Interactions between recreational drugs and antiretroviral agents
Antoniou T, Tseng AL. HIV Program/Inner City Health, St. Michael's Hospital, Toronto, Ontario, Canada.
Ann Pharmacother 2002 Oct;36(10):1598-613
The October 2002 issue of the Annals of Pharmacotherapy contains a pretty thorough review of the literature available addressing concerns of drug interactions between recreational drugs and HIV antiretrivirals (HAART medications). Immediately below is a brief summary of the article, followed by extensive excerpts from the article.
This review collected information on HAART medications and:
--MDMA, also known as ecstasy, XTC, Adam, and Essence (Amphetamines)
--GHB (liquid ecstasy)
--Ketamine (also known as special K or kit kat)
--PCP (known on the street as angel dust, rocket fuel, or killer weed)
--Meperidine and Other Opiates
--Cocaine and Heroin
--Benzodiazepines (among the most commonly prescribed psychotropic drugs)
--Tetrahydrocannabinol (THC), the active ingredient of smoked marijuana
Brief Summary: OBJECTIVE: To summarize existing data regarding potential interactions between recreational drugs and drugs commonly used in the management of HIV-positive patients. DATA SOURCES: Information was obtained via a MEDLINE search (1966-August 2002) using the MeSH headings human immunodeficiency virus, drug interactions, cytochrome P450, medication names commonly prescribed for the management of HIV and related opportunistic infections, and names of commonly used recreational drugs. Abstracts of national and international conferences, review articles, textbooks, and references of all articles were also reviewed. STUDY SELECTION AND DATA EXTRACTION: Literature on pharmacokinetic interactions was considered for inclusion. Pertinent information was selected and summarized for discussion. In the absence of specific data, prediction of potential clinically significant interactions was based on pharmacokinetic and pharmacodynamic properties. RESULTS: All protease inhibitors (PIs) and nonnucleoside reverse transcriptase inhibitors are substrates and potent inhibitors or inducers of the cytochrome P450 system. Many classes of recreational drugs, including benzodiazepines, amphetamines, and opioids, are also metabolized by the liver and can potentially interact with antiretrovirals. Controlled interaction studies are often not available, but clinically significant interactions have been observed in a number of case reports. Overdoses secondary to interactions between the "rave" drugs methylenedioxymethamphetamine (MDMA) or gamma-hydroxybutyrate (GHB) and PIs have been reported. PIs, particularly ritonavir, may also inhibit metabolism of amphetamines, ketamine, lysergic acid diethylmide (LSD), and phencyclidine (PCP). Case series and pharmacokinetic studies suggest that nevirapine and efavirenz induce methadone metabolism, which may lead to symptoms of opiate withdrawal. A similar interaction may
exist between methadone and the PIs ritonavir and nelfinavir, although the data are less consistent. Opiate metabolism can be inhibited or induced by concomitant PIs, and patients should be monitored for signs of toxicity and/or loss of analgesia. PIs should not be coadministered with midazolam and triazolam, since prolonged sedation may occur. CONCLUSIONS: Interactions between agents commonly prescribed for patients with HIV and recreational drugs can occur, and may be associated with serious clinical consequences. Clinicians should encourage open dialog with their patients on this topic, to avoid compromising antiretroviral efficacy and increasing the risk of drug toxicity.
Comments by Jules Levin: It's important to bear in mind several points. Just because a drug interaction is found between drugs in the laboratory that does not mean such an affect will definitely occur in the patient taking the drugs. For example, although it has been found that protease inhibitors can affect the levels of methadone in patients taking HAART and methadone they do not always find methadone levels are affected. Patients have individual responses and need to be evaluated individually after starting HAART if they are taking methadone. NNRTIs efavirenz and nevirapine doses often have to be altered after starting HAART regimens containing these drugs if a person is taking methadone. There have been a number of reports of deaths of individuals taking recreational "Club" drugs along with HAART medications. But actual studies of the specific effects of these drugs together have not been conducted and probably will not be conducted. Combining such "Club" drugs with HAART medications is risky. What about Heroin and Cocaine. There have not been studies of interaction between these illicit drugs and HAART medications. Such studies are difficult to conduct and not likely to be conducted at least in the near future. However, individuals who use Heroin and are taking HAART appear to be able to achieve undetectable HIV viral load if the person is adherent. Individuals actively using IV Heroin or crack are usually not considered eligible for treatment for hepatitis C by physicians.
EXCERPTS FROM ARTICLE (pdf of full article containing tables and references can be downloaded here: PDF recreational drugs HIV).
The advent of potent new therapies for HIV has seemingly turned the tide in the battle against this disease. Specifically, combinations of antiretroviral drugs that include a member of the protease inhibitor (PI) or nonnucleoside reverse transcriptase inhibitor (NNRTI) family offer new hope that the progression of the disease and death can be delayed.1-3 However, the addition of combination therapies to already complex medication regimens dramatically increases the likelihood of drug interactions.4-7 PIs and NNRTIs, in particular, have a propensity for causing drug interactions as a result of their ability to either inhibit or induce the cytochrome P450 (CYP450) enzyme system.8-13 The effects of medications commonly used by patients with HIV/AIDS on the body's various drug-metabolizing pathways are summarized in Table 1 .8-13
While numerous interactions of varying clinical significance have been well described10-12 with these antiretrovirals, less is known about the potential for drug interactions with recreational drugs. This is an issue of concern, since drug use by injection remains a significant risk factor for the acquisition of HIV infection.14, 15 According to the Centers for Disease Control and Prevention,16 the proportion of AIDS cases in the US associated with injection drug use has increased from 12% in 1981 to 25% of all new cases through June 2001. In Canada, the proportion of new HIV infections attributable to injection drug use increased from 24% in 1987-1990 to 34% in 1999.17
A report18 of a suspected fatal interaction between ritonavir and 3,4-methylenedioxymethamphetamine (MDMA, ecstasy) has sparked demands for increased awareness and research in this area. Realistically, however, it is unlikely that pharmacokinetic interactions between drugs used in HIV pharmacotherapy and most recreational agents will be formally studied, due to legal and ethical constraints. However, it is often possible to predict potential interactions using in vitro and in vivo drug metabolism data.19 Since many recreational drugs are metabolized to some degree by the CYP450 system, it is reasonable to anticipate that concomitant use with PIs and delavirdine could possibly result in drug accumulation and/or toxicity. Similarly, treatment with enzyme inducers such as the NNRTI nevirapine may precipitate withdrawal reactions to recreational agents metabolized by the CYP450 system. Interactions between the NNRTI efavirenz and recreational drugs may be more difficult to predict, given that efavirenz can both inhibit (3A4, 2C9/19) and induce (3A4) selected isoenzymes of the CYP450 system, although induction of CYP3A4 appears to predominate over inhibition of this particular isoenzyme.7, 13
The purpose of this review is to summarize data on drug interactions between recreational drugs and antiretrovirals. In the absence of such data, the potential for an interaction is addressed based on the metabolic fate of the recreational drug. General information regarding the steps involved in drug metabolism has been reviewed elsewhere.19
Information was collected on documented or suspected interactions and metabolic pathways of both commonly prescribed HIV medications and commonly used recreational drugs. Information was retrieved via a MEDLINE search (1966-August 2002) using the MeSH headings human immunodeficiency virus, drug interactions, cytochrome P450, names of antiretrovirals, and chemical and common names of frequently used recreational drugs including methylenedioxymethamphetamine (MDMA), methamphetamine, -hydroxybutyrate (GHB), ketamine, phencyclidine (PCP), lysergic acid diethylamide (LSD), cocaine, heroin, methadone, meperidine, codeine, morphine, oxycodone, benzodiazepines, marijuana, and alcohol. Abstracts of international and national conferences, review articles, textbooks, and references of all articles were also searched. All literature on pharmacokinetic or pharmacodynamic interactions was considered for inclusion. When data on a particular combination were unavailable, a possible or potential interaction was predicted based on the metabolic fate of the involved agents.
Many prescription, nonprescription, and recreational drugs undergo extensive hepatic metabolism via CYP450 isoenzymes and/or glucuronidation. Thus, there is potential for significant interactions between these agents and antiretrovirals, particularly PIs and NNRTIs. Concentrations of many recreational drugs may be significantly increased or decreased in the presence of these antiretrovirals and may be associated with serious adverse outcomes.
MDMA, also known as ecstasy, XTC, Adam, and Essence, is a commonly used substance at all-night dance parties known as raves and is also increasingly being used recreationally by young professionals. When MDMA is taken orally as a capsule or tablet at average doses of 75-100 mg,20 users cite enhanced feelings of empathy for others, anxiolysis, and strong feelings of euphoria. MDMA is an amphetamine-like compound that undergoes demethylenation principally by CYP2D6.21-23 Concomitant administration with CYP2D6 inhibitors could lead to significant increases in MDMA exposure with potentially dangerous and even fatal consequences, as illustrated by a case report.18 Within a few hours of taking 180 mg of MDMA, a 32-year-old man with AIDS experienced symptoms suggestive of a heightened serotonergic state including tachypnea, tachycardia, cyanosis, and profuse sweating. He then experienced an apparent tonic-clonic seizure, tachypnea, and tachycardia (carotid pulse 200 beats/min), and subsequently died from cardiorespiratory arrest. This patient had previously taken similar amounts of MDMA on several occasions without adverse effects, but this was the first time he had taken MDMA since adding ritonavir 600 mg twice daily to his antiretroviral regimen. At autopsy, the patient's blood concentrations of MDMA were approximately tenfold higher than expected given the amount ingested. Since ritonavir is a well-known potent inhibitor of many hepatic isoenzymes including CYP2D6, the clinicians concluded that the patient likely experienced a fatal serotonergic reaction to MDMA as a result of an interaction with ritonavir.
The danger associated with this interaction may be magnified due to the large variability in the actual amount of MDMA between tablets and the presence of other chemicals (e.g., amphetamines, ephedrine) in some MDMA tablets whose metabolism can also be inhibited by ritonavir, leading to a life-threatening consequence.24 Thus, the combination of MDMA and ritonavir should be avoided. Other isoforms of the CYP450 system may also be involved in the metabolism of MDMA, notably 1A2, 2B6, and 3A4.23 All PIs can inhibit CYP3A4 activity to varying degrees, and ritonavir, nelfinavir, and the NNRTI efavirenz also demonstrate inhibitory activity against 2B625 ; therefore, individuals using MDMA should be warned about the potential for an interaction with these agents and advised to take appropriate precautions (e.g., use 25% of the usual amount of MDMA, take breaks from dancing, ensure rave or party has medical team on site, maintain adequate hydration by avoiding alcohol and replenishing fluids regularly).
Other amphetamines, particularly methamphetamine (crystal meth, speed), may be used at raves. These drugs are also mainly metabolized by CYP2D6.26-28 Thus, potentially dangerous interactions with ritonavir may occur, and the combination should be avoided if possible.
GHB, also known as liquid ecstasy, grievous bodily harm, or G, is a naturally occurring metabolite of the neurotransmitter -aminobutyric acid (GABA) that is used at raves for its euphoric effects. Colorless, odorless, and tasteless, GHB has also been used in the context of date rape when slipped into beverages. The pharmacokinetics of GHB have not been well characterized. The major route of elimination is expired breath as carbon dioxide, although animal data29, 30 suggest that first-pass metabolism may also play a large role in GHB clearance. Since first-pass metabolism is often mediated by the CYP450 system, it is possible that inhibitors of this system could predispose patients to GHB-related toxicity. As the precise metabolic pathway involved in the metabolism of GHB is unknown, patients who use this substance should be warned about the potential dangers of a drug interaction with PIs (especially ritonavir) and the NNRTIs delavirdine and, possibly, efavirenz.
The potential for an interaction is highlighted by a report31 of an HIV-positive patient taking ritonavir and saquinavir who developed symptoms consistent with GHB toxicity shortly after ingesting a small amount of GHB (10 mg/kg). The patient had taken GHB to counter the agitating effects of 2 MDMA tablets, which had lasted much longer (29 h) than when he had used MDMA prior to initiating antiretroviral therapy. Since the man had taken similar doses of both MDMA and GHB without incident prior to initiating therapy with ritonavir and saquinavir, the authors concluded that PI-mediated inhibition of MDMA and GHB was responsible for the adverse reactions.
Ketamine, also known as special K or kit kat, may be used at raves for its dissociative, intoxicating, and amnesic properties. Users may inhale the powder form, while ketamine liquid is usually added to drinks and ingested orally. The main route of ketamine metabolism is N-demethylation to norketamine, a metabolite with approximately one-third the anesthetic activity of its parent compound. Norketamine is then hydroxylated and conjugated to water-soluble conjugates that are excreted in the urine.32 CYP2B6 appears to be the main enzyme involved in ketamine metabolism, with 3A4 and 2C9 involved to a lesser extent.33 There are no studies or case reports describing interactions between ketamine and antiretroviral agents. However, since ritonavir, nelfinavir, and efavirenz are potent inhibitors of CYP2B6, patients who use ketamine recreationally may be at risk for ketamine toxicity due to drug accumulation. Animal studies34, 35 suggest that ketamine may be a weak inhibitor of CYP3A4, although the clinical significance of this is unclear in the absence of human data. Still, until such results can be confirmed, it may be prudent to avoid recreational ketamine use while taking drugs that are CYP3A4 substrates and have narrow safety thresholds (e.g., cisapride, terfenadine, astemizole).
PCP, known on the street as angel dust, rocket fuel, or killer weed, may be used at raves for its hallucinogenic or dissociative properties. Users may also report feelings of empowerment and invulnerability with PCP use. PCP is metabolized in the liver through oxidative hydroxylation, with up to 5 metabolites being formed. CYP3A4 appears to play a major role in the hydroxylation of PCP.36 Results from rat model studies also suggest that CYP2C11 may be involved in PCP metabolism37 and that CYP2B1 may be inhibited in vitro.38 Thus, it would be expected that concurrent use of PCP with PIs, delavirdine, and possibly efavirenz may result in elevated PCP concentrations and resultant toxicity. Patients using PCP who are also receiving treatment with antiretrovirals should be cautioned to use less than what they would normally use given the potential for a drug interaction.
LSD is also known popularly as acid or blotters, since it may be used in the form of paper microdots for its hallucinogenic and mild euphoric properties. Although the CYP450 system may be involved in the metabolism of LSD, the exact contribution of this system in overall LSD clearance and the isoenzymes involved have not been detailed.39, 40 Thus, anticipating drug interactions with LSD is extremely difficult. Patients who use LSD recreationally and who receive treatment with antiretrovirals should be cautioned about the possibility of an interaction and to be familiar with signs of LSD toxicity, and perhaps consider using a smaller amount than normal. Table 2 18, 21-23, 26-40 summarizes the interactions between rave drugs and antiretrovirals.
Since methadone is metabolized primarily by CYP3A4, with additional contributions by 2D6, 2C19, and 2B6, the likelihood of interactions with NNRTIs and PIs is high.41-44 Several such interactions have been described in the literature and are summarized in Table 3 .45-72
As expected, patients maintained on methadone who are subsequently treated with either efavirenz or nevirapine are at risk of developing opiate withdrawal symptoms due to NNRTI-mediated enzyme induction. Such patients may require an increase in their methadone dose, although the magnitude of the dose increase may not always parallel the reduction in total methadone exposure. For example, data reported by Clarke et al.45 suggest that, despite a decrease of >50% in methadone AUC seen with the addition of efavirenz, a mean increase in methadone dose of only 22% (in 10-mg increments) was required to counteract symptoms consistent with opiate withdrawal. A similar interaction has been described68 between nevirapine and methadone, in that a mean increase in methadone dose of 16% was required to compensate for a 50% reduction in methadone AUC.
Interactions between PIs and methadone have been even less predictable. In vitro, the AUC for methadone increased twofold when the drug was administered with ritonavir and 30% when administered with indinavir.73 A later study74 in healthy volunteers did not confirm these findings, noting a decrease in the AUC of methadone of 36% with concomitant ritonavir. However, these results are somewhat limited since only a single 5-mg dose of methadone was studied. Similarly, reduced methadone concentrations have been noted in the presence of lopinavir/ritonavir70 and nelfinavir.53 These observations suggest that ritonavir, nelfinavir, and possibly lopinavir may be inducing an alternative route of methadone metabolism.52, 57, 58
Reduced methadone concentrations have not always been accompanied by symptoms of opiate withdrawal. This lack of correlation between concentrations and clinical withdrawal may be related to a disproportionately larger induction in the metabolism of methadone's inactive S(+)-enantiomer as opposed to the R()-enantiomer, which harbors essentially all opiate activity.54 Further studies need to be conducted comparing methadone with PIs to better clarify the nature of these interactions. Clinicians should be prepared for the possibility that some patients stabilized on methadone might require a dose increase when nelfinavir or ritonavir is introduced.
Interactions between methadone and the nucleoside reverse transcriptase inhibitors zidovudine, didanosine, and stavudine have also been described.59-61 Overall, methadone appears to increase total exposure to zidovudine. The mechanisms underlying this interaction appear to involve inhibition of zidovudine glucuronidation and, to a lesser extent, decreased renal clearance of zidovudine. Although the clinical implications of these findings are unclear, patients receiving the combination of methadone and zidovudine should be monitored for zidovudine-related toxicities such as nausea, vomiting, headaches, and myelosuppression.60, 61 Since many of these symptoms may mirror those of opiate withdrawal, patients may confuse the symptoms of zidovudine toxicity with a requirement for a higher methadone dose. However, methadone concentrations do not appear to be altered by concomitant zidovudine administration, thereby discounting the association of such symptoms with opiate withdrawal.
In contrast to zidovudine, methadone appears to decrease concentrations of both stavudine and didanosine (buffered tablet formulation), possibly by delaying absorption of these agents and thereby allowing enhanced time for enzymatic or acid-catalyzed degradation. Since didanosine is more prone to acid-catalyzed degradation than is stavudine, the impact of methadone on didanosine concentrations is more pronounced than for stavudine.59 This theory is corroborated by recent evidence72 which indicates that the impact of methadone on didanosine concentrations is negligible when didanosine is administered as an enteric-coated capsule preparation, as such a coating would be expected to protect the drug from degradation until it has cleared the stomach. As well, the significance of reductions in didanosine concentrations is unclear, since intracellular concentrations of dideoxyadenosine triphosphate were not measured, and neither virologic nor immunologic outcomes were addressed. Although an increase in the dose of didanosine may be necessary when the buffered tablet formulation is taken with methadone, there are currently no guidelines for dosage adjustment.
As well as being a substrate of the CYP450 system, methadone can also act as an inhibitor of the 2D6 and 3A isoforms.75-77 It is therefore possible that concomitant use of methadone and PIs or NNRTIs may result in increased antiretroviral concentrations and predispose patients to drug-specific adverse events. However, methadone did not alter the pharmacokinetics of delavirdine, a CYP3A4 substrate.63 In addition, aside from a reduction in concentrations of the pharmacologically active M8 metabolite, significant changes to the pharmacokinetics of nelfinavir were not observed with concomitant administration of methadone.65 The metabolism of nelfinavir to its M8 metabolite is mediated by CYP2C19, suggesting that methadone may inhibit this isoenzyme as well. Although virologically active, a reduction in M8 concentrations does not appear to be clinically significant.78 Thus, significant elevations in the concentrations of PIs and NNRTIs may not occur with methadone. Still, the impact of methadone on other members of these classes is unknown and, as with zidovudine, it may be difficult to discriminate between symptoms associated with PI toxicity (e.g., nausea, vomiting, diarrhea) and methadone withdrawal. However, since enzyme inhibition is an acute process, while enzyme induction occurs following several days of drug administration, it may be possible to distinguish the 2 interactions based on the time course of symptom development. Specifically, symptoms that develop within 2-3 days of concomitant administration may be due to PI toxicity, whereas those that develop after 6 days are more likely to be related to opiate withdrawal.
Meperidine and Other Opiates
Two pathways are involved in meperidine metabolism: hydrolysis to meperidinic acid by liver carboxylesterases and demethylation to normeperidine by microsomal enzymes. Demethylation to normeperidine may be mediated by the CYP450 system, although the exact isoenzyme involved is unknown.79, 80 In patients with renal failure or with frequent dosing, normeperidine can accumulate, leading to central nervous system (CNS) excitatory toxicity.
In an open-label study,81 8 HIV-negative volunteers received meperidine 50 mg prior to treatment and following 10 days of treatment with escalating doses of ritonavir. Meperidine AUC decreased 67% in the presence of ritonavir (p < 0.005), while normeperidine AUC increased 47%, suggesting that ritonavir induces the metabolism of meperidine to normeperidine. However, since normeperidine has some pharmacologic activity, the potential for decreased analgesic effect and risk of opiate withdrawal may be lessened. On the other hand, because normeperidine possesses CNS excitatory effects, patients who use meperidine and ritonavir concomitantly may be at increased risk for seizures. Patients with renal failure may also be at increased risk for CNS excitatory toxicity due to normeperidine accumulation.
Reports detailing interactions between antiretrovirals and commonly used opiate analgesics such as codeine, morphine, or oxycodone are lacking. Postulated interactions between these opiates and antiretrovirals are described in Table 4 .79-91
Cocaine and Heroin
The significant role played by cocaine in the transmission of HIV cannot be underestimated. While injecting cocaine or heroin puts users at risk of acquiring HIV through contaminated syringes, smoking "crack" cocaine may independently be associated with acquisition of HIV infection through its association with high-risk sexual practices such as the exchange of drugs for sex.92-94 Since patients who acquire HIV in the context of crack or cocaine use may continue their drug use practices, an understanding of the potential for interactions with antiretrovirals is important.
Cocaine is metabolized chiefly by 1 of 3 pathways.95 Spontaneous hydrolysis of cocaine to benzoylecgonine accounts for approximately 39%, 30%, and 16% of a single dose of cocaine administered by the intravenous, intranasal, and smoked routes, respectively.96 Degradation by serum and hepatic cholinesterases to ecgonine methyl ester can account for up to 32-49% of an administered cocaine dose.95, 97 Finally, N-demethylation to norcocaine, mediated by CYP3A4, makes up <10% of cocaine's biotransformation.95, 98, 99 Other metabolites (e.g., anhydroecgonine methyl ester, p-hydroxy cocaine) are also produced in the metabolism of cocaine, although in smaller amounts.
Interactions between cocaine and antiretrovirals have not been described. Theoretically, inhibition of CYP3A4 may increase concentrations of the parent compound by blocking a route of cocaine metabolism. However, given that N-demethylation is a relatively small component of cocaine metabolism, such an interaction would not be expected to increase the risk of cocaine toxicity. An exception may occur in patients who are also cholinesterase deficient, since they lack the complementary enzymes necessary to metabolize the excess cocaine burden.100
Inhibition of the CYP3A4 isoform would consequently result in decreased production of norcocaine; norcocaine is thought to play a critical role in mediating the hepatotoxicity of cocaine.101, 102 In vitro studies103 documenting the protective effect of 3A4 inhibitors against cocaine-elicited hepatotoxicity lend credence to this notion. Thus, it is possible that inhibition of CYP3A4 by some antiretrovirals may theoretically ameliorate the hepatotoxicity associated with cocaine, although it should be stressed that there are no clinical data to support this. Furthermore, such postulated effects may not be clinically significant in the context of other factors, such as concomitant hepatitis B or C infection.
However, if inhibition of CYP3A4 is theoretically protective against cocaine-mediated liver injury, the reverse may be true. That is, induction of CYP3A4 by nevirapine or efavirenz may lead to increasing amounts of norcocaine being formed, potentially increasing the risk of hepatotoxicity. Again, further research is necessary to clarify the nature and consequences of interactions between enzyme inducers and cocaine.
Heroin is rapidly metabolized to 6-monoacetylmorphine and morphine by plasma and liver esterases, respectively. Maximal blood concentrations of heroin and 6-monoacetylmorphine are attained within minutes and are cleared rapidly, while morphine concentrations increase and decrease more slowly.104-107 Thus, potential interactions of concern may be similar to those noted with morphine (Table 4) .
Benzodiazepines remain among the most commonly prescribed psychotropic drugs. In Canada, the overall prevalence of benzodiazepine use for anxiolysis in the 1990s was estimated at roughly 8% of the adult population, while about 2.5% of adults were prescribed this group of drugs for insomnia.108 Benzodiazepines may be used recreationally either alone or, more commonly, in the setting of multiple drug abuse. Potential abuses of benzodiazepines include moderating the effects of stimulants, allaying withdrawal symptoms from other recreational substances, acting as disinhibitory agents, or augmenting the effects of other recreational drugs. As a class, benzodiazepines are extensively metabolized by the liver, with individual agents metabolized predominantly by either the CYP450 system or glucuronyltransferases.
Midazolam, triazolam, and alprazolam are metabolized mainly by CYP3A4.109, 110 Interactions with PIs, delavirdine, and, possibly, efavirenz are thus likely to produce increased concentrations of these compounds and place patients at risk of toxicity such as extreme sedation and respiratory depression. Pharmacokinetic studies and case reports documenting such interactions are summarized in Table 5 .111-115 It is interesting to note that conflicting data exist regarding the interaction between alprazolam and ritonavir. While Frye et al.111 noted a reduction in alprazolam exposure and relatively little change in pharmacodynamic effect following 12 days of ritonavir, subsequent work by Greenblatt et al.112 found that acute exposure to ritonavir reduced alprazolam clearance and enhanced alprazolam's pharmacodynamic properties. This discrepancy may be accounted for by the fact that ritonavir, over time, may induce as well as inhibit CYP3A4.116 Thus, acute exposure to ritonavir may place patients at increased risk for alprazolam toxicity, while longer-term exposure to ritonavir may result in a loss of anxiolysis and possible withdrawal in patients who are using alprazolam recreationally. A longer-term study is necessary to further clarify the time course and nature of the interaction between alprazolam and ritonavir.
Additional information is also required to clarify the safety of using midazolam with PIs. Palkama et al.113 concluded that, aside from the possibility of a longer sedative effect, the use of bolus doses of intravenous midazolam with saquinavir is likely safe. However, other investigators114 reported on a patient who experienced prolonged sedation secondary to the combination of midazolam and saquinavir; their experience warrants that patients receiving the combination should be closely monitored. Data with other PIs are lacking. The use of midazolam with PIs and delavirdine should be avoided if possible, given the risk of prolonged sedation and respiratory depression associated with large increases in midazolam concentrations. Although formal pharmacokinetic studies are lacking, similar interactions between clonazepam and flunitrazepam and PIs are possible, since both agents are substrates of CYP3A4.117, 118 As well, caution should be exercised with diazepam, particularly in combination with ritonavir, since both the 3A4 and 2C19 systems appear to be important in its metabolism.119, 120 In contrast, nevirapine and efavirenz may put patients who are using midazolam, triazolam, alprazolam, clonazepam, and flunitrazepam at risk for loss of effect and/or withdrawal due to their 3A4 inductive potential.
Interactions between lorazepam, oxazepam, or temazepam and antiretrovirals differ from those described above, since these members of the benzodiazepine family are metabolized primarily by glucuronidation.121, 122 Thus, drugs that increase the activity of glucuronyltransferases (i.e., ritonavir, nelfinavir) may accelerate the metabolism of these compounds, resulting in lower drug exposure. Although reports are lacking, concomitant use of lorazepam, oxazepam, or temazepam with either ritonavir or nelfinavir may decrease the anxiolytic effect of these agents or precipitate symptoms consistent with benzodiazepine withdrawal reaction due to the aforementioned interaction. A higher dose of the benzodiazepine may be necessary to compensate for the interaction.
Tetrahydrocannabinol (THC), the active ingredient of smoked marijuana, remains a commonly used recreational agent. In Canada, 23.1% of surveyed adults had used marijuana more than once in their lives, and current use was estimated at 7.4%.123 In the context of HIV/AIDS, smoked marijuana or THC-containing preparations may also be used for antiemetic or appetite stimulation purposes.
THC is metabolized in humans by microsomal oxidation to several hydroxylated metabolites, among them 11-hydroxy-THC, which is pharmacologically active. Concentrations of 11-hydroxy-THC vary with the route of administration, with oral administration generally producing more of the active metabolite than inhaled THC due to significant first-pass effect. Limited data suggest that CYP3A and 2C9 isoenzymes are involved in microsomal oxidation of THC.124-127 Although inhibition of CYP3A4 or 2C9 may decrease the formation of pharmacologically active metabolites, the effects of THC are unlikely to be significantly attenuated, as THC itself is active and will be more bioavailable. Increased THC concentrations may lead to dose-related effects including frank hallucinations, delusions, paranoid thinking, accentuation of altered time sense, anxiety, panic, depersonalization, loss of insight, orthostatic hypotension, and increased heart rate. Furthermore, inhibition of THC metabolism to 11-hydroxy-THC may be important only in the setting of oral administration, since just trace amounts of the active metabolite are present following smoking.
Induction of CYP3A4 may increase the formation of the pharmacologically active metabolite; however, the conversion of active metabolite to its inactive counterparts may also be accelerated, thereby decreasing the duration of THC effect. This action may be more clinically important with oral THC administration due to its large first-pass effect. The impact of THC on the pharmacokinetics of indinavir and nelfinavir has been evaluated in a small, randomized, placebo-controlled study.128 Patients on stable indinavir or nelfinavir therapy were randomized to receive either THC 3.95% cigarettes, THC 2.5-mg capsules, or placebo, each administered 3 times a day. Nelfinavir and indinavir concentrations were determined prior to and on day 14 of THC use. A statistically significant 14% reduction in indinavir maximum concentration was observed with smoked THC. As well, smoked THC significantly reduced the ratio of M8 (active metabolite of nelfinavir) to nelfinavir by 18% (p = 0.039). However, as mentioned previously, reductions in M8 concentrations do not appear to be clinically important. Furthermore, a significant reduction (p = 0.025) in M8 concentrations relative to baseline was observed in patients receiving placebo. Other variables did not change significantly, nor did oral THC produce significant changes in indinavir or nelfinavir pharmacokinetics. The long-term clinical consequence of these changes is likely negligible, especially with the increasing use of boosted PI regimens. There are no reports documenting the impact of antiretrovirals on THC pharmacokinetics or pharmacodynamics. The nature of such an interaction would be difficult to predict, as several variables, including route of administration and the concentration of THC smoked, may confound the outcome.
Considering the widespread use of smoked and oral THC derivatives for appetite stimulation and control of nausea and vomiting, and the lack of reports documenting deleterious effects secondary to the combination of THC and PIs, a clinically significant drug interaction may not exist when THC is used in moderate amounts. Patients who use THC and are beginning antiretroviral therapy should be warned about possible accentuation of the effects of THC, and that they may need to use less THC for the same effect following treatment initiation.
Ethanol metabolism is mediated chiefly by the enzymes alcohol dehydrogenase (formation of acetaldehyde) and aldehyde dehydrogenase. Since 1 of the 2 main metabolites of abacavir is a carboxylate derivative, the formation of which is catalyzed by the alcohol dehydrogenase enzyme, an interaction between ethanol and abacavir is possible due to competition for metabolism. A randomized, open-label, crossover study129 confirmed the existence of such an interaction. Twenty-five HIV-positive patients were randomized to receive either a single dose of abacavir 600 mg, ethanol 0.7 g/kg, or the combination of abacavir and ethanol, with a washout period of 7 days between treatments. Concomitant administration of ethanol and abacavir resulted in a statistically significant 41% increase in abacavir AUC (CI 1.35 to 1.48); no changes in ethanol blood concentrations were observed. The increase in abacavir AUC is unlikely to be clinically significant, as the concentrations were within the ranges observed in previous pharmacokinetic studies of abacavir that employed higher abacavir doses and did not demonstrate additional safety issues.128
Acute administration of alcohol may increase plasma concentrations of other substrates by inhibiting isoforms such as CYP2D6 and 2C19.130 On the other hand, chronic administration may reduce plasma concentrations of drugs metabolized by CYP2E1 and 3A.131, 132 Thus, there is potential for induction of PI and NNRTI metabolism with chronic alcohol use. Such an interaction may result in subtherapeutic concentrations of these agents, predisposing to resistance and compromising antiretroviral efficacy over time. However, there are currently no data documenting such an interaction. Appropriately conducted pharmacokinetic studies are necessary to confirm the existence of an interaction between antiretrovirals and chronic alcohol use and to clarify appropriate management strategies.
The increasing numbers of available PIs and NNRTIs and the identification of various isoforms of the CYP450 enzyme system have heightened awareness about the significance of drug interactions in the HIV population. However, recreational drugs are often not considered by both clinicians and patients when reviewing a particular medication regimen for potential interactions. One of the inherent concerns associated with recreational drug use is that the margin of safety for many of these substances is often poorly defined, and quality control is often highly variable. Thus, factors that may lead to unpredictable drug concentrations can further increase the risk of adverse outcomes. Given the increasing incidence of HIV infection among substance users and the increasing use of complex combination antiretroviral regimens, the risk of adverse drug interactions with possibly fatal consequences cannot be overlooked or ignored. Clinicians should, therefore, strive to gather information about recreational drug use as part of a comprehensive medication history. Reassuring the patient that confidentiality will be respected and the use of open-ended questions directed in a nonthreatening and nonjudgmental manner will facilitate the information-gathering process.
Much of the information presented in this article is largely extrapolated from in vitro pharmacokinetic experiments, case reports, or animal model studies. There are obviously many limitations in applying such data to clinical practice settings. With case reports, information is often anecdotal in nature. Patients' own recall bias is an obvious limitation, making direct causality difficult to establish. Even when in vitro or in vivo data are available, results often may not be directly extrapolated to clinical situations. For instance, much of the interaction information for ritonavir is based on full-dose (i.e., 600 mg twice daily) studies. However, ritonavir is now frequently used at lower doses (e.g., 100-200 mg twice daily) as a pharmacokinetic boosting agent. Ritonavir can inhibit CYP450 activity and increase protease trough concentrations in a dose-related manner.133 Therefore, the frequency, extent, and/or clinical significance of interactions with ritonavir 100 mg twice daily may be lower compared with higher doses of ritonavir. As an example, when efavirenz was added to a combination of amprenavir 600 mg twice daily plus ritonavir 100 mg twice daily, amprenavir concentrations were decreased by almost 80%; however, when the ritonavir dose was increased to 200 mg twice daily, amprenavir concentrations remained stable in the presence of efavirenz.134
These confounding factors highlight the importance of designing interaction studies that accurately reflect situations encountered in clinical practice. However, due to legal and ethical constraints, it is highly unlikely that rigorous, prospective, controlled interaction studies between antiretrovirals and recreational drugs will ever be conducted. As such, these limited data may serve as a tool for clinicians in anticipating and hopefully averting potential detrimental interactions with recreational drugs.
Adverse interactions between agents commonly prescribed in HIV and recreational drugs can occur, and may possibly be associated with serious clinical consequences. This issue highlights the need for clinicians to obtain thorough patient histories on both prescription as well as recreational drug use and to counsel and/or adjust therapeutic regimens when required to minimize the risk of morbidity or mortality.