Subject selection for the placebo- and comparator-controlled trials of neuroleptics in schizophrenia

It has been suggested that inclusion of a placebo treatment arm in controll ed clinical trials might bias the selection of study subjects. Presumably, patients in the placebo-controlled studies are more stable, but there are n o data available to support such an assumption. The authors tested the hypo thesis in a set of randomized trials of neuroleptics in treating schizophre nia by comparing placebo-controlled (PCTs) and comparator-controlled trials (CCTs) in terms of basic patient characteristics. The results, based on a total of 296 studies, showed that the patients in BCTs, compared with those in CCTs, mere older (p < 0.002), had a longer duration of illness (p < 0.0 01), and a loa er initial symptom severity (p < 0.02). No difference was fo und in the number of subjects per treatment arm or in the proportion of fem ale subjects. However, investigation of studies which used same-gender stud y subjects revealed that female-only populations mere more likely to be tes ted in PCTs (p < 0.03) than in CCTs, To investigate current trends in psych opharmacologic research, the authors tested separately a subset of trials o f new atypical antipsychotics. The results indicated a significantly smalle r number of females participating in the latest PCTs (p < 0.0003). Moreover , our findings suggest that the characteristics of patients in the current controlled trials are rather uniform; thus, the generalizability of new stu dy findings for certain groups of patients,vith schizophrenia (e.g., with e arly or late onset or brief duration of illness) may be compromised.