Most patients are asymptomatic for many years after becoming infected. But once symptomatic with more advanced HIV disease, they may become high users of medical care, including mental health services (270). They may need frequent assessment and treatment visits, undergo many medical tests, require hospitalization, and take multiple medications (often as many as 10–15 different medications at different times of day). Many patients receive medical care from an HIV primary care or infectious disease clinic or from an infectious disease specialist in the community. A recent study reported better medical outcomes for patients with HIV infection who receive care from physicians with more experience treating HIV diseases (271).

The Ryan White CARE (Comprehensive AIDS Resources Emergency) Act was enacted in 1990 to help states, communities, and families provide care for people with HIV/AIDS who lack adequate health insurance or other resources. Since its inception, CARE Act grants have totaled $6.4 billion and have served approximately 500,000 individuals with HIV/AIDS annually. Administered by the Health Resources and Services Administration of the U.S. Department of Health and Human Services, CARE Act grants were reauthorized in 1996 and comprise several programs. Title I grants provide funds to cities for low-income or under- or uninsured persons to cover the cost of health care and medications, as well as support services like counseling and home and hospice care. Title II grants are for the 50 states, the District of Columbia, Puerto Rico, and U.S. territories to improve the quality, accessibility, and organization of both health care and support for patients with HIV/AIDS. Title III grants are given to public or nonprofit entities that provide comprehensive primary health care services for persons with AIDS or at-risk populations. Title IV grants fund efforts to coordinate HIV services and access to research for children, youth, women, and families. An additional component of the CARE program includes training for health care professionals, grants for innovative service delivery models for special populations, and dental services.

Despite CARE Act funding, lack of access to care remains a problem that affects morbidity and mortality for certain groups. One study documented that among urban patients who receive medical care, there are no differences in HIV disease progression or survival associated with sex, race, injection drug use, or socioeconomic status. This suggests access to medical care is a main determinant of how quickly a patient develops AIDS (272). More recently, investigators from the HIV Cost and Services Utilization Study, a national population-based sample, found that African American, Hispanic, female, uninsured, and Medicaid-insured individuals all had inferior patterns of medical care (e.g., later initiation of antiretroviral medication) than did Caucasian and privately insured individuals (273).

Persons with HIV infection have higher rates of most psychiatric disorders than the general population. Psychiatric patients with common medical problems, especially those with severe mental illness, receive suboptimal medical care and suffer greater morbidity and mortality from medical illnesses than the general population. Studies are now under way to determine the access of patients with severe mental illness to HIV-related medical treatments; it can be assumed that these patients need help to access medical services.

Psychiatric disorders can be underdetected, misdiagnosed, undertreated, or treated improperly in primary care settings. For example, while a substantial majority of HIV-infected persons presenting for intake to an HIV primary care clinic suffered from depression, substance dependence, or another current psychiatric disorder, most cases were not detected by clinicians (66). Underscoring the importance of mental health care is the finding that ongoing high-risk sexual behavior is predicted by higher levels of depression and recreational drug use (274).

There are several patient barriers to meeting the mental health needs of HIV-infected persons. These barriers include adherence difficulties (e.g., missed appointments or inability to follow treatment recommendations); fear of additional stigmatization (the "double stigma" of HIV and a mental disorder); negative prior experiences with the mental health system; lack of funds for transportation, payment of care, or medications; unstable housing and frequent relocation; and the complex comorbid presentations of patients with HIV infection (e.g., depression complicated by concurrent substance dependence and a personality disorder). On the other hand, health care professionals should not assume that depression is "normal" in a person with HIV infection.

Some innovative HIV clinics integrate medical, mental health, and substance abuse treatment within their programs (275). These model programs increase the opportunities for prevention and treatment and may reach a more impaired population (276). One model uses an outpatient psychiatric consultation-liaison clinic that specializes in the treatment of HIV-infected patients. The clinic can be physically off-site from an HIV clinic but receives most of its referrals from that clinic. Other models include HIV programs within community mental health centers, programs located within residential substance abuse treatment facilities, freestanding community-based agencies, and referrals to private practitioners with expertise in treating HIV patients. All of these models appear to be effective (67).

In general, the use of psychotropic agents in HIV patients follows similar principles for using these medications in such populations as geriatric patients or those with comorbid medical illnesses. This is particularly true for patients with more advanced HIV disease and those on complex antiretroviral regimens who may be more sensitive to medication dosages and side effects and are at higher risk for drug-drug interactions (277). Because some standard medications used to treat HIV infection or HIV-related conditions can potently inhibit or induce the cytochrome P450 system, psychotropic medications that share metabolic pathways should be used judiciously; however, this concern should not preclude appropriate pharmacologic intervention.

General guidelines, particularly for patients with symptomatic HIV disease, include 1) using lower starting doses and slower titration, 2) providing the least complicated dosing schedules possible, 3) focusing on drug side effect profiles to avoid unnecessary adverse events (e.g., anticholinergic effects from tricyclic antidepressants, leukopenia from carbamazepine), and 4) maintaining awareness of drug metabolism/clearance pathways to minimize drug-drug interactions and possible end organ damage.

Published reports of adverse events and side effects can offer specific guidance for the use of some psychotropic agents in HIV-infected patients. However, clinicians must be cautioned that many studies that have examined psychotropic effectiveness in HIV patients were conducted on relatively asymptomatic patients, often did not include women or youth, and were conducted before the availability of combination antiretroviral therapy. Liver disease, in particular hepatitis C, is becoming increasingly important as a comorbid condition. Thus, routine liver function testing is necessary when prescribing psychotropic medications for patients with HIV infection.

Combination therapy with multiple medication classes constitutes the standard of care for successful treatment of HIV infection. Although limited clinical information is available concerning drug-drug interactions involving psychotropic and antiretroviral medications, effective psychotropic medication management in the setting of HIV requires attention to possible interactions.

Potential drug-drug interactions in the setting of antiretroviral therapy include the effects of antiretroviral agents on psychotropic medications and vice versa. Protease inhibitors and nonnucleoside reverse transcriptase inhibitors are metabolized by the cytochrome P450 system and may inhibit or induce multiple isoenzymes. For example, the protease inhibitor ritonavir inhibits the cytochrome P450 isoenzymes 3A, 2D6, and 2C9/19; thus, most psychotropic medications may be affected when concomitantly used with ritonavir. Other protease inhibitors tend to only inhibit the cytochrome P450 isoenzyme 3A, thus limiting possible drug-drug interactions to psychotropic medications metabolized via this pathway (e.g., some benzodiazepines, citalopram, and nefazodone).

The nonnucleoside reverse transcriptase inhibitors also require careful attention. For example, nevirapine and efavirenz are metabolized via the cytochrome P450 isoenzymes 3A and 2B6 and induce cytochrome P450 activity, which could result in decreased psychotropic concentrations at standard doses (278). Delavirdine, another nonnucleoside reverse transcriptase inhibitor, inhibits the cytochrome P450 isoenzyme 3A and reduces hepatic clearance of psychotropic agents primarily metabolized by this pathway (279). The protease inhibitors and nonnucleoside reverse transcriptase inhibitors are among the most problematic of the antiretrovirals when coadministered with psychotropic medication. Table 15 indicates the relative rank order of inhibition potency of the currently available protease inhibitors.

For combination antiretroviral therapy that includes protease inhibitors and nonnucleoside reverse transcriptase inhibitors, an overarching concern in terms of coadministration with psychotropic agents is not only the possibility of plasma levels of psychotropic medications outside the targeted therapeutic concentration range but also the possible reduction of antiretroviral levels to the point of risking their effectiveness. While most interactions inhibit the cytochrome P450 system, some clinically important drug interactions also induce this system with medications used to treat co-occurring medical conditions in HIV disease (e.g., glucocorticoids, rifampin, phenytoin). Table 16 provides an overview of psychotropic agents and HIV therapies, routes of metabolic inhibition and induction, and potential clinical concerns.

Although the aforementioned risks for drug-drug interactions may seem substantial, to date, clinical data have not revealed severe side effects among most patients receiving antiretrovirals in conjunction with psychotropic medications (50). A few reports have documented the inhibitory effects of protease inhibitors. For example, when coadministered with ritonavir, desipramine levels have been shown to increase 145% in vitro (280) and to increase to some extent when coadministered with other protease inhibitors (281). In a clinical setting, coadministration of saquinavir and midazolam has caused prolonged sedation (282). These in vitro and in vivo data should lead to careful monitoring for potentially problematic or dangerous side effects in patients taking antiretrovirals.

Drugs of abuse can also cause drug-drug interactions. MDMA, an amphetamine derivative also known as "ecstasy" that is primarily metabolized via the cytochrome P450 isoenzyme 2D6, reportedly led to a lethal overdose when taken by a patient being treated with ritonavir (283). Other drugs of abuse, particularly those metabolized via the cytochrome P450 isoenzymes 2D6 or 3A (e.g., amphetamines, ketamine, heroin, cocaine, and -hydroxybutyrate) may lead to toxic events when taken by patients being treated with protease inhibitors.

Although some drug-drug interactions may be only theoretical, it is prudent for clinicians to explain these concerns during the informed consent process of psychotropic prescribing. Because of the extent of comorbid substance abuse among some HIV patients, a general warning of the possible interactions of combining antiretroviral agents, psychotropic medications, and recreational drugs is often indicated. Web sites are available to clinicians and patients that address antiretroviral drug interactions and may be useful educational and reference tools (http://www.aegis.com; http://www.drug-interactions.com; and http://www.dml.georgetown.edu/depts/pharmacology/davetab.html).

Psychiatrists should remember that HIV patients who have been clinically stable on regimens of maintenance psychotropic medications and who later initiate combination antiretroviral therapies may require adjustments or changes in their psychotropic regimen. The rational choice of psychotropic medications must also include critical considerations of adherence. Generally, it is important to minimize the number of drug doses and to tie dosing to times of the day with natural cues (e.g., meals, work schedule, bedtime). Outlined here are unique factors to consider in administering specific psychotropic agents to HIV patients, particularly those patients simultaneously receiving antiretroviral therapies.

1. Antidepressant agents

The prevalence of mood disorders in HIV patients is higher than general population rates; evidence from effectiveness studies and clinical trials indicates that antidepressants are generally well tolerated by HIV patients, even those with symptomatic HIV infection or AIDS (284–290). Thus, antidepressants are commonly prescribed to HIV patients for the treatment of depression as well as for other HIV-related conditions, such as chronic pain (e.g., peripheral neuropathy) (291). However, clinicians must be cautioned that two factors may limit the generalizability of some antidepressant studies. First, many studies of antidepressants in HIV patients were conducted before the availability of combination antiretroviral therapy and thus did not take into account possible pharmacokinetic complications (292). Second, a minority of published psychopharmacology studies has included patients with advanced stages of HIV infection or AIDS. There is, to date, no evidence to suggest adverse effects of antidepressants on the immune system (293).

a) SSRIs

Several studies have indicated that tricyclic antidepressants are less well tolerated than SSRIs (287, 294), leading some to recommend SSRIs over tricyclic antidepressants (277). One study found that despite efficacy for treating depression, fluvoxamine was poorly tolerated by HIV patients (295). Because similar metabolic pathways, particularly the cytochrome P450 isoenzyme 2D6, are utilized by SSRIs and ritonavir, the psychiatrist should maintain communication with other clinicians treating the patient. Although SSRIs are generally well tolerated across a broad range of blood levels, the untoward outcome of serotonin syndrome may be possible in some cases in which the cytochrome P450 isoenzyme 2D6 is inhibited (296). While the SSRIs themselves inhibit this isoenzyme, the potency of their inhibition is less than that of ritonavir. However, the potential impact of SSRI-related inhibition of the cytochrome P450 isoenzyme 2D6 should be remembered when coadministering SSRIs with other agents that are substrates of the cytochrome P450 2D6 isoenzyme.

b) Tricyclic antidepressants

While tricyclic antidepressants are metabolized by other cytochromes, they rely principally on the cytochrome P450 isoenzyme 2D6 for clearance. Therefore if tricyclic antidepressants are coadministered with ritonavir, monitoring of ECG results and tricyclic antidepressant plasma levels is recommended because of the possibility of increased tricyclic antidepressant plasma levels from cytochrome P450 inhibition and associated tricyclic antidepressant side effects (e.g., delayed cardiac conduction, anticholinergic effects, orthostasis).

c) Other antidepressants

Nefazodone and the protease inhibitors are both potent inhibitors of the cytochrome P450 isoenzyme 3A, and their combination may be problematic. Nefazodone administration may theoretically decrease the metabolism of some protease inhibitors, leading to increased protease inhibitor plasma levels and increased side effects. However, one study that examined the efficacy of nefazodone in HIV patients found that only one of 15 patients had a clinically significant drug interaction that was related to coadministration with ritonavir (288). Of general concern is the coadministration of nefazodone or protease inhibitors with the antipsychotic pimozide, which can cause cardiac arrhythmias at increased plasma levels (296). Other potentially dangerous interactions with inhibitors of the cytochrome P450 isoenzyme 3A include the coadministration of most benzodiazepines (respiratory depression), clozapine (seizures), ergot alkaloids (sustained systemic vasoconstriction), and sildenafil (priapism).

Venlafaxine has been reported to decrease the concentration of indinavir significantly even though a lack of such interaction was predicted based on each drug's pharmacokinetic parameters. At this time, venlafaxine should be avoided in those patients receiving indinavir, since decreases in concentration of protease inhibitors can affect treatment efficacy and increase the chances for viral resistance to develop (297).

Previously it was thought that bupropion was eliminated via the cytochrome P450 isoenzyme 2D6. However, recent data have shown that it is actually cleared via the 2B6 isoenzyme. Therefore, bupropion has been removed from the list of drugs contraindicated with ritonavir, and this is reflected in the revised product monograph.

No adverse events have been reported with the agents citalopram and mirtazapine. Citalopram has minimal inhibitory actions on the cytochrome system and may be particularly useful for patients taking multiple medications where drug-drug interactions are an ongoing concern. Mirtazapine, because of its sedating profile, may be particularly useful for depressed patients with insomnia. It may also be helpful for patients with poor appetite because of its tendency to increase appetite and also helpful for patients who may be prone to sexual dysfunction with other antidepressant agents.

2. Antipsychotic medications

Generally, the use of antipsychotic medication among patients with advanced HIV infection is associated with an increased incidence of extrapyramidal side effects. High-potency standard neuroleptics (haloperidol is the most commonly used in published reports) can be associated with severe dystonia, rigidity, akathisia, and parkinsonism (31, 298–301). Low-potency standard neuroleptics can also result in severe extrapyramidal side effects but overall appear to be less problematic than high-potency neuroleptics (302–305).

A study that contrasts with the perspective that standard neuroleptics are associated with severe extrapyramidal side effects in advanced HIV infection was conducted in hospitalized patients with AIDS who were randomly assigned pharmacologic treatment for delirium; 13 received chlorpromazine, and 11 received haloperidol (46). The study stated that "no clinically significant medication-related side effects were noted." However, patients were treated with very low doses of neuroleptic (i.e., average maintenance doses were 36 mg/day for chlorpromazine and 1.4 mg/day for haloperidol). The study is consistent with the rest of the literature in stressing the importance of treating patients with advanced HIV infection with the lowest possible effective dose when using a typical neuroleptic.

Neuroleptic malignant syndrome has been noted with standard neuroleptics in five reports. In one case, neuroleptic malignant syndrome occurred within 24 hours; in the four other cases, neuroleptic malignant syndrome occurred within days of initiating treatment (300, 306–309). All patients recovered with usual management. Onset of tardive dyskinesia has been reported within a period as short as 6 weeks or within months (31, 310, 311). Confusion has also been reported with standard neuroleptics (312).

Most patients with extrapyramidal side effects respond to typical treatments, but one case report of a patient with parkinsonian side effects noted a slow response (299). Another patient had no response to multiple antiparkinsonian agents (313). Scurlock et al. (298) reported an AIDS patient who developed a grand mal seizure while taking chlorpromazine, and Jones et al. (314) reported a grand mal seizure in one man with symptomatic HIV infection who was taking amitriptyline and chlorpromazine.

Atypical antipsychotic agents appear to be better tolerated with fewer serious side effects, as evidenced by two case series of 30 combined patients who received risperidone (310, 315), one report of four patients who received remoxipride (298), and another of four patients who received molindone (312). Some of these patients had previously experienced severe side effects while taking typical neuroleptics. All were able to tolerate the atypical agents, and most but not all patients had a good therapeutic response. The most common side effects seen with risperidone were drowsiness and drooling (315). Extensive clinical experience suggests that olanzapine is also well tolerated for the treatment of psychosis, delirium, and behavioral disturbances among HIV patients.

It is of interest that movement disorders, including dystonia and parkinsonism, have been reported in advanced HIV infection without exposure to neuroleptics (316, 317). In the one case report of irreversible extrapyramidal side effects in a man with HIV-associated dementia following exposure to standard neuroleptics, an autopsy revealed numerous microscopic changes, including neuronal loss in the basal ganglia (313). It may be that damage to the basal ganglia is a factor that increases the susceptibility of HIV patients to extrapyramidal side effects after neuroleptic administration (318). A similar pattern of extrapyramidal side effect susceptibility in concert with neuronal damage has been described in the elderly (311, 319).

For all antipsychotic agents, the possibility of drug-drug interactions must be considered for those patients also taking antiretroviral medication. Antipsychotic dosage adjustments may be required to minimize side effects such as extrapyramidal side effects. Because of the rare dose-related side effect of agranulocytosis among patients taking clozapine, this particular agent should be monitored closely if given to any HIV patient (320). Likewise, clinicians should monitor for risk of seizures if clozapine is administered with ritonavir, since ritonavir inhibits the cytochrome P450 isoenzyme 2D6, clozapine's primary metabolic pathway. Pimozide should also be used cautiously in combination with any inhibitor of the cytochrome P450 isoenzyme 3A because of risk of fatal arrhythmias at elevated doses. Depot antipsychotic medication should be avoided in advanced HIV disease.

3. Psychostimulants

Psychostimulants, such as methylphenidate and dextroamphetamine, are frequently used in HIV patients with neurocognitive impairment, fatigue, and, to a lesser extent, in some patients with depression (321, 322). These drugs are generally safe and well tolerated with no drug-drug interactions reported to date among HIV patients. The possibility of dopamine agonism inducing psychosis should be considered in vulnerable patients, such as those with a history of HIV-related CNS complications or a history of psychosis. Psychostimulants should be avoided or used judiciously among patients with a past or current history of amphetamine abuse.

4. Mood stabilizers

a) Valproate

Use of valproate in patients with bipolar disorder is not uncommon, particularly among HIV patients who develop mania secondary to HIV-related complications or psychoactive substance use. Concerns have been raised about the potential for valproate to possibly accelerate disease progression in HIV-positive patients. This concern is based on the fact that valproate has been shown in vitro to lower glutathione levels, which leads to activation of HIV replication (323). However, a retrospective review by Maggi and Halman (324) of 11 patients with HIV infection who were also treated with valproate found no evidence of increased viral load in those receiving adequate antiretroviral therapy.

The metabolic pathways of valproate are not fully delineated, but this agent has been associated with hepatotoxicity. There is a case report of valproic acid–induced hepatotoxicity when valproate was used together with ritonavir and nevirapine (325). Valproate inhibits glucuronyl transferase. Zidovudine uses glucuronyl transferase as its primary pathway of metabolic clearance. Therefore, increased plasma levels of zidovudine and associated side effects could result when the two agents are combined.

Generally, divalproex sodium is better tolerated than valproic acid, since it produces fewer gastrointestinal side effects and can be dosed less frequently (two or three times a day rather than four), particularly among HIV patients who may have preexisting gastrointestinal illness. Increased side effects may be caused by increased concentrations of free valproate in HIV patients because of factors such as the reduction of plasma protein levels, the increased percentage of free drug due to chronic illness, or drug-drug interactions that may decrease protein binding of valproate (326).

b) Lithium carbonate

Lithium is commonly used in HIV patients with primary bipolar disorder. Because it does not require hepatic metabolism, it is a rational choice for many patients taking antiretroviral agents who require extensive hepatic metabolism. Lithium should be used with great caution in patients who develop HIV-related nephropathy, a generally irreversible complication that can lead to decreased lithium clearance and possible lithium toxicity. When given at doses necessary to maintain serum lithium concentrations between 0.5 and 1.5 meq/liter, lithium has been associated with signs of toxicity in some HIV patients, even among patients without renal disease (327). Thus, clinical experience suggests that lithium is best avoided in patients with probable AIDS mania or advanced HIV disease due to risk of toxicity.

c) Carbamazepine

Although not absolutely contraindicated, carbamazepine is used less commonly in HIV patients because of its potential to cause bone marrow complications (e.g., leukopenia or aplastic anemia). This risk should be avoided in immunocompromised patients. Carbamazepine, like other anticonvulsants, phenobarbital, and phenytoin, is known to induce activity of the cytochrome P450 isoenzyme 3A. This induction can accelerate antiretroviral drug metabolism and possibly cause decreased plasma levels and diminished efficacy. It is recommended that carbamazepine concentrations be closely monitored.

d) Newer anticonvulsants

Lamotrigine, gabapentin, and topiramate are newer anticonvulsants that have recently been used off-label to treat selected populations of patients with bipolar disorder, typically those with rapid cycling, mixed mood states, or bipolar depression. While no data currently support their use for mood disorders in HIV patients, gabapentin may offer some advantages over other anticonvulsants because it is neither protein bound nor metabolized by the liver. Gabapentin is thus less likely than other anticonvulsants to lead to drug-drug interactions. Gabapentin has been successfully used in HIV patients to treat peripheral neuropathy, which can be both a complication of HIV disease and a side effect of some antiretroviral agents.

5. Anxiolytic and sedative-hypnotic medications

a) Buspirone

Buspirone is commonly used and generally well tolerated in HIV patients as a treatment for generalized anxiety disorder (328). There are no reports of HIV-related drug-drug interactions with buspirone. However, one case report describes induction of manic psychosis following a single dose of buspirone in a patient with asymptomatic HIV illness (329).

b) Benzodiazepines

Benzodiazepines are commonly used in HIV patients and are best used for short periods to minimize the risk of tolerance or addiction and to maximize effectiveness. This class of drugs is generally best avoided in patients with alcoholism and those with a history of substance abuse. Benzodiazepines are metabolized primarily by the cytochrome P450 isoenzyme 3A; however, oxazepam, lorazepam, and temazepam are metabolized by direct conjugation via glucuronyl transferase, resulting in shorter elimination half-lives and less likelihood of metabolite accumulation and associated side effects.

When some benzodiazepines are coadministered with inhibitors of the cytochrome P450 isoenzyme 3A (such as protease inhibitors), one would expect a decrease in benzodiazepine clearance with increased sedation, potentially resulting in respiratory depression. Even saquinavir's relatively low-potency inhibition of the cytochrome P450 isoenzyme 3A has been shown to prolong the sedative effect of midazolam (282). Although one study that examined the chronic effect of ritonavir administration on alprazolam metabolism reported a 10%–15% decrease in expected alprazolam levels (330), a more recent study found that short-term, low-dose ritonavir administration significantly impaired alprazolam clearance (331). These studies call attention to the need for judicious use of alprazolam in combination with ritonavir.

An additional consideration when using benzodiazepines that are metabolized by glucuronidation (e.g., oxazepam, lorazepam, temazepam) involves ritonavir's inductive effect on glucuronyl transferase activity (332). The concentration of glucuronyl transferase substrates may be decreased when these agents are coadministered with ritonavir. Therefore in some cases, patients taking one of these three benzodiazepines may require higher doses for symptom control. Benzodiazepine-dependent patients may also require higher doses to avoid the onset of withdrawal (296).

6. Medications for HIV-related wasting

Some patients with wasting syndrome can be helped by medications that serve as appetite stimulants to combat loss of appetite and associated weight loss. Four agents currently have been approved by the Food and Drug Administration (FDA). Two of these agents, dronabinol and oxandrolone, are Schedule II controlled substances. Because these agents have unique psychiatric side effect profiles, they should be monitored closely among patients who receive psychiatric care.

Dronabinol is a cannabinoid indicated for HIV-associated anorexia and nausea. It is a major active metabolite in marijuana and, like marijuana, has complex effects on the CNS, including sympathomimetic activity. Oxandrolone is an anabolic steroid synthetically derived from testosterone that is used for weight gain and promotion of lean body mass (333–335). Its side effect profile is similar to that of other anabolic steroids, including mood effects. Another agent, megestrol acetate, is a synthetic derivative of progesterone that increases fat rather than muscle gain. It also lowers testosterone and may produce side effects (e.g., mood disturbances) similar to those associated with glucocorticoids. Somatotropin is an injectable form of recombinant human growth hormone indicated for the treatment of AIDS wasting and cachexia (336–340). It is an anabolic and anticatabolic treatment that combats the loss of lean body mass and is generally well tolerated by patients. The effects of somatotropin are often lost when it is discontinued.

These medications can be extremely beneficial for patients with significant anorexia, weight loss, and nausea. However, because they are usually used in more advanced stages of HIV disease, when patients are more medically frail and receiving multiple other medications, psychiatrists should actively monitor response, side effects, and drug interactions. Limited data do not indicate a high incidence of psychiatric side effects.

Thalidomide, a potent teratogen once marketed as a sedative and morning sickness agent in the 1950s and 1960s, is currently available on a restricted basis for HIV patients for the treatment of wasting syndrome (341, 342). The agent is FDA-approved as an immunomodulatory drug for the treatment of severe HIV-related oral ulcers. Prescribing physicians must register under a special restricted distribution program with the FDA. Primary psychiatric side effects include CNS depression; few data are available concerning drug-drug interactions, although thalidomide does not appear to be metabolized by the liver to any large extent.

7. Testosterone

A number of clinical trials have found testosterone to be effective in the treatment of a number of HIV-related symptoms (e.g., fatigue, wasting, impaired sexual functioning, and mild depression) often associated with more advanced HIV disease (126, 343, 344). Although typically used to treat hypogonadism in the general population, clinical response to testosterone in HIV patients is not correlated with serum testosterone levels at baseline. HIV-infected men with normal levels of testosterone respond as well as those with lower than normal levels (126). The most common psychiatric side effects of testosterone are changes in mood and irritability.

8. Opiate analgesics

Many HIV patients are treated with opiate analgesics for both acute pain syndromes, such as that associated with Candida esophagitis, as well as chronic pain conditions (e.g., neuropathic and myopathic pain syndromes). Chronic pain is also common in advanced HIV disease, particularly among cachectic patients with wasting syndrome. Because opioids are metabolized primarily via the cytochrome P450 isoenzymes 2D6 and 3A, they should be used cautiously in conjunction with antiretroviral agents that are likely to increase their plasma levels.

Some opioids, such as codeine and hydrocodone (as well as the nonopiate tramadol), must be converted into active metabolites to produce analgesic effects. This conversion can be blocked by cytochrome P450 inhibitors, which results in reduced pain control in addition to adverse reactions from the buildup of unmetabolized drug (345). Oxycodone and morphine, which are primarily glucuronyl transferase substrates, are alternatives for pain control. However, in the presence of a glucuronyl transferase inducer such as ritonavir, their analgesic effect can be diminished as well (296, 332). As HIV disease progresses, the provision of safe, effective pain control requires clear understanding of the above factors.

9. Drugs used in the treatment of substance use disorders

a) Opioid dependence

There is limited information available on the extent and clinical implications of drug interactions in HIV-positive patients with comorbid substance use disorders. However, several studies have been published. Methadone has been shown to significantly increase zidovudine levels, but zidovudine has not been shown to affect methadone levels (346–348). Interim results from a study of the interaction of zidovudine with LAAM, buprenorphine, or naltrexone indicate that LAAM does not affect zidovudine levels, whereas buprenorphine may decrease levels nonsignificantly (349). Another report indicates that abacavir, a newer nucleoside analogue reverse transcriptase inhibitor, increases methadone clearance, although the proposed mechanism remains unclear (350). Preclinical studies of the coadministration of protease inhibitors and methadone revealed that ritonavir causes a twofold increase in the area under the curve for methadone level and that indinavir causes a 30% increase, compared to no increase with saquinavir (351). One case report has indicated that nelfinavir may induce cytochrome P450 enzymes leading to decreased methadone levels (352). Nevirapine and efavirenz induce cytochrome P450 enzymes important to methadone metabolism and are reported to precipitate opiate withdrawal symptoms in methadone maintenance patients (278, 353). The consequences of undetected drug interactions among opioid-dependent patients are complicated and include nonadherence with treatment regimens, the potential development of resistant HIV strains, increased illicit drug abuse, drug toxicity, and loss of efficacy. In administering methadone to HIV patients taking protease inhibitors or nonnucleoside reverse transcriptase inhibitors, clinicians should adjust doses to compensate for cytochrome P450 enzyme effects.

b) Alcohol dependence

There is no clinical information regarding the use of disulfiram in patients with HIV infection. Psychiatrists should therefore follow general guidelines for the use of disulfiram cautiously. Naltrexone, an opioid antagonist, is indicated in the treatment of alcohol dependence and for the blockage of the effects of exogenously administered opioids. Because naltrexone can cause hepatocellular damage when given in excessive doses, it should be used cautiously in patients with liver disease. It should also be noted that naltrexone's opioid antagonism might block the clinical effects of some commonly used medications in HIV patients that contain opioids (e.g., antidiarrheal medications, cough and cold preparations).

10. Alternative/complementary agents

Patients with HIV infection may use a wide range of alternative/complementary agents for preventive and symptomatic treatment of HIV-related complications. Of particular interest to psychiatrists are those agents used to treat psychiatric symptoms. For example, patients may take Hypericum perforatum (St. John's wort) or the nutritional supplement S-adenosylmethionine (or SAM-e) to treat depression, kava kava for anxiety, melatonin or valerian root for insomnia, and ginkgo biloba for memory problems. One published report found that St. John's wort, which contains an inducer of the cytochrome P450 isoenzyme 3A, dramatically lowered levels of indinavir (and presumably other substrates of the cytochrome P450 isoenzyme 3A) (354). The FDA thus issued a public health advisory that concomitant use of St. John's wort with protease inhibitors and nonnucleoside reverse transcriptase inhibitors is not recommended. There are no data regarding other alternative/complementary agents at this time.

Since the beginning of the AIDS epidemic, psychosocial support has emerged as a cornerstone of the comprehensive treatment for many persons infected and affected by HIV/AIDS. For many HIV patients, psychotherapy and psychosocial interventions have been invaluable in the search for meaning during the course of living with HIV (355). Psychotherapy can be an important intervention to address conditions that may interfere with a patient's acceptance of HIV illness or their ability to work cooperatively with their health care team. With the advent of new antiretroviral treatments, new psychotherapy themes have emerged, including changes in role definitions and life trajectory as well as treatment adherence challenges (356–358).

For patients who are socially disenfranchised, psychosocial treatments may consist of basic services that assist them in gaining access to food, shelter, transportation, and child care—essential elements of a care plan that diminish health care delivery barriers and promote mental health care utilization.

Generally, psychosocial treatments are those treatments, often offered in the community, that provide direct assistance along such parameters as social adjustment, social support, coping, and overall adjustment to HIV infection. Psychosocial interventions often occur in group formats, such as informal support groups, although more formal groups may tailor their membership (i.e., women, men, mothers, caregivers) or their mission (HIV risk reduction, treatment adherence) (359, 360). Other formats include peer counseling, a supportive intervention typically offered by HIV-positive consumers based in community AIDS service organizations.

Other psychosocial treatments have been developed that incorporate a service model (e.g., psychosocial rehabilitation programs for persons with HIV/AIDS, day treatment programs for HIV-positive persons with neurocognitive impairment) or a service theme (e.g., permanency planning for parents with advanced AIDS, HIV prevention, case management, disclosure of HIV status) (67, 79, 355, 361).

Researchers have described a wide range of psychological and psychosocial sequelae from HIV/AIDS (251). In their review of empirical findings, Kalichman and Sikkema (362) showed that psychological complications include distress, anger, guilt, bereavement, and the full spectrum of psychiatric disorders. Syndromes such as depression may appear, subside, and recur during the course of living with HIV/AIDS. Moreover, certain psychosocial conditions, such as increased social support and fewer incidents of HIV-related discrimination, have been shown to predict greater life satisfaction among HIV patients (143).

Psychosocial treatment of HIV-related complications will often involve referral of patients to community AIDS service organizations that specialize in the provision of psychosocial support. While many of the treatment issues are also themes that will emerge in psychotherapy (i.e., stigma, coping strategies, or disclosure of HIV status), support groups and other interventions provide a model for structuring social support over the course of living with HIV infection (362). Psychosocial services may take on added importance for patients living in rural regions, in which more community HIV-related stigma has been documented (143).

Psychosocial interventions have been shown to enhance adaptive coping strategies, social support, perceived sense of meaning and purpose, and self-esteem and also have been shown to lower anxiety and HIV risk behaviors among HIV-positive adults (359, 363–365). Because of the spectrum of social, psychological, and neuropsychiatric consequences of HIV/AIDS, psychiatric treatment should consider the relevance for a given individual of a wide variety of potential psychosocial interventions, spanning the spectrum from individual to family or group psychotherapy as well as psychodynamic/psychoanalytic, interpersonal, behavioral, or supportive approaches.

Psychiatrists caring for HIV-infected patients can be faced with complex legal and ethical issues, some of which are unique to HIV infection. The AIDS Litigation Project conducted a national review of cases that involved individuals with HIV/AIDS in the federal and state courts in the United States between 1991 and 1997. The review identified important subsets of litigation, which included issues of "testing and reporting; privacy, the duty to warn, and the right to know; physician standards of care in prevention and treatment; and discrimination and access to health care" (366).

Most U.S. states have legal requirements surrounding HIV testing that require written informed consent and a minimum amount of pre- and posttest counseling—both to those who test positive and to those who test negative. Furthermore, many states are required to report positive HIV test results to public health agencies. Some states mandate that the report include the name of the individual who tested HIV positive. All states have mandatory name reporting when a patient is diagnosed with an AIDS-defining condition.

For those infected with HIV, confidentiality of HIV status is often paramount, especially for patients who are asymptomatic or who are concerned about possible discrimination by their employers or by insurance companies. Psychiatrists are typically attuned to confidentiality issues and regard patient-doctor confidentiality as fundamental to effective treatment. Yet, there are limits to confidentiality, such as when there is potential for others to be harmed, and these limitations should be discussed with patients when treatment is initiated.

For persons infected with HIV, particularly those recently diagnosed, their recent sexual partners and/or those with whom they share injection equipment should be informed of their potential exposure to HIV. These partners can benefit from HIV counseling and testing and knowing their HIV infection status. Many health departments offer services known either as "partner notification" or prevention counseling and referral services (367–369).

Another important issue involves the dangerousness of HIV patients. A patient with HIV infection is not a danger to others in the absence of imminent threats to harm others. In general, there are no legal requirements for physicians to inform partners of HIV-infected patients of their HIV status, especially if the patient agrees to notify their partner or if they are clearly practicing "safer sex." However, because HIV-infected patients are able to transmit the virus and potentially put others in danger of infection, psychiatrists who care for HIV patients have an ethical requirement to counsel their patients about modes of HIV transmission and promote safer behavior, such as the consistent use of condoms and/or the use of clean needles. Psychiatrists can offer to assist patients in notifying individuals at risk or help refer a patient to public health officials for partner notification, if necessary.

On the other hand, if a patient has not disclosed their HIV-positive status to a partner and continues high-risk behavior, psychiatrists may have a duty to warn identifiable sexual partners or other at-risk individuals. APA policy reads that, "If a patient refuses to change behavior that places others at risk for HIV infection or to inform individuals at ongoing risk, or if the psychiatrist has good reason to believe that the patient has failed to or is unable to cease such behaviors or to inform those at risk, it is ethically permissible for the psychiatrist to notify identifiable individuals at risk or to arrange for public health authorities to do so" (370). Psychiatrists need to be mindful that while it may be ethically permissible to notify, it may not be legally permissible, and therefore psychiatrists should obtain legal advice before breaching confidentiality.

For patients who are impulsive or psychotic, protective measures such as voluntary or involuntary psychiatric hospitalization are options for treatment of psychiatric disorders and protection of others. For antisocial individuals, hospitalization should not be used merely as a means of social control.

During the course of HIV illness, and especially after the onset of AIDS, patients may experience cognitive impairment or symptoms of dementia. Such disturbances may affect a patient's ability to give informed consent to accept or refuse treatment. When psychiatrists are asked to conduct competency evaluations for patients with HIV/AIDS, a comprehensive mental status examination that includes an evaluation of cognition is critical, followed—if indicated—by formal neuropsychological testing and assessment of specific treatment decision-making capacity.

The opposite of a patient's refusal of treatment is the refusal on the part of physicians to provide HIV treatment to psychiatric patients. Because antiretroviral therapies involve complex medication regimens and adherence to treatment is critical, physicians may be biased against prescribing these treatments to patients with a comorbid mental illness such as a substance use disorder. While refusal of treatment due to poor adherence is appropriate, refusal of treatment solely on the grounds of a mental illness is not. When the psychiatrist is a consultant who does not intervene directly in the care of the patient, he or she will need to work with the primary clinician to help ensure that appropriate psychiatric treatment is received (371).

Potent antiretroviral treatment has given hope to patients infected with HIV yet simultaneously has increased the complexities of managing HIV illness (372). For instance, the combination of three or more medications used in most antiretroviral treatment regimens makes adherence difficult. In addition, some medications must be taken with food, others without food; dosing is usually twice a day but sometimes can be more frequent. Side effects can be significant, requiring discontinuation or the initiation of a new combination of medications. Because antiretroviral treatment should begin as an aggressive broad attack on HIV replication, doses of the drugs are generally not titrated upward but rather are initiated at the full dose. Thus, patients may experience many initiation side effects.

Adherence to antiretroviral therapy regimens is critical, as evidenced by a study that measured adherence by using a microelectronic monitoring system (95). The authors found that 95% adherence to treatment was associated with complete viral suppression but that failure rates increased sharply with less than 95% adherence. Researchers have documented that patients who receive potent antiretroviral treatment but who are not enrolled in clinical trials show viral suppression rates one-half those of study patients, which suggests the real-world difficulty of maintaining strict adherence (373).

Adherence to any kind of medical treatment is a well-established problem. Overall adherence rates vary from 20% to 80% and average 50% (374). Psychiatrists should make every effort to treat and control psychiatric disorders that affect a patient's ability to maintain close adherence to an HIV treatment regimen. Each year, about 7% of patients with AIDS develop HIV-associated dementia (375). This condition can interfere with ability to follow and understand treatment regimens. In addition, researchers have found that depression can be a cause of nonadherence and that treatment of depression might thus improve medical outcomes (96). Strategies that have proven useful in improving adherence for both psychotropic and antiretroviral medication have been reported and are summarized below (376).

1. Institutional factors

How a health care system is structured will strongly influence whether clinicians and patients can collaborate and enhance adherence. One important element is a respectful and culturally sensitive environment. The availability and continuity of clinicians over time with 24-hour/day phone coverage to handle crises that might disrupt a regimen (severe side effects, loss of medication), as well as clinician caseloads that permit time for adequate discussion with patients, all promote adherence. Hospitals do not always dispense medications on the 8- or 12-hour schedules required for antiretroviral dosing. Last, social service mechanisms such as case management for handling financial and logistical barriers to obtaining and following the regimens provide crucial support.

2. Physician factors

When initiating antiretroviral treatment, it is essential for the physician to prepare the patient. Because the best chance at viral suppression occurs with the first regimen, it makes sense to wait until the patient is ready. For a psychiatrist, it is important to treat or recommend treatment for comorbid alcohol/substance use disorders, depression, psychotic illnesses, or personality disorders that are likely to interfere with adherence (377). At the same time, if the patient is motivated to begin antiretroviral treatment, this can be a good opportunity to introduce psychiatric or substance use treatment if indicated.

3. Clinical strategies

Multiple approaches to enhancing adherence are often needed (Table 17). For longer-term adherence, individual medication management, in which a staff member provides assessment and problem-solving support, may improve adherence (378).

There are numerous psychiatric difficulties that can occur in the context of HIV infection, yet studies of the treatment of HIV-related psychiatric disorders are limited. This section summarizes the existing scientific literature on the treatment of specific psychiatric disorders for patients with concurrent HIV/AIDS.

1. Dementia and the spectrum of cognitive disorders

Most research studies on the treatment of cognitive disorders associated with HIV infection have focused on HIV-associated dementia. Fewer studies have looked at the benefits of intervention at earlier stages of cognitive decline despite the possibilities for treating disorders, such as minor cognitive motor disorder, that involve neuronal cell dysfunction rather than cell death. Treatment of neuropsychological impairment without clinical disorder has been studied to some extent, and protection against neuropsychological impairment has been studied least.

Pharmacologic treatment strategies for cognitive disorders can be divided into four types: 1) antiretroviral therapies, 2) therapies aimed at immunological measures or inflammatory mediators, 3) therapies aimed at bolstering the response of the brain to the onslaught of the infection (e.g., neurotransmitter manipulation), and 4) nutritional therapies. Most controlled studies have investigated the efficacy of antiretroviral therapies, and while these studies have advanced our knowledge about interventions for cognitive disorders, several key factors must be kept in mind. First is the fact that most published studies to date report on the treatment strategy of administering a single antiretroviral agent. Their findings are therefore difficult to interpret in light of the multidrug regimens that are now the standard of care in developed countries. Second, the reports vary widely with regard to the study population, since some studies enrolled subjects on the basis of established criteria for HIV-associated dementia or minor cognitive motor disorder but other studies enrolled cognitively impaired subjects without specifying whether they also met criteria for a clinical disorder. Third, the range of HIV clinical severity also varied widely in study subjects.

Whether antiretroviral agents penetrate the blood-brain barrier sufficiently to adequately suppress viral replication is a key issue that requires further study (379, 380). We know that near-perfect adherence to antiretroviral therapy is required to maintain plasma viral load at undetectable levels. Even in the best-case scenario, if decreased plasma viral load is achieved and levels of antiretroviral agents in CSF or brain tissue are optimized, it is not yet known whether complete suppression of replication in the CNS can be achieved (381). Another theoretical concern is that if antiretroviral resistance develops in the CNS, it is possible that resistant HIV could then reseed the peripheral circulation. This mechanism could potentially lead to increased prevalence of neurocognitive disorders as well as systemic progression of HIV disease.

a) Zidovudine treatment of cognitive disorders or dysfunction

Two randomized, placebo-controlled clinical trials have provided evidence for the effectiveness of zidovudine in reducing neuropsychological impairment for patients with HIV infection. Sidtis and colleagues (382) treated a group of 40 homosexual men (early-stage dementia and mean CD4 cell counts of 500 cells/mm³) with 1000 or 2000 mg/day of zidovudine or placebo. They found that both treatment groups demonstrated significant combined mean neuropsychological z score improvements when compared to the placebo group. In another randomized, placebo-controlled clinical trial, Schmitt and colleagues (383) used zidovudine to treat 281 subjects who did not necessarily have clear-cut signs of neuropsychological impairment or cognitive disorder but were stratified into two groups, AIDS or early asymptomatic HIV infection, on the basis of CD4 cell count. Significant improvements were found in attention, memory, visual-motor, and simple motor function at 4 months in both groups, which led to premature discontinuation of the study.

An open study of zidovudine in injection drug users found improvements in neuropsychological functioning on a number of subtests (384). Four case control studies retrospectively examined whether zidovudine treatment was associated with outcomes such as reduced leukoencephalopathy (385, 386), lower CSF ₂-microglobulin levels (which are related to neuropsychological improvement) (387), or improved neuropsychological measures (388). Three of the four studies found that the zidovudine-treated patients had significantly positive findings compared to untreated patients.

While the above studies support the efficacy of zidovudine, they do not resolve the issue of optimal dosage nor zidovudine resistance (380). Doses of up to 2000 mg/day of zidovudine were used in some of the above studies, which are often associated with significant toxicity, including neutropenia and anemia. A typical zidovudine dose used now is 500–600 mg/day, and many patients cannot tolerate even these reduced doses (375). Patients can develop resistance to zidovudine, especially if used alone, and because zidovudine-resistant strains can be transmitted from one person to another, some HIV experts have considered the need for testing for zidovudine resistance before choosing which agents to use in treatment (389).

b) Other antiretroviral treatments of cognitive disorders or dysfunction

Treatment studies of antiretroviral agents other than zidovudine are fewer but illuminating. De Ronchi and colleagues (390) conducted a clinical trial (N=88) that compared zidovudine to didanosine to refusal of therapy in both symptomatic and asymptomatic HIV patients. They found that for subjects who were symptomatic for HIV disease, both zidovudine and didanosine treatment resulted in better performance on all neuropsychological subtests. In general, asymptomatic subjects did not show significant differences in neuropsychological performance compared to the untreated group. The findings suggest differences in effectiveness based on HIV clinical status.

A more recent placebo-controlled study involved the use of abacavir, a nucleoside analogue reverse transcriptase inhibitor with good CSF penetration, in 99 subjects with moderate to severe HIV-associated dementia (391). This study did not demonstrate significant differences in neuropsychological scores between abacavir and placebo groups, but a major limiting factor in this study design was that abacavir alone was added to an already failing antiretroviral treatment regimen.

Of currently available antiretroviral agents, the CSF penetration of stavudine is relatively high (0.4 ratio with serum), but a recent study indicated that stavudine nevertheless penetrates brain tissue poorly (392). The individual protease inhibitors have not been well studied for effects on neurocognitive impairment, and all but indinavir (0.16 ratio with serum) penetrate the CSF poorly (393). Regarding the nonnucleoside reverse transcriptase inhibitors, nevirapine has been shown to have a 0.45 CSF-to-serum ratio and therefore may be a good candidate to include in future controlled studies of HIV-associated cognitive disorders (381).

Ferrando and colleagues (394) studied the effects of various combination antiretroviral therapy regimens on measures of cognitive motor function. They found that tests of attention, concentration, psychomotor speed, learning, and memory were significantly better in those subjects taking antiretroviral therapy than in those who were untreated. Furthermore, they found that those without neuropsychological impairment had lower mean viral load levels, which suggests that antiretroviral treatment benefits neuropsychological functioning through lowering viral load. A similar conclusion was offered by Sacktor et al. (40).

c) Antiretroviral treatments to prevent HIV-associated dementia

A few studies have examined the use of antiretroviral medications to prevent HIV-associated dementia. Moore and colleagues (395) conducted a study of 863 homosexual men treated with zidovudine who were evaluated every 2 months for 2 years or until death. Development of HIV-associated dementia continued with zidovudine treatment but was significantly correlated with low CD4 cell count. Montforte and colleagues (396) used didanosine in a study of 1,047 subjects (median CD4 cell count of 47 cells/mm³ and mean prior zidovudine treatment of 19 months). Subjects were followed every 2 months to a clinical endpoint (AIDS-defining condition, severe adverse event, or death). They reported a very low incidence of HIV-associated dementia (11 subjects, or 1%), although no formal criteria for defining HIV-associated dementia were offered. One randomized clinical trial with 32 subjects that compared zidovudine treatment to placebo found no significant difference in rate of developing dementia by treatment assignment (397). HIV-associated dementia prevention requires further study, but it appears that treatment with antiretroviral medications may prevent the development of HIV-associated dementia and other cognitive motor disorders.

d) Interventions that affect inflammatory mediators

Although peptide T can be thought of as "antiretroviral," since it blocks the binding of envelope glycoprotein to CD4 receptors, it more likely exerts its effects through decrements in the deleterious effects of inflammatory mediators known as cytokines, such as tumor necrosis factor. A multisite, randomized, placebo-controlled clinical trial with intranasal peptide T was conducted on a sample of 215 patients with HIV-associated neurocognitive impairment, most of whom were taking a single antiretroviral agent or a two-drug combination regimen (398). The study found no statistically significant differences between treatment and placebo groups on global neuropsychological z scores. However, subgroup analyses showed that subjects with a CD4 count of 201–500 cells/mm³ and subjects with at least moderate neuropsychological impairment at baseline improved significantly with peptide T compared with those given placebo (p<0.05). In addition, peptide T was well tolerated with no clinically significant toxic effects.

Studies of inflammatory mediators are indicated because researchers have postulated that HIV-related neuronal injury involves the activation of voltage-dependent calcium channels and N-methyl-d-aspartate (NMDA) receptor-operated channels (399). Therefore, studies are underway using memantine (an NMDA antagonist), inhibitors of tumor necrosis factor-, thalidomide, pentoxifylline, and naloxone (an inhibitor of interferon-). Mapou and colleagues (400) conducted a clinical trial with interferon--n3 in 20 asymptomatic subjects with HIV infection and reported significant improvements in neuropsychological performance in selected subtests. While interferon has been associated with neuropsychological side effects, including depression, Skillman et al. (401) reported that such side effects are minimal.

Nimodipine is a voltage-dependent calcium channel antagonist that has been postulated to prevent HIV-related neuronal injury. Two controlled studies of nimodipine, one in combination with zidovudine (402) and one with nimodipine alone (403), found no significant differences in overall neuropsychological performance.

e) Treatments involving neurotransmitter manipulation

Stimulant drugs have been used as palliative agents to help manage symptoms of fatigue, decreased concentration, or memory deficits among patients with HIV-associated dementia or minor cognitive motor disorder. With regard to the specific effect of stimulants on neurocognitive impairment, there has been one randomized, placebo-controlled clinical trial of a small sample of eight opioid-dependent patients maintained on a regimen of methadone (404). This 7-day crossover study with sustained-release methylphenidate (20–40 mg/day) resulted in nonsignificant neuropsychological z score improvements from baseline with methylphenidate but not placebo. Fernandez and colleagues (321) presented case reports of treating patients with HIV-related disease with methylphenidate (30–90 mg/day) or dextroamphetamine (30–60 mg/day). All patients improved significantly; however, there was no placebo control group and the sample size was small. Similar data from a small trial are offered by Holmes et al. (405).

Studies of dopaminergic agonists are supported by preliminary data in the form of case studies of carbidopa and l-dopa (303). The Dana Consortium conducted a randomized, placebo-controlled clinical trial of 36 subjects with HIV-associated cognitive motor impairment who were given l-deprenyl, a putative antiapoptotic agent, and found that deprenyl-treated subjects showed an improvement in neuropsychological test performance, specifically in verbal memory (406). Investigation of the dopaminergic agent pramipexole is currently under way.

Other neurotransmitters such as serotonin and 5-hydroxyindoleacetic acid have been found to be decreased in CSF as HIV disease progresses (407). It is possible that increasing CNS serotonin levels through administration of agents such as SSRIs may be useful in the treatment of cognitive motor impairment.

f) Nutritional therapies and HIV-associated cognitive disorders

Nutritional therapies need to be considered as potential interventions for the cognitive motor symptoms associated with HIV infection. A case report of a subject with apparent late-stage HIV-associated dementia and a low cobalamin (vitamin B₁₂) level due to decreased intrinsic factor secretion responded over 2 months to B₁₂ replacement therapy with complete symptom resolution (408). Oxidative free radical scavengers such as vitamin E, the experimental antioxidant OPC-14117, and the trace mineral and antioxidant selenium may prove to be therapeutically useful.

2. Delirium

Antipsychotic medications have been the treatment of choice in treating delirious states that develop in patients with HIV infection, with most evidence arising from case reports or uncontrolled trials (409, 410). For patients with HIV infection, a prominent issue is the observed increased sensitivity to side effects (299, 300). This sensitivity applies to extrapyramidal side effects with higher-potency agents and also to cognitive side effects with lower-potency antipsychotics.

To date, only one randomized clinical trial has studied the treatment of delirium associated with HIV infection. Breitbart and colleagues (46) approached 244 hospitalized AIDS patients and monitored them for the development of delirium. Among the 30 subjects who went on to develop delirium, 310 episodes of delirium occurred that met both DSM-III-R and Delirium Rating Scale criteria (411). The 30 subjects were randomly assigned to receive haloperidol (N=11), chlorpromazine (N=13), or lorazepam (N=6). Because the six subjects taking lorazepam developed signs of toxicity (oversedation, disinhibition, ataxia, and increased confusion), this treatment was terminated prematurely. Both haloperidol and chlorpromazine treatment resulted in significantly reduced scores on the Delirium Rating Scale in the first 24 hours. The researchers did not report significant problems with extrapyramidal side effects in study patients, but mean daily doses were low.

Scattered case reports or series suggest that other antipsychotic agents may be useful for treating HIV-associated delirium. For instance, molindone (312), risperidone (315), and droperidol (412, 413) have been suggested as alternatives to the standard antipsychotic medication because of differences in side effect profile, onset of action, or potency.

3. Mood disorders

Disturbances of mood occur frequently in the context of HIV infection. Treatment studies are quite varied with regard to variables such as severity of mood disorder, methodology to categorize or diagnose mood disorders, outcome measures, and stage of HIV illness.

a) Acute treatment of depressive disorders—somatic treatments

Many studies, including randomized controlled trials, open-label trials, case series, and case reports, support the efficacy and safety of many antidepressants in the acute treatment of depressive disorders, including major depression, in a wide range of HIV-infected populations at all disease stages.

Wagner and colleagues (286) reported on a series of 6-week studies that used random assignment and placebo control to examine the efficacy of standard and alternative antidepressants for major depression in HIV-infected patients. The agents used at standard doses were fluoxetine, sertraline, imipramine, dextroamphetamine, and testosterone. Response rates to the different agents ranged from 70% to 93% and were significantly higher than placebo.

Similar findings were reported in a randomized, placebo-controlled clinical trial conducted by Elliot and colleagues (287) in 75 HIV patients (45% with AIDS) who had major depression. Both imipramine- and paroxetine-treated groups improved significantly compared to placebo at weeks 6, 8, and 12. The dropout rate due to side effects was higher for those receiving imipramine (48%) than it was for placebo (24%) or paroxetine (20%).

One randomized, placebo-controlled clinical trial (N=120) found that fluoxetine was an efficacious antidepressant agent among depressed HIV patients; the response rate for fluoxetine was 74% versus 47% for placebo (289). The same study reported significant differences between ethnic groups in terms of antidepressant efficacy; the response rates were 84% for white, 67% for Latino, and 50% for black subjects (414). Despite the effectiveness of fluoxetine in treating depression, because both placebo response and attrition were high, the researchers suggested that the addition of another medication in patients with serious medical illness who require multiple concomitant medications may be a significant barrier to the treatment of depression.

In a double-blind, placebo-controlled trial that compared desipramine to fluoxetine in 14 women with advanced HIV disease, both agents were linked to symptom improvement; however, for most of the women significant depressive symptoms remained after the 6-week trial (294). These findings are consistent with those of an open-label fluoxetine trial that compared HIV-positive and HIV-negative depressed men, which found that improvement in the HIV-positive cohort was significantly delayed beyond the 8-week trial compared to the HIV-negative patients (415).

The randomized trials are supported by uncontrolled studies of depression in HIV populations and confirm the effectiveness of fluvoxamine (295) and fluoxetine (293, 416, 417) in HIV-infected asymptomatic patients and of paroxetine, sertraline, and fluoxetine in symptomatic HIV patients (285, 418). In an open trial in which HIV-positive depressed women were given either fluoxetine or sertraline, both drugs were found to be effective treatments (419). Another open-label study found nefazodone effective in the treatment of depressed HIV-positive patients (288). One open-label trial of 32 patients with advanced HIV disease found venlafaxine effective among depressed patients who also had minor cognitive motor disorder (420).

The use of psychostimulants for treatment of major depression in HIV-infected outpatients has been supported in smaller studies, including five case reports, one open-label trial, and one placebo-drug-placebo clinical trial (321, 322, 421, 422). These studies suggest that methylphenidate, up to 35 mg/day, or dextroamphetamine, up to 10 mg/day, can improve major depression.

An open trial of testosterone replacement for hypogonadal men found that of 34 study participants with major depressive disorder, 79% had significant improvement in mood (126). There is a case report of a patient with HIV infection and comorbid major depressive disorder whose depression resolved with treatment with the antiretroviral agent zidovudine (423). For patients with severe major depression who are delusional or have failed medication treatment, ECT has been found to be safe and effective (424).

b) Acute treatment of depressive disorders—psychotherapy

Several studies of psychotherapeutic treatment of depression in HIV patients have been conducted. In a randomized, controlled trial, Targ and colleagues (425) compared structured group therapy plus fluoxetine to structured group therapy with placebo in 20 asymptomatic homosexual men with major depression or adjustment disorder and Hamilton Depression Rating Scale scores 16. Both groups showed significant improvements in depression after 6 weeks of treatment, but there were no differences between treatment groups (425). In contrast, another randomized, placebo-controlled clinical trial that compared the efficacy and safety of fluoxetine plus group psychotherapy versus group psychotherapy alone in 47 depressed HIV-positive men for 7 weeks found that fluoxetine in addition to group therapy was more efficacious than group psychotherapy alone. Differences were particularly apparent for patients whose initial depressive episodes were rated as severe (Hamilton depression score 24) (290).

Markowitz and colleagues (426) conducted a randomized, placebo-controlled clinical trial that compared 16-week interventions with interpersonal psychotherapy, cognitive behavior therapy, supportive psychotherapy, and supportive psychotherapy plus imipramine in 101 HIV patients with depressive symptoms (Hamilton depression score 15). Patients randomly assigned to interpersonal psychotherapy and supportive psychotherapy plus imipramine had significantly greater symptom improvement than those receiving supportive psychotherapy alone or cognitive behavior therapy.

In an earlier study (427), the Markowitz group demonstrated the benefits of interpersonal psychotherapy in HIV-positive, asymptomatic outpatients in an open study of men and women with DSM-III-R major depression or dysthymia. Twenty (87%) out of 23 patients achieved a full remission after 6 weeks by clinical assessment; mean Hamilton depression scores at baseline and follow-up were 25 and 6.8, respectively. Levine et al. (428) reported that four outpatients with major depression remained euthymic for 9 months with structured group psychotherapy after acute remission was achieved with fluoxetine, 20–40 mg/day.

Other studies such as two retrospective chart reviews (284, 429) and a prospective effectiveness study (430) support the findings from controlled and uncontrolled treatment studies of major depression that treatment with a number of different antidepressant agents is effective and generally tolerated in HIV patients.

c) Treatment of other types of depression

Other studies have investigated the efficacy of treatment for other types of depression in HIV patients. These studies generally use screening instruments (e.g., the Center for Epidemiologic Studies Depression Scale [CES-D Scale]) rather than structured interviews or clinical evaluations to diagnose specific depressive disorders. All involved ambulatory patients who had no significant clinical manifestations of HIV disease. In a study of patients with depression (CES-D Scale score >16) but with unknown psychiatric diagnoses, investigators compared cognitive behavior therapy and brief supportive group therapy to a waiting list control condition. Outpatients (N=68) were randomly assigned to an 8-week intervention of cognitive behavior therapy, support group, or a control group. While both types of psychotherapy were superior to the control condition, neither led to significant reductions on the CES-D Scale into the nondepressed range (431).

Eller (432) conducted a randomized, placebo-controlled clinical trial of alternative therapies for depression and assigned 81 male and female outpatients with HIV infection to guided imagery, relaxation therapy, or usual care. The sample had a mean CES-D Scale score of 19. There were no significant declines in CES-D Scale scores during treatment in any of the three groups.

d) Treatment of mania/bipolar disorder

One chart review (433) and one case report (434) have addressed the treatment of manic syndromes in HIV-infected patients. The chart review study identified seven male outpatients with mania who did not improve with lithium or neuroleptic treatment due to toxicity. Three patients had been treated with valproic acid (750–1750 mg/day), two with clonazepam (2 mg/day), one with phenytoin, and one with carbamazepine. All seven patients experienced remission of their manic syndrome with the agent used (433). The case report described the effectiveness of valproic acid in the treatment of acute mania in a patient with AIDS complicated by dementia (434). There are no published studies of maintenance treatments for mania or bipolar disorders in HIV-infected populations.

4. Substance use disorders

Clinical consensus has established the necessity of treating substance use disorders in order to produce optimal psychiatric and medical outcomes for patients with HIV infection. Despite this knowledge, remarkably few studies have investigated psychiatric treatment of substance use disorders in HIV-infected patients. This section will review the literature of the treatment of substance use disorders and associated syndromes.

a) Entry into a treatment program

Entry into a treatment program is probably one of the more critical steps for the assessment and treatment of substance use disorders. Guydish and colleagues (435) described a consultation-education-triage model program that referred drug users with HIV infection to follow-up care for substance use disorders. Of the 86 patients referred over an 8-month period, 81% were referred for further care to a substance use program, with 58% actually contacting the referral resource. Eighty percent had AIDS or symptomatic HIV disease; 37% of patients were dependent on alcohol, 36% were dependent on amphetamines, and 26% were dependent on heroin, marijuana, or cocaine.

b) Methadone maintenance

The Program for AIDS Counseling and Education in San Francisco is considered a model program for inner-city, poor, opiate-dependent, mostly minority men (436, 437). The program provides medical assessment and care, psychiatric treatment, methadone treatment, counseling, detoxification, and tolerance management for opiate addicts. Outcomes for patients enrolled in the Program for AIDS Counseling and Education have been reported in two conference papers. In one report, 29 opioid-dependent HIV patients with a mean of 28 days of drug use in the past month reduced their use to a mean of 5.9 days of use per month after 3 months in the program (438). The other report focused on clinical events related to drug use before and after enrollment for 62 patients in the program (439). Clinical events included medical complications and lapses in compliance with treatment. At baseline, patients experienced 0.72 clinically significant medical complications per 100 patient years, whereas the rate decreased significantly to 0.16 during enrollment in the Program for AIDS Counseling and Education. There was also a reduction in preventable HIV clinical events from 0.21 per 100 patient years to 0.02 per 100 patient years after starting the program. The reports provide some evidence that both the medical care as well as substance use behaviors of HIV-infected opioid addicts can benefit from a multidisciplinary, structured intervention.

c) Pharmacologic treatment of substance use disorders

One study of HIV-infected patients investigated the effect of antidepressant medications on symptoms of drug dependency, such as drug craving and comorbid depression. Batki (440) conducted a controlled trial that compared fluoxetine to placebo in 37 patients with cocaine and opiate dependency who were on a regimen of methadone maintenance. Of this sample, 41% also had major depression. Participants treated with fluoxetine (20–40 mg/day) had lower median measures of cocaine metabolites in urine and fewer mean days of cocaine use per week (1.6) compared to those given placebo (2.6); the fluoxetine-treated patients also reported reduced cocaine craving. Those with major depression showed significant improvement after the 12-week treatment.

d) Overall outcomes of psychiatric treatment

Lyketsos and colleagues (430) conducted a 14-month prospective effectiveness study of clinical outcomes in HIV patients who were referred to psychiatric treatment. Within the cohort, 110 patients had current (N=66) or recent (N=44) substance use disorders; there were high rates of polysubstance dependence, with only a small proportion of patients who abused alcohol. The psychiatric treatment took place within a primary care clinic and included treatment for substance abuse and concurrent psychiatric disorders. Psychiatric intervention led to a decrease in the use of substances and better clinical condition.

5. Anxiety disorders

Anxiety disorders are prevalent in the population of persons who have or are at risk for HIV infection, but studies of specific treatments for anxiety disorders in HIV patients are lacking. The literature provides only one retrospective chart review and two case reports that address the management of anxiety disorders in HIV-positive patients. There are no treatment studies of PTSD among HIV-infected patients, although PTSD rates are high in some HIV-affected groups (441).

A chart review study found that buspirone was an effective anxiolytic in opioid-dependent, HIV-positive patients who were on a regimen of methadone maintenance and who also had a comorbid anxiety disorder. The study additionally showed a modest reduction of urine tests positive for substances of abuse, from 42% to 30% (328).

McDaniel and Johnson (442) found fluoxetine to be effective in the treatment of obsessive-compulsive disorder (OCD) in two HIV-positive patients. Both patients tolerated the fluoxetine with minimal side effects and responded with resolution of OCD symptoms within 6–8 weeks.

A male patient with advanced AIDS complicated by comorbid trichotillomania and major depression was reported to have responded well to sertraline at a maximum dose of 150 mg/day. The patient reported experiencing a 70%–80% decrease in hair pulling and a modest improvement in depressed mood, with minimal medication side effects (443).

6. Psychotic disorders

Nine literature citations, two of them clinical trials and the remainder case reports, present evidence concerning the efficacy of psychotropic medication for the management of psychosis in the presence of HIV infection. The nine reports contain fewer than 100 subjects, approximately 90% of them male, and most with symptomatic HIV infection, usually AIDS. Where stated, the majority did not have a history of psychosis before HIV infection, and, where stated, the majority had cognitive impairment.

Antipsychotic medications were effective in treating psychosis whether or not cognitive dysfunction or delirium was present. One large case history series found that patients with manic psychosis showed more improvement than patients with schizophreniform psychosis (315). The two clinical trials (31, 444) found that positive symptoms responded better than negative symptoms. Across studies the subjects, who largely had advanced HIV disease, required lower doses of antipsychotic medication, similar to the pattern seen in the elderly. Lower doses were also useful due to increased sensitivity to side effects. If a patient did not respond to or tolerate one antipsychotic, it was useful to try another from a different class (300, 301, 310, 312).

In case reports, the use of adjunctive medications, including antianxiety drugs, antidepressants, and mood stabilizers, was a common practice and often useful in individual cases, but there have been no controlled studies on the efficacy of these combinations. Two case reports, each of a single male subject with AIDS, found that catatonia responded rapidly to lorazepam (445, 446). No side effects were noted.

Maintenance medication was not always necessary to sustain remission of psychotic symptoms (42, 312). In some cases it was specifically noted that symptoms of psychosis diminished as severe late-stage medical illness progressed (42). While it has been hypothesized that there is a window of vulnerability to psychosis or mania occurring early in the course of HIV-associated dementia (447), no data support this theory, and most patients treated for psychosis have continued their maintainance medication regimens.

There are few data on the treatment of psychosis in early-stage HIV infection or among women nor is there information concerning potential modifications in the treatment of preexisting psychotic illness, which might be necessary in the presence of advancing HIV infection.

7. Adjustment disorders

Two open-label clinical trials have assessed the effectiveness of psychotropic medications in the management of adjustment disorders in HIV-positive patients. As part of a larger study that examined the effectiveness of paroxetine in the treatment of depression in the context of HIV infection, Grassi and colleagues (418) included five patients with a diagnosis of adjustment disorder with depressed mood. All five patients received paroxetine (20 mg/day) and showed significant recovery as measured by the Hamilton depression scale at the 6-week endpoint.

Five male patients with neurocognitive impairment as well as a diagnosis of adjustment disorder with depressed mood were treated as part of a larger study of the effectiveness of psychostimulants in patients with HIV-associated dementia. Four of the five patients who received methylphenidate had moderate to marked improvement in symptoms; the one patient with minimal response was switched to dextroamphetamine and subsequently had a moderate response (405).

8. Sleep disorders

Numerous reports have documented disrupted sleep and altered sleep architecture in HIV patients (29, 448, 449), including one study which linked fatigue and sleep disturbances to morbidity and disability in homosexual men (450). While the etiologies remain unclear, some investigators have documented growth hormone dysregulation associated with sleep pathology (451), and one study found obstructive sleep apnea due to adenotonsillar hypertrophy to be a primary cause of sleep disruption in a cohort of HIV patients with excessive daytime sleepiness (452).

The treatment of sleep disorders in HIV-positive persons has been examined in only one open-label clinical trial of flurazepam in 12 patients (453). All patients were given a one-time dose of flurazepam, 30 mg, to examine changes in the patients' electroencephalographic recordings during sleep. Flurazepam mainly affected non-REM parameters, such as reduction of times awake during the night and increases in stage two and effective sleep time. As this study did not address treatment with flurazepam beyond a one-time dose, actual treatment recommendations cannot be made.

9. Disorders of infancy, childhood, and adolescence

Since HIV infection in childhood is associated with developmental delays or loss of acquired skills, it is hoped that early treatment will promote normal development. Currently, it appears that antiretroviral drug therapy reduces morbidity in children, especially in those with HIV-associated neurocognitive impairment. Several studies have reported a significant increase in mental ability, as measured by the WISC-R, the McCarthy Scales of Children's Abilities, or the Bayley Scales of Infant Development, in over 450 children during 12- and 36-month trials with zidovudine (58, 246, 454, 455). None of the studies included a control group.

The safety of zidovudine for children is similar to that for adults, although up to 40% of children experience hematological side effects that require dose alterations (246). The one published study of 54 children that did not find a significant improvement in cognitive functioning with zidovudine treatment did find a significant decrease in CD4 levels over the treatment period. No data were provided on adherence to treatment or zidovudine resistance (456). Other antiretroviral agents, such as didanosine, have been studied, with a multisite project showing that combination zidovudine and didanosine therapy was superior to either used as monotherapy for most of the CNS outcomes evaluated (457). Although combination antiretroviral therapy with three agents is standard practice with adults, there is only one case report of its use with a child (458). Caregivers noted the positive impact of this regimen on an 8-year-old boy's neurocognitive development.

The effectiveness of the treatment of other mental and behavioral disorders associated with HIV infection in children and adolescents is largely unstudied. Case reports of single patients suggest that attention deficit disorders and depressive disorders in those with HIV infection may be treated by using regimens that are standard for those without HIV infection (459). As in the treatment of adults, children with HIV infection may be especially sensitive to some treatment side effects, such as insomnia or appetite suppression associated with psychostimulant use.

Pain frequently accompanies pediatric HIV and can involve multiple organ systems (460). Clinicians should be aware that the inadequate verbal skills of younger children or the fact that some children are from non-English-speaking families can lead to the underrecognition and undertreatment of pain (461).

10. HIV-associated syndromes with psychiatric implications

A number of HIV-associated clinical syndromes have psychiatric implications, such as wasting syndrome, fatigue, pain, and sexual dysfunction. Because there are overlapping symptoms with these conditions and psychiatric disorders, it is useful for psychiatrists to be aware of treatment studies for such syndromes so that overall clinical outcomes can be improved.

a) Wasting syndrome

Several studies have focused on the use of testosterone for treatment of wasting syndrome in both patients with low testosterone levels and patients with "clinical hypogonadism" (normal hormone levels). Grinspoon and colleagues (462) conducted a double-blind, placebo-controlled study of hypogonadal HIV-positive men with wasting syndrome that demonstrated the safety and efficacy of testosterone replacement therapy (300 mg i.m. every 3 weeks). The treatment resulted in a significant increase of lean body mass, overall quality of life, and self-perception of appearance in treated subjects. Similar findings have been reported by Rabkin and colleagues (126, 343, 463), who conducted several randomized, placebo-controlled trials as well as open trials of testosterone replacement in men with clinical hypogonadism that led to significant positive effects on mood and weight (lean body mass). Another study that used oxymetholone, a testosterone derivative, demonstrated weight gain in cachectic men with AIDS (464). A controlled study of hypogonadal women with HIV infection that used transdermal testosterone replacement also reported improvement in lean body mass and quality of life (465).

Wagner and colleagues (466) reported an additional improvement of measures on the Brief Symptom Inventory and nutritional status for patients who exercised in addition to receiving testosterone replacement. In a double-blind, randomized, placebo-controlled trial, researchers found that resistance exercise in addition to testosterone and the anabolic steroid oxandrolone substantially increased lean body mass (467). Two open-label studies have reported that oxandrolone, 20 mg/day, in HIV patients with wasting syndrome led to significant weight gain, in particular, increases in body cell mass (334, 335).

Several clinical trials have used the recombinant growth hormone somatotropin for treatment of HIV-related wasting (336, 340). A large multicenter, double-blind, placebo-controlled study that used growth hormone in patients with AIDS wasting reported significant improvement in overall quality of life and weight gain, with no negative effect on virological or immunological markers (339). Treatment was well tolerated.

Since tumor necrosis factor- has been postulated to have a role in the pathogenesis of wasting syndrome, the effect of thalidomide, a selective inhibitor of tumor necrosis factor-, has been studied in patients with AIDS wasting syndrome. Two reports (a mail survey of thalidomide users and a small randomized, placebo-controlled clinical trial) have supported the positive effects of thalidomide on weight gain, appetite, and quality of life (341, 342).

b) Fatigue

Breitbart and colleagues (468) conducted a randomized, placebo-controlled clinical trial that used two psychostimulant medications, methylphenidate and pemoline, in treating fatigue among HIV patients. They found that both treatments were superior to placebo, with a stronger effect in the methylphenidate group. In addition, they reported that improvement in fatigue was associated with improvement in depressive symptoms and psychological distress. In an open trial of depressed patients with CD4 counts below 200, dextroamphetamine administration led to improvement of fatigue in 95% of patients along with reduction in depression scores (322). Testosterone replacement with dehydroepiandrosterone (DHEA) in HIV-infected hypogonadal men has also been found to improve symptoms of fatigue (469).

c) Pain

Studies of the frequency and adequacy of treatment for pain among patients with AIDS find wide variations in the effectiveness of interventions to manage pain. A study by Breitbart et al. (123) found that 85% of HIV patients in their ambulatory sample treated at a cancer center received inadequate pain therapy. Yet a different study reported that only 20% of their sample of AIDS patients treated in the hospital, home care, or nursing facility did not receive effective pain control (470).

While agents such as tricyclic antidepressants and anticonvulsant medications are often used to manage chronic pain from peripheral neuropathy, published treatment studies for HIV-associated peripheral neuropathy have not supported their use. Two randomized, controlled trials have focused on HIV-associated peripheral neuropathy; both trials used the agent mexiletine, an antiarrhythmic drug with local anesthetic properties that had been used in the treatment of diabetic neuropathy (317, 471). Neither study found pain relief different from that of placebo. Furthermore, Kieburtz et al. (317) did not find that treatment with amitriptyline provided significant pain relief compared to placebo. Shlay and colleagues (472) conducted a randomized, controlled trial that compared acupuncture, amitriptyline, and placebo for pain due to HIV-related peripheral neuropathy and found that neither acupuncture nor amitriptyline was more effective than placebo.

d) Sexual dysfunction

Sexual dysfunction is experienced commonly by HIV patients, both men and women, particularly with HIV disease progression. While no treatment trials are currently published, a cautionary note regarding sildenafil is clinically important. Sildenafil should be used judiciously in patients with erectile dysfunction who are also taking protease inhibitors, since the primary route of sildenafil metabolism is via the cytochrome P450 isoenzyme 3A4. One death has been reported in a 47-year-old male taking ritonavir and saquinavir who was administered 25 mg of sildenafil (473). The manufacturer of sildenafil has recommended that patients receiving ritonavir should not take more than a single sildenafil dose of 25 mg in a 48-hour period.