The incidence of cancer was studied in a cohort of all first admitted 9156 patients in Denmark with a diagnosis of schizophrenia in the period 1970-1987. The overall incidence of cancer was reduced particularly in the males. Adjustment for the smoking habits of psychiatric patients enhanced this risk reduction. Fewer than expected had been diagnosed with cancer prior to the first schizophrenia admission. This tendency was limited to the female patients. The reduced cancer incidence was particularly marked for genital cancers, in particular testicular cancer, and skin cancers including malignant melanoma. Breast cancer risk was not increased, thus not substantiating concerns that neuroleptics would increase breast cancer risk through the elevation of serum prolactin levels. Some available evidence in the literature supports the hypothesis of an antineoplastic effect of neuroleptics as an explanation for the low occurrence of cancer in schizophrenic patients. Further large sample studies including an extension of the follow-up of this cohort are needed to establish the reduced risk of cancer in schizophrenic patients as well as exploring the causes for this reduction.