INTRODUCTION — Antipsychotic medications were introduced more than 50 years ago. Although initially targeted to the treatment of schizophrenia, these medications are effective against psychosis irrespective of cause. They are also of benefit for manic episodes, agitation, delirium, impulse control, dissociation, and as augmentation therapy in other psychiatric disorders.
This topic reviews drug selection, treatment protocols, side effects and management of antipsychotic medications. The pharmacological characteristics of these medications in adults is discussed separately, as are identification and treatment of overdose with atypical antipsychotic agents, the target symptoms, the treatment of psychosis in special populations, and the non-pharmacologic treatment of psychosis.
PHARMACOTHERAPY FOR PSYCHOSIS — As discussed below, both conventional and atypical antipsychotic medications are associated with significant side effects, including sudden death. While antipsychotics are clearly indicated for patients with schizophrenia and other psychoses, their use in all patients should be limited to situations for which the benefits exceed the risks.
Drug selection — No one antipsychotic medication has emerged as the preferred choice for first-line treatment . They have comparable efficacy for psychosis and manic symptoms, and carry similar side effect burdens in reviews of large clinical trials [2-4].
The most comprehensive of these trials was the National Institute of Mental Health (NIMH)-sponsored CATIE study (Clinical Antipsychotic Trials of Intervention Effectiveness), a double-blind assessment of 1493 patients in the maintenance phase of treatment for chronic schizophrenia . Patients were randomly assigned to receive the conventional antipsychotic perphenzazine or one of the newer agents (olanzapine,quetiapine,risperidone orziprasidone) . The study defined effectiveness as time to discontinuation of treatment for any cause over 18 months. The overall discontinuation rate was 74 percent, with olanzapine having the lowest rate at 64 percent.
Olanzapine demonstrated a modest advantage in treatment duration in the original data analysis. However, subsequent reanalysis suggested this may have been due to more patients taking olanzapine than other drugs prior to study entry; reanalysis found only a trend toward longer treatment duration for olanzapine . This suggests that patients in long-term treatment do better when continued on the same medication than when switched. Thus, dose optimization or addition of psychosocial or behavioral therapy may be preferable to a switch in medication for patients stabilized on one medication.
In phase 2 of CATIE, patients who did not complete the study were given the choice ofclozapine or randomization to a different atypical drug. Among patients receiving a first-line drug, those givenrisperidone andolanzapine stayed in treatment longer than those receivingquetiapine orziprasidone, although 74 percent of these patients discontinued the medication prior to study completion . Only 9 percent of eligible patients chose clozapine, but of this group only 56 percent discontinued the drug prematurely . The average duration of treatment with clozapine was 10.5 months versus 2 to 3 months with olanzapine, quetiapine, or risperidone. Clozapine alone demonstrated clearly superior efficacy, a finding already well established in previous studies [9-14].
In general, studies comparing usual doses ofhaloperidol (5 to 20 mg daily) with atypical antipsychotics have reported higher rates of medication discontinuation and patient relapse for the conventional drug . However, it has been argued that these studies used higher doses of haloperidol than necessary, increasing the incidence of side effects. A multicenter non-blinded randomized study conducted in Europe in 500 patients newly diagnosed with schizophrenia compared low dose haloperidol (1 to 4 mg daily) with four atypical antipsychotics (amisulpride,olanzapine,quetiapine, andziprasidone) . At 12 months, patients were more likely to discontinue haloperidol than the atypical drugs (72 percent versus 33 to 53 percent), although, with an unblinded study, this may in part reflect a tendency to switch to a perceived more effective treatment. There were no differences in measured symptom response or hospital admission rates among the five medications.
A meta-analysis of 150 randomized trials directly comparing antipsychotics in patients with schizophrenia found similar efficacy for atypical and conventional drugs, but heterogeneity among the atypical drugs . Four atypical drugs (clozapine, amisulpride,olanzapine, andrisperidone) had better efficacy than other antipsychotics, with small to medium effect size. Most studies usedhaloperidol as a comparator, and did not evaluate lower potency conventional agents.
Another meta-analysis evaluated 78 randomized head-to-head trials of nine atypical antipsychotics (including three not available in the US) . Better efficacy (primarily improvement in positive symptom scores), with small to moderate effect sizes, was also found forolanzapine,risperidone, andclozapine . The authors noted that differences in side effects and costs may outweigh differences in efficacy.
In contrast to relative modest differences in efficacy, studies consistently show statistically and clinically significant differences among antipsychotics in side effects, including risk of EPS, weight gain, metabolic effects, sedation, hypotension, and GI effects. In the CATIE study,olanzapine discontinuation was primarily related to weight gain and metabolic side effects (increased blood glucose, hyperlipidemia);quetiapine to sedation; andperphenazine to extrapyramidal effects . Additional differences among drugs include routes of administration, pharmacokinetics, and cost. Individual patients may respond preferentially to one of these agents, but there are no indications or biological markers to predict which drug will work for a specific person .
In general, atypical antipsychotic medications have been considered first-line, because of their lower risk for extrapyramidal side effects compared to conventional antipsychotics. However, atypical antipsychotics are more expensive, and several cause more metabolic side effects and weight gain than conventional agents. Additionally, unnecessarily high doses of conventional antipsychotics in earlier studies may have increased the incidence of side effects.
In practice, patients often have had previous medication exposure, and treatment selection can be based on their prior response. Patients with a history of good response to a drug should be given that medication. Patients who had intolerable side effects to one drug should be treated with a drug with the lowest propensity for that adverse effect. For patients who initially failed to respond to a drug, any of the others may reasonably be selected. Individuals who responded well but were habitually nonadherent to treatment should be considered for depot medication (table 1).
For new patients, drug selection may be based on the expectation of a specific side effect, such as weight gain in a patient with obesity or diabetes, or EPS in a patient with parkinsonian symptoms. When no obvious risk exists, factors of cost, convenience, and clinician familiarity with the drug should be considered (table 2). For the practitioner who only occasionally uses these medications, it is reasonable to develop greater experience and comfort with one or two medications and preferentially use those.
Treatment course and duration — Dosages of the antipsychotic drugs should be titrated into the average therapeutic range as quickly as tolerated. Resolution of psychotic symptoms generally occurs over several days and may take as much as four to six weeks, although there is an immediate calming effect for agitated patients. Clinicians should avoid the impulse to change the medication or dose prematurely.
The decision to increase the dose beyond the average range should follow at least several days of treatment at moderate doses. Higher dosing should be accompanied by careful observation for side effects, but is usually well tolerated. In most instances, an increase in dose of the current medication is preferable to a switch to an alternative drug.
The duration of treatment depends on the diagnosis. Patients with delirium, agitation, brief psychotic disorders, or impulsivity may require only a few days or weeks of treatment. Schizophrenic patients, in contrast, benefit from continuous treatment and are at risk for relapse with medication discontinuation or even dose reduction .
The role of antipsychotics in maintenance therapy for bipolar disorder is uncertain; usual practice is to withdraw antipsychotics and continue mood stabilizers after several weeks of mood stability . Maintenance treatment may also be considered in patients treated for persistent impulsivity or behavioral problems, but there are limited data.
Medications should be periodically reviewed for patients on long-term medication. Routine dose reduction in stable patients to establish the “lowest effective dose” is associated with high rates of relapse and should be avoided [1,21]. Dose reduction may be justified in patients at risk for side effects and in patients on high doses of medication following an acute episode of illness.
Medication changes — Two basic strategies have been recommended for switching antipsychotic medications [23,24]. A standard cross-titration involves simultaneous taper of the current medication with titration of the replacement drug in three to four steps over several days to several weeks. For patients at higher risk of relapse, the current medication can be maintained at its full dose as the new medication is titrated. Once the second drug has reached its target dose, the former medication may be tapered and discontinued.
Discontinuation of antipsychotic medications is generally well tolerated, except forclozapine, for which both cholinergic rebound and withdrawal-emergent movement disorders have been reported [25-27]. A slow taper of clozapine (over one to two weeks) is recommended. No specific recommendations for discontinuation of other antipsychotics have been made.
Treatment-refractory cases — Up to 30 percent of schizophrenic patients do not respond to first-line antipsychotic treatment. Three basic strategies address resistant cases: use ofclozapine, augmentation, or simultaneous multiple antipsychotics. Of these options, only clozapine treatment is of proven efficacy for treatment-refractory patients [27-32], a finding confirmed in the CATIE study .
Clozapine — There is no consensus as to how many other antipsychotics should be tried before beginningclozapine; guidelines variously recommend two to four other drugs, and differ as to specific medications.
Clozapine is effective in 30 to 50 percent of patients with psychosis who have not responded to other antipsychotic drugs. Life-threatening agranulocytosis has been reported in one to two percent of patients during the first six months of treatment. Consequently, a schedule for monitoring has been designed requiring weekly CBC for the first six months, biweekly CBC for the second six months, and subsequent testing every four weeks. We suggest that clozapine be prescribed by a specialist familiar with this medication.
Augmentation — Augmentation strategies include using anticonvulsants,lithium, benzodiazepines, and electroconvulsive therapy . Augmentation may be most helpful if a specific secondary symptom, such as mood instability, irritability, or anxiety is prominent, allowing rational drug selection to target those symptoms.
Combination therapy — The simultaneous use of two or more antipsychotic drugs is more controversial . This is difficult to justify pharmacologically, since the antipsychotic effectiveness appears to be related to the same dopaminergic receptor system for all of these drugs.
Data on combination therapy are unclear. Three randomized trials that combineclozapine withrisperidone for treatment of refractory schizophrenia came to differing conclusions [30-32]. One industry-sponsored study (n = 40) showed modest benefit when risperidone 6 mg/day was added to clozapine for 12 weeks, compared to placebo plus clozapine . A second industry-sponsored study of 30 patients treated for six weeks with risperidone 6 mg/day plus clozapine or placebo plus clozapine showed better symptom response in the placebo/clozapine group . A third study (n = 68), not industry-sponsored, compared an eight week course of clozapine plus risperidone at 3 mg/day with clozapine plus placebo and also showed no symptom benefit with the addition of risperidone; there was mild memory impairment in the risperidone-treated group .
ADVERSE EFFECTS AND MANAGEMENT — Side effects are usually mild and of limited duration (table 3). If side effects do not improve with time, dose reduction is recommended, particularly for higher doses. If dose reduction is ineffective, a switch to another medication may be required. A small number of cases require adjunctive medication. The side effects of greatest concern in long-term treatment are tardive dyskinesia (TD) and weight gain.
Tardive dyskinesia — Tardive dyskinesia (TD) is the late-onset of choreoathetotic movements associated with prolonged exposure to antipsychotic medications. Major risk factors include the use of conventional antipsychotic medications, older age, and longer duration of treatment . (See“Tardive dyskinesia: Etiology and epidemiology”.)
The course of TD symptoms is variable. Although older textbooks describe TD as chronic and progressive, more recent studies have emphasized the self-limiting quality of many cases, with the majority of patients showing at least intermittent improvement for both conventional and atypical drugs. (See“Tardive dyskinesia: Clinical features and diagnosis”.)
Patients on long-term maintenance treatment should be monitored periodically for TD. When clinically appropriate, dose reduction may be considered to reduce symptoms. Patients on conventional antipsychotics should be considered for an atypical agent.Clozapine should be considered for patients with severe or progressing symptoms . Numerous other treatments, includingvitamin E, dopaminergic agents, benzodiazepines, and branched-chain amino acids, have not been shown to be effective. (See“Tardive dyskinesia: Prevention and treatment”.)
Weight gain — Weight gain is a significant issue with several of the atypical antipsychotics and should be monitored carefully (table 4). Clozapine carries the greatest risk, followed byolanzapine [35-37]. The CATIE study found that over 18 months, patients taking olanzapine gained an average of two pounds per month and 30 percent gained 7 percent or more of their initial body weight .
Studies in mice have demonstrated that weight gain with atypical antipsychotics results from stimulation of AMP-activated protein kinase in the hypothalamus; deletion of the H-1 histamine receptor in knockout mice blocks this effect . This suggests the potential to develop new antipsychotic agents with differential activity on dopamine and histamine receptors, to avoid weight gain.
No adjunctive medication has been shown to neutralize weight gain, although studies ofmetformin suggest that it may modulate the effect.
A 12 week study of 80 patients stabilized onolanzapine who were randomly assigned to (850 to 2250 mg daily) or placebo showed modest weight loss (1.4 kg) in themetformin group, but none with placebo .
A 12 week trial in 128 patients who had gained at least 10 percent of baseline body weight with antipsychotic treatment foundmetformin (750 mg daily) significantly more effective than placebo plus lifestyle intervention in reducing BMI and waist circumference; metformin plus lifestyle intervention was more effective than metformin alone .
Metformin also attenuated weight change in a randomized trial of patients who were receiving initial therapy witholanzapine for first episode schizophrenia, and therefore had not already demonstrated weight gain .
Several other agents have been evaluated. Modest weight loss was reported in randomized trials fortopiramate [43-45],amantadine , orsibutramine , although sibutramine was not more effective than placebo in a study ofclozapine-induced weight gain . Amantadine may worsen psychosis and sibutramine led to elevated blood pressure and anticholinergic symptoms.Fluoxetine was not better than placebo in preventing weight gain in patients takingolanzapine .
Althoughmetformin shows promise for patients with significant weight gain, the preponderance of evidence supports switching to a different antipsychotic (especiallyaripiprazole orziprasidone) in preference to the addition of an adjunctive medication.
Diabetes and other metabolic abnormalities — Antipsychotic medications, and atypical antipsychotics in particular, have been associated with hyperglycemia, dyslipidemia and hypertension [50-52].
The risk of new-onset diabetes was equivalent for patients treated for one year witholanzapine,risperidone, orquetiapine, and significantly greater than in patients treated withhaloperidol (HR 1.60 to 1.67) . One-third of new cases of diabetes could be attributed to use of atypical antipsychotics in the VA patients treated with these drugs.
In the CATIE study, at three months, waist circumference was most increased byolanzapine andquetiapine, was not changed withziprasidone, and decreased withperphenazine . Fasting triglyceride levels increased with olanzapine and decreased with ziprasidone.
Atypical antipsychotic drugs have also been associated with a small increased risk of diabetic ketoacidosis [56,57].
In a case control study comparing patients treated with antipsychotic drugs to control patients, the odds ratio for hyperlipidemia ranged from 1.82 forclozapine to 1.26 for first generation antipsychotics;aripiprazole was the only antipsychotic that did not significantly increase the risk for hyperlipidemia .
A consensus statement from the American Psychiatric Association, American Diabetes Association, and others recommends baseline assessment of weight, blood pressure, fasting plasma glucose, and fasting lipid profile, and reassessment after 12 weeks of treatment for patients treated with atypical antipsychotic medications . Patients with impaired fasting glucose should be tested for diabetes with a 2 hour oral glucose tolerance test . Weight should be followed monthly for the first three months and quarterly thereafter (table 4). Unfortunately, adoption of these guidelines has been limited. Among 17,800 patients initiating treatment with atypical antipsychotics in one study, serum lipids were tested in 10 percent and glucose monitored in 22 percent at baseline and 18 percent at 12 weeks .
No adjunctive medication has been conclusively shown to reduce the metabolic effects. Even for drugs that mitigate weight gain, such asmetformin, metabolic effects have been modest [40,41] or absent . Patients who develop antipsychotic-induced metabolic disturbance should be switched to an agent with less propensity for this side effect, specificallyaripiprazole orziprasidone.
Extrapyramidal symptoms — EPS are common with conventional agents and occasionally encountered at higher doses of the atypical drugs. If dose reduction or change in medication are not effective or not well tolerated, several approaches are recommended. Anticholinergic drugs, such asbenztropine 1 to 2 mg twice daily to four times daily ortrihexyphenidyl 2 to 5 mg twice daily to four times daily are most commonly used. Other possible agents includeamantadine 100 mg twice daily to four times daily ordiphenhydramine 25 to 50 mg twice daily to four times daily.
Akathisia — Akathisia is a subjective sense of restlessness, often accompanied by voluntary movements of the limbs or trunk. It is among the more common movement disorders associated with antipsychotic medications, and may be mistaken for anxiety or insomnia. It should be suspected in any patient who is unable to sit quietly or who complains of feeling nervous, uncomfortable, or unable to sleep.
Approaches to the treatment of akathisia include antipsychotic dose reduction, change to a different atypical drug, or addition of a beta-blocker , antihistamine , benzodiazepine , anticholinergic , or serotonin antagonist (eg,mirtazapine) . Beta-blockers have been the traditional “gold standard” for akathisia treatment but are not more efficacious than other agents . There are no studies demonstrating efficacy for combination therapy. Severe akathisia is an indication forclozapine .
Neuroleptic malignant syndrome — The neuroleptic malignant syndrome (NMS), a tetrad of clinical features (fever, rigidity, mental status changes, and autonomic instability), is associated with medications that block dopamine transmission. NMS is seen with every class of antipsychotic drug, including “atypical” antipsychotics . Treatment involves withdrawal of medication, and intensive management for cardiovascular support, control of hyperthermia, and fluid and electrolyte balance. (See“Neuroleptic malignant syndrome”.)
Stroke risk — Evidence regarding stroke risk associated with atypical antipsychotic drugs is conflicting [28,69-72]. A large cohort study of adults aged≥65 years found a similar risk of ischemic stroke among users of atypical versus typical antipsychotics . A study of the time concurrence of antipsychotic use and stroke in patients who both used antipsychotics and had a stroke found the risk of stroke was greater during the time period an antipsychotic was used, was greater for atypical than typical antipsychotics, and was greatest for patients with dementia .
Sudden death — Both conventional and atypical antipsychotics increase risk for sudden death . A case-controlled study of 90,000 Medicaid recipients (mean age 45.7) found that, compared to nonuse, current use of antipsychotics doubled the risk of sudden death (incidence rate ratios 1.99, 95% CI 1.68-2.34, and 2.26, 1.88-2.72 for typical and atypical drugs respectively). Risk increased with higher medication dose. The rate of sudden death with antipsychotics (1.8 per 1,000 person years) is almost ten-fold the rate of death related toclozapine-associated agranulocytosis .
The mechanism underlying sudden death was not determined, but is presumed related to the known effect of antipsychotics on cardiac repolarization and QTc prolongation. It is unclear whether ECG monitoring will identify patients at increased risk for sudden death. Drugs and conditions that prolong the QTc interval should be avoided, when possible, in patients who are treated with antipsychotics (table 5). (See“Acquired long QT syndrome”.)
Elderly patients with dementia — Both conventional and atypical antipsychotic medications have been shown to be associated with an increased risk of mortality when used to treat elderly patients with psychosis or agitated behavior resulting from dementia [77-79]. The US Food and Drug Administration (FDA) issued a public health advisory in 2008 highlighting these findings and emphasizing that antipsychotics are not approved for the treatment of dementia-related psychosis . They advised that the mortality risk should be discussed with patients, families, and other caregivers if antipsychotics are prescribed.
Major studies informing this determinaton include:
A meta-analysis of 15 studies (nine unpublished) of atypical antipsychotic drug therapy for dementia found that mortality was increased in treated patients compared with controls (3.5 versus 2.3 percent, OR 1.54) . Risks did not differ for the medications studied (aripiprazole,olanzapine,quetiapine, and risperidone). Most studies were less than three months duration.
A retrospective cohort study of 22,890 elderly patients receiving antipsychotic medications compared mortality risks with conventional versus atypical agents . Significantly higher mortality was seen in patients taking conventional agents (OR 1.37). The increase in risk was greatest soon after therapy was initiated and with higher doses of conventional agents.
A Canadian retrospective study found increased mortality risk at 30 days for patients receiving atypical antipsychotics, compared to no antipsychotics, for both community-dwelling and long-term care patients (HR 1.31 and 1.55, respectively) . Conventional antipsychotics were associated with greater 30-day mortality than atypical antipsychotics.
A trial in the UK compared mortality for 165 patients with Alzheimer disease who were randomly assigned to continue their antipsychotic medication or switch to placebo . Survival at 12 and 24 months was significantly greater for the group assigned to placebo (at 24 months, 71 versus 46 percent for antipsychotic continuance).
These studies evaluated mortality when antipsychotic medications are used to manage behavioral issues associated with dementia in the elderly. Behavioral disturbance and psychosis are separate problems, although the FDA-mandated warning specifies increased risk for “dementia-related psychosis.” Other treatment options are available for behavioral problems (eg, mood stabilizers, benzodiazepines), but there are no other medication options for delusions or hallucinations . The only recourse for clinicians is to inform the patient, family, and caregivers of the risk associated with antipsychotic medications in the elderly, and to decide with them whether to treat with antipsychotic medications.
Schizophrenia — Although it has been suggested that treatment with antipsychotic medication may contribute to the higher mortality rates associated with schizophrenia, a population-based study drawing on Finland’s national data registries found that long-term treatment of schizophrenia with any antipsychotic was associated with a lower rate of mortality compared with not receiving these medications (hazard ratio (HR) = 0.81, 95% CI , 0.77–0.84) . The relationship between antipsychotics and mortality varied by medication. Clozapine was associated with the greatest reduction in mortality (HR = 0.74, 95% CI 0.60–0.91), while quetiapine was associated with the largest increase (HR = 1.41, 95% CI 1.09–1.82).
The use of antipsychotic medications is widespread and often extends beyond approved indications, populations, and dosing strategies. The quality and quantity of evidence for these practices is variable, but often the medications are useful and well tolerated in these situations.
The conventional and atypical antipsychotics are comparable in efficacy in the treatment of psychosis, with the exception ofclozapine which is more effective but has serious side effects.Olanzapine may also be somewhat more effective in long-term maintenance therapy for schizophrenia but carries greater risk of metabolic side effects. Genericrisperidone may have cost-effectiveness advantages as the lowest cost atypical drug.
For patients requiring long-term therapy with antipsychotic medication, we suggest starting with an atypical agent other thanclozapine, and basing this choice on the side effect profile preferred by the patient, and (when available) the patient’s previous history of medication response (Grade 2B). (See‘Drug selection’ above.)
Clozapine is effective in 30 to 50 percent of patients with psychosis who have not responded to other antipsychotic drugs. Life-threatening agranulocytosis has been reported in one to two percent of patients. A schedule for monitoring has been designed requiring weekly CBC for the first six months, biweekly CBC for the second six months, and then testing every four weeks. We suggest that clozapine be prescribed by a specialist familiar with this medication. (See‘Clozapine’ above.)
Augmentation strategies using drugs such aslithium, anticonvulsants, or benzodiazepines may be helpful for patients refractory to antipsychotic monotherapy. We suggest not treating patients with refractory schizophrenia with two or more antipsychotic agents simultaneously (Grade 2B) (see‘Augmentation’ above and‘Combination therapy’ above.
Tardive dyskinesia is the late onset of choreoathetotic movements.Clozapine treatment is effective for patients with severe tardive dyskinesia symptoms (see‘Tardive dyskinesia’ above.
Patients under treatment with antipsychotic medications should be monitored closely for weight gain and switched to another antipsychotic medication if indicated (see‘Weight gain’ above.
Antipsychotic medications, both conventional and atypical, have been associated with increased mortality in the elderly when used to manage dementia-related behavioral issues (see‘Elderly patients with dementia’ above.