Wondering About the Wonder Drugs: A Second Look at the Atypical Antipsychotics

Although 10 times as expensive as the older neuroleptics, atypical antipsychotic drugs rapidly captured market share and became first-line treatments for schizophrenia and for psychosis and agitation in Alzheimer disease (AD). In manufacturer-designed trials that compared these drugs to placebo or haloperidol in patients with relatively uncomplicated illnesses and that measured outcomes with a single symptom-rating scale, the atypicals were found to be better tolerated, superior for negative symptoms, and more likely to be taken long-term. More recently, government-sponsored, randomized, controlled, double-blind studies have raised questions about the advantages of these medications.

In an 18-month, multicenter study (published in fall 2005) from the NIMH-sponsored Clinical Antipsychotic Trials of Intervention Effectiveness (CATIE), 1493 patients with chronic schizophrenia received either the neuroleptic perphenazine or an atypical antipsychotic (olanzapine, risperidone, quetiapine, or ziprasidone). Clinicians could adjust doses according to side effects and response. The primary outcome measure was time to treatment discontinuation, which was believed to be a real-life indicator of how patients fare. Seventy-four percent of patients discontinued their medication for any reason (range, 64% [olanzapine] to 82% [quetiapine]). When corrections were made for multiple statistical tests, patients were found to stay on olanzapine longer than quetiapine or risperidone, but not significantly longer than perphenazine or ziprasidone. The discontinuation rate for lack of efficacy was significantly lower with olanzapine than with other medications (except ziprasidone), but these rates were not enormously different (olanzapine, 15%; other medications, 24%–28%). The groups showed no differences in likelihood of medication discontinuation because of patient preference or adverse effects, including extrapyramidal symptoms (EPS). However, significantly more patients taking olanzapine than those taking other medications had clinically important weight gain (olanzapine, 30%; other medications, 10%–15%), and cholesterol and triglyceride levels increased significantly more with olanzapine than with other medications.

These results are compatible with those from the British Cost Utility of the Latest Antipsychotic Drugs in Schizophrenia Study (CUtLASS I), in which 227 patients with chronic schizophrenia received either a neuroleptic or an atypical chosen by their psychiatrists and dose-adjusted according to clinical response. The primary outcome measure was a well-validated, blindly rated quality-of-life scale assessing social, psychological, and occupational functioning. Dropout rates were lower than in CATIE, but after a year of treatment, there were no significant between-group differences in quality of life or in EPS, compliance, attitudes toward medication, positive symptoms, negative symptoms, depression, or global functioning. The average cost of care was statistically equivalent for patients taking neuroleptics and those taking atypicals, because most of the costs were due to inpatient hospital admissions.

Another CATIE study examined antipsychotic efficacy in 421 community-dwelling patients with AD and prominent psychosis, aggression, or agitation who received either placebo or an atypical antipsychotic (olanzapine, risperidone, or quetiapine) for 36 weeks. Most patients were taking other medications for various illnesses and required skilled care. The primary outcome measure was, again, time to study discontinuation, and there were no significant between-group differences (range, 5–8 weeks). Fewer than 20% of patients stayed on their initial medication. The average time to discontinuation because of ineffectiveness, however, was significantly longer for risperidone (27 weeks) and olanzapine (mean, 22 weeks) than for quetiapine (9 weeks) or placebo (9 weeks). Still, significantly more patients taking olanzapine or risperidone than placebo had EPS and confusion. All active drugs had higher rates of sedation and weight gain than did placebo. There were no significant between-group differences in rates of global improvement (21%–32%) or serious adverse effects. Although somewhat higher doses might have produced better results (and even more adverse effects), these findings suggest that antipsychotic drugs might not be first-line treatments for behavioral disorders in AD. Nonpharmacologic therapies (e.g., bright light in the morning for patients who are agitated at night) and other medications (e.g., anticonvulsants) should be considered first.

The advantages of the atypical antipsychotics, apparent in the original manufacturer-sponsored registry trials, do not seem as predictable when more heterogeneous patient samples are studied in practical clinical trials in which there is more uncertainty and in which doses are adjusted according to clinical judgment rather than a protocol. Atypicals are probably better tolerated acutely than neuroleptics, but their long-term acceptability is still disappointing. The only atypical that is clearly more effective in refractory schizophrenia is clozapine. In nonpsychotic AD patients with intermittent agitation, neuroleptics are no better, and atypicals seem more effective, than placebo. Interestingly, medication costs, even for atypicals, comprise only a small portion of the overall cost of care.