Antiretroviral agents

Antiretroviral drugs are the mainstay of HIV treatment. Table 1 lists currently available drugs, grouped according to their mechanism of action. Many other agents are at a developmental stage.

Table 1 Common psychiatric side-effects of licensed antiretroviral agents

Agent	Psychiatric side-effects
Protease inhibitors
Amprenavir	Mood changes
Indinavir	Mood changes
Lopinavir+ritonavir	Mood changes, agitation, anxiety
Nelfinavir	None reported
Ritonavir	Anxiety
Saquinavir	Depression, anxiety, sleep disturbances
Atazanavir	Depression, insomnia
Fosamprenavir	None reported
Nucleoside reverse transcriptase inhibitors (NRTIs)
Abacavir	None reported
Didanosine	Lethargy, nervousness, anxiety, confusion, sleep disturbances, mood disorders, psychosis
Lamivudine	Insomnia, mood disorders
Stavudine	Sleep and mood disorders, delirium
Zalcitabine	Somnolence, impaired concentration, mood disorders, delirium
Zidovudine	Sleep disturbance, vivid dreams, agitation, mania and depression, psychotic symptoms, delirium
Coformulated NRTIs
Zidovudine+lamivudine	See above
Zidovudine+lamivudine+ abacavir	See above
Nucleotide reverse transcriptase inhibitors (NtRTI)
Tenofovir	None reported
Non-nucleoside reverse transcriptase inhibitors (NNRTIs)
Delavirdine	None reported
Efavirenz	Agitation, depersonalisation, hallucinations, disturbed dreams, mood disorders, depression, suicidality, antisocial behaviour, psychosis, catatonia, delirium
Nevirapine	None reported
Fusion or entry inhibitors
Enfuvirtide	Anxiety, depression

The treatment of HIV infection is based on a combination of at least three antiretroviral drugs and is known as highly active antiretroviral therapy (HAART). Hence there is a large potential for drug interactions and subsequent toxicity (Reference TolosaTreisman & Kaplan, 2002).

The degree of CNS drug penetration may be of particular importance. Lack of penetration could leave the brain as a reservoir for the virus, resulting in CNS infection and lack of final viral clearance (Reference Gonzalez and EverallGonzalez & Everall, 1998). Also, the degree of CNS penetration may relate to the antiretrovirals’ effectiveness in treating neuropsychiatric manifestations of HIV (Reference Gonzalez and EverallGonzalez & Everall, 1998). The mechanisms underlying the development of neuropsychiatric symptoms in HIV are complex. Several factors interact, including the direct effect of the virus, immune mediators and medications (Reference TolosaTreisman & Kaplan, 2002). There is increasing evidence that neuronal damage induced by HIV may be mediated by immune activation and viral infection of brain macrophages and microglia (Reference StockleySwindells et al, 1999). Although effective against CNS infection, antiretrovirals are themselves increasingly recognised as a source of neuropsychiatric disorders. Some side-effects were identified during the trial phases and others have come to light in published case reports. How these drugs exert their deleterious effects on the brain remains poorly understood.

Drugs from the five groups of antiretrovirals have a variable capacity to induce psychiatric disorders (Table 1).

Protease inhibitors have limited CNS penetration and therefore less-pronounced CNS (neurological and psychiatric) side-effects (Reference Harry, Matthews and SalvaryHarry et al, 2000; Reference TolosaTreisman & Kaplan, 2002). Within this group, ritonavir and saquinavir are more likely to produce neurological side-effects (Reference TolosaTreisman & Kaplan, 2002), the others inducing mostly mood disturbances.

Zidovudine was the first drug of the nucleoside reverse transcriptase inhibitor (NRTI) group introduced for the treatment of HIV. Psychiatric disturbances are usually dose related (Reference TolosaTreisman et al, 2002). Didanosine and lamivudine can cause psychiatric complications. Abacavir does not cause prominent psychiatric side-effects, although when used in combination, it may induce psychosis and catatonia (Reference Foster, Olajide and EverallFoster et al, 2003). Tenofovir, a nucleotide reverse transcriptase inhibitor (NtRTI), has not been associated with psychiatric side-effects.

Within the non-nucleoside reverse transcriptase inhibitor (NNRTI) group, efavirenz induces a variety of CNS side-effects in up to 50% of patients (Reference Colebunders and VerdonckColebunders & Verdonck, 1999), which may be severe and sudden in onset (Reference Lang, Halleguen and PicardLang et al, 2001; Reference Peyriere, Mauboussin and RouanetPeyriere et al, 2001; Reference Sano, Stern and CoteSanz de la Garza et al, 2001; Reference Riley and LangSabato et al, 2002). Symptoms include twilight states, personality changes, with increased hostility, and cognitive disturbances (Reference Lang, Halleguen and PicardLang et al, 2001; Reference TolosaTreisman & Kaplan, 2002). Efavirenz treatment may be particularly associated with major depression and severe suicidal ideation (Reference PuzantianPuzantian, 2002). Side-effects of the drug tend to present within the first 4 weeks of treatment and may subside spontaneously despite continuation of treatment (Reference Colebunders and VerdonckColebunders & Verdonck, 1999). A possible association with high drug plasma levels remains unconfirmed. Side-effects have been described in patients without psychiatric antecedents, but those with a previous psychiatric history are more vulnerable and should be monitored closely (Reference Peyriere, Mauboussin and RouanetPeyriere et al, 2001). Side-effects are reduced with bedtime dosing (Reference TolosaTreisman & Kaplan, 2002), although patients often experience vivid dreams. Side-effects are less common with delavirdine and nevirapine (Reference TolosaTreisman & Kaplan, 2002).

Treatment of psychiatric complications

The diagnosis of antiretroviral-induced psychiatric side-effects requires the exclusion of other causes, as well as consideration of the additive effects of medications. Antiretroviral drug-level monitoring may be helpful.

The treatment of antiretroviral-induced psychiatric disturbance varies according to its severity. Mild symptoms should be monitored, and psychotropics may be added if required. Severe cases may require switching or discontinuing the HAART regimen, which could be reinstated (often with different medication) once the patient improves.

Current guidelines recommend aggressive treatment of HAART-induced depression, especially if there is a previous psychiatric history, as well as considering switching the HAART regimen (Reference TolosaTreisman & Kaplan, 2002). In severe cases, efavirenz has to be discontinued (Reference Peyriere, Mauboussin and RouanetPeyriere et al, 2001), which can result in rapid improvement (Reference Sano, Stern and CoteSanz de la Garza et al, 2001). However, efavirenz has a long half-life and therefore side-effects may persist for several weeks.

In patients with depression about to start HAART, guidelines suggest first an active treatment of the depression, with HAART initiation depending on the degree of immunosuppression, HIV load and CD4 T-cell counts. Effective treatment of psychiatric complications may improve HAART adherence (Reference TolosaTreisman & Kaplan, 2002).

Psychotropic medication may be required to treat psychiatric symptoms resulting either from the HIV CNS infection or from the HAART treatment (Reference Sano, Stern and CoteSanz de la Garza et al, 2001). However, patients with HIV have increased sensitivity to psychotropics and care is required because of the additive side-effects of the different medications (Reference Everall, Drummond and CatalanEverall et al, 2004). Psychotropics and antiretrovirals have complex metabolic interactions at the level of the cytochrome P450 family of hepatic enzymes (Reference Trosch, Friedman and LannonTseng & Fosy, 1999). Some of these interactions are bidirectional, leading to changed levels of both groups of drugs. Within the protease inhibitors, ritonavir is a moderate inhibitor of CYP2D6 isoenzyme, which metabolises tricyclic antidepressants (TCAs), newer antidepressants (paroxetine, venlafaxine and fluvoxamine) and many antipsychotics, including risperidone. Concomitant administration of ritonavir with these psychotropics may lead to toxic blood levels of both. Protease inhibitors and NNRTIs can either increase or decrease levels of a wide variety of psychotropics by their action on the CYP3A4 isoenzyme. This enzyme is a more common site of protease inhibitor effects.

Psychiatrists are advised to consult pharmacy services for advice on possible interactions when prescribing psychotropics to an HIV-infected patient. Low doses of psychotropics should be used initially (between a quarter and a half of the usual starting dose) and increased gradually (Reference Everall, Drummond and CatalanEverall et al, 2004).

Several antidepressants are useful in these patients (Box 2). Selective serotonin reuptake inhibitors (SSRIs) are generally well tolerated, although they may induce gastrointestinal disturbances (Reference Everall, Drummond and CatalanEverall et al, 2004). Some SSRIs (e.g. fluoxetine, fluvoxamine) interact with specific antiretrovirals, so possible interactions should be checked (for further information see Reference Starkstein, Preziosi and BerthierStockley, 2002; British Medical Association & Royal Pharmaceutical Society of Great Britain, 2004; www.emims.net; www.epocrates.com).

Of major importance is the interaction between antiretrovirals and St John’s wort, a herbal remedy commonly used to treat anxiety and depression. This compound significantly reduces levels of protease inhibitor and NNRTIs and may lead to treatment failure (Reference JamesJames, 2000; Reference Piscitelli, Burstein and ChaittPiscitelli et al, 2000). Protease inhibitors and NNRTIs interact with diazepam, midazolam, alprazolam and zolpidem, causing marked benzodiazepine effects.

Patients with HIV have increased sensitivity to neuroleptics, with frequent emergence of extrapyramidal side-effects (Reference Meyer, Marsh and SimpsonMeyer et al, 1998). Although similar extrapyramidal reactions have been occasionally described with atypical antipsychotics (Reference Meyer, Marsh and SimpsonMeyer et al, 1998), these are usually easier to use (Reference TolosaTreisman & Kaplan, 2002). Clozapine has shown good results in HIV-infected patients with associated psychosis, although there are concerns regarding higher risk of bone marrow toxicity (Reference Lera and ZirulnikLera & Zirulnik, 1999). Clozapine is contraindicated in patients receiving protease inhibitors (Reference Everall, Drummond and CatalanEverall et al, 2004).

Clinically significant interactions have been observed between antiretrovirals in general and many classes of recreational drugs. Protease inhibitors inhibit metabolism of many of these drugs, particularly ‘rave’ drugs such as methylene dioxymethamphetamine (MDMA), amphetamine and ketamine, resulting in toxic overdoses (Reference Antoniou and TsengAntoniou & Tseng, 2002). Protease inhibitors and NNRTIs induce methadone metabolism, leading to methadone withdrawal symptoms (Reference Antoniou and TsengAntoniou & Tseng, 2002). It is still unclear whether this metabolic interaction is bidirectional (Reference TolosaTreisman & Kaplan, 2002).

Treatment of psychiatric complications

Isolated visual hallucinations and psychotic symptoms usually respond to general support and reassurance and adjustment of total daily dose of antiparkinsonian medications (Box 3) (Reference Wolters, Jansen and Tuynman-QuaWolters, 1999; Reference Friedman and FactorFriedman & Factor, 2000; Reference Catalan-Alonso and Del ValCatalan-Alonso & Del Val, 2001). In rare instances psychotic symptoms occur only in the ‘off’ stage and therefore may require an increased dose of dopaminomimetics (Reference Nissenbaum, Quinn and BrownNissenbaum et al, 1987).

Antipsychotic drugs are required if psychotic symptoms do not respond to these measures. The older antipsychotics have D₂-receptor antagonism and therefore induce deterioration of mobility in Parkinson’s disease. Atypical antipsychotics should be used cautiously in patients with the disease, starting at low doses to minimise drowsiness, orthostatic hypotension and delirium and to allow monitoring of mobility. Clozapine is the only drug with confirmed benefit for psychosis in Parkinson’s disease (Reference Melamed, Friedberg and ZoldanMelamed et al, 1999; Reference Friedman and FactorFriedman & Factor, 2000). In daily doses of up to 50 mg it effectively reduces drug-induced psychotic symptoms without inducing a marked motor deterioration (Parkinson Study Group, 1999; Reference Wolters, Jansen and Tuynman-QuaWolters, 1999). Beneficial effects, even in patients with dementia, have been shown to be sustained for up to 5 years (Reference Klein, Gordon and PollakKlein et al, 2003). Additionally, clozapine has been shown to reduce tremor, hypersexuality and sleep disturbances in Parkinson’s disease (Reference Trimble, Rusch and BettsTrosch et al, 1998).

Results with other atypical antipsychotics are less impressive. Even low doses of risperidone (up to 3mg/day) and olanzapine (up to 7.5mg/day) may be detrimental to mobility (Reference Wolters, Kuiper, Zwaan, Wolters and ScheltensWolters et al, 1996; Reference Aarsland, Larsen and LimAarsland et al, 1999b ; Reference Wolters, Jansen and Tuynman-QuaWolters, 1999; Reference Goetz, Blasucci and LeurgansGoetz et al, 2000; Reference Manson, Schrag and LeesManson et al, 2000; Reference Mohr, Mendis and HildebrandMohr et al, 2000).

Limited data suggest that quetiapine (average dose of 60mg/day) may be beneficial and better tolerated than other atypicals (Reference Friedman and FactorFriedman & Factor, 2000; Reference Swindells, Zheng and GendelmanTargum & Abbott, 2000; Reference Fernandez, Trieschmann and BurkeFernandez et al, 2002, Reference Fernandez, Trieschmann and Burke2003). Patients with dementia are less likely to respond (Reference Fernandez, Trieschmann and BurkeFernandez et al, 2003). In such patients, preliminary studies with cholinesterase inhibitors such as rivastigmine and donepezil have shown improvement of psychotic symptoms without motor deterioration (Reference Reading, Luce and McKeithReading et al, 2001; Reference Bergman and LernerBergman & Lerner, 2002; Reference Tiffani and CummingsTolosa, 2003). On occasions the only alternative is to reach a compromise with patient and carers regarding psychotic symptoms and degree of mobility.

Treatment of psychotic symptoms in dementia with Lewy bodies needs to be instigated cautiously. Up to 50% of patients with Lewy body dementia treated with the older neuroleptics experience life-threatening exacerbation of parkinsonism and confusion (Reference Barber, Panikkar and McKeithBarber et al, 2001). They may experience extrapyramidal side-effects even with small doses of atypical antipsychotics and a minority develop features of neuroleptic malignant syndrome (Reference Barber, Panikkar and McKeithBarber et al, 2001). It has been suggested that quetiapine can be beneficial (Reference Fernandez, Trieschmann and BurkeFernandez et al, 2002). Cholinesterase inhibitors are a valid therapeutic alternative, as they improve non-cognitive as well as cognitive symptoms in Lewy body dementia (Reference Barber, Panikkar and McKeithBarber et al, 2001). This therapeutic effect possibly results from enhanced cortical muscarinic activation. In Lewy body dementia these cholinergic post-synaptic receptors are relatively better preserved than the cholinergic ascending presynaptic projections (Reference Barber, Panikkar and McKeithBarber et al, 2001).

Symptoms of depression greatly affect quality of life for patients with Parkinson’s disease and their relatives. They require active treatment. Mood fluctuations associated with the ‘off’ periods must be carefully explained to patient and caregivers. Dopaminergic therapy itself does not have a major antidepressant role. MAO–B and COMT inhibitors might enhance mood, but there are no systematic trials to support their use as antidepressants (Reference Thomas, Trimble and SchmitzTiffani & Cummings, 1998). Most antidepressants can be used to treat depression in Parkinson’s disease although there are isolated case reports of worsened parkinsonism with SSRIs (Reference Tseng and FosyValldeoriola et al, 1997).

Combination of SSRIs or tricyclics and MAO–B inhibitors such as selegiline could precipitate a serotonin syndrome and should not be used (Reference Tseng and FosyValldeoriola et al, 1997; Reference Thomas, Trimble and SchmitzTiffani & Cummings, 1998).

Electroconvulsive therapy (ECT) is effective in the treatment of severe depression in patients with Parkinson’s disease (Reference CoteCote, 1999). Both motor and depressive symptoms respond to ECT, but relapse can occur soon after it is discontinued (Reference Holcomb, Sternberg and HeningerHolcomb et al, 1983; Reference Thomas, Trimble and SchmitzTiffani & Cummings, 1998). Parkinsonian patients have an increased susceptibility to developing interictal delirium perhaps secondary to basal ganglia lesions or to ECT premedication (Reference Thomas, Trimble and SchmitzTiffani & Cummings, 1998). Psychotherapy and support groups occupy a central role in the treatment of depression in these patients (Reference CoteCote, 1999).

Anti-epileptic medication

All anti-epileptics can induce psychiatric disturbances (Table 3). Delirium and other cognitive changes are particularly common (Reference Kanner and Rivas NietoKanner & Rivas Nieto, 1999). Patients may require more than one anti-epileptic to control seizures, with increased risk of toxic levels due to metabolic interactions.

Table 3 Psychiatric side-effects of anti-epileptics

Drug	Psychiatric side-effects
Phenobarbital	Depression, sedation, sleep disturbances, psychosis, cognitive impairment, paradoxical agitation, delirium
Phenytoin	Agitation, insomnia, delirium
Primidone	Sedation, mood lability, psychotic symptoms, delirium
Benzodiazepines	Agitation, sedation, hallucinations, psychosis, cognitive impairment, delirium, withdrawal syndrome
Hydantoins	Similar to phenobarbital
Ethosuximide	Mood changes, irritability, sleep disturbances, psychosis, delirium
Sodium valproate	Sedation, hallucinations, depressive symptoms, delirium
Carbamazepine	Depression, agitation, sedation, psychosis, cognitive impairment, delirium
Vigabatrin	Agitation, lethargy, irritability, agitation, major depression, psychosis (‘schizophrenia-like’, in 2–4 % of treated patients), cognitive impairment
Topiramate	Psychosis (6 % of treated patients), depression, emotional lability, cognitive difficulties
Tiagabine	Psychosis (0.8 % of treated patients), depressive symptoms, sedation
Levetiracetam	Irritability, sedation and psychosis
Gabapentin	Sedation, agitation, fatigue
Lamotrigine	Sedation, depression, agitation, psychosis (0.3% of treated patients)

Of the most established anti-epileptics, ethosuximide, clobazam, phenytoin, carbamazepine, phenobarbital, primidone and benzodiazepines have all been associated with development of psychotic symptoms (Reference LancmanLancman, 1999). Phenobarbital and primidone have been associated with depression as have, occasionally, sodium valproate and carbamazepine (Reference Kanner and Rivas NietoKanner & Rivas Nieto, 1999).

The newer anti-epileptics are also associated with the development of psychiatric complications. Two molecules closely related to established anti-epileptics are oxcarbazepine and fosphenytoin, which have a profile of side-effects similar to that of their predecessors. Other recently introduced anti-epileptics include vigabatrin, topiramate, tiagabine, gabapentin, lamotrigine and levetiracetam.

Vigabatrin causes few cognitive side-effects, but can result in transient psychosis (Reference GubermanGuberman, 1996; Reference Levinson and DevinskyLevinson & Devinsky, 1999). Increased risk has been associated with a right-sided epileptic focus and acute suppression of seizures (Reference Targum and AbbottThomas et al, 1996).

Topiramate, and to a lesser degree tiagabine, can induce psychosis (Reference Ketter, Post and TheodoreKetter et al, 1999; Reference Khan, Faught and GilliamKhan et al, 1999). Side-effects are less likely using lower starting doses and slow titration (Reference Adkins and NobleAdkins & Noble, 1998; Reference KalviainenKalviainen, 2001; Reference Sabato, Wesselingh and FullerSackellares et al, 2002).

Gabapentin and lamotrigine do not appear to be strongly associated with development of psychiatric disturbances, including cognitive dysfunction (Reference Ketter, Post and TheodoreKetter et al, 1999).

The withdrawal of anti-epileptics may cause prominent psychiatric symptoms, including psychosis. This is well recognised for barbiturates and benzodiazepines. However, withdrawal of phenytoin, carbamazepine and valproate has been implicated in the development of psychiatric symptoms in up to 40% of patients with epilepsy (Reference Ketter, Post and TheodoreKetter et al, 1999).

Of the newer drugs, gabapentin, lamotrigine, levetiracetam and oxcarbazepine appear to have mood-stabilising effects; there is insufficient information regarding other compounds (Reference Ketter, Post and TheodoreKetter et al, 1999; Reference BesagBesag, 2001).

Treatment of psychiatric complications

The treatment of psychiatric symptoms requires establishing whether these are temporally associated with seizures, result from iatrogenic effects of anti-epileptic medication or have an alternative cause. Psychiatric symptoms may emerge after a recent withdrawal of medication or a change in dose. A psychiatric disorder induced by an anti-epileptic drug may respond to a reduced dosage or a change of medication.

Psychotropic drugs should be used with caution in patients with epilepsy, because of their tendency to increase seizure frequency. Overall, seizures occur in 1% of patients taking antipsychotics (Reference LancmanLancman, 1999). Treatment with antipsychotics should be initiated at low doses and increased slowly. Butyrophenones, risperidone and olanzapine are less likely to induce seizures than are phenothiazines. Clozapine has a dose-related higher risk for seizures (Reference LancmanLancman, 1999).

Tricyclic antidepressants are more proconvulsant than MAO–A inhibitors or SSRIs, although this action appears to relate to the antidepressants’ plasma levels, but not to their intrinsic action over brain concentrations of monoamines (Reference Kanner and Rivas NietoKanner & Rivas Nieto, 1999). Antidepressants should be started at small doses and increased cautiously, to minimise risk of seizures. Owing to their side-effects profile, SSRIs are recommended as first-line treatment of depression in epilepsy (Reference Kanner and Rivas NietoKanner & Rivas Nieto, 1999). Newer antidepressants such as moclobemide and venlafaxine do not appear to be proconvulsant (Reference Lambert and RobertsonLambert & Robertson, 1999).

Antidepressants and anti-epileptics, which are both metabolised in the liver, may have pharmacological interactions and these should be checked with regard to specific combinations. Electroconvulsive therapy can be used if antidepressants are ineffective or inappropriate (Reference Kanner and Rivas NietoKanner & Rivas Nieto, 1999; Reference Lambert and RobertsonLambert & Robertson, 1999).

The role of carbamazepine and sodium valproate in the treatment of bipolar disorder in epilepsy has not been established. Withdrawal of these medications can unmask a bipolar disorder (Reference Kanner and Rivas NietoKanner & Rivas Nieto, 1999). Lithium is proconvulsant at therapeutic serum concentrations (Reference Kanner and Rivas NietoKanner et al, 1999). Treatment response can be optimised by behavioural and cognitive therapies combined with education of patients and carers (Reference Kanner and Rivas NietoKanner & Rivas Nieto, 1999).

Treatment of psychiatric side-effects

The general principles of treatment for antibiotics and the remaining medications mentioned below are similar to those already discussed, namely reducing dosage or discontinuing the responsible medication. On occasions, patients may need additional use of psychotropics to control symptoms.