Numerous studies compare side effects or adverse drug reactions (ADRs) of the various typical and newer atypical neuroleptics in patients with schizophrenia. However, these studies, as controlled randomized trials, represent an artificial setting of drug administration and do not easily relate to the "real-life" setting of psychiatric treatment. In contrast, the AMSP drug safety program allows the monitoring of ADRs of all types of psychopharmacological agents in the naturalistic setting of routine clinical practice. In the present study, the data on neuroleptics acquired in the AMSP program from 1993 to 2000 are analyzed. In this period, 86,439 patients treated with at least one neuroleptic agent were monitored. In 1.1 % of the patients severe ADRs occurred. In contrast to the results from controlled trials, atypical neuroleptics caused more severe ADRs than did typical neuroleptics. This result was mainly caused by the high number of severe ADRs in patients treated with clozapine and concerned delirium and non-EPS neurological, gastrointestinal, hepatic, dermatological, hematological, and endocrinological ADRs. Atypical neuroleptics were found to be superior in EPS and urological ADRs. Excluding the data on clozapine, we found typical and atypical neuroleptics to be similar in the occurrence of severe ADRs, although the profiles differ between these two groups as well as between the single substances. Our findings provide valuable information on the type and frequency of ADRs in psychiatric practice, thus enabling differential indication of neuroleptics based not only on the efficacy and tolerability data of controlled trials but also on their differential ADR profile occurring in the "real-life" setting of routine clinical treatment.