As many as half of all schizophrenic patients have abused alcohol or illicit drugs. This study determines the extent of substance abuse in a treatment-resistant population and assesses the response of this population to olanzapine treatment. Sixty patients with a DSM-III-R diagnosis of schizophrenia were included in an open 7-week trial of up to 25 mg/day of olanzapine. A history of substance abuse was present in 23 (38%) of the patients. At baseline evaluation, patients with a history of substance abuse had lower CGI scores and less negative symptomatology while having a higher rate of tardive dyskinesia. The overall group improved significantly over time. There were no differences in response between the substance-abusing (SA) and non-substance-abusing (NSA) patients as measured by the total BPRS, GGI and SANS ratings. The NSA group had significantly greater improvement in negative symptoms as measured by the BPRS negative symptom factor. Sixty-nine per cent (16/23) of the SA group and 60% (22/37) of the NSA were considered olanzapine responders by a priori criteria (p = NS). Extrapyramidal symptoms declined significantly in the overall group, but did not significantly differ between the SA and NSA groups. Treatment-refractory patients with prior substance abuse had a comparable outcome on olanzapine therapy to those with no history of abuse, as well as no increase in adverse effects. This suggests that olanzapine may be of benefit to SA patients who have a greater tendency for antipsychotic side effects and tardive dyskinesia.