Introduction
The co-occurrence of human immunodeficiency virus (HIV) with psychiatric disorders is commonplace, as each serves as an independent risk factor for the other.1 The seroprevalence of HIV in patients with psychiatric illness is estimated between 5–23%, which is significantly higher than the 0.3–0.4% seen in the general U.S. population.2 In particular, patients with substance use disorders are at an increased risk of engaging in risky behaviors such as needle-sharing and unprotected sexReference Angelino3—behaviors that have been directly associated with the transmission and development of blood-borne pathogens, such as HIV.4
Similarly, being diagnosed with HIV has been associated with an increased risk of developing mood, anxiety, and cognitive disorders.1 Depression and anxiety, the 2 most common psychiatric disorders diagnosed in patients with HIV, have prevalence estimates as high as 72% and 82%, respectively.Reference Chaudhury, Bakhla and Saini5 Symptoms of depression may be related to life-changing decisions, such as informing their partner of HIV status, selecting treatment options, and weighing whether or not to have children or any future sexual partners.Reference Chaudhury, Bakhla and Saini5, Reference Forstein6 In addition, symptoms of anxiety may arise from apprehension about disease progression and death.Reference Chaudhury, Bakhla and Saini5, Reference Forstein6
These co-occurring disease states are of particular concern, as alcohol,Reference Silva, Ximenes and Miranda Filho7 illicit drug use,Reference Silva, Ximenes and Miranda Filho7 and depressionReference Beer and Skarbinski8–Reference Starace, Ammassari and Trotta12 have each been associated with decreased adherence to antiretroviral therapy (ART) for HIV. Nonadherence to ART may lead to disease progression, treatment resistance, and other immediate health risks resulting from suboptimal viral suppression such as life-threatening opportunistic infections.Reference Schaecher13 In addition, psychiatric conditions in patients with HIV can negatively impact quality of life, which may lead to poorer outcomes overall.Reference Sherbourne, Hays and Fleishman14
Treatment of HIV using antiretrovirals has evolved significantly since their introduction, with a number of new additions to the clinician’s armamentarium. These include the increased prominence of members of the integrase strand transfer inhibitor (INSTI) class, which boast a more favorable adverse effect profile and, in some cases, a decreased potential for drug–drug interactions (DDIs). Utilization of highly toxic agents such as didanosine and stavudine has dropped significantly in favor of safer alternatives; however, despite strong guideline recommendations, many patients may still remain on these agents due to patient or provider reluctance to therapeutic interchange.Reference Boillat-Blanco, Darling and Taffe15 Thus, it is imperative for clinicians encountering members of the HIV population to continue to maintain a working knowledge of all commercially available ART agents, even as the list of guideline-preferred regimens becomes more streamlined.
An important feature that distinguishes antiretrovirals from one another is their pharmacokinetic profile, specifically with respect to hepatic metabolism via the cytochrome P450 (CYP) enzyme system (Table S1, available online in the Supplementary Material). Antiretrovirals that inhibit or induce the activity of these enzymes can result in the potential for clinically meaningful DDIs with other drugs, including other antiretrovirals, whose clearance is dependent on biotransformation by one or more of these enzymes. In some cases, this DDI is minor and clinically insignificant. However other times the DDI may be of such magnitude as to affect the drug’s efficacy or safety. These DDIs may be beneficial, such as with the use of the pharmacokinetic enhancers ritonavir and cobicistat in ART. These agents inhibit select CYP enzymes in the liver and gut to “boost” concentrations of concomitantly administered antiretrovirals metabolized by these enzymes, most commonly protease inhibitors (PIs) and the INSTI elvitegravir, to therapeutic levels.16 However, DDIs associated with ART may also complicate therapy. All currently available PIs are metabolized by CYP 3A4, the most common enzymatic pathway for hepatically metabolized drugs,Reference Ogu and Maxa17 and most are additionally inhibitors, particularly when co-administered with ritonavir.Reference Ogu and Maxa17 PIs are also substrates and inhibitors of P-glycoprotein, a transmembrane drug efflux protein found in various areas of the body including the intestinal epithelium, and are subject to DDIs by this pathway as well.Reference Lund, Petersen and Dalhoff18 P-glycoprotein is involved in first pass metabolism, and inhibition of P-glycoprotein results in impaired efflux of P-glycoprotein substrates, potentially resulting in increased substrate serum concentrations.Reference Lund, Petersen and Dalhoff18 Among the non-nucleoside reverse transcriptase inhibitors (NNRTIs), effects on the CYP enzymes vary significantly among the agents, and may include induction, inhibition, or both.16 Management of an actual or potential DDI will depend on the specific drugs involved and the nature of the interaction, and may include a need for closer patient monitoring, dose alterations, or avoidance of the combination altogether.
The co-dependent relationship between HIV and psychiatric disorders likely translates to a high incidence of clinicians encountering medication regimens containing treatment modalities for both disease states. These regimens, composed of both ART for HIV and psychotropics for psychiatric conditions, are complex, and in many cases involve clinically significant DDIs and compounding adverse effects. Given that barriers to successful treatment of these conditions are already present, it is imperative that clinicians optimize medication regimens to be efficacious, safe, and tolerable. With this, the objective of this article was to review DDIs, compounding adverse effects, and other clinical caveats that must be considered for medication regimens containing both antiretrovirals and psychotropics.
Methods
Literature search
The PubMed/MEDLINE database was searched for articles from inception through December 2017 using search terms for each psychiatric condition of interest (eg, for depression, “depression, major depressive disorder, depressive illness”) and/or the drugs utilized in the treatment of each condition with the names of commonly prescribed HIV medications and HIV medication classes. Guideline documents for the management of HIV, as well as the package inserts of antiretrovirals with anticipated DDI potential, were also examined. Pertinent studies were selected based on the titles and abstracts retrieved. Studies of particular interest are detailed in the online appendix (Tables S2–S7). Human pharmacokinetic or pharmacodynamic studies were prioritized; however, in the absence of clinical data, relevant in vitro data were utilized to predict additional potential DDIs.
Antidepressants
Depression is the most common psychiatric comorbidity among HIV-positive individuals, with rates as high as 37%, or double that of the general population.Reference Wagner, Goggin and Remien19 Rates among intravenous drug users with HIV are even higher, and have been reported up to 70%.Reference Watkins, Pieper and Treisman20 Depression has been associated with ART nonadherence,Reference Gonzalez, Batchelder and Psaros10 and thus poses a significant health risk to affected individuals, making the selection of an appropriate antidepressant regimen paramount. However, the majority of the newer antidepressant classes are extensively metabolized by the CYP450 system, and therefore may potentiate metabolically mediated DDIs with antiretroviral agents (Table S2).
Pharmacokinetic interactions
Selective serotonin reuptake inhibitors
The selective serotonin reuptake inhibitors (SSRIs) are the most widely prescribed class of antidepressantsReference Bauer, Bschor and Pfennig21, and are utilized in the treatment of major depressive disorder, anxiety, and other mood disorders. All SSRIs are substrates for CYP isoenzymes, with most additionally having CYP inhibitory properties, creating the potential for DDIs with antiretrovirals. However, the particular isoenzymes affected and the clinical significance of the interaction varies by the specific agent. Generally, SSRIs should be slowly titrated to antidepressant effect, starting with the lowest initial dose if the DDI potential is not known.
Citalopram and escitalopram, the active S-enantiomer of citalopram, are substrates for CYP 2C19 and 3A4, and are both substrates and weak inhibitors of CYP 2D6.Reference von Moltke, Greenblatt and Giancarlo22 Due to their multiple enzymatic pathways, and limited effects on the CYP enzyme system, minimal interaction potential is anticipated. In an open-label, crossover study of 21 healthy volunteers who were administered escitalopram 20 mg and a single dose of ritonavir 600 mg, a CYP 2D6 and 3A4 inhibitor and 2C19 inducer, no significant pharmacokinetic interactions were observed, with the exception of an increased time to maximum escitalopram concentration and a 10% reduction in the apparent volume of distribution, which the authors stated was unlikely to be clinically relevant.Reference Gutierrez, Rosenberg and Abramowitz23 This was consistent with a previous study that found no impact when citalopram was administered with the strong CYP 3A4 inhibitor ketoconazole.Reference Gutierrez and Abramowitz24 Studies evaluating the impact of CYP 2C19 inhibition on escitalopram pharmacokinetics suggest that co-administration of the antiretrovirals efavirenz and etravirine, CYP 2C19 inhibitors, is also unlikely to be of concernReference Malling, Poulsen and Søgaard25; however, it should be noted that the maximum recommended dose of citalopram with co-administration of CYP 2C19 inhibitors is capped at 20 mg daily per manufacturer recommendations.26 Based on their favorable pharmacokinetic profile, these agents may be preferable antidepressant choices for the patients with HIV, particularly for those receiving PIs.
Fluoxetine, along with its active metabolite norfluoxetine, potently inhibits CYP 2D6, with additional inhibitory effects on CYP 2C9, 2C19, and 3A4.Reference Spina, Pisani and de Leon27 Paroxetine and sertraline additionally inhibit CYP 2D6, with minimal inhibition of other CYP isoenzymes.Reference Spina, Pisani and de Leon27 All are metabolized by multiple CYPs, most predominantly CYP 2D6 and 3A4.Reference Spina, Pisani and de Leon27 Although co-administration of antiretrovirals metabolized by CYP 2D6 theoretically poses a risk of increased antiretroviral concentrations and subsequent toxicity, few antiretrovirals are processed via this pathway and the magnitude of the interaction appears to be limited. Ritonavir, a CYP 2D6 and 3A4 substrate, had its area under the curve (AUC) increased by 19% with concomitant fluoxetine, a relatively minor DDI not requiring dose adjustments.Reference Ouellet, Hsu and Qian28 Extrapolation of these results to cobicistat is likely reasonable, given the similarity of CYP interplay (Table S1). The NNRTI delavirdine is also metabolized by CYP 3A4 and 2D6.29 Population pharmacokinetic data from efficacy studies indicated that trough delavirdine concentrations increased by 50% with fluoxetine administration, similar to the effect seen by the addition of the strong CYP 3A4 inhibitor ketoconazole.29 However, no dose changes were recommended.29 Similarly, antidepressant concentrations have also been noted to be affected by ART. In 1 study, both fluoxetine and norfluoxetine concentrations were significantly reduced when nevirapine, an inducer of CYP 3A4 and 2B6, was used concomitantly.Reference de Maat, Huitema and Mulder30 Additionally, ritonavir-boosted darunavir and fosamprenavir have been shown to decrease concentrations of paroxetine and sertraline by 39–55%, with efavirenz decreasing concentrations of sertraline, but not paroxetine, by 39%.16 Additionally, 4 cases of serotonin syndrome, or serotonin toxicity, were reported after patients with HIV receiving fluoxetine had changes made to their ART.Reference DeSilva, Le Flore and Marston31 Three of the 4 patients had newly started ritonavir, a CYP 3A4 and 2D6 inhibitor, and 1 received efavirenz. For 3 of the patients, symptoms resolved following reductions in the fluoxetine or ritonavir dose. It should be noted that the ritonavir doses used in these patients were 400–1200 mg daily, which are significantly higher than those used in practice today.16 CYP inhibition by ritonavir is stronger with increased doses, and also affects CYP 2D6 to a greater extent, resulting in an increased DDI potential than would be expected with conventional dosing.Reference Bertz, Foit and Chiu32 Titrating antidepressant dose to clinical response and monitoring for loss of antidepressant response in patients newly started on ART with CYP induction potential is recommended.
Fluvoxamine, a CYP 1A2 and 2D6 substrate, potently inhibits CYP 1A2 and 2C19 and moderately inhibits CYP 2C9 and 3A4, with weak effects on CYP 2D6.Reference Spina, Pisani and de Leon27 A wide range of antiretrovirals have the potential to be affected by fluvoxamine therapy. In 1 study, a 34% decrease in nevirapine clearance was found when fluvoxamine was co-administered, which appeared to be dose-dependent and mediated by inhibition of CYP 3A4.Reference de Maat, Huitema and Mulder30 Guideline recommendations also suggest consideration of either an alternative antidepressant or an HIV provider consult when fluvoxamine therapy is being considered due to the potential for altered PI or elvitegravir concentrations.16 In a case series of 16 patients with HIV and depression treated with fluvoxamine, 10 patients discontinued fluvoxamine due to intolerable adverse effects, including insomnia, gastrointestinal (GI) disturbances, anorexia, excessive sedation, and episodes of aggressive and impulsive behavior, though the remainder of patients reported good clinical response.Reference Grassi, Gambini and Scarone33 Given the availability of alternative SSRIs with less potential to cause clinically significant DDIs with ART, it is reasonable to consider other SSRIs before initiating fluvoxamine therapy.
Vortioxetine and vilazodone both undergo metabolism by a variety of CYP enzymes while having no or weak inhibitory effectsReference Spina, Pisani and de Leon27 and thus are unlikely to cause clinically significant DDIs, though studies evaluating this are not available at this time.
Serotonin-norepinephrine reuptake inhibitors
Serotonin-norepinephrine reuptake inhibitors (SNRIs) are similar to the SSRIs with respect to undergoing metabolism via the CYP system, resulting in the potential for DDIs. The SNRIs venlafaxine, desvenlafaxine, levomilnacipran, and duloxetine have FDA approvals for the treatment of major depression. Milnacipran is currently only FDA approved for fibromyalgia. Among the SNRIs, duloxetine moderately inhibits CYP 2D6, while the other SNRIs are weak or negligible inhibitors.Reference Spina, Pisani and de Leon27
Venlafaxine and its isomer desvenlafaxine are metabolized by CYP 3A4 and 2D6 with only weak CYP inhibitory activity.Reference Spina, Pisani and de Leon27 In a healthy volunteer study, extended release formulations of venlafaxine and desvenlafaxine failed to have an effect on the pharmacokinetics of indinavir, a CYP 3A4 and P-glycoprotein substrate, on day 1 or day 12 of antidepressant therapy. Venlafaxine and desvenlafaxine pharmacokinetics were also unaffected.Reference Jann, Spratlin and Momary34 However, it should be noted that in an earlier healthy volunteer study using venlafaxine, indinavir AUC and peak concentration (Cmax) were decreased by 28% and 36%, respectively, theorized by the authors to represent possible P-glycoprotein induction.Reference Levin, Nelson and DeVane35 This was supported by an in vitro model, which demonstrated increased P-glycoprotein expression when venlafaxine and desvenlafaxine were present, though not to a statistically significant degree with desvenlafaxine.Reference Bachman, Beauleiu-Abdelahad, Ganey, Mullan and Levin36 Although the clinical significance of this interaction is unknown, given the potential for reduced antiretroviral concentrations leading to virologic resistance, other antidepressants should be first considered for patients on PIs or other P-glycoprotein substrates. If the benefit of venlafaxine or desvenlafaxine therapy is believed to outweigh the risks, extended-release formulations of venlafaxine or desvenlafaxine may be preferred, as the reduced drug exposures in the GI tract may be less likely to promote P-glycoprotein–mediated efflux.Reference Jann, Spratlin and Momary34
Duloxetine is a substrate of CYP 2D6 and 1A2, with increased concentrations possible upon co-administration of CYP 2D6 inhibitors. Co-administration of duloxetine with paroxetine, a strong CYP 2D6 inhibitor, increased duloxetine concentrations by 60%,Reference Skinner, Kuan and Pan37 which poses a theoretical concern for similar elevations in duloxetine concentrations when duloxetine is given with antiretrovirals inhibiting CYP 2D6 such as ritonavir (Table S1). Concentrations increased strikingly (6-fold AUC increase) with co-administration of CYP 1A2 inhibitors38; however, strong CYP 1A2 inhibition is not anticipated with the currently approved antiretrovirals. Monitoring of adverse effects for patients receiving duloxetine with the pharmacokinetic enhancers cobicistat and ritonavir and the antiretrovirals tipranavir and delavirdine is advised. Given the potential for CYP 1A2 induction, patients on ritonavir or tipranavir may alternatively have decreased duloxetine concentrations and an alternative antidepressant should be considered for nonresponders. In addition to being a substrate, duloxetine is a moderate CYP 2D6 inhibitorReference Spina, Pisani and de Leon27 and may increase concentrations of CYP 2D6 metabolized drugs. This effect may be additive with other CYP 2D6 inhibitors.Reference Aung, O’Brien and Tien39
Milnacipran and levomilnacipran undergo minimal metabolism by CYP 3A4 and may require dose reductions only when administered with strong CYP 3A4 inhibitors.Reference Spina, Pisani and de Leon27 Neither strongly affects CYP enzymes and overall neither is unlikely to be involved in clinically significant pharmacokinetic DDIs, though studies evaluating this are not available at this time.
Serotonin modulators
Trazodone, a CYP 3A4 and 2D6 substrate, has no appreciable inhibitory effect on the activity of CYP isoenzymes.Reference Spina, Pisani and de Leon27 However, its concentrations can be significantly affected by antiretrovirals that induce or inhibit CYP 3A4 (Table S1). In vitro studies showed strong inhibition of CYP 3A4-mediated trazodone clearance by the PIs ritonavir and indinavir, with weaker inhibition by saquinavir and nelfinavir.Reference Zalma, von Moltke and Granda40 This interaction has been associated with a clinically significant increase in adverse effects. In a blinded crossover study, when ritonavir was administered with trazodone to healthy volunteers, increased sedation, fatigue, hypotension, and syncope were reported compared to the group receiving trazodone and placebo. This increase in adverse effects was correlated with an approximate 240% increase in the AUC of trazodone.Reference Greenblatt, von Moltke and Harmatz41 Thus, alternative antidepressants should be first considered for patients receiving PIs or cobicistat-boosted elvitegravir. If trazodone is indicated, use the lowest initial dose and monitor for central nervous system and cardiovascular side effects. An exception is saquinavir, which is contraindicated with any dose of trazodone due to an increased risk for cardiac arrhythmias.16
Nefazodone has been largely discontinued due to its black box warning for hepatotoxicity and liver failure.Reference Voican, Corruble and Naveau42 While overall response rates were high in an 11-patient open-label trial of nefazodone for individuals with HIV and depression, 1 patient taking ritonavir discontinued therapy due to emergence of headache, confusion, dizziness, and anxiety attributed to an elevation in nefazodone levels due to ritonavir-mediated CYP inhibition, as nefazodone is metabolized by CYP 3A4 while its metabolite is metabolized by CYP 2D6.Reference Elliott, Russo and Bergam43 The authors proposed that doses of nefazodone should be limited to a maximum of 50–100 mg daily in patients taking ritonavir. Additionally, nefazodone strongly inhibits CYP 3A4 in vitro, posing a concern for toxic levels of CYP 3A4-metabolized antiretrovirals, such as the PIs, and caution is advised when prescribing or continuing nefazodone for a patient receiving PI therapy.Reference Elliott, Russo and Bergam43 Given this interaction potential, and nefazodone’s unfavorable safety profile, an alternative antidepressant is recommended.
Norepinephrine-serotonin modulator
Mirtazapine has been shown to be efficacious for the treatment of depression in individuals with HIV while also beneficially promoting weight gain for those with HIV wasting syndrome.Reference Elliott and Roy-Byrne44 Mirtazapine is a substrate of CYP 1A2, 3A4, and 2D6, while having minimal inhibitory or inductive properties.Reference Spina, Pisani and de Leon27 Clinically notable interactions between mirtazapine and antiretrovirals have not been well-described. However, use of the lowest effective mirtazapine dose should be considered when it is co-administered with potent inhibitors or inducers of these enzymes, such as PIs or efavirenz.
Norepinephrine-dopamine reuptake inhibitor
Bupropion is indicated for major depressive disorder, seasonal affective disorder, and smoking cessation. It is metabolized by CYP 2B6 to an active metabolite, hydroxybupropion.27Although hydroxybupropion maintains some activity, it is only half as potent as bupropion.45 Therefore, the package insert for bupropion cautions that dose titration may be necessary if co-administered with CYP 2B6 inducers such as ritonavir, lopinavir, nevirapine, and efavirenz.45 Indeed, in healthy volunteer studies measuring the effect of ritonavir-boosted lopinavir and efavirenz at standard doses, bupropion exposures were reduced by 57% and 55%, respectively.Reference Hogeland, Swindells and McNabb46, Reference Robertson, Maldarelli and Natarajan47 The authors estimated that this interaction may necessitate as much as a 100% dose increase for bupropion to achieve therapeutic concentrations. All ritonavir-boosted PIs have the potential to lower bupropion concentrations, which can happen rapidly, occurring after only 3 days of treatment with ritonavir.Reference Kharasch, Mitchell and Coles48, Reference Park, Vousden and Brittain49 There is also a dose-dependent correlation, with higher ritonavir doses (600 mg twice daily) decreasing both the AUC and Cmax of bupropion by 62% to 67%, a statistically significant decrease compared to lower ritonavir doses (100 mg twice daily), which resulted in 21–22% reductions in bupropion concentrations.Reference Kharasch, Mitchell and Coles48, Reference Park, Vousden and Brittain49 Although this is concerning, ART regimens requiring ritonavir doses exceeding 400 mg per day are now uncommon.16 In 1 prospective study of sustained-release bupropion in 20 patients with HIV and depression, the mean effective bupropion dose eliciting a 60% response rate was 265 mg/day (standard deviation [SD] = 65), near that of the maximum dose studied, 300 mg/day. Of these patients, 70% were taking a PI; however, only 1 was on ritonavir.Reference Currier, Molina and Kato50 Patients managed on both bupropion and a CYP 2B6 inducer may require dose titrations as needed until clinical response, though the maximum recommended dose provided by the manufacturer should not be exceeded as bupropion has been associated with dose-dependent seizures.
Tricyclic antidepressants
Tricyclic antidepressant (TCA) use for the treatment of depression has fallen out of favor versus other classes owing to their increased anticholinergic, sedative, hypotensive, and cardiotoxic side effects51; however, members of the class continue to be used for off-label indications, including among the HIV population.Reference Fernandez52 TCAs are primarily metabolized via CYP 3A4 or 2D6,Reference Gillman53 and caution and close monitoring is advised when co-administered with inhibitors such as ritonavir.54 Desipramine clearance was reduced 59% by ritonavir in 1 pharmacokinetic study; however, a large daily ritonavir dose of 1000 mg was used,Reference Bertz, Cao and Cavanaugh55 which is higher than doses in modern practice.16 Nortryptiline, a weak inhibitor of CYP 2D6,Reference Gillman53 increased nevirapine Cmax over 2-fold in an in vitro modelReference Usach, Melis and Gandía56; however, these results have yet to be replicated in humans, and therefore the clinical significance is uncertain. Despite this, an advantage of TCAs is that therapeutic drug monitoring is possible, allowing for identification of supratherapeutic concentrations if an interaction is suspected.
Monoamine oxidase inhibitors
Monoamine oxidase inhibitors (MAOIs), including phenelzine and in particular tranylcypromine, with its additional stimulant-like effects, have been shown to be effective in major depressive disorder and successfully utilized for this purpose in patients with HIV.Reference Pieper and Treisman57 However, MAOIs are generally reserved for patients not responsive to SSRIs or TCAs owing to their many adverse DDIs and risk of hypertensive crisis.51 Although these medications generally do not pose a concern with ART from a pharmacokinetic perspective, some experts recommend avoiding MAOIs in patients receiving the nucleoside reverse transcriptase inhibitor (NRTI) zidovudine, which may inhibit catechol-O-methyltransferase, a major enzyme responsible for catecholamine catabolism in the brain.Reference Maldonado, Fernandez and Levy58 Despite their seemingly benign DDI profile with ART, their narrow therapeutic index, necessity of adherence, and strict dietary precautions preclude these agents as preferred antidepressants in patients with HIV.
Over-the-counter
Hypericum perforatum, more commonly known as St. John’s wort, is an herbal medicine associated with antidepressant properties.Reference Linde, Berner and Kriston59 Although often well-tolerated from an adverse effect perspective,Reference Knuppel and Linde60 St. John’s wort is a potent inducer of CYP 3A4 and P-glycoprotein and has also been reported to induce the activities of CYP 1A2, 2C9, 2C19, and 2D6.Reference Izzo and Ernst61 With the majority of antiretrovirals being substrates for CYP 3A4, P-glycoprotein, or both (Table S1), co-administration of St. John’s wort has the potential to cause many clinically significant DDIs, which may result in subtherapeutic ART concentrations with subsequent treatment resistance or treatment failure.
In a nonlinear mixed-effect modeling analysis of 5 men with HIV who were taking St. John’s wort and nevirapine-based ART regimens, the clearance of nevirapine, a substrate for both CYP 3A4 and P-glycoprotein, was 35% faster in comparison to control patients on nevirapine without St. John’s wort.Reference de Maat, Hoetelmans and Mathôt62 In another open-label healthy volunteer study, St. John’s wort decreased concentrations of the PI indinavir by a mean of 57% (range 49–99%), with an 81% decrease in concentration at 8 hours.Reference Piscitelli, Burstein and Chaitt63 The authors concluded that these reduced blood concentrations were sufficient to result in the possible loss of HIV suppression. Additionally, an increase in HIV ribonucleic acid (RNA) viral load was detected in a patient taking indinavir-based ART and concomitant St. John’s wort,Reference Henderson, Yue and Bergquist64 supporting the possibility of St. John’s wort–mediated HIV treatment failure.
Currently, HIV guidelines classify St. John’s wort as contraindicated for patients receiving PIs, NNRTIs, the INSTIs elvitegravir or dolutegravir, maraviroc, or tenofovir alafenamide.16 With this, patients should be strongly advised to avoid St. John’s wort while taking these agents; similarly, patients who have been chronically managed on ART with concomitant St. John’s wort should be closely monitored.
Compounding adverse effects
The different antidepressant classes vary in their side effect profiles, and when feasible it is desirable to avoid overlapping adverse effects with patients’ ART regimens. Insomnia occurs commonly among patients taking efavirenz, rilpivirine, or INSTIs,16 and for patients reporting this side effect, a more sedating antidepressant such as trazodone, mirtazapine, or a TCA may be preferred over an SSRI or SNRI. Sexual dysfunction has been reported in as high as 30% of all patients taking SSRIs and is a common reason for antidepressant nonadherence.Reference Rosen and Marin65 This is of particular concern for the population with HIV, with sexual dysfunction, including erectile dysfunction, highly prevalent, particularly among men who have sex with men (MSM). Reported rates of erectile dysfunction in the MSM population range from 33–74%, compared to 0–18% in the general population, and may be increased by antidepressant use.Reference Hart, Mustanski and Ryan66 Monitoring for antidepressant adherence and side effects, including sexual dysfunction, is highly advised. Antidepressants such as vilazodone or bupropion, which are associated with significantly less sexual dysfunction, may be recommended. Before initiating pharmacologic therapy for sexual dysfunction, the potential for DDIs should be assessed. For example, reduced doses of phosphodiesterase type 5 (PDE5) inhibitors such as sildenafil are required when co-administered with PIs.16
TCAs are notable for their anticholinergic side effects, including dry mouth, blurred vision, constipation, and urinary retention.Reference Fernandez52 Drowsiness, confusion, weight gain, hypotension, and tachycardia are also commonly reported.Reference Fernandez52 In a study of 75 HIV-positive patients diagnosed with depression comparing treatment with imipramine, paroxetine, and placebo, significantly more patients in the imipramine group discontinued therapy due to adverse events, with a 38% dropout rate.Reference Elliott, Uldall and Bergam67 The orthostatic hypotension associated with TCAs, as well as trazodone, may be particularly problematic for patients with HIV-related dysautonomia or who are taking the CCR5 co-receptor antagonist maraviroc, which has also been associated with this side effect.68 Baseline hyposalivation is common among HIV-positive individuals (29% self-reported rate), which may make the increased mouth dryness caused by TCAs especially intolerable.Reference Nittayananta, Chanowanna and Pruphetkaew69, Reference Younai, Marcus and Freed70 Although the extensive side effect profile of TCAs is generally regarded as detrimental, certain effects, such as weight gain and sedation, may be desirable in some patients, with mirtazapine additionally often selected specifically for its sedative properties and to promote weight gain as an off-label treatment for HIV wasting syndrome.Reference Badowski and Pandit71 It should be noted, however, that weight gain should not be considered a universally desirable antidepressant characteristic among those with HIV. As HIV has increasingly become a chronic disease, obesity rates have also been increasing, and may be compounded by use of ART such as the PIs, which have been associated with lipodystrophy and insulin resistance.Reference Koethe, Jenkins and Lau72
Certain antidepressants including citalopram, escitalopram, venlafaxine, and mirtazapine, have been linked to an increased risk for torsades de pointes.Reference Jasiak and Bostwick73 Ritonavir-boosted saquinavir has been associated with significant QT interval prolongation in healthy volunteers, and its use is contraindicated in patients with a baseline QTc interval of greater than 450 msec.74 Other PIs did not appear to be independently associated with QT prolongation in a large prospective analysisReference Soliman, Lundgren and Roediger75; however, nelfinavir, ritonavir, and lopinavir have shown the ability to block human potassium channels in vitro.Reference Anson, Weaver and Ackerman76 The NNRTIs rilpivirine and efavirenz also carry warnings in their labeling noting a potential QT prolongation risk.77, 78 Use of higher-risk antidepressants or other agents that may prolong the QT interval with saquinavir should be avoided; however, routine electrocardiogram monitoring for patients on other antiretrovirals is not advised in the absence of other significant risk factors for arrhythmia.
The NNRTI efavirenz is associated with a high rate of neuropsychiatric side effects, particularly in the first month of therapy.78 Antidepressants such as mirtazapine with overlapping central nervous system toxicities may exacerbate these effects. All antidepressants carry a boxed warning for potential increased risk of suicidality in adolescents and young adults. Of concern, a doubled risk of suicidality was reported with efavirenz-containing ART compared to those without efavirenz based on data from 4 randomized trials, 3 of which were open-label.Reference Mollan, Smurzynski and Eron79 However, no association was found between efavirenz-containing regimens and suicidality in a later propensity score-adjusted analysis of real-world data and nearly 20,000 patients, though the authors acknowledged that channeling bias could not be ruled out as a confounder.Reference Nkhoma, Coumbis and Farr80 Antidepressant therapy should not be withheld from patients receiving efavirenz; however, close initial monitoring, particularly for patients with a history of suicide attempts, remains warranted. TCAs should be avoided in patients with active suicidality given their severe toxicity with overdose.
Overall, antidepressant tolerance and effectiveness will vary by patient, and no one class can be universally recommended as first-line in the HIV population. Clinicians should consider individual risk factors and potential DDIs in determining the optimal antidepressant selection.
Stimulants
Psychostimulants such as methylphenidate and amphetamine products have been utilized in the treatment of a diverse range of psychiatric conditions, including attention deficit hyperactivity disorder (ADHD), narcolepsy, chronic fatigue, binge eating disorder, and as an adjunct for depression and obesity treatment.Reference Fernandez52 Among patients with HIV, stimulants have demonstrated efficacy for the treatment of depression and fatigue,Reference Breitbart, Rosenfeld and Kaim81–Reference Wagner and Rabkin84 and have additionally been shown to improve cognitive function among patients suffering from HIV-associated cognitive impairment,Reference Hinkin, Castellon, Hardy, Farinpour, Newton and Singer85 which affects up to 90% of those with symptomatic infection.Reference Maldonado, Fernandez and Levy58 Among children with HIV, over one-quarter are affected by ADHD.Reference Scharko86, Reference Zeegers, Rabie and Swanevelder87 Approximately half of these children are expected to continue to manifest symptoms into adulthood and require continued treatment.Reference Dopheide and Pliszka88 Additionally, recent research has suggested that certain genotypes associated with ADHD may increase the risk of future HIV acquisition,Reference Koutsilieri, Riederer and du Plessis89 emphasizing the importance of psychoeducation and behavioral intervention in this population in addition to pharmacologic therapy and possibly prophylactic HIV medication in high-risk individuals.
Pharmacokinetic interactions
Overall, the interactions between stimulants and the CYP P450 system have not been well-defined, with no studies on pharmacokinetic interactions among patients with HIV on ART and stimulant therapy identified. However, CYP 2D6 appears to play a role in the metabolism of amphetamines, in addition to oxidation and deamination.Reference Landino, Buckley and Roy90 This suggests that drugs inhibiting CYP 2D6, such as the pharmacokinetic enhancers ritonavir and cobicistat, may increase serum concentrations of these agents with the concern of increasing adverse effects. However, this interaction is currently theoretical, as currently no data have shown clinically apparent adverse effects with these combinations. Methylphenidate may also inhibit CYP 2D6 and 2B6 activity based on in vitro data, but again, the clinical significance is currently unknown.Reference Markowitz, Straughn and Patrick91 Careful monitoring should be performed when initiating a psychostimulant with CYP P450 interaction potential, using the lowest initial dose and titrating to clinical effect. Alternatively, dexmethylphenidate and lisdexamfetamine do not undergo CYP P450 metabolism and may be appropriate lower-risk options if adverse drug reactions suspected to be due to a DDI occur.Reference Domnitei and Madaan92, Reference Krishnan and Moncrief93
Compounding adverse effects
The main adverse effects of psychostimulants are related to their sympathomimetic activity and neuropsychiatric effects such as insomniaReference Cortese, Holtmann and Banaschewski94; it is unknown if these effects are additive with ART regimens that include INSTIs or NNRTIs such as efavirenz and rilpivirine, which additionally have been associated with these effects.77, 78 A potential for seizure exists for patients with HIV central nervous system involvement receiving high doses of stimulants; these patients should be titrated carefully to the lowest effective stimulant dose and be more closely monitored.Reference Fernandez52 The potential for increased cardiovascular effects with stimulants has also been a long-standing concern. Small, but statistically significant, increases in blood pressure and heart rate without electrocardiogram changes have been reported in small, short-term (<2 year) studies, though the rate of discontinuation due to cardiovascular effects was low at 2% and most of the observed effects resolved spontaneously.Reference Hennissen, Bakker and Banaschewski95 The impact of long-term amphetamine use on cardiovascular health is still not well characterized, with population-based studies producing conflicting results on whether or not an association is present.Reference Awudu and Besag96 In 2007, the FDA required all stimulants used in ADHD treatment to be dispensed with a medication guide notifying patients of this potential risk.97 This is of particular concern for individuals with HIV, who are known to be at increased risk for cardiovascular disease, including myocardial infarction (MI), atherosclerosis, and coronary heart disease compared with uninfected persons.Reference Burkholder, Tamhane and Salinas98 With respect to ART, the negative metabolic effects of PIs are well-established, though the association of specific antiretrovirals (abacavir being the most commonly implicated) with myocardial infarction is controversial. No increased short-term or long-term risk of MI was observed with ART containing abacavir in a long-term observational cohort of 5056 individuals over 17,404 patient-years (HR over a 1-year period = 0.7 [CI 0.2–2.4]).Reference Ribaudo, Benson and Zheng99 However, by contrast, in an even larger study of 8,265 individuals contributing 29,077 patient-years, combined risk of MI was greater for patients who received abacavir in the previous 6 months (adjusted HR = 1.84 [CI 1.17–2.91]).Reference Elion, Althoff and Zhang100 Some analyses suggest that among the PIs, atazanavir may be preferred as a lower-risk option.Reference Chow, Shikuma and Ritchings101, Reference LaFleur, Bress and Rosenblatt102 Careful assessment of cardiovascular risk factors in these patients, including initiation of primary prevention therapies as indicated, is warranted.
Other considerations
Stimulant onset of effect occurs during the absorption phase.Reference Wolraich and Doffing103 It should be noted that for certain stimulant formulations (eg, immediate-release methylphenidate and dexmethylphenidate, lisdexamfetamine), high-fat meals may delay the time to peak concentration by up to an hour, which may be an issue for patients taking their medication at the same time as antiretrovirals requiring food for absorption, such as rilpivirine and elvitegravir. If a faster onset of stimulant action is desired, the patient may be instructed to take their stimulant at least an hour before meals, or alternatively the patient may be switched to an extended-release formulation that is less impacted by food.
Antipsychotics
Patients diagnosed with HIV have been shown to have an increased risk for both schizophrenia and acute psychosis,Reference Helleberg, Pedersen and Pedersen104 and the prevalence of HIV in those with severe mental illness is much higher than the general population, estimated to be approximately 6%.Reference Hughes, Bassi and Gilbody105 There is a paucity of available data for clinical information about DDIs between ART and antipsychotics, especially with first-generation antipsychotics.Reference Hill and Lee106 Available data are summarized below. Generally, many of the first- and second-generation antipsychotics are metabolized via the CYP P450 system, in particular CYP 3A4 and CYP 2D6,Reference Hill and Lee106 and the potential for DDIs exists with concomitant administration of medications affecting these enzymes.
Pharmacokinetic interactions
Aripiprazole
Aripiprazole is an atypical antipsychotic, unique in its mechanism of action, in that it is a D2 receptor partial agonist.107 In addition, it has agonist activity at the 5-HT1A receptor.107 Aripiprazole undergoes hepatic metabolism via CYP 3A4 and CYP 2D6 isoenzymes, and has demonstrated increased levels when given concomitantly with CYP 3A4 inhibitors (specifically ketoconazole) and CYP 2D6 inhibitors (specifically quinidine).Reference Aung, O’Brien and Tien39, 107 Based on these data, aripiprazole requires dosage adjustments for both the oral formulation as well as the depot long-acting intramuscular formulation with concomitant administration of strong CYP 3A4 and/or CYP 2D6 inhibitors.107 Ritonavir is known to be a strong CYP 3A4 and 2D6 inhibitor, and co-administration with other medications highly dependent on these CYP isoenzymes for elimination is specifically cautioned against.54 Cobicistat, which additionally is utilized in ART regimens for CYP 3A4 and 2D6 inhibition, also carries precautionary verbiage against concomitant administration with drugs dependent on CYP 3A4 or CYP 2D6 metabolism.108 Apart from ritonavir- or cobicistat-boosted ART regimens, the unboosted PIs or NNRTIs affecting CYP 3A4 (eg, efavirenz) also pose a theoretical concern for altered aripiprazole levels (Table S1).
Despite these theoretical interactions, there is little information available documenting the clinical effects of these expected DDIs. Two case reports are available describing conflicting adverse outcomes related to co-administration of aripiprazole and ART. One such case describes an ART regimen including darunavir boosted with ritonavir, and concomitant aripiprazole, duloxetine, and buspironeReference Aung, O’Brien and Tien39 (Table S3), which led to multiple hospital admissions as well as supratherapeutic aripiprazole serum concentrations attributed to a DDI between aripiprazole and the patient’s ART. Conversely, in a case of a patient receiving ART containing a ritonavir-boosted PI, the patient experienced lower-than-expected aripiprazole serum levels with clinically worsening psychiatric symptoms after being converted from oral to depot aripiprazole using the manufacturer’s recommended conversion.Reference Hahn and Roll109 The patient was ultimately stabilized on double the recommended monthly dose. These cases highlight that very little is known about the DDI between ritonavir and aripiprazole. While hypotheses can be made about tolerability based on expected interactions, careful monitoring of both side effects as well as efficacy is warranted. Although there are no reports available regarding concomitant use of aripiprazole with cobicistat-containing ART regimens, due to similarities in CYP inhibition between cobicistat and ritonavir, similar precautions would be prudent for cobicistat, as well as for other drugs affecting CYP 3A4.
Olanzapine
Olanzapine is an atypical antipsychotic that is metabolized via CYP 1A2 and uridinediphosphate-glucuronosyltransferase (UGT) 1A4.Reference Jacobs, Colbers and Velthoven-Graafland110 Ritonavir, although typically implicated as a CYP isoenzyme inhibitor, also displays induction effects on CYP 1A2 as well as on glucuronidation,Reference van der Lee, Dawood and ter Hofstede111, Reference Yeh, Gaver and Patterson112 thus raising concern for the possibility of subtherapeutic olanzapine levels with co-administration. An open-label crossover study with healthy volunteers investigated this possibility. Analysis of olanzapine serum levels showed that, when compared to olanzapine administration alone, concomitant administration of ritonavir resulted in a significant increase in olanzapine clearance, a decrease in olanzapine half-life from 32 hours to 16 hours, and a 53% decrease in olanzapine AUC.Reference Penzak, Hon and Lawhorn113 It is worth noting that the ritonavir doses utilized in this particular study (300–500 mg twice daily)Reference Penzak, Hon and Lawhorn113 are much higher than modern practice.16
Another similar study in healthy volunteers investigated the interaction between olanzapine and ritonavir, this time employing a dose of ritonavir 100 mg twice daily, which more accurately reflects modern day practice.16, Reference Jacobs, Colbers and Velthoven-Graafland110 Analysis of serum levels of olanzapine revealed a shorter olanzapine half-life by approximately 32%. The AUC of olanzapine 10 mg dose was found to be comparable to the AUC of olanzapine 15 mg dose when given with fosamprenavir/ritonavir. The authors concluded that a 50% increase in olanzapine dose is warranted when given concomitantly with ritonavir-containing ART regimens. It is worth noting that even though the ritonavir dose of 100 mg twice daily is more reflective of ritonavir doses used in clinical practice, the more commonly used and recommended ritonavir-boosted PI regimens are once-daily regimens utilizing a 100 mg once daily ritonavir dose.16 As olanzapine exposure appears to differ depending on the ritonavir dose,Reference Jacobs, Colbers and Velthoven-Graafland110, Reference Penzak, Hon and Lawhorn113 the effect of a 100 mg daily ritonavir dose on olanzapine exposure is unknown. Careful monitoring for olanzapine effect is recommended. Cobicistat, on the other hand, another pharmacokinetic enhancer, is not known to have the same effects on CYP 1A2 or the UGT system108; as such the same DDI with cobicistat-containing ART regimens is not anticipated.
Quetiapine
Quetiapine is an atypical antipsychotic metabolized via CYP 3A4Reference Hantson, Di Fazio and Wallemacq114; ritonavir, a known CYP 3A4 and 2D6 inhibitor,54 is expected to affect quetiapine metabolism. A case of quetiapine overdose (total 8000 mg ingested) in a patient concomitantly taking an atazanavir/ritonavir-containing ART regimen experienced an increased quetiapine half-life of 62 hours, which is 10 times the half-life per quetiapine’s package insert, and still nearly 3 times the 22-hour terminal elimination half-life calculated for a patient following quetiapine overdose (3000 mg) in the absence of CYP 3A4 inhibitors.Reference Pollak and Zbuk115 Accordingly, there are reports available describing adverse events correlated to a DDI between quetiapine and ritonavir.
Two cases of adverse events due to DDIs between quetiapine and atazanavir/ritonavir have been reported.Reference Pollack, McCoy and Stead116 In the first case, approximately 6 months after quetiapine was added to an ART regimen containing ritonavir and atazanavir, the patient experienced weight gain of approximately 50 pounds and developed hyperglycemia.Reference Pollack, McCoy and Stead116 Approximately 5 months after all 3 agents were discontinued, the patient returned to his baseline weight. The second case involves a patient who, upon the start of co-administration of quetiapine with an ART regimen containing ritonavir-boosted atazanavir, began to experience a marked increase in sedation and mental confusion, though the symptoms completely resolved once the patient self-discontinued quetiapine. Like ritonavir, atazanavir is also known to inhibit CYP 3A4 (Table S1),Reference Pollack, McCoy and Stead116 and thus likely contributed to these reported quetiapine-related reactions.
Priapism as a result of a suspected DDI between ritonavir and quetiapine has also been reported. The patient had quetiapine and perphenazine both added to his medication regimen, in addition to an ART regimen including lopinavir/ritonavir 400/100 mg twice daily.Reference Geraci, McCoy and Crum117 Perphenazine, as a substrate of CYP 2D6,Reference Geraci, McCoy and Crum117 could also have been affected by the ritonavir in the patient’s ART regimen; given that the quetiapine and the perphenazine were taken at the same time, it is difficult to say which of the two played a larger role in the development of priapism.Reference Geraci, McCoy and Crum117 Nevertheless, prescriber awareness of this possible side effect and appropriate patient education are warranted.
While there are no specific recommendations in the quetiapine labeling for dose adjustments with PIs, HIV treatment guidelines recommend that if a PI-containing regimen is started in a patient on a stable quetiapine dose, the quetiapine dose should be decreased to 1/6th of the original dose.16
Risperidone
Risperidone is metabolized largely via CYP 2D6, and to a smaller extent, via CYP 3A4.Reference Kelly, Beique and Bowmer118 As such, concomitant administration of indinavir or ritonavir would be expected to alter metabolism of risperidone, as ritonavir is known to inhibit both CYP 3A4 and 2D6 and indinavir is known to inhibit CYP 3A4.Reference Kelly, Beique and Bowmer118 Illustrating this interaction, 1 case report describes a patient who developed extrapyramidal symptoms while receiving indinavir 800 mg twice daily with ritonavir 200 mg twice daily and risperidone, which was hypothesized to be the result of a DDI.Reference Kelly, Beique and Bowmer118 Development of neuroleptic malignant syndrome due to a DDI between indinavir/ritonavir and risperidone has also been reported.Reference Lee, Klesmer and Hirsch119 Another case report described the development of altered mental status that ultimately progressed to a comatose state upon co-administration of indinavir/ritonavir with risperidone.Reference Jover, Cuadrado and Andreu120 The patient returned to his baseline mental status 24 hours after discontinuation of all of his medications.
Multiple cases of angioedema secondary to a suspected DDI with risperidone have also been reported.Reference Gonzalez, Kothari and Kasle121 In all instances, the angioedema was noted to occur months after the interacting medications were started (range 5–15 months), though swelling improved once risperidone was discontinued. Typically, symptoms resulting from DDIs between antipsychotics and ART are more acutely evident, though cases such as those reporting angioedema demonstrate a need for prolonged adverse effect monitoring. Use of the lowest initial dose is recommended when initiating risperidone, titrating to effect while monitoring for toxicities.
Lurasidone
Lurasidone is another relatively new antipsychotic and is primarily metabolized via CYP 3A4, which makes administration of strong CYP 3A4 inhibitors contraindicated.122 Ritonavir is specifically mentioned in the lurasidone package insert as an example of a strong CYP 3A4 inhibitor,122 and while there is no specific mention of cobicistat, based on expected CYP inhibition by cobicistat,108 theoretically, the same contraindication would also apply to cobicistat-containing ART regimens. Atazanavir is a moderate CYP 3A4 inhibitor,123 and would also be expected to affect lurasidone metabolism.122 As such, the lurasidone dose should be halved when given with unboosted atazanavir, which is consistent with HIV treatment guideline recommendations.16, 122 Demonstrating this potential DDI, extrapyramidal symptoms following receipt of lurasidone and concomitant atazanavir have been reported, with symptom cessation following atazanavir discontinuation.Reference Naccarato, Hall and Wai124
Compounding adverse effects
Antipsychotics, particularly the second generation antipsychotics, have the known long-term adverse effect of metabolic syndrome, which has been linked to an increased cardiovascular risk.Reference Vergara-Rodriguez, Vibhakar and Watts125 While some individual antipsychotics such as olanzapine, risperidone, and quetiapine may have a stronger association with development of this side effect, it is considered a class effect.Reference Vergara-Rodriguez, Vibhakar and Watts125 ART has also been linked to development of metabolic syndrome; the degree of effect appears to be agent specific, although definitive relationships have been difficult to conclusively establish.Reference Safrin and Grunfeld126 PIs have classically been associated with the development of metabolic changes,Reference Safrin and Grunfeld126 but other antiretrovirals, such as stavudine, have also been implicated.Reference Safrin and Grunfeld126 Darunavir and atazanavir, 2 newer PIs, have both been shown to have a more favorable effect on metabolic parameters compared to earlier agents within the class.Reference Aberg, Tebas and Overton127, Reference Molina, Andrade-Villanueva and Echevarria128 While there is a wealth of data describing metabolic effects of ART and antipsychotic agents individually, there are very little data available describing their compounded effects. This gap in knowledge was addressed by a recently published study that looked to retrospectively establish a relationship between concomitant use of ART and second-generation antipsychotic agents.Reference Ferrara, Umlauf and Sanders129 The study found that patients taking both ART and second generation antipsychotics were more likely to have higher blood pressure, higher diabetes prevalence, and higher serum triglycerides as compared to those only taking a second generation antipsychotic. These associations remained present after controlling for use of ART regimens containing ritonavir, stavudine, zidovudine, and indinavir. Most patients were receiving only 1 antipsychotic; quetiapine (57%), risperidone (25%), and olanzapine (24%) were the most commonly prescribed agents, which are also the antipsychotics that carry the largest risk of metabolic syndrome. However, it should be noted that this study drew on data from patients receiving earlier antiretroviral agents that are no longer commonly used due to poor tolerability and the availability of more effective agents.16, Reference Ferrara, Umlauf and Sanders129 At this point, careful monitoring when these medications are used concomitantly is warranted.
Clozapine is a highly effective second-generation antipsychotic with its use restricted to those with treatment-resistant schizophrenia due to concerns for agranulocytosis.Reference Nejad, Gandhi and Freudenreich130 Of the antiretrovirals, zidovudine has been associated with clinically significant bone marrow suppression.Reference Nejad, Gandhi and Freudenreich130 Early studies found rates of anemia (defined as hemoglobin<8 g/dL) and neutropenia (defined as absolute neutrophil count (ANC) <750 cells/mm3) associated with zidovudine to be up to 6.3% each.Reference Volberding, Lagakos and Koch131 These high rates of anemia and neutropenia that were initially observed were associated with a 1500 mg daily zidovudine dose, higher than that of modern day practice.16, Reference Volberding, Lagakos and Koch131 This same study found rates of anemia and neutropenia associated with a 500 mg daily zidovudine dose to be far lower at 1.1% and 1.8%, respectively. There are limited data available describing the compounding myelosuppressive effects of concomitant use of clozapine with zidovudine. However, there are 2 case reports describing successful concomitant use of clozapine with a zidovudine-containing ART regimen.Reference Nejad, Gandhi and Freudenreich130, Reference Gerken, McGahee and Keuroghlian132 Both cases provided long-term follow up for each patient showing successful and safe chronic use of both agents. The few reports available suggest that with careful monitoring, patients may do well if the combination of clozapine and zidovudine is warranted.
Another possible overlapping toxicity is QT prolongation. As a class, antipsychotic medications have a risk of prolonging the QT interval.Reference Hill and Lee106 Of the antiretrovirals, the PIs lopinavir/ritonavir and saquinavir have precautions in their labeling regarding QTc prolongation.16 Saquinavir is not recommended in patients who have a baseline QTc interval greater than 450 milliseconds (msec), as torsades de pointes has been reported with saquinavir use.16 Lopinavir/ritonavir also carries a warning for QT prolongation with reported torsades de pointes, but causality has not been definitively established.16 Nevertheless, caution is warranted given the cardiomodulatory effects observed in vitro for both drugs, as well as nelfinavir.Reference Anson, Weaver and Ackerman76 Of the NNRTI class, rilpivirine and efavirenz carry warnings in their labeling regarding QTc prolongation.16 Of note, rilpivirine has been noted to prolong the QTc interval at doses of 75 mg daily and 300 mg daily, which are much higher than the 25 mg daily dose used in clinical practice.77 While there are no clinical data available for degree of QT prolongation with concomitant antipsychotic use with these antiretrovirals, caution and monitoring of high-risk patients are warranted.
The potential to exacerbate psychiatric conditions, particularly psychosis, is another consideration. Of the antiretrovirals, NNRTIs have classically been associated with neuropsychiatric adverse effects, with efavirenz being the biggest culprit.Reference Aberg, Tebas and Overton127, Reference Arendt, de Nocker and von Giesen134 Some of the neuropsychiatric side effects seen with efavirenz early in treatment include nightmares and vivid dreams, fatigue, and insomnia, and typically resolve without intervention within 6–8 weeks if therapy is continued. These early side effects are important to be aware of, as they may lead some providers to avoid the NNRTI class. In regard to new-onset psychosis, a large cohort study found that initiation of ART was protective against development of psychosis, and did not find an increased risk of psychosis associated with efavirenz-based ART.Reference Helleberg, Pedersen and Pedersen104 However, the authors acknowledged that it is possible that a link between psychosis and efavirenz-based ART was not seen due to avoidance of efavirenz use in patients with a known psychiatric history. Nonetheless, multiple cases are available demonstrating new-onset psychosis temporally related to initiation of efavirenz-based therapy.Reference de la Garza, Paoletti-Duarte and Garcia-Martin135, Reference Peyriere, Mauboussin and Rouanet136 In all reported cases, cessation or dose reduction of efavirenz therapy resulted in complete resolution of symptoms and no recurrences of symptoms were seen. Reports of new-onset psychosis related to therapy with nevirapine have also been reported.Reference Wise, Mistry and Reid137 No reports of new-onset psychosis with rilpivirine or etravirine, 2 other NNRTIs, have been published. While clinical trials also show association with neuropsychiatric effects such as insomnia and abnormal dreams with rilpivirine and etravirine, the incidence with these 2 NNRTIs is much lower.Reference Molina, Andrade-Villanueva and Echevarria128, Reference Nelson, Stellbrink and Podzamczer138
In addition to the potential compounding toxicities mentioned above, augmentation of expected adverse events due to underlying pharmacology of a particular antipsychotic agent may also theoretically occur. For example, agents with a high affinity for muscarinic receptors, such as clozapine and olanzapine, are expected to have anticholinergic effects such as dry mouth, constipation, and sedation.Reference Miyamoto, Duncan and Marx139 Concomitant administration of agents that inhibit their metabolism would be expected to exacerbate these side effects. A potential increase in sedation may also be seen when agents with strong H1 and alpha 2 receptor activity (such as clozapine and olanzapine)Reference Miyamoto, Duncan and Marx139 are given with concomitant CYP inhibiting agents. Agents with strong alpha 2 receptor affinity, such as risperidone, often cause hypotension as an adverse effect; this hypotensive effect would be particularly concerning with concomitant administration of maraviroc, as it can also independently cause hypotension.68 Aripiprazole, owing to its unique mechanism of action as a partial D2 agonist, has raised concerns about a theoretical increase in suicidality when used for treatment of schizophrenia. When prescribing aripiprazole concomitantly with ART that could increase aripiprazole levels, careful monitoring and assessment for suicidality in patients is prudent. Awareness of these potential adverse event exaggerations are important, as they may necessitate a change in therapy for some patients for maximal tolerability.
Anxiolytics
Anxiety is a condition that commonly accompanies other psychiatric diagnoses, and is estimated to occur in up to 12% of patients with HIV infection.Reference Brogan and Lux140, Reference Sewell, Goggin and Rabkin141 Those with symptomatic HIV illness have been noted to have worsened anxiety symptoms than those with asymptomatic HIV infection.Reference Sewell, Goggin and Rabkin141 It has also been observed that severe anxiety was a predictor of non-adherence to ART in patients with HIV.Reference Campos, Guimaraes and Remien142 Thus, it is not unexpected that patients would require concomitant administration of ART and anxiolytic agents.
Pharmacokinetic interactions
Buspirone
Buspirone is an anxiolytic agent that is metabolized via CYP 3A4,143 and caution is advised when given concomitantly with agents inhibiting CYP 3A4 (Table S1), such as ritonavir, though specific dosing adjustments are not available. There are sparse data available to provide further information about this DDI, though 1 case report detailed parkinsonian symptoms when a total of 75 mg daily of buspirone was combined with a ritonavir-containing ART regimenReference Clay and Adams144 (Table S4). The patient’s symptoms resolved 15 days after the buspirone dose was decreased and the indinavir/ritonavir portion of the ART regimen was changed. The buspirone dose of 75 mg daily is also worth noting in this case, as the maximum recommended daily dose is 60 mg daily.143 Nevertheless, concomitant administration of ritonavir-containing regimens with buspirone warrants careful monitoring. While there are no reports available describing DDIs between cobicistat and buspirone, because cobicistat is also a potent CYP 3A4 inhibitor,108 the same precautions would be advised with concomitant buspirone and cobicistat-containing ART regimens.
Benzodiazepines
Although human studies are lacking, midazolam has a wealth of data examining the effect of CYP 3A4 inhibitors on its metabolism, as midazolam is commonly used as the representative CYP 3A4 substrate in pharmacokinetic studies.Reference van der Lee, Dawood and ter Hofstede111, Reference Ieiri, Tsunemitsu and Maeda145, Reference Schmitt, Hofmann and Riek146 These studies unanimously demonstrate a marked increase in midazolam exposure associated with ritonavir use. The same has also been shown with cobicistat.Reference Mathias, German and Murray147 In accordance with these studies, the HIV treatment guidelines recommend against concomitant administration of oral midazolam with either ritonavir or cobicistat, and recommend parenteral midazolam be used as a single dose only and with careful monitoring.16 In addition, HIV treatment guidelines provide the same recommendations for midazolam use with concomitant efavirenz use, which may similarly inhibit the metabolism of midazolam.16
Diazepam is another benzodiazepine metabolized via CYP 3A4, with additional metabolism via CYP 2C19.Reference Kakuda, Scholler-Gyure and Hoetelmans148 Metabolism of diazepam also results in many active metabolites, which help contribute to its prolonged effect.Reference Kakuda, Scholler-Gyure and Hoetelmans148 Ritonavir and cobicistat, both strong CYP 3A4 inhibitors, are theorized to slow diazepam metabolism, thus prolonging its effects, and HIV treatment guidelines recommend avoiding co-administration of diazepam with these agents. Instead, use of alternate benzodiazepines not dependent on CYP metabolism, such as lorazepam or oxazepam, is recommended.16 The NNRTI class of antiretrovirals typically induces CYP activity, and so there is a possible decrease in diazepam exposure with concomitant use of efavirenz or nevirapine warranting close monitoring and dose adjustments as needed.16 Etravirine, however, in addition to its CYP induction properties, is known to inhibit CYP 2C19, which is one of the pathways of diazepam metabolism.Reference Kakuda, Scholler-Gyure and Hoetelmans148 There are no published reports available to provide further detail on the extent of this DDI. It is therefore recommended to proceed with caution when prescribing both etravirine and diazepam16; it may be prudent to start diazepam at a lower dose and carefully titrate to effect.
Triazolam, although utilized in the treatment of insomnia rather than as an anxiolytic, is also metabolized via CYP 3A4. In vitro studies show a significant impairment of triazolam metabolism in the presence of ritonavir,Reference von Moltke, Greenblatt and Grassi149 and research in healthy human subjects confirms this.Reference Greenblatt, von Moltke and Daily150 The increased triazolam exposure associated with ritonavir-mediated CYP inhibition was also noted to increase sedation and psychomotor impairment in healthy subjects. Accordingly, co-administration of triazolam with either ritonavir or cobicistat is contraindicated.16 Concomitant use of triazolam and efavirenz is also not recommended.16
Alprazolam is also metabolized via CYP 3A4.Reference Frye, Bertz and Granneman151 Data regarding the effect of CYP 3A4 inhibitors on the metabolism of alprazolam are conflicting. One study found that ritonavir had no effect on the metabolism of a single dose of alprazolam,Reference Frye, Bertz and Granneman151 and based on this study, the contraindication for concomitant administration of alprazolam and ritonavir was removed.Reference Greenblatt, von Moltke and Daily150 Since then, another study conducted in healthy subjects receiving ritonavir and alprazolam found that the alprazolam AUC and half-life more than doubled, with more sedation and psychomotor impairment in subjects receiving ritonavir as compared to those receiving placebo.Reference Greenblatt, von Moltke and Daily150 Despite the lack of guidance regarding alprazolam co-administration with ritonavir (or cobicistat), it may be prudent to start at a lower alprazolam dose and to carefully titrate to effect if it is to be used with either ritonavir or cobicistat. Of the NNRTIs, as they are expected to induce CYP metabolism, HIV treatment guidelines recommend monitoring for alprazolam efficacy and potentially increasing alprazolam doses with concomitant use of either efavirenz, etravirine, or nevirapine.16 This interaction, however, is theoretical, as there are no clinical data to offer insight on the magnitude of this interaction.
Compounding adverse effects
Due to the DDIs discussed above, there is concern for the potential augmentation of expected adverse events of anxiolytic agents. Benzodiazepines particularly pose a concern for excessive sedation, as kinetic studies in healthy volunteers noted an impairment in motor function as well as excessive sedation for concomitant use of ritonavir with triazolam or with alprazolam.Reference Greenblatt, von Moltke and Daily150 Although there are sparse clinical data illustrating this interaction, HIV treatment guidelines nonetheless suggest using caution with several of the benzodiazepines, as discussed above.16
Mood Stabilizers
There is nearly 4-times higher prevalence of bipolar disorder among adult patients with HIV than in the general U.S. population (8.1% versus 2.1%).Reference de Sousa Gurgel, da Silva Carneiro and Barreto Rebouças152 Additionally, patients with bipolar disorder are more likely to engage in risky sexual behaviorsReference de Sousa Gurgel, da Silva Carneiro and Barreto Rebouças152 and illicit substance use,Reference Hunt, Malhi and Cleary153 potentially increasing their risk for acquisition of HIV. Mood stabilizers have also been utilized in the treatment of manic syndromes in patients with HIV infection secondary to HIV encephalitis.Reference Halman, Worth and Sanders154
Pharmacokinetic interactions
Carbamazepine
Carbamazepine is a CYP 3A4 substrate and inducer in addition to a CYP 1A2 inducer, which increases the likelihood of pharmacokinetic interactions.155 The most commonly documented DDI with carbamazepine and ART is with ritonavir-containing regimens. Carbamazepine, when combined with ritonavir, a strong CYP 3A4 inhibitor,54 may reach toxic plasma levels, which could manifest as blood dyscrasias, ocular abnormalities, cardiovascular instability, neurologic deficits, and severe dermatologic reactions.Reference Bates and Herman156–Reference Mateu-de Antonio, Grau and Gimeno-Bayón160 Close monitoring of carbamazepine levels, with dose adjustments as indicated, is recommended within the first 2 weeks of initiating, withdrawing, or modifying CYP 3A4 inhibitor therapy.
In turn, carbamazepine is contraindicated with antiretroviral agents that are CYP 3A4 substrates (Table S1), as this would result in suboptimal antiretroviral plasma levels, such as with indinavirReference Hugen, Burger and Brinkman161 and dolutegravir,Reference Song, Weller and Patel162 where reduced plasma levels were observed in the presence of carbamazepine (Table S5). These agents may require a dose titration (ie, dolutegravir 50 mg daily to twice daily) to maintain virologic suppression, and close monitoring of viral load should occur to prevent subtherapeutic levels. Last, an interesting 2-way interaction was observed between carbamazepine and efavirenz, such that plasma levels of both medications decreased in the presence of one anotherReference Ji, Damle and Xie163; their co-administration is therefore contraindicated.
In addition, carbamazepine, as a CYP3A4 inducer, has been associated with hastened clearance of nevirapine,Reference Baranyai, Muro and Gödtel-Armbrust164, Reference Muro, Fillekes and Kisanga165 a CYP3A4 substrate,166 which may also increase the risk of nevirapine resistance.Reference Cressey, Jourdain and Lallemant167–Reference Muro, Droste and Hofstede170
Divalproex/valproic acid
Valproic acid displayed clinically significant DDIs when co-administered with zidovudine. In 2 case reports and 1 small study, plasma and cerebral spinal fluid (CSF) levels of zidovudine were found to be elevated,Reference Akula, Rege and Dreisbach171–Reference Lertora, Rege and Greenspan173 and 1 patient experienced severe anemia that was correlated with increased zidovudine levels.Reference Antoniou, Gough and Yoong172 The mechanism for this interaction was proposed to be due to valproic acid–induced inhibition of zidovudine glucuronidation. The American Academy of Neurology (AAN) issued a Level C recommendation that patients may require a decreased zidovudine dose if co-prescribed with valproic acid,Reference Birbeck, French and Perucca174 though the Department of Health and Human Services (DHHS) guidelines do not reciprocate this recommendation.16 The AAN also issued a Level C recommendation that co-administration of valproic acid and efavirenz may not require a dose adjustment,Reference Birbeck, French and Perucca174 based on 1 report where patients taking efavirenz and valproic acid displayed no significant pharmacokinetic alterations.Reference DiCenzo, Peterson and Cruttenden175 However, 1 conflicting case report suggests that a DDI exists between valproic acid and efavirenz, where the patient experienced decreased valproic acid concentrationsReference Saraga, Preisig and Zullino176; as such, DHHS guideline recommendations do not align with those of AAN and do suggest dose adjustment. Last, reports of valproic acid DDIs with ritonavir-containing regimens are mixed. One case study examined ritonavir, saquinavir, stavudine, and nevirapine with valproic acid and found no significant changes despite development of hepatic dysfunction that was correlated to either ritonavir or nevirapine.Reference Cozza, Swanton and Humphreys177 Another prospective study containing ritonavir and atazanavir found no significant results when co-administered with valproic acid.Reference DiCenzo, Peterson and Cruttenden178 However, 2 studies did find DDIs with valproic acid when the PI combination product lopinavir/ritonavir was co-administered. In the case report, the patient experienced decreased valproic acid concentrations, likely due to ritonavir-induced metabolism, which resulted in clinical symptoms of mania/hypomania.Reference Sheehan, Brouillette and Delisle179 In a prospective study, patients experienced increased lopinavir concentrations, and although this was deemed not clinically significant,Reference DiCenzo, Peterson and Cruttenden175 DHHS guidelines recommend monitoring for lopinavir toxicity in the presence of valproic acid.16 Ultimately, empiric dose adjustments are not usually warranted with the addition of a potentially interacting medication in absence of prior toxicity; however, close level monitoring with valproic acid dose titrations as indicated is advised.
Lamotrigine
Lamotrigine displayed decreased half-life and overall exposure when co-administered with lopinavir/ritonavir in 1 studyReference Jacobs, Colbers and Velthoven-Graafland110 and atazanavir/ritonavir in another study,Reference Burger, Huisman and Van Ewijk180 each likely due to ritonavir-mediated induction of glucuronidation. Decreased overall exposure of lamotrigine in patients with bipolar disorder may result in subtherapeutic concentrations, placing patients at risk of exacerbation to bipolar depression or mania/hypomania. With this, the AAN has issued a Level C recommendation that patients receiving this combination of medications may require a 50% increase in lamotrigine dosing to maintain therapeutic serum concentrations,Reference Birbeck, French and Perucca174 and this recommendation is consistent with DHHS guidelines.16 Currently, no data are available with respect to the administration of lamotrigine to patients receiving cobicistat-boosted PIs; however, DHHS guidelines recommend an alternative agent in this circumstance, or close monitoring of lamotrigine levels given the theoretical potential for reduced lamotrigine exposures with the combination.16
Lamotrigine, when co-prescribed with atazanavir alone or raltegravir alone, did not appear to have any significant DDIs, and as such dose adjustments with co-administration are not routinely required.
Lithium
Overall, lithium appears to have an underwhelming amount of literature relevant to pharmacokinetic concerns when co-administered with antiretrovirals. However, with no known effects on CYP enzymes or transporters, such hepatically mediated DDIs are unlikely.
Compounding adverse effects
Carbamazepine
Some antiretroviral agents, mainly NNRTIs, may independently cause Stevens–Johnson syndrome (SJS),Reference Luther and Glesby181 a potentially fatal dermatologic reaction that may also be caused by carbamazepine.155 It is possible that the risk of developing SJS may be compounded in patients receiving both carbamazepine and NNRTIs, and caution is advised. Similarly, antiretroviral agents such as zidovudine may cause bone marrow suppression, which may be problematic in patients receiving carbamazepine, as carbamazepine has been associated with aplastic anemia and agranulocytosis.155 Hematologic monitoring should be increased when carbamazepine is combined with antiretroviral agents known to disrupt hematologic parameters.
Divalproex/valproic acid
Another concern that may be compounded in the presence of antiretrovirals is the association between valproic acid and boxed warnings for hepatotoxicity and pancreatitis.182 Many antiretrovirals are associated with hepatotoxicity,Reference Surgers and Lacombe183 and agents such as didanosine, stavudine, and lopinavir/ritonavir are associated with pancreatitis.Reference Badalov, Baradarian and Iswara184 Administering these agents with valproic acid may result in further increased risk. Similarly, valproic acid has displayed dose-related thrombocytopenia and alopecia,182 both of which may also be caused by indinavir. Last, valproic acid is associated with dermatologic rash, which has also been associated with many antiretroviral agents,Reference DiCenzo, Peterson and Cruttenden175 and the risk of development may be compounded.
Lamotrigine
High serum concentrations of lamotrigine have also been associated with SJS.Reference Mockenhaupt185 The development of SJS in patients taking lamotrigine is mostly correlated to inappropriately rapid titrations or DDIs resulting in high plasma concentrations.Reference Raymond, Lori and Neil186 Therefore, it is critical that the prescribing clinician stay abreast of any changes in a patient’s ART, such as removal of inducing agents such as ritonavir, to avoid lamotrigine toxicity.
Lithium
Lithium remains a medication with a narrow therapeutic index that can become toxic in cases of dehydration and impaired renal function.187 Numerous antiretroviral agents are associated with nausea, vomiting, and diarrhea,Reference Anastasi and Capili188 which could increase a patient’s risk of fluid loss resulting in possible dehydration and risk for lithium toxicity. In addition, night sweats have been associated with advanced stages of HIV,Reference Rabeneck, Crane and Risser189 which could be another source of fluid loss and potential dehydration. Another concern is combining lithium with cardiotoxic antiretroviral agents, as lithium has been associated with atrioventricular block, bradyarrhythmia, and Brugada syndrome, though currently the increased risk posed by co-administration is theoretical.
More positively, lithium has demonstrated a potential pharmacodynamic benefit at ameliorating zidovudine-induced myelosuppression by taking advantage of its propensity to cause sustained leukocytosis and neutrophilia. Bone marrow toxicity is a treatment-limiting adverse effect of zidovudine therapy190; however, in small studies when lithium was co-administered, the incidence of these adverse effects is reduced despite the use of higher zidovudine doses than used in modern ART regimens.Reference Hernández, Pérez and Wilder191, Reference Roberts, Berman and Nakasato192 Although intriguing, recommendations against or in favor of this medication combination as a therapeutic strategy are beyond the scope of this review.
Last, lithium may independently impair renal function with long-term use as a result of damaging the renal tubules.187 Several antiretrovirals of the NRTI class have recommended dose reductions for renal impairment16 that should be checked for patients with declining renal function. However, the NRTIs emtricitabine and lamivudine, although renally eliminated, do not have dose-related toxicities and thus renal dose adjustments are less of a concern than for agents such as tenofovir disoproxil fumarate and zidovudine.Reference Berns and Kasbekar193 It should also be noted that in the past few years, 2 new medications with potential effects on serum creatinine laboratory values, dolutegravir and cobicistat, have been approved for utilization in HIV treatment. These agents have been shown to cause nonprogressive increases in serum creatinine due to inhibition of tubular secretion of creatinine without affecting renal glomerular function. This effect should be considered when interpreting changes in estimated creatinine clearance for patients receiving ART including these agents. The increase is typically modest (+0.1–0.2 mg/dL), but may exceed +0.5 mg/dL or more for some patients.108, 194 These increases typically occur in the first 2 months of treatment, after which they stabilize, and do not necessitate ART discontinuation or changes in therapy. By contrast, tenofovir disoproxil fumarate, an NRTI with nephrotoxic properties, like lithium is also associated with true renal tubular toxicity, and concern for additive toxicity is reasonable. However, in a randomized, placebo-controlled trial investigating the safety of lithium co-administration to patients with HIV receiving ART containing tenofovir disoproxil fumarate, no increased nephrotoxicity was found in the lithium group versus placebo after 24 weeks of therapy.Reference Decloedt, Lesosky and Maartens195 This result is reassuring; however, given the study’s small sample size of 53, closer monitoring for changes in renal function for these patients is still advisable.
Other adverse effects that occur independent of the co-administration of mood stabilizers and antiretrovirals, but may be additive upon concomitant use, include hypotension, constipation, nausea, vomiting, dizziness, and somnolence. Hypotension, as previously mentioned, may be particularly problematic for patients receiving carbamazepine with HIV-related dysautonomia or those who are taking the CCR5 co-receptor antagonist maraviroc, which has also been associated with this side effect.68 Constipation and other GI-related side effects such as nausea and vomiting are present with NRTIs and PIs, which may worsen in the presence of carbamazepine, lamotrigine, valproic acid, or lithium. Dizziness is a known adverse effect of efavirenz, carbamazepine, lamotrigine, and valproic acid. Although the combination of carbamazepine and efavirenz is contraindicated, this adverse effect would likely be compounded if co-administered. Lastly, somnolence and fatigue are associated with NNRTIs, which may be more pronounced in the presence of carbamazepine, lamotrigine, or valproic acid. The extended-release formulation of lamotrigine has been associated with insomnia, and this would likely worsen with antiretrovirals that are also associated with insomnia, such as efavirenz, rilpivirine, or INSTIs. Although many of these side effects are benign, they can intensify upon co-administration. Efforts to alleviate side effects should be actively sought out, and selection of alternative mood stabilizers or antiretrovirals that are less likely to compound these effects should be considered.
Treatments for Opioid Use Disorder
Intravenous drug use is a significant risk factor for the acquisition of HIV,Reference Lansky, Brooks and DiNenno196 as well a principle factor in nonadherence to ART.Reference Arnsten, Demas and Grant197 Methadone and buprenorphine are effective treatments for suppression of opioid withdrawal symptoms, and methadone utilization has been associated with decreases in HIV incidence rates.Reference Bruce, Altice and Moody198
Pharmacokinetic interactions
Methadone
The mechanism of methadone inactivation involves as many as 6 CYP enzymes, though hepatic clearance via CYP 3A4 and 2B6 play the largest role,199 which introduces the possibility of DDIs with antiretrovirals. However, a large body of evidence suggests that the role of CYP 3A4 metabolism is relatively minor for methadone compared to CYP 2B6,Reference Kharasch, Bedynek and Hoffer200–Reference Kharasch and Stubbert202 which limits the interacting drugs of greatest concern to those with substantial impact on CYP 2B6 function.
Integrase strand transfer inhibitors
No clinically significant DDIs have been identified between methadone and members of the INSTI class. In pharmacokinetic studies of HIV-negative volunteers receiving methadone maintenance therapy, no statistically significant effects on methadone pharmacokinetic parameters were observed with the addition of either raltegravir or dolutegravir dosed twice daily.Reference Anderson, Mabalot Luk and Hanley203, Reference Song, Mark and Chen204 As the INSTI elvitegravir requires co-administration with cobicistat, a CYP 3A4 inhibitor, a concern for drug interactions is reasonable. However, among HIV-seronegative patients stabilized on at least 2 weeks of methadone, no significant increases in methadone concentrations were observed, suggesting that the combination may be administered safely without dose adjustments.Reference Bruce, Winkle and Custodio205 Additionally, among 25 treatment-experienced patients with HIV initiating therapy with the 2-drug regimen of etravirine plus raltegravir, 36% had discontinued their previous antiretroviral regimen due to DDIs, including DDIs with methadone.Reference Bañón, Moreno and Quereda206 No clinical failures were observed, supporting the use of INSTI-containing regimens for patients with previous issues with their ART related to DDIs.
Protease inhibitors
The potential for PIs to affect methadone concentrations is primarily driven by whether or not ritonavir is present as a boosting agent. As ritonavir is a CYP 2B6 inducer, lower concentrations of methadone are possible due to increased metabolism, with a concern for opioid withdrawal if newly added, or opioid excess upon ritonavir discontinuation.Reference Kharasch, Bedynek and Park201 This effect appears to be dose-dependent. For example, a methadone patient initiated on saquinavir/ritonavir 400/400 mg twice daily experienced pronounced opioid withdrawal symptoms 7 days later, including diaphoresis, abdominal pain, and hypotension, necessitating a 40% increase in his methadone doseReference Geletko and Erickson207 (Table S6), and another pharmacokinetic study of this interaction found a 32% decrease in methadone AUC with the same saquinavir/ritonavir dose (400/400 mg twice daily).Reference Gerber, Rosenkranz and Segal208 By contrast, a pharmacokinetic study of saquinavir/ritonavir 1000/100 mg twice daily found a significant methadone AUC reduction of only 19% that was deemed not clinically relevant,Reference Jamois, Smith and Morrison209 and no statistically significant differences in methadone concentrations were found when a 1600/100 mg daily dosing scheme was used.Reference Shelton, Cloen and DiFrancesco210 This result was upheld in a multicenter prospective study of HIV-positive patients receiving saquinavir/ritonavir 1500/100 mg daily with 2 NRTIs (24% on methadone) that found no relevant pharmacokinetic DDIs.Reference López-Cortés, Viciana and Ruiz-Valderas211 Minimal (<25% reduction) or no changes in methadone pharmacokinetic parameters were observed with darunavir/ritonavir (600/100 twice daily),Reference Sekar, Tomaka and Lefebvre212 fosamprenavir/ritonavir (700/100 twice daily),Reference Cao, Smith and Wire213 or boosted (100 mg ritonavir twice daily) or unboosted indinavir.Reference Kharasch, Bedynek and Hoffer200, Reference Kharasch, Hoffer and Whittington214 With respect to lopinavir/ritonavir (400/100 twice daily), no opioid withdrawal was observed in 2 observational studies of HIV-positive patients receiving lopinavir/ritonavir-based regimens while on methadone,Reference Clarke, Mulcahy and Bergin215, Reference Stevens, Rapaport and Maroldo-Connelly216 despite 1 study reporting 36% reductions in methadone AUC.Reference Clarke, Mulcahy and Bergin215 However, contradicting these findings was a pharmacokinetic study where 27% of methadone patients administered lopinavir/ritonavir reported experiencing symptoms of opioid withdrawal.Reference McCance-Katz, Rainey and Friedland217 Most concerning was 5 days after stopping antiretroviral therapy with lopinavir/ritonavir, a patient experienced torsades de pointes as a result of supratherapeutic methadone concentrations from failing to adjust his methadone dose to account for the removal of a CYP inducer.Reference Lüthi, Huttner and Speck218 Ultimately, opioid withdrawal is uncommon but may occur with the addition of a boosted PI to a patient receiving methadone. Dose adjustment is not usually required; however, patients should be monitored for signs and symptoms of opioid withdrawal, increasing the methadone dose as clinically indicated. More caution is warranted with the removal of a boosted PI, and upfront dose reductions should be considered.
Nelfinavir, an unboosted PI, also has CYP 2B6 inducing properties, and has been associated with 40–50% reductions in methadone concentrations despite substantial CYP 3A4 inhibition.Reference Hsyu, Lillibridge and Daniels219–Reference McCance-Katz, Rainey and Smith221 Although no opioid withdrawal was observed in these studies, a case report of a 40-year-old male with a 13-year history of receiving methadone was newly initiated on nelfinavir and experienced new onset musculoskeletal pain, abdominal cramping, and rhinorrhea.Reference McCance-Katz, Farber and Selwyn222 His methadone levels were found to be subtherapeutic, and he required a 185% increase in his methadone dose from 100 mg to 285 mg daily to alleviate his symptoms. Nelfinavir was discontinued, and methadone slowly titrated down to a stable dose of 125 mg daily. This case highlights the importance of monitoring for any changes in ART for patients on methadone, keeping in mind that the effect of drug inducers may not be fully realized until days or even weeks after initiation or removal.
Another PI that has been studied in combination with methadone is atazanavir. When administered unboosted, atazanavir has no effect on CYP 2B6 and as expected no significant effect on methadone concentrations.Reference Friedland, Andrews and Schreibman223 Of interest, 1 study did note decreased concentrations of atazanavir when co-administered with methadone compared to patients not receiving methadone, though the effect was only statistically significant for Cmax, and an increase in virologic failure was not observed.Reference Haberl, Moesch and Nisius224 The authors proposed that this result may be specific to the oral solution of methadone formulated in Germany, which is particularly viscous and may more readily complex with atazanavir to impair its absorption. In another real-world survey and pharmacokinetic study of patients with HIV receiving atazanavir- or lopinavir-based regimens, the patients who were also receiving methadone had lower troughs of atazanavir and lopinavir than their counterparts.Reference Higgins, Zingman and Slish225 It is unclear whether this is the result of a pharmacokinetic interaction or a reflection of a decreased likelihood of medication adherence in this population. Ultimately, although there were no significant differences in self-reported adherence rates, nonadherence is the most likely explanation, as methadone’s effects on the CYP system are inhibitory in nature, which would be expected to result in increased antiretroviral concentrations, if anything.
Non-nucleoside reverse transcriptase inhibitors
Opioid withdrawal is common with the addition of the CYP 2B6 inducers efavirenz and nevirapine to patients receiving methadone maintenance therapy, and in case reports and prospective cohort trials, a dose increase for methadone was often warranted to counteract the decreased plasma levels. In trials detailed in Table S6, whether methadone was administered with efavirenz, nevirapine, or both agents, methadone plasma levels decreased significantly, eliciting symptoms of clinical opioid withdrawal in these patients.Reference Altice, Friedland and Cooney226–Reference Stocker, Kruse and Kreckel239 Almost all patients required methadone dose titrations; however, the magnitude of this increase varied widely from 15–157%. Accordingly, guideline recommendations note that increased methadone doses are often necessary with these combinations, but do not advocate for a specific percent increase.16 Due to the delayed nature of CYP enzyme induction, patients receiving methadone maintenance should be closely monitored for at least 2 weeks following the initiation of efavirenz or nevirapine therapy and advised to report any withdrawal symptoms such that timely methadone dose increases may be made as indicated. Additionally, for patients switching from efavirenz- or nevirapine-based therapy to an alternative ART regimen, or who require suspension of their ART, proactive reductions in these patients’ methadone doses is likely needed to prevent serious and potentially fatal methadone toxicity in anticipation of increasing methadone levels following the removal of strong CYP 2B6 inhibition.
No statistically significant differences in methadone concentrations were observed with etravirine.Reference Schöller-Gyüre, van den Brink and Kakuda240 The minor statistically significant effects on methadone disposition observed with rilpivirine (16% AUC reduction) are unlikely to be clinically relevant and do not require proactive dose adjustments; however, clinical monitoring for opioid withdrawal for patients newly initiated on therapy would still be prudent.Reference Crauwels, van Heeswijk and Vandevoorde241
Nucleoside reverse transcriptase inhibitors
NRTIs have not been shown to impact methadone concentrations, with the exception of a minor 22% increase in methadone clearance by abacavir not requiring dose adjustment.16, Reference Rainey, Friedland and Snidow242, Reference Smith, Kearney and Liaw243 However, methadone was shown to reduce concentrations of didanosine and stavudine in 1 study, which the authors theorized was primarily due to decreases in the antiretrovirals’ bioavailability and delayed drug absorption.Reference Rainey, Friedland and McCance-Katz244 As such, switching to alternative ART is advised. Additionally, methadone has been shown to increase zidovudine exposure up to 2-fold in some patients as a result of methadone-mediated inhibition of zidovudine glucuronidation and decreased renal clearance.Reference Burger, Meenhorst and ten Napel245–Reference Schwartz, Brechbühl and Kahl247 Patients should be monitored for zidovudine-related adverse effects such as anemia, and symptomatic patients may require zidovudine discontinuation. Reduced zidovudine doses (eg, 200 mg twice daily) have be used by some clinicians in an attempt to manage this DDI for patients with anemia; however, this practice is not supported by the guidelines and no clinical studies to date have evaluated the safety or efficacy of this practice.
Buprenorphine
Buprenorphine is mainly metabolized by CYP 3A4, whereby it is converted into its 2 primary metabolites: norbuprenorphine and buprenorphine-3GA.248 Fluctuations of buprenorphine plasma levels and any of its relative metabolites may occur when it is co-administered with antiretrovirals affecting CYP 3A4, such as ritonavir. As such, studies have shown a statistically significant increase in buprenorphine plasma levels in subjects concomitantly administered buprenorphine and ART regimens involving ritonavir.Reference Bruce, Altice and Moody249–Reference McCance-Katz, Moody and Smith252 This type of DDI increases the risk of buprenorphine toxicity and can create the potential for opioid excess. Despite this risk, subjects in these studies did not show any signs of opioid excess or withdrawalReference Bruce, Altice and Moody198, Reference Sekar, Tomaka and Lefebvre212, Reference Bruce, Altice and Moody249–Reference Gruber, Rainey and Moody253 (Table S6). However, patients administered buprenorphine with unboosted atazanavir did experience increased sedation and reversible increases in total bilirubin.Reference McCance-Katz, Moody and Morse251 HIV guideline recommendations advise against co-administration of buprenorphine with unboosted atazanavir, and also close monitoring of patients receiving buprenorphine with atazanavir/ritonavir for signs of excessive sedation and overmedication, with similar caution advised for other agents within the PI class.16 When patients with HIV are newly initiated on buprenorphine and are receiving a PI or the INSTI elvitegravir, it is advisable to select the lowest initial dose, titrating up to effect as needed. Although theoretically CYP 3A4 inducers such as the NNRTI efavirenz may reduce buprenorphine concentrations, dose adjustments are not recommended by the guidelines,16 though monitoring for withdrawal symptoms or opioid excess for patients initiated or discontinued on efavirenz, respectively, is advised.
Compounding adverse effects
Buprenorphine has been associated with a potential for hepatotoxicity, with the ability to produce necrosis of hepatocytes in vitro Reference Berson, Fau and Fornacciari254 and transaminase elevations observed in patients following intravenous injection.Reference Peyrière, Tatem, Bories, Pageaux, Blayac and Larrey255 Patients with HIV newly receiving buprenorphine/naloxone and atazanavir, a PI associated with a high rate of hepatic enzyme abnormalities, were compared prospectively in a cohort study to patients not receiving ART with atazanavir.Reference Vergara-Rodriguez, Tozzi and Botsko256 Encouragingly, no significant changes in AST or ALT values were observed between the groups over a 6-month follow-up period, and no signs of clinical toxicity were noted. PIs in general have been associated with AST/ALT elevations, though atazanavir was the only PI that has been directly studied. Theoretically, caution is advised when co-administering buprenorphine with PIs, and monitoring for liver enzyme elevations should still be employed.
Both methadone and buprenorphine have the potential to prolong the QT interval, a risk factor for life-threatening torsades de pointes ventricular tachycardia. An estimated 83% of patients receiving long-term methadone maintenance treatment have a QTc interval exceeding normal reference values.Reference Maremmani, Pacini and Cesaroni257 Within the HIV population, prolonged QTc is present at a frequency of approximately 10%.Reference Chinello, Lisena and Angeletti258 Although no study has directly evaluated the safety of the combination, significant caution is advised when methadone or buprenorphine is co-administered with ritonavir-boosted saquinavir, which has been associated with significant QT interval prolongation, even in patients without a history of cardiac disease.74 Although the association of other PIs with QT prolongation is controversial,Reference Soliman, Lundgren and Roediger75 nelfinavir, ritonavir, atazanavir, and lopinavir have demonstrated cardiomodulatory effects in vitro,Reference Anson, Weaver and Ackerman76, Reference Gallagher, Kieran and Sheehan259 and nelfinavir and the NNRTI efavirenz were associated with an increased risk for QTc prolongation in a nested case-control study,Reference Chinello, Lisena and Angeletti258 with efavirenz’s package insert additionally noting a QTc prolongation potential.78 One study evaluated the risk of QTc prolongation in subjects who received buprenorphine and antiretrovirals such as efavirenz, nelfinavir, delavirdine, ritonavir, and lopinavir/ritonavir.Reference Baker, Best and Pade260 Results showed that the presence of both buprenorphine and ART showed a modest increase in the QTc interval for all therapies, with delavirdine showing the greatest increase of 13.12 msec. However, researchers did not believe that this interaction was clinically significant enough to warrant a dose reduction. With respect to methadone, 3 patients receiving ART and methadone (doses ranging from 80–125 mg daily) experienced prolonged QTc, leading to sustained ventricular tachycardia.Reference Gallagher, Kieran and Sheehan259, Reference Prosser, Mills and Rhim261 One patient had taken an overdose of flunazepam prior to his episode,Reference Gallagher, Kieran and Sheehan259 while another was positive for cocaine per urine toxicology.Reference Prosser, Mills and Rhim261 Two patients had their current ART discontinued,Reference Gallagher, Kieran and Sheehan259 while another was discharged with a plan to be tapered off of methadone.Reference Prosser, Mills and Rhim261 In general, the benefits of therapy should be weighed against the potential for adverse effects, taking into consideration patient-specific risk factors and whether increased vigilance with respect to cardiac monitoring is warranted.
Other adverse effects that occur independent of the co-administration of maintenance medications for opioid use disorder and antiretrovirals, but may be additive upon concomitant use, include hypotension, constipation, nausea, vomiting, dizziness, sedation, and diaphoresis. Hypotension, as previously mentioned, may be particularly problematic for patients with HIV-related dysautonomia or those who are taking the CCR5 co-receptor antagonist maraviroc, which has also been associated with this side effect.68 Constipation and other GI-related side effects such as nausea and vomiting are present with NRTIs and PIs, which may worsen in the presence of methadone or buprenorphine. Dizziness is a known adverse effect of efavirenz as well as methadone and buprenorphine, and this adverse effect would likely be compounded if co-administered. Somnolence and fatigue are associated with NNRTIs, which may be more pronounced in the presence of methadone or buprenorphine. Methadone and buprenorphine have been associated with diaphoresis and night sweats, which has also been associated with advanced stages of HIV.Reference Rabeneck, Crane and Risser189 Although many of these side effects are benign, they can intensify upon co-administration. Efforts to alleviate side effects should be actively sought out given the limited availability of alternative agents for maintenance treatment of opioid use disorder.
Treatments for Alcohol Use Disorder
Co-occurring alcohol use disorder (AUD) and HIV is a multifaceted healthcare concern. AUD may negatively impact overall physical health, including a heighted risk of hepatic dysfunction and damage.Reference Rocco, Compare and Angrisani262 Patients with co-occurring HIV are even more likely to experience hepatic damage due to increased risk of hepatitis B or C virus co-infectionReference Rocco, Compare and Angrisani262 or from concurrent hepatotoxic antiretrovirals.Reference Abrescia, D’Abbraccio and Figoni263 In addition, heavy alcohol consumption has been negatively associated with medication non-adherence, with patients being 2.5 times less likely to adhere to ART.Reference Harris, Pillinger and Fromstein264 Medication adherence is instrumental in decreasing drug resistance and preventing disease progression of an otherwise lethal disease. With this, it is absolutely essential to treat AUD in this patient population to ensure optimal outcomes.
Pharmacokinetic interactions
The 4 FDA-approved agents used to maintain abstinence are acamprosate, disulfiram, oral naltrexone, and intramuscular naltrexone.265–268 None of these medications have significant effects on CYP enzymes, and thus co-administration of these medications with antiretrovirals is mostly considered safe, as CYP-mediated DDIs are not anticipated. However, the few pharmacokinetic clinical caveats are detailed in Table S7 and described below.
Atazanavir, when co-administered with disulfiram, rendered disulfiram unable to inhibit aldehyde dehydrogenase, resulting in drug ineffectiveness; however, DDIs with ritonavir and efavirenz were either absent or not clinically significant.Reference McCance-Katz, Gruber and Beatty269 Clinicians should avoid prescribing disulfiram to patients receiving atazanavir given this risk for therapeutic futility.
With oral naltrexone, 1 study assessed for pharmacokinetic DDIs with zidovudine and found no statistically significant changes.Reference McCance-Katz, Rainey and Friedland270 Acamprosate and intramuscular naltrexone did not return any literature results related to co-administration with antiretrovirals.
Compounding adverse effects
An important clinical caveat includes the possibility of a reaction when disulfiram is co-administered with combination lopinavir/ritonavir.Reference Cvetkovic and Goa271, 272 The solution product contains ethanol, in which case an accumulation of acetaldehyde would occur if disulfiram were co-administered. This accumulation, known as a disulfiram-like reaction, may result in vomiting, flushing, diaphoresis, and palpitations, or may be severe enough to cause respiratory depression, cardiovascular collapse, arrhythmias, MI, acute congestive heart failure, unconsciousness, convulsions, and death.265 Another concern that may be compounded in the presence of antiretrovirals is the association between disulfiram and naltrexone with hepatitis or hepatic failure. Many antiretrovirals are associated with hepatotoxicity,Reference Núñez and Soriano273 though the risk for compounded liver toxicity is unknown. This theoretical increased risk does not preclude dual therapy, but may influence monitoring depending on each individual patient’s specific risk factors.
Other adverse effects that occur independent of the co-administration of maintenance medications for alcohol use disorder and antiretrovirals, but may be additive upon concomitant use, include nausea, vomiting, dizziness, insomnia, loss of appetite, and increased creatinine kinase levels. Nausea and other GI-related side effects such as vomiting are present with NRTIs and PIs, which may worsen in the presence of acamprosate, disulfiram, and both formulations of naltrexone. Dizziness is a known adverse effect of efavirenz, as well as acamprosate, disulfiram, and both formulations of naltrexone, and this adverse effect would likely be compounded if co-administered. Insomnia occurs commonly among patients taking efavirenz, rilpivirine, or INSTIs and is a known side effect of naltrexone products. Loss of appetite can occur with naltrexone products as well as abacavir, and can especially be problematic for patients with HIV, as this is possible with disease progression. Last, elevated creatinine kinase levels may occur with naltrexone products as well as zidovudine or other NRTIs, and, although not common, elevations can occur as a result of HIV-associated myopathy, which is often multifactorial.Reference Koubar, Estrella and Warrier274 Although many of these side effects are benign, they can intensify upon co-administration. Efforts to alleviate side effects should be actively sought out, and selection of alternative maintenance medications for alcohol use disorder or antiretrovirals that are less likely to compound these effects should be considered.
Conclusions
While the management of HIV has evolved over the past decade, the potential for DDIs remains high with many commonly utilized antiretroviral regimens. Many psychotropic agents are primarily dependent on CYP metabolism, and thus, are sensitive to CYP alterations and susceptible to reduced therapeutic efficacy or increased toxicity as a result of altered plasma concentrations or overlapping adverse effect profiles with antiretrovirals. However, in most cases, psychotropics and antiretrovirals may be used safely, particularly when care is taken to either select agents with a reduced potential to interact or known DDIs are proactively managed. Clinicians should be aware of the pharmacokinetic and pharmacodynamic properties of these agents to best direct therapy and to provide optimal patient care.
Supplementary materials
To view supplementary material for this article, please visit https://doi.org/10.1017/S109285291800113X
