To the Editor: Caution is needed when interpreting the results of the study by Dr. Volavka et al. The article and the accompanying editorial (1) acknowledged that the cohort effect cannot be ruled out, the dose of risperidone was too high, and 18% of the funding was obtained from Eli Lilly and Company. The olanzapine arm of the study was included 15 months after the study had started; there was no quetiapine arm, although both drugs became available around the same time. Use of haloperidol in comparison with either loxapine or molindone as a comparator first-generation antipsychotic drug is questionable. Both offer certain advantages over other first-generation antipsychotics.

The authors’ concept of suboptimal response is less rigorous than the more widely accepted criteria of Kane et al. for treatment resistance (2). The score on the Positive and Negative Syndrome Scale required for study entry was >60. These factors may have contributed to entry of treatment-responsive patients and those with milder illness.

High doses for haloperidol (mean=25.7 mg/day, SD=5.7; approximately 1720 chlorpromazine-equivalent units) during the second phase of variable dosing cannot be justified. Low doses (mean=3.4 mg/day, SD=2.3) of haloperidol are efficacious in patients with acute schizophrenia; higher doses cause a significant increase in extrapyramidal side effects (3). Positron emission tomography experiments performed by Farde et al. (4) suggested that there is sufficient dopamine D₂ receptor occupancy with doses of haloperidol as low as 4–6 mg/day. High doses of haloperidol in treatment-resistant patients with schizophrenia do not provide any advantage based on D₂ receptor occupancy (5).

The advantages of atypical antipsychotics in terms of efficacy and dropouts disappear when doses below 12 mg/day of haloperidol are used (6). The optimal dose for risperidone is 4 mg/day, and there is a therapeutic window: poor response results at higher doses (Love et al., 1999; Williams, 2001). No incremental clinical improvement in chronic psychosis is seen at doses above 375 mg/day in chlorpromazine-equivalent doses, although a significant increase in adverse reactions is observed (7).

The Clinical Antipsychotic Trials of Intervention Effectiveness study, sponsored by the National Institute of Mental Health, may offer some more insight; however, this study is also limited by including only one comparator from the first-generation drug category. No study on the comparative effectiveness of first-generation versus newer antipsychotics can be definitive without using loxapine or molindone in treatment arms.