3rd International Workshop on Clinical Pharmacology of HIV Therapy

This is Archived Content. This content is available for historical purposes only. It may not reflect the current state of science or language from the National Institute on Drug Abuse (NIDA).
View current meetings on nida.nih.gov.


Washington, DC


United States

Meeting Summary

NIDA-SPONSORED MINI-SYMPOSIUM on Interactions Between Drugs of Abuse and Pharmacotherapeutics Used in the Treatment of Infections, Drug Abuse, and Mental Disorders.


Published and unpublished research presented in national and international forums (NIDA/NIH 1999 workshop; 2nd International Workshop on Clinical Pharmacology of HIV Therapy, Noordwijk, the Netherlands, April 2001) suggests that there are significant pharmacokinetic and pharmacodynamic interactions among various anti-infective medications used in the treatment of infections including HIV, hepatitis C, and tuberculosis, resulting in significant potential for adverse health consequences. However, very few studies have been published or presented on interactions between drugs of abuse (e.g., cocaine, marijuana, opiates, amphetamines, and misused benzodiazepines), agents used to treat substance abuse (e.g., methadone, LAAM, buprenorphine), and antiretroviral pharmacotherapeutic agents. Based on these facts, Dr. Charles Flexner of Johns Hopkins School of Medicine and Jag Khalsa, Ph.D. of NIDA's Center on AIDS and Other Medical Consequences of Drug Abuse (CAMCODA), in collaboration with Dr. Frank Vocci of the Division of Treatment and Research Development (DTRD) and Dr. Steve Gust of the Office of the Director (OD, NIDA), organized a mini-symposium at the 3rd International Workshop on Clinical Pharmacology of HIV Therapy, Washington, DC, April 13, 2002, where experts in HIV pharmacology, substance abuse, pharmacogenomics, and pharmacokinetic modeling presented current data on the subject of drug-drug interactions. The article below summarizes the six presentations made at the well-attended mini-symposium.


Charles W. Flexner, MD
Johns Hopkins University School of Medicine, Baltimore, Maryland

Dr. Flexner opened the symposium by describing a case that raised several issues the meeting aimed to address. A 34-year-old African-American woman on methadone maintenance therapy began a regimen of nelfinavir plus Combivir (AZT/3TC) for HIV infection. Before long she reported opioid withdrawal symptoms to her clinician. She also revealed that she began taking Chinese herbs when starting antiretroviral therapy because a friend told her they would make the antiretrovirals more tolerable. Did her withdrawal symptoms represent?

  1. An interaction between the antiretrovirals and methadone
  2. A drug interaction mediated by the Chinese herbs
  3. An attempt to get a higher dose of methadone

Answering that question remains difficult because people with HIV infection typically take many medications. Documenting all the interactions would be a formidable task, according to Dr. Flexner. Among 45 recent discharges from The Johns Hopkins Hospital, Dr. Flexner noted, the number of pharmacologically active medications averaged 7 per patient. Among those discharged while taking a protease inhibitor (PI), the number of medications averaged 8.6 per patient.

Three overarching questions

Dr. Flexner posed three overarching questions about drug-drug interactions in people taking antiretrovirals and other medications. He outlined some possible answers:

  1. How to study the problem
    • In-vitro models
    • In-vivo models
    • Observational versus prospective studies
    • Database and population pharmacokinetics
    He suggested that observational studies have become more attractive because mounting prospective studies of all potential interactions would be virtually impossible.
  2. How to manage the problem
    • Putting more information in each drug's package insert
    • Disseminating information via the Web
    • Proposing general guidelines versus striving for individualized patient management
    • Practicing therapeutic drug monitoring‚Äč
    Dr. Flexner doubted that more information in package inserts would help because the existing information is often overlooked. He endorsed a move toward individualized management versus general treatment guidelines.
  3. How to address special topics related to substance abuse
    • Interactions between prescription medications and illicit drugs
    • Interactions between complementary and alternative medications and prescription drugs
    • Contaminants of illicit drugs
    • Metabolites and enantiomers
    • Role of plasma protein binding
    • Role of drug transport proteins
    • Pharmacogenomics
    Research has produced little information on interactions between prescription and illicit drugs, Dr. Flexner noted. Although the role of drug transporters may have an important impact on antiretroviral treatment outcome, these issues remain complicated and difficult to study.

Effect of staggered versus simultaneous PI dosing

Not all drug interactions are harmful, Dr. Flexner reminded attendees. Some can be beneficial; many are complex. One element of that complexity is whether the timing of drug administration affects two-way drug interactions. AIDS Clinical Trials Group (ACTG) investigators addressed that question in a study of staggered versus simultaneous administration of nelfinavir, saquinavir, and ritonavir. ACTG Protocol 378 enrolled 18 healthy volunteers and had a complex design involving six treatment periods for each person. The ACTG team reported several findings:

  • When saquinavir was given 4 hours before ritonavir, saquinavir's area under the concentration-time curve (AUC) was reduced by 63% compared with saquinavir's AUC when the PIs were given simultaneously.
  • When saquinavir was given 4 hours before nelfinavir, saquinavir's AUC was reduced by 68% compared with saquinavir's AUC when the PIs were given simultaneously.
  • When nelfinavir was given 4 hours before saquinavir, saquinavir's AUC was increased by 76% compared with when the PIs were given simultaneously.
  • Staggered dosing did not affect the AUCs of ritonavir or nelfinavir.

When given 48 hours before a second PI, Dr. Flexner added, ritonavir boosted concentrations of the later PI even though ritonavir had disappeared from the circulation.

Trying to correct interactions by dose adjustment

Another ACTG study, Protocol 365, determined whether interactions between rifabutin and the PI indinavir could be corrected by adjusting doses of those drugs. Rifabutin at a dose of 300 mg once daily decreases the AUC of indinavir by 32% when the PI is given at its standard dose of 800 mg every 8 hours. Co-administration at those doses triples rifabutin's AUC. ACTG 365 tested 1000 mg of indinavir every 8 hours given with 150 mg of rifabutin once daily in 10 people with HIV infection and in 18 healthy volunteers. The study yielded the following findings.

  • The 1000-mg dose of indinavir coadministered with the 150-mg dose of rifabutin produced an indinavir AUC similar to that of standard-dose indinavir without rifabutin.
  • The 150-mg dose of rifabutin coadministered with the higher dose of indinavir produced a rifabutin AUC 60% higher and a DA-rifabutin (major metabolite) AUC 125% higher than the AUCs of standard-dose rifabutin without indinavir.
  • The clinical significance of the higher rifabutin and DA-rifabutin concentrations is unknown.

The take-home message from this study, Dr. Flexner proposed, is that dose adjustment can compensate for only some unwanted drug-drug interactions. Many interactions are unavoidable. For example, it may not be possible to adjust an antiretroviral dose that causes a lower methadone concentration. But drug interactions can be better understood, he concluded, and many can be managed better.

The HIV-Infected Substance Abuse Patient as a Research Subject

Gerald H. Friedland, MD
Yale University School of Medicine, New Haven, Connecticut

Baseline demographic features, including geography, can sharply separate HIV-infected injection drug users (IDUs) from the population of HIV-infected gay men. Dr. Gerald Friedland illustrated that point by showing two maps of AIDS incidence in New York City in the 1980s with zip codes of highest proportions of new AIDS cases. One showed incident cases of AIDS in IDUs, the other in gay men. Except for one or two neighborhoods, there was no overlap. This underlies the great differences in race, sex, socioeconomic status among these two populations of people with HIV disease in the United States.

The estimated number of IDUs with AIDS rose from about 50,000 in 1991 to about 110,000 in 2000. Dr. Friedland expects that increase to continue. HIV affects IDUs beyond the US as well. IDU is the major risk for HIV infection in parts of Latin America, Europe and Asia, and high percentages of IDUs in these areas have HIV infection. IVDU and HIV infection as a consequence has now appeared in parts of southern Africa.

Under-representation of IDUs in clinical trials

Despite representing a high proportion of infected people on every continent, IDUs have not been proportionately represented in clinical trials. As a result, Dr. Friedland noted, problems unique to drug use and antiretroviral therapy have not been well studied. The same is largely true for diseases specific to substance abusers.

Dr. Friedland traced the low rate of IDU participation in clinical trials to four factors.

  • Structural, demographic, or attitudinal concerns
  • Behavioral issues
  • Sexual or reproductive issues (a reasonable proportion of IDUs are women, who were excluded from many early trials)
  • Medical exclusion criteria

The rationale for excluding IDUs from trials has evolved over the past two decades, Dr. Friedland noted. At first a history of substance abuse was enough to keep an HIV-infected person out of a trial. Eventually that proscription changed to "active substance abuse," usually as judged by the investigator. Most recently, exclusion from trials has rested on the investigator's anticipation of a person's poor compliance with the study protocol.

Dr. Friedland listed the percentage of IDUs enrolled in a few North American and European trials, as well as in studies conducted by two US research groups, the AIDS Clinical Trials Group (ACTG) and the Community Programs for Clinical Research on AIDS (CPCRA).

Percentage of IDUs enrolled in specific trials and in studies of two trials groups

Network or trial Percentage of IDUs enrolled
ACTG 13.2
CPCRA 23.3
Delta (Europe) 12 (n = 3207)
Caesar (Canada, Australia, Europe, South Africa) 13 (n = 1840)
DuPont 006 (US, Canada, Germany, Puerto Rico) 12 (n = 1266)

Those percentages do not match the proportion HIV-infected people in the US who acquired their infection by injection drug use - about 30%. Despite the low percentage of IDUs in the Delta study, Dr. Friedland noted, IDUs make up about 20% to 30% of more recent European trial populations, proportions reflecting the high percentage of HIV-infected IDUs in southern Europe.

Burden of comorbidities among IDUs

Dr. Friedland outlined the health status and comorbidities of IDUs:

  • The population is heterogeneous, but most IDUs have a low socioeconomic status.
  • Overall age-specific morbidity and mortality among IDUs is 10 to 20 times higher than among non-IDUs.
  • IDUs have an increased rate of a wide array of diseases, including:
    • Trauma
    • Neurologic disorders
    • Hepatic, renal, and pulmonary diseases
    • Psychiatric problems
    • Drug-related complications
    • Infectious diseases

In one survey of 375 IDUs, 24% had been diagnosed at least once with skin abscesses, 23% with pneumonia, and 10% with endocarditis.

Approximately 20% to 50% of people with HIV/AIDS have severe mental illness, five times the rate in the general population. The prevalence of mental illness is even higher among substance abusers with HIV infection. From 20% to 40% of HIV-infected substance abusers suffer from major depression.

A survey of a cohort of 4042 HIV-infected people selected to represent the entire US HIV population documented high rates of substance abuse [1]. About 45% of the cohort had a history of drug abuse; over 30% had used marijuana; over 20% had injected drugs; and about 10% had abused sedatives, cocaine, or alcohol. More than 35% had at least one mental disorder, and over 25% had depression or received mental health care. Approximately 10% had a history of generalized anxiety disorder or panic disorder.

The IDU-HIV-HCV triad

In many IDU populations, substance abuse, HIV infection, and hepatitis C virus (HCV) infection constitute a challenging triad, with these characteristics:

  • The prevalence of HCV infection ranges from 50% to 80% among IDUs.
  • Hepatic disease has become the major cause of morbidity and mortality among HIV-infected IDUs.
  • HCV accelerates the natural history of HIV infection.
  • HIV accelerates the natural history of HCV infection.
  • Treatment of HIV infection complicates treatment for hepatitis, and vice versa.

Treatment of both HIV infection and HCV infection improved dramatically with the introduction of potent antiretroviral combinations and interferon plus ribavirin. But research has yet to yield much information on coinfection with these two viruses, partly because hepatitis often excludes HIV-infected people from antiretroviral trials. At the same time, active drug use excludes people from HCV therapy trials. Dr. Friedland called for more research of coinfected individuals.

Pain management remains a problem in people with AIDS and can be a particular problem for IDUs. One study of nearly 600 people with AIDS and 600 with cancer determined that 16% of the AIDS patients had adequate pain control compared with 58% of the cancer patients [2]. Poor pain management breeds distrust in patients and inclines them to avoid clinical trials.

Treatment adherence among IDUs

Injection drug use can negatively influence adherence with treatment. In Vancouver injection drug use decreased adherence by 58%, and poor adherence correlated with increased mortality [3]. Another study isolated two independent predictors of antiretroviral treatment acceptance, trust in physicians and trust in medications. In this study a survey of 200 IDUs recorded the following findings:

  • 77% believe antiretrovirals to be harmful when taken with heroin or cocaine.
  • 67% will not take antiretrovirals if planning to get high on drugs.
  • 61% believe antiretrovirals are harmful when taken with methadone.
  • 60% have seen people suffer side effects and die after starting antiretrovirals.
  • 59% have seen people suffer side effects and die after stopping street drugs.

Other work has confirmed the intuitive assumption that the tolerability of antiretrovirals directly affects the success of these regimens [4]. And unfortunately many clinical problems among substance abusers overlap antiretroviral side effects:

Overlap between substance abuse problems and antiretroviral toxicities

Clinical problem/toxicity Antiretroviral involved
Psychiatric Efavirenz, AZT
Gastrointestinal Protease inhibitors, AZT
Hepatic Protease inhibitors and NNRTIs
Neurologic ddI, d4T
Renal Indinavir
Hematologic AZT

Design issues and recommendations

Dr. Friedland outlined nine design and methodologic issues that must be considered when planning drug interaction research addressing the needs of HIV-infected substance abusers:

  • In-vitro versus In-vivo studies
  • Within-subject versus between-subject studies
  • Opiate-dependent versus non-opiate-dependent trial participants
  • Single-dose versus steady-state pharmacokinetic studies
  • Restriction of other medications (including smoking and coffee)
  • Inpatient versus outpatient studies
  • Urine drug screening
  • Payment and coercion
  • Intravenous access

Dr. Friedland closed with the following recommendations:

  1. More studies of HIV therapeutics are needed in drug users, and the array of abused substances studied should be broadened to include both prescribed and illicit substances of abuse.
  2. When incorporating substance abusers into clinical trials, sufficient numbers should be obtained to enable stratification by substance abuse status.
  3. The overall clinical care of drug users should be improved with integration of medical, psychiatric, and substance abuse services. This in turn, will bring more drug users into HIV drug interaction and therapeutic trials.
  4. Side effect and toxicity profiles of HIV therapies in drug users should be better defined.
  5. The cadre of interested and expert researchers who bridge both substance abuse and HIV therapeutic areas should be increased.


  1. Turner BJ, Fleishman JA, Wenger N, et al. Effects of drug abuse and mental disorders on use and type of antiretroviral therapy in HIV-infected persons. J Gen Intern Med 2001;16:625-633.
  2. Rosenfeld B, Breitbart W, McDonald MV, et al. Pain in ambulatory AIDS patients. II: Impact of pain on psychological functioning and quality of life.Pain 1996;68:323-328.
  3. Altice FL, Mostashari F, Friedland GH. Trust and the acceptance of and adherence to antiretroviral therapy. J Acquir Immune Defic Syndr 2001;28:47-58.
  4. Girard PM, Guiguet M, Bollens D, et al. Long-term outcome and treatment modifications in a prospective cohort of human immunodeficiency virus type 1-infected patients on triple-drug antiretroviral regimens. Clin Infect Dis 2000;31:987-994.

Antiretroviral Therapy in Injection Drug Users With HIV Disease: Using Drug Interactions to Design More Effective Treatment

Elinore F. McCance-Katz, MD, PhD
Albert Einstein College of Medicine of Yeshiva University and Montefiore Medical Center, New York, New York

There are between 800,000 and 1 million heroin users in the US, most of them injection drug users (IDUs). Dr. McCance-Katz outlined three treatment options for heroin addiction:

  1. Detoxification (which fails in more than 90%)
  2. Naltrexone (opiate antagonist) maintenance
  3. Opiate agonist maintenance
    • Methadone
    • 1-acetyl-methadol (LAAM)
    • Buprenorphine (not currently available in the US)

Only about 185,000 US heroin users are enrolled in a maintenance program.

IDUs account for 31% of prevalent HIV infections in the US. In northeastern urban areas they account for more than 50%. The treatment of choice for most of these people is long-term maintenance therapy.

For IDUs, consistent adherence with antiretroviral therapy is not possible without treatment for substance abuse. Opiate maintenance therapy stabilizes a chaotic lifestyle and provides a platform for other interventions, such as modified directly observed antiretroviral therapy. But Dr. McCance-Katz noted that cotreatment of HIV and substance abuse is complicated by interactions between many antiretrovirals and the drugs used for opiate maintenance. Opioids, protease inhibitors (PIs), and nonnucleoside reverse transcriptase inhibitors (NNRTIs) are all metabolized via isoforms of the cytochrome P450 (CYP) system. These interactions have several possible consequences in people on opiate maintenance therapy:

  • Nonadherence with antiretrovirals
  • Emergence of resistance to antiretrovirals
  • Lack of efficacy of antiretroviral and opiate maintenance therapy
  • Drug toxicity

AZT and NNRTI interactions with opioids

Most antiretroviral-opioid interaction studies reported to date involve methadone. Several years ago Dr. McCance-Katz showed increased AZT concentrations in people on methadone maintenance, because methadone inhibits the metabolism and renal clearance of AZT [1]. These findings raised concerns about AZT toxicity in people taking methadone.

To determine whether the AZT-methadone interaction extends to other opioids, Dr. McCance-Katz gave AZT to HIV-negative adults taking LAAM, buprenorphine, or naltrexone, and to HIV-negative controls not on opiate maintenance therapy [2]. Among people taking naltrexone, AZT levels were similar to those in controls. Burprenorphine and LAAM significantly decreased concentrations of AZT.

Further work by Dr. McCance-Katz showed that efavirenz given by directly observed therapy significantly lowered methadone's area under the concentration-time curve (AUC) by about 50%. Study participants experienced substantial opioid withdrawal symptoms.

Delavirdine, an inhibitor of the CYP3A4 isoform, modestly elevated methadone's AUC, without clinical consequences. Delavirdine significantly raised concentrations of LAAM and nor-LAAM, the major metabolite of LAAM. But delavirdine significantly lowered levels of dinor-LAAM, the second major LAAM metabolite. Dr. McCance-Katz concluded that delavirdine could be used by patients on opioid maintenance therapy without dose adjustment.

PI interactions with opioids

Methadone levels were significantly lower when coadministered with nelfinavir, [3] a finding confirming the work of others. Even so, Dr. McCance-Katz said, the methadone dose rarely needs to be adjusted when given with nelfinavir. Nelfinavir had little effect on LAAM concentrations. Nor-LAAM levels were higher when coadministered with nelfinavir, but dinor-LAAM levels were lower.

Methadone and LAAM had little effect on nelfinavir concentrations. LAAM increased levels of M8, nelfinavir's major metabolite, but that increase is not clinically meaningful.

In another study methadone levels fell 30% to 40% when given with lopinavir/ritonavir (Kaletra) and resulted in a significant increase in opioid withdrawal symptoms.

Clinical implications and conclusions

Dr. McCance-Katz concluded that differences between In-vitro and In-vivofindings on antiretroviral-opioid interactions underscore the need to conduct pharmacokinetic studies in humans when possible. Findings of such studies can help select an appropriate opioid and appropriate antiretrovirals, she added, stressing that treatment must be individualized based on clinical needs of the patient.


  1. McCance-Katz EF, Rainey PM, Jatlow P, Friedland G. Methadone effects on zidovudine disposition (AIDS Clinical Trials Group 262). JAIDS 1998;18:435-443.
  2. McCance-Katz EF, Rainey PM, Friedland G, et al. Effect of opioid dependence pharmacotherapies on zidovudine disposition. Am J Addict 2001;10:296-307.
  3. McCance-Katz EF, Farber S, Selwyn PA, O'Connor A. Decrease in methadone levels with nelfinavir mesylate. Am J Psychiatry 2000;157:481.
  4. McCance-Katz EF, Gourevich MN, Arnsten J, Sarlo J, Rainey P, Jatlow P: Modified Directly Observed Therapy (MDOT) For Injection Drug Users with HIV Disease. Am J Addiction, in press.

Interactions Between Methadone and Antiretroviral Medications

John G. Gerber, MD
University of Colorado Health Sciences Center, Denver, Colorado

Methadone remains the drug most commonly used to treat narcotic addiction because it has withstood the test of time for safety and efficacy. Of its two isomers, R-methadone and S-methadone, only R-methadone is active. As a result, Dr. Gerber explained, methadone interaction studies must stereoselectively determine the effect of the interacting drug on these isomers.

Are In-vitro metabolism predictions reliable?

The metabolism of methadone was originally attributed primarily to the cytochrome P450 (CYP) 3A4 isoform. All protease inhibitors (PIs) inhibit CYP3A4 in the following order of potency:

Ritonavir > indinavir = nelfinavir = amprenavir > saquinavir > lopinavir

Among the nonnucleoside reverse transcriptase inhibitors (NNRTIs), delavirdine inhibits CYP3A4 as well as CYP2C19. Efavirenz induces CYP3A4 and CYP2B6 while inhibiting CYP2C19. Nevirapine induces CYP3A4 and CYP2B6, and the antimycobacterial rifabutin induces CYP3A4.

On the basis of In-vitro estimates:

  • Ritonavir should inhibit methadone metabolism, possibly causing toxicity.
  • Rifabutin, efavirenz, and nevirapine should induce methadone metabolism.

But studies of people on methadone maintenance therapy contradicted some of these predictions. Although methadone maintenance patients in a rifabutin interaction study complained of opioid withdrawal symptoms, rifabutin actually had no effect on methadone concentrations [1]. Apparently study participants perceived withdrawal symptoms because they had been cautioned that they might experience withdrawal.

In AIDS Clinical Trials Group (ACTG) protocol 401, Dr. Gerber and colleagues measured 24-hour concentrations of methadone in 12 people before and 14 days after they began ritonavir/saquinavir at a dose of 400/400 mg twice daily [2]. The PIs lowered the area under the concentration-time curve (AUC) of both R- and S-methadone. But when results were corrected for plasma protein binding, the decrease in methadone exposure was much less. Concentrations of the inactive S-methadone decreased slightly more than concentrations of the active R-methadone, but these decreases were not large enough to cause opioid withdrawal symptoms. The ACTG 401 team concluded that methadone metabolism does not depend mainly on CYP3A4.

A study of methadone-nelfinavir interactions also found a decrease in concentrations of R-methadone and S-methadone [3]. Again, the decrease was greater for the inactive S-methadone, and study participants had no opioid withdrawal symptoms.

A study of 12 HIV-seronegative individuals on methadone maintenance compared the effects of indinavir for 8 days with the effects of placebo in a crossover design [4]. Indinavir did not affect methadone levels, while methadone decreased the maximum and increased the minimum concentrations of indinavir without any changes in AUC.

In methadone maintained patients amprenavir had little effect on concentrations of active R-methadone while significantly lowering exposure to inactive S-methadone [5]. Study participants experienced no withdrawal symptoms.

At steady state lopinavir/ritonavir decreased single dose methadone exposure, but this study did not measure levels of the R and S isomers [6]. Another study confirmed lower methadone concentrations with lopinavir/ritonavir but also failed to measure R-methadone and S-methadone [7]. Study participants were receiving methadone maintenance and had no narcotic withdrawal symptoms.

The CYP3A4 inducer efavirenz significantly lowered methadone concentrations in people on maintenance therapy and withdrawal symptoms were common [8]. In another study nevirapine had similar effects [9].

Making sense of the In-vivo data

On the basis of these findings, Dr. Gerber concluded that CYP3A4 could not be the major cytochrome P450 isoform involved in methadone metabolism. Neither rifabutin, a CYP3A4 inducer, nor indinavir, a CYP3A4 inhibitor, had an effect on methadone levels. In addition, ritonavir, the most potent CYP3A4 inhibitor, induced rather than inhibited methadone metabolism.

Although inducer PIs accelerated methadone metabolism and the NNRTIs efavirenz and nevirapine potently induced methadone metabolism, these drugs certainly induce CYPs other than CYP3A4. For example, both efavirenz and nevirapine potently induce CYP2B6.

Which CYP is mainly responsible for methadone metabolism? Using microsomes from a baculovirus-infected insect cell system expressing human specific human CYPs, Dr. Gerber screened all the drug-metabolizing enzymes. He found that only three isoenzymes metabolized methadone to EDDP to a great extent. He then performed full enzyme kinetics to calculate Vmax and Km in order to estimate which enzyme may be most important to methadone metabolism.

Dr. Gerber determined that CYP2B6 rapidly metabolizes methadone. CYP2C19 metabolizes methadone less than CYP2B6, and CYP3A4 less than CYP2C19. He concluded that CYP2B6 is the major isoform involved in methadone metabolism. Inactive S-methadone is metabolized faster with CYP2B6. With CYP2C19, active R-methadone is metabolized faster.

Dr. Gerber proposed the following conclusions:

  1. Methadone is not mainly metabolized by CYP3A4 because:
    • Rifabutin has no effect on methadone pharmacokinetics;
    • Indinavir, a pure CYP3A4 inhibitor, has no effect on methadone pharmacokinetics;
    • Ritonavir induces rather than inhibits methadone metabolism.
  2. Methadone is mainly metabolized by CYP2B6 because:
    • Methadone drug interactions are most prominent with CYP2B6 inducers (efavirenz, nevirapine, phenobarbital);
    • In-vitro, CYP2B6 has a higher Vmax and a lower Km than other CYP isoforms; methadone metabolism is stereoselective; and livers with higher CYP2B6 expression metabolize methadone more rapidly.
  3. CYP2C19 could also play an important role in methadone metabolism.
  4. PIs are safer than NNRTIs in methadone-using patients.

Dr. Gerber added that future drug-drug interaction studies must consider not only chiral metabolism issues, protein binding displacement, and pharmacodynamic correlates to pharmacokinetic alterations, but also the key CYPs that can effectively metabolize methadone at clinically relevant substrate concentrations.


  1. Brown LS, Sawyer RC, Li R, et al. Lack of a pharmacologic interaction between rifabutin and methadone in HIV-infected former injecting drug users. Drug Alcohol Depend 1996;43:71-77.
  2. Gerber JG, Rosenkranz S, Segal Y, et al. Effect of ritonavir/saquinavir on stereoselective pharmacokinetics of methadone: results of AIDS Clinical Trials Group (ACTG) 401. JAIDS 2001;27:153-160.
  3. Hsyu PH, Lillibridge JH, Maroldo L, et al. Pharmacokinetic and pharmacodynamic interactions between nelfinavir and methadone. 7th Conference on Retroviruses and Opportunistic Infections. January 30- February 2, 2000, San Francisco. Abstract 245.
  4. Cantilena L, McCrea J, Blazes D, et al. Lack of a pharmacokinetic interaction between indinavir and methadone. Clin Pharmacol Ther 1999;65:135. Abstract PI-74.
  5. Hendrix C, Wakeford J, Wire MR, et al. Pharmacokinetic and pharmacodynamic evaluation of methadone enantiomers following co-administration with amprenavir in opioid-dependent subjects. 40th Interscience Conference on Antimicrobial Agents and Chemotherapy.September 17-20, 2000, Toronto. Abstract 1649.
  6. Bertz R, Hsu A, Lam W, et al. Pharmacokinetic interactions between lopinavir/ritonavir (ABT-378r) and other non-HIV drugs. AIDS 2000;14(suppl 4):S100. Abstract P291.
  7. Clarke S, Mulcahy F, Bergin C, et al. Absence of opioid withdrawal symptoms in patients receiving methadone and the protease inhibitor lopinavir-ritonavir.Clin Infect Dis 2002;34:1143-1145.
  8. Clarke SM, Mulcahy FM, Tjia J, et al. The pharmacokinetics of methadone in HIV-positive patients receiving the non-nucleoside reverse transcriptase inhibitor efavirenz. Br J Pharmacol 2000;51:213-217.
  9. Clarke SM, Mulcahy FM, Tjia J, et al. Pharmacokinetic interactions of nevirapine and methadone and guidelines for use of nevirapine to treat injection drug users. Clin Infect Dis 2001;33:1595-1597.

Mechanisms for Pharmacokinetic and Pharmacodynamic Interactions Between Psychoactive and Antiretroviral Medications

David J. Greenblatt, MD
Tufts University School of Medicine, and Tufts-New England Medical Center, Boston, Massachusetts

Antiretroviral drugs complicate the management of drug abusers for several reasons:

  • Antiretrovirals can influence their own kinetics.
  • They can enhance the abuseability and toxicity of abused drugs.
  • They can enhance the abuseability and toxicity of agents used to treat drug abuse.
  • They can diminish the effect of agents used to treat substance abuse or psychiatric disorders.

The dilemma facing pharmacologists studying these interactions is that there are too many interactions to be studied clinically. One solution to this dilemma is refining In-vitro methods to study drug-drug interactions [1,2].

What can be learned from In-vitro studies?

Dr. Greenblatt began his consideration of In-vitro studies involving antiretrovirals and psychotropic agents with the example of triazolam, a triazolo-benzodiazepine hypnotic drug. Pharmacologic research predicted that the chance of interactions with triazolam is high if the plasma concentrations of the inhibitor greatly exceed the 50% inhibitory concentration (IC50). That prediction proved true in a study assessing ritonavir-triazolam interactions [3,4]. Ritonavir significantly increased and prolonged the concentration of triazolam. Increased triazolam levels can significantly elevate beta EEG amplitude and produce excessive sedation.

To further demonstrate interactions between antiretrovirals and psychotropic agents, Dr. Greenblatt described a four-part study involving trazadone and ritonavir at a dose of ritonavir high enough to inhibit cytochrome P450 (CYP) 3A but not to induce it.

Plan for a trazadone-ritonavir interaction study

Study arm Trazadone Ritonavir
Arm 1 50 mg Placebo
Arm 2 Placebo 200 mg x 4
Arm 3 50 mg 200 mg x 4
Arm 4 Placebo Placebo

The study showed that ritonavir significantly increased the area under the concentration-time curve (AUC) of trazadone, whereas trazadone did not affect plasma levels of ritonavir.

Potential problems with In-vitro studies

Dr. Greenblatt noted that these studies show how In-vitro data can yield reasonable clinical predictions. But certain factors may complicate the analysis of In-vitro interaction studies:

  • Substrate/inhibitor binding and depletion [5,6]
  • Mechanism-based inhibition [7]

As they do in In-vivo, protein concentrations can lower drug concentrations in In-vitroIn-vitro protein concentrations influence:

  • Substrate kinetic parameters and estimation of inhibitory potency
  • Predictions of clinical drug interactions based on In-vitro data

Dr. Greenblatt specified two operational correlates of mechanism-based inhibition [8]:

  • Preincubation, in which the inhibitor is exposed to microsomes and cofactors before the addition of substrate, increases the potency of a mechanism-based inhibitor
  • Preincubation reduces inhibitor potency

The clinical implications of mechanism-based inhibition have not been established, Dr. Greenblatt added.

Two clinical issues that may complicate interpretation of In-vitro drug interaction studies are the role of physiologic variables and mixed induction and inhibition of CYP3A by a single drug. Ritonavir, for example, both inhibits and induces CYP3A. As a result, in the short term ritonavir inhibits the metabolism of the anxiolytic alprazolam [9]. But over the long term ritonavir induces alprazolam metabolism.

Drug transporter and animal models

Much recent attention has focused on the role of drug transporters such as P-glycoprotein (P-gp) on the disposition of antiretroviral drugs [10,11]. Because P-gp is expressed in intestinal epithelial cells, in liver and kidney, and at various blood-tissue barriers, its expression could limit antiretroviral entry into critical compartments and may especially limit the systemic availability of protease inhibitors (PIs).

Dr. Greenblatt noted that there are In-vitro models for P-gp interactions, including Caco-2 cells and LS-180 intestinal cells, a human colon carcinoma cell line [12-15]. In the LS-180 system, the PIs ritonavir, nelfinavir, and amprenavir strongly induce P-gp expression [13-15].

Animal models can also be used to understand drug interactions. Ritonavir at a dose of 20 mg/kg significantly induced CYP3A in male Sprague-Dawley rats. Sprague-Dawley rats have also been used to measure ritonavir activity at the blood-brain barrier and in intestinal mucosa. Dr. Greenblatt anticipates that this animal model could be useful in further antiretroviral studies.


  1. von Moltke LL, Greenblatt DJ, Schmider J, Wright CE, Harmatz JS and Shader RI. In vitro approaches to predicting drug interactions in vivo. Biochemical Pharmacology 1998;55:113-122.
  2. Venkatakrishnan K, von Moltke LL and Greenblatt DJ. Human drug metabolism and the cytochromes P450: application and relevance of in vitro models. Journal of Clinical Pharmacology 2001;41:1149-1179.
  3. von Moltke LL, Greenblatt DJ, Grassi JM, Granda BW, Duan SX, Fogelman SM, Daily JP, Harmatz JS and Shader RI. Protease inhibitors as inhibitors of human cytochromes P450: high risk associated with ritonavir. Journal of Clinical Pharmacology 1998;38:106-111.
  4. Greenblatt DJ, von Moltke LL, Harmatz JS, Durol ALB, Daily JP, Graf JA, Mertzanis P, Hoffman JL and Shader RI. Differential impairment of triazolam and zolpidem clearance by ritonavir. Journal of Acquired Immune Deficiency Syndromes 2000;24:129-136.
  5. Venkatakrishnan K, von Moltke LL, Obach RS and Greenblatt DJ. Microsomal binding of amitriptyline: effect on estimation of enzyme kinetic parameters in vitro. Journal of Pharmacology and Experimental Therapeutics 2000;293:343-350.
  6. Gibbs MA, Kunze KL, Howald WN and Thummel KE. Effect of inhibitor depletion on inhibitory potency: tight binding inhibition of CYP3A by clotrimazole. Drug Metab Dispos 1999;27:596-599.
  7. von Moltke LL, Durol ALB, Duan SX and Greenblatt DJ. Potent mechanism-based inhibition of human CYP3A in vitro by amprenavir and ritonavir: comparison with ketoconazole. European Journal of Clinical Pharmacology 2000;56:259-261.
  8. Silverman R. Mechanism-based enzyme inactivators. Methods in Enzymology 1995;249:240-283.
  9. Greenblatt DJ, von Moltke LL, Harmatz JS, Durol ALB, Daily JP, Graf JA, Mertzanis P, Hoffman JL and Shader RI. Alprazolam-ritonavir interaction: implications for product labeling. Clinical Pharmacology and Therapeutics 2000;67:335-341.
  10. Kim RB, Fromm MF, Wandel C, Leake B, Wood AJJ, Roden DM and Wilkinson GR. The drug transporter P-glycoprotein limits oral absorption and brain entry of HIV-1 protease inhibitors. Journal of Clinical Investigation 1998;101:289-294.
  11. Schuetz EG and Schinkel AH. Drug disposition as determined by the interplay between drug-transporting and drug-metabolizing systems. J Biochem Molecular Toxicology 1999;13:219-222.
  12. Störmer E, von Moltke LL, Perloff MD and Greenblatt DJ. P-glycoprotein interactions of nefazodone and trazodone in cell culture. Journal of Clinical Pharmacology 2001;41:708-714.
  13. Perloff MD, von Moltke LL, Marchand JE and Greenblatt DJ. Ritonavir induces P-glycoprotein expression, multidrug resistance-associated protein (MRP1) expression, and drug transporter-mediated activity in a human intestinal cell line. Journal of Pharmaceutical Sciences 2001;90:1829-1837.
  14. Perloff MD, von Moltke LL, Sturmer E, Shader RI and Greenblatt DJ. Saint John's wort: An in vitro analysis of P-glycoprotein induction due to extended exposure. British Journal of Pharmacology 2001;134:1601-1608.
  15. Perloff MD, von Moltke LL, Fahey JM, Daily JP and Greenblatt DJ. Induction of P-glycoprotein expression by HIV protease inhibitors in cell culture. AIDS 2000;14:1287-1289.

Pharmacogenomics of Antiretrovirals and Drugs of Abuse Interactions

Thomas Kakuda, PharmD
Abbott Laboratories

Pharmacogenomics-the study of genetic correlates that may explain individual differences in treatment response-has attracted the attention of HIV pharmacologists seeking to explain variable responses to antiretrovirals. The ultimate goal of such research, Dr. Kakuda proposed, is "predictive medicine": using genetic testing to predict predisposition to specific diseases or variations in response to therapy. Simply put, pharmacogenomics may improve the odds of "therapy with the right drug, at the right dose, in the right patient."

Genetic polymorphisms may affect both the pharmacokinetics and pharmacodynamics of a given drug. Alone or combined, such polymorphisms produce heterogeneity in the therapeutic response. Foe example, a recent study demonstrated the frequency of mutant for six drug metabolizing enzymes from a large ethnically diverse population [1].

Genetic polymorphisms also affect drug receptor targets for numerous pharmaceutical agents, including:

  • ACE inhibitors
  • Antiretrovirals
  • Beta agonists
  • Chemotherapeutic agents
  • Estrogen
  • Tacrine

Pharmacogenomics and drugs of abuse

Opiate addiction is a genetically complex disease. Although a large number of genes are associated with opiate addiction, phenotype is not dependent on genotype. Dr. Kakuda listed the following genetic correlates from a recent study:

Genetic correlates of opiate dependence

Gene Protein Results
CYP2D6 Cytochrome Protective effect of homozygous deletion
DRD2-4 Dopamine receptor Dependence associated with DRD4 allele 7 and DRD2 alleles A1 and B1
SLC6A4 Solute carrier Dependence associated with allele 10

Source: Lichtermann D et al [2].

Dr. Kakuda cited estimates that 40% to 60% of alcoholism is attributable to genetics. At the same time, excessive drinking correlates closely with gender and age:

Percentage of alcoholics by gender and age

Age Men (%) Women (%)
18 to 24 24 8.6
25 to 44 19.3 4.8
45 to 64 12.7 2.0
65+ 2.9 0.4

In addition, the genetics of alcohol metabolism vary by ethnic group. The metabolic process can be described as:

EtOH Æ acetaldehyde Æ acetate

Among Israeli Jews, a polymorphism in alcohol dehyrogenase (ADH2 R47H) increases formation of acetaldehyde, the compound that causes toxic reactions to alcohol [3]. Another pathway that leads to alcohol toxicity is impairment of aldehyde dehydrogenase (ALDH). The prevalence of the mutant allele (ALDH2 G487K) is ~30% among Chinese and Japanese. Because this mutant allele is dominant, heterozygotes have similar expression as homozygote mutant leading to impaired ALDH in almost half of the Chinese and Japanese population. Partly because Chinese and Japanese have such a high risk of delayed acetaldehyde metabolism, the risk of alcoholism among them has been estimated as 4- to 10-fold lower than in other populations.

Amphetamine derivatives such as MDMA ("ecstasy"), MDA ("love pills"), and MDE ("Eve") have become popular recreational drugs among some people with or at risk for HIV infection. Their interactions with antiretrovirals remain poorly characterized but case reports of fatal interactions with protease inhibitors have been documented. One randomized, double-blind, crossover study of 8 people receiving 75 of 125 mg of MDMA determined that MDMA has non-linear kinetics and that the CYP2D6 genotype (see first table above) is not associated with MDMA metabolism [4].

Cannabinoids, taken as an appetite stimulant by some people with HIV infection, are metabolized by CYP2C and CYP3A4 isoforms. But a recent study found that marijuana cigarettes and dronabinol have little effect on protease inhibitor concentrations [5].

HLA haplotype and antiretrovirals

Two recent studies associated specific HLA haplotypes with hypersensitivity to the nucleoside analog abacavir. The studies are instructive in suggesting the effect of genetic diversity on antiretroviral toxicity. A study of the Western Australia HIV Cohort linked abacavir hypersensitivity to three haplotypes with a positive predictive value of 100% and a negative predictive value of 97% [6].

HLA typing in a North American population taking abacavir found that 39 (46%) of those with hypersensitivity reactions were HLA-B57 positive [7]. Although that rate was much higher than the 4% of abacavir-tolerant people who were HLA-B57 positive, the association with hypersensitivity is much weaker than in the Australian cohort. Dr. Kakuda speculated that the Australian cohort might be much less genetically heterogeneous than the North American cohort and that the utility of HLA typing for abacavir is currently unwarranted.

A separate study by the Australian researchers found that certain HLA haplotypes independently increased or decreased the risk of eight reverse transcriptase-associated resistance mutations and seven protease-associated mutations [8].


Dr. Kakuda outlined the following genetic determinants in HIV disease:

Disease susceptibility or progression

  • Chemokines (RANTES)
  • Chemokine receptors
  • HLA polymorphisms (B*35)
  • IL-4 589T polymorphism

Response to drugs

  • CYP2D6
  • HLA-B*5701 (abacavir hypersensitivity)
  • MDR1
  • SREPB-1 (hyperlipidemia)

(Adapted from Telenti A, Aubert V, Spertini F. Individualizing HIV treatment-pharmacogenetics and immunogenetics. Lancet 2002; 359: 722-723)

Dr. Kakuda concluded that pharmacogenomics is an emerging discipline applicable to all drugs and diseases. Ultimately, pharmacogenomic research could greatly enhance individualized approaches to drug therapy. He predicted that high-throughput microarrays might eventually provide simple and quick results. But cost and availability are likely to remain important issues.


  1. Wilson JF, Weale ME, Smith AC, et al. Population genetic structure of variable drug response. Nat Genetics 2001;29:265-269.
  2. Lichtermann D, Franke P, Maier W, Rao ML. Pharmacogenomics and addiction to opiates. Eur J Pharmacol 2000;410:269-279.
  3. Radel M, Goldman D. Pharmacogenetics of alcohol response and alcoholism: the interplay of genes and environmental factors in thresholds for alcoholism.Drug Metab Dispos 2001;29(4 pt 2):489-494.
  4. de la Torre R, Farre M, Ortuno J, et al. Non-linear pharmacokinetics of MDMA ('ecstasy') in humans. Br J Clin Pharmacol 2000;49:104-109.
  5. Kosel BW, Aweeka FT, Benowitz NL, et al. The effects of cannabinoids on the pharmacokinetics of indinavir and nelfinavir. AIDS 2002;16:543-550.
  6. Mallal S, Nolan D, Witt C, et al. Association between the presence of HLA-B*5701, HLA-DR7, and HLA-DQ3 and hypersensitivity to HIV-1 reverse-transcriptase inhibitor abacavir. Lancet 2002;359:727-732.
  7. Hetherington S, Hughes AR, Mosteller M, et al. Genetic variations in HLA-B region and hypersensitivity to abacavir. Lancet 2002;359:1121-1122.
  8. Moore C, John M, James I, Mallal S. The influence of host HLA on antiretroviral drug resistance mutation in HIV-1. 9th Conference on Retroviruses and Opportunistic Infections. February 24-28, 2002, Seattle. Abstract 554.

NIDA Support for Drug Interaction Research

Jagjitsingh H. Khalsa, PhD
Center on AIDS and Other Medical Consequences of Drug Abuse (CAMCODA)
National Institute of Drug Abuse, NIH
6001 Executive Blvd., Room 5198, MSC 9593
Bethesda, Maryland
E-mail: jk98p@nih.gov

Dr. Khalsa closed the symposium by inviting attendees to contact him with queries about financial support for research on interactions involving drugs of abuse. NIDA will consider support for both US and non-US researchers. He listed the following examples of research NIDA may support:

  • Underlying mechanisms of drug interactions and metabolic pathways
  • In-vitro or In-vivo models to study drug-drug interactions
  • Factors in and mechanisms of drug induction
  • Exploratory clinical pharmacology and diagnostic screening studies
  • Interactions between drugs of abuse and pharmacotherapeutics (antiretrovirals, antibiotics) used in the treatment of infections including HIV, hepatitis, and STDs, drug addiction (e.g, methadone, LAAM, buprenorphine), co-morbid mental disorders, and oral contraceptives.