ng/ml⋅h (range, 1.1–9.0 ng/ml⋅h). In two subjects the delta-9-THC levels in plasma were below the limit of quantitation at all measured time points but were detectable in a single urine collection. Therefore, as for NFV subjects, it was not possible to calculate the AUC.
To our knowledge, this is the first study to evaluate the metabolic interaction of a Schedule I controlled substance with antiretroviral therapy. A cohort study in the San Francisco area found no association between the use of recreational drugs, including marijuana, by HIV-infected subjects and progression to AIDS . However, reports describing recreational drug interactions with the PI, including a fatal interaction, have started to emerge [12,13]. Cannabinoids are characterized by a large therapeutic index, which minimizes the risk of adverse clinical events due to elevated cannabinoid plasma levels. There have been no confirmed deaths reported from marijuana overdose . However, moderate changes in PI concentrations can impact the clinical efficacy and toxicity of antiretroviral therapy [15–18].
In our study, the addition of smoked marijuana to an antiretroviral regimen containing either IDV or NFV resulted in diminished PI plasma concentrations. In contrast, comparisons between baseline and day 14 results for the dronabinol and placebo arms revealed no obvious alterations in the pharmacokinetics of either PI. Concurrent marijuana or dronabinol administration had no negative short-term clinical effects or impact on viral load or CD4 T-lymphocyte count . Determination of the long-term clinical effects of concomitant smoked marijuana and PI would require separate evaluation.
Alterations in plasma IDV concentrations impacted all three pharmacokinetic parameters. However, the only statistically significant effect, which occurred in the IDV-marijuana arm, was a 14% decrease in median Cmax. This minor change is unlikely to be clinically relevant and the impact of diminished Cmax on viral load and CD4 T-lymphocyte count has been examined only briefly . Interestingly, the 34% decrease in median Cmin observed did not achieve statistical significance, because of the large variability in Cmin values. Six of the nine patients in the IDV-marijuana arm demonstrated decreases in Cmin from baseline and the median IDV Cmin plasma level on day 14 was 97 ng/ml. A minimal reduction of −14.5% in IDV AUC8 from baseline to day 14 was also observed and approached significance (P = 0.074) but is unlikely to haven an impact on clinical outcome. As with IDV, similar reductions in NFV plasma concentrations occurred in the smoking group albeit to a lesser extent with none of these changes achieving statistical significance.
IDV and NFV results for the dronabinol and placebo treatment arms exhibited extensive variability with respect to plasma concentrations, and no changes achieved significance. Furthermore, there were no changes > 10% in the median values of AUC8, Cmax, or Cmin in either treatment arm.
The majority of in vitro and animal data regarding cannabinoid involvement with the cytochrome P-450 enzyme system suggests an inhibitory effect on the CYP3A and CYP2C subfamilies [20–22]. These findings were further established in mouse and human microsomes, which demonstrated the potential for inhibition of cyclosporine metabolism resulting in increased levels of cyclosporine, a known substrate of CYP3A4 hepatic metabolism . Contrary to these data, the results of this study suggest an induction of PI metabolism in the presence of high plasma concentrations of delta-9-THC resulting in the observed decreases in IDV and NFV plasma concentrations in the marijuana arm. The impact of cannabinoids on hepatic induction of CYP3A and CYP2C has been investigated only in mice .
Another possible explanation for diminished PI plasma concentrations could be attributed to decreased gastrointestinal motility and altered absorption caused by the cannabinoids [24,25]. There were no significant alterations in Tmax (data not shown) in any of the treatment arms suggesting that gastrointestinal motility was unchanged. However, the decreases in Cmax in both IDV and NFV treatment arms could be attributed to altered absorption in the gut. The mechanism of diminished IDV and NFV plasma levels cannot be fully elucidated.
NFV is metabolized via multiple isozymes including cytochrome 3A4. M8 is the primary oxidative metabolite with several minor oxidative metabolites also identified. There were significant decreases in the M8 : NFV AUC8 ratio for the marijuana and placebo arms at day 14 but non-significant decreases in the dronabinol arm. These changes may be attributable to an effect of time, given the significant changes in the placebo group. However, previous examination of NFV metabolism over an extended time period demonstrated no significant changes in M8 formation . Consequently, there exists a potential for CYP2C19 (responsible for M8 formation) inhibition by the cannabinoids resulting in a decreased M8 : NFV ratio.
One limitation in study design was the absence of a smoking placebo arm as the act of smoking itself can be a confounding factor in the evaluation of drug interaction studies. Aromatic hydrocarbons found in cigarette smoke are known inducers of cytochrome P450 enzymes, primarily CYP1A1. A review of clinically significant drug interactions with cigarette smoking describes the potential effects on multiple metabolic isoenzymes including CYP1A1, CYP1A2, CYP2E1, and glucuronosyltransferase . Based on the lack of CYP3A4 involvement, we anticipated no drug interactions with the PI due to smoking alone.
Large inter-patient variability complicates the interpretation of PI plasma levels and confounds the ability to translate PI concentrations into a pharmacodynamic relationship. Given the large variability of the pharmacokinetic measures, any justified adjustments in PI dosing should be determined on an individual basis. Based on our short-term findings, the use of marijuana and dronabinol is unlikely to effect antiretroviral efficacy.
We sincerely thank the following for their important contributions to our study: A. Jaywaredene, J. Stone, R. Leiser, and the research nurses in the General Clinical Research Center at the San Francisco General Hospital for their indebted help.
1. Abrams D. Short Term Safety of Cannabinoids in HIV Patients. Eighth Conference on Retroviruses and Opportunistic Infections
. Chigaco, February 2001 [abstract 744].
2. Doblin RE, Kleiman MA. Marijuana as antiemetic medicine: a survey of oncologists’ experiences and attitudes. J Clin Oncol 1991, 9: 1314–1319.
3. Barry M, Mulcahy F, Merry C, Gibbons S, Back D. Pharmacokinetics and potential interactions amongst antiretroviral agents used to treat patients with HIV infection. Clin Pharmacokinet 1999, 36: 289–304.
4. Bornheim LM, Correia MA. Selective inactivation of mouse liver cytochrome P-450IIIA by cannabidiol. Mol Pharmacol 1990, 38: 319–326.
5. Yamamoto I, Watanabe K, Narimatsu S, Yoshimura H. Recent advances in the metabolism of cannabinoids. Int J Biochem Cell Biol 1995, 27: 741–746.
6. Jaeger W, Benet LZ, Bornheim LM. Inhibition of cyclosporine and tetrahydrocannabinol metabolism by cannabidiol in mouse and human microsomes. Xenobiotica 1996, 26: 275–284.
7. Benowitz NL, Jones RT. Effects of delta-9-tetrahydrocannabinol on drug distribution and metabolism. Antipyrine, pentobarbital, and ethanol.
Clin Pharmacol Ther 1977, 22: 259–268.
8. Huestis MA, Henningfield JE, Cone EJ. Blood cannabinoids. I. Absorption of THC and formation of 11-OH-THC and THCCOOH during and after smoking marijuana.
J Anal Toxicol 1992, 16: 276–282.
9. Jayewardene AL, Kearney B, Stone JA, Gambertoglio JG, Aweeka FT. An LC-MS-MS method for the determination of indinavir, an HIV-1 protease inhibitor, in human plasma. J Pharm Biomed Anal 2001, 25: 309–317.
10. Foltz RL, McGinnis KM, Chinn DM. Quantitative measurement of delta 9-tetrahydrocannabinol and two major metabolites in physiological specimens using capillary column gas chromatography negative ion chemical ionization mass spectrometry. Biomed Mass Spectrom 1983, 10: 316–323.
11. Di Franco MJ, Sheppard HW, Hunter DJ, Tosteson TD, Ascher MS. The lack of association of marijuana and other recreational drugs with progression to AIDS in the San Francisco Men's Health Study. Ann Epidemiol 1996, 6: 283–289.
12. Harrington RD, Woodward JA, Hooton TM, Horn JR. Life-threatening interactions between HIV-1 protease inhibitors and the illicit drugs MDMA and gamma-hydroxybutyrate. Arch Intern Med 1999, 159: 2221–2224.
13. Henry JA, Hill IR. Fatal interaction between ritonavir and MDMA. Lancet 1998, 352: 1751–1752.
14. Gurley RJ, Aranow R, Katz M. Medicinal marijuana: a comprehensive review. J Psychoactive Drugs 1998, 30: 137–147.
15. Acosta EP, Henry K, Baken L, Page LM, Fletcher CV. Indinavir concentrations and antiviral effect. Pharmacotherapy 1999, 19: 708–712.
16. Kerr B. Virologic response-plasma drug concentration relationship in Phase III study of nelfinavir mesylate. VII International Conference on AIDS.
Geneva, June 1998 [abstract 12304].
17. Marzolini C, Telenti A, Decosterd LA, Greub G, Biollaz J, Buclin T. Efavirenz plasma levels can predict treatment failure and central nervous system side effects in HIV-1-infected patients. AIDS 2001, 15: 71–75.
18. Burger DM, Hoetelmans RM, Hugen PW. et al
. Low plasma concentrations of indinavir are related to virological treatment failure in HIV-1-infected patients on indinavir-containing triple therapy. Antivir Ther 1998, 3: 215–220.
19. Anderson P, Brundage R, Kakuda T, Fletcher C. Indinavir (IDV) maximum plasma concentration (Cmax) predicts CD4+ response in persons with highly suppressed HIV RNA. Eighth Conference on Retroviruses and Opportunistic Infections
. Chicago, February 2001 [abstract 731].
20. Bornheim LM, Lasker JM, Raucy JL. Human hepatic microsomal metabolism of delta 1-tetrahydrocannabinol. Drug Metab Dispos 1992, 20: 241–246.
21. Bornheim LM, Everhart ET, Li J, Correia MA. Characterization of cannabidiol-mediated cytochrome P450 inactivation. Biochem Pharmacol 1993, 45: 1323–1331.
22. Bornheim LM, Grillo MP. Characterization of cytochrome P450 3A inactivation by cannabidiol: possible involvement of cannabidiol-hydroxyquinone as a P450 inactivator. Chem Res Toxicol 1998, 11: 1209–1216.
23. Bornheim LM, Everhart ET, Li J, Correia MA. Induction and genetic regulation of mouse hepatic cytochrome P450 by cannabidiol. Biochem Pharmacol 1994, 48: 161–171.
24. Izzo AA, Mascolo N, Capasso F. Forgotten target for marijuana: the endocannabinoid system in the gut. Trends Pharmacol Sci 2000, 21: 372–373.
25. Izzo AA, Mascolo N, Capasso F. Marijuana in the new millennium: perspectives for cannabinoid research. Trends Pharmacol Sci 2000, 21: 281–282.
26. Zhang KE, Wu E, Patick AK. et al
. Circulating metabolites of the human immunodeficiency virus protease inhibitor nelfinavir in humans: structural identification, levels in plasma, and antiviral activities. Antimicrob Agents Chemother 2001, 45: 1086–1093.
27. Zevin S, Benowitz NL. Drug interactions with tobacco smoking. An update.
Clin Pharmacokinet 1999, 36: 425–438.
Keywords:© 2002 Lippincott Williams & Wilkins, Inc.
Cannabinoids; HIV protease inhibitors; pharmacokinetics; interaction; HIV; indinavir; nelfinavir