OBJECTIVE The pharmacotherapy of 61 suicide victims (0.24 % of 27,078 admissions from January 1, 1980 to December 31, 1999) was compared to that of a control group matched for age, gender and diagnosis at the time of discharge. RESULTS Both groups were also comparable regarding stay in hospital, history of psychiatric disease, and frequency of hospitalisations during the year preceding the index evaluation. Multiple but not single suicide attempts were significantly more frequent in patients who were later to complete the suicide than in controls. Schizophrenia (ICD-9, ICD-10) was the most frequent diagnosis among suicide victims (44.3 %). Affective psychosis (ICD-9, ICD-10) bore the highest relative risk (0.8 %). 50 % of the schizophrenic patients in the suicide group had been continuously treated with full-dose tricyclic antidepressants. The CPZ-equivalents in the patients treated with antipsychotics were not of discriminating value. Four of 27 schizophrenic patients in the suicide group had been off neuroleptics for ten days or more; this was never observed among the controls. Lorazepam applied in 40% of the schizophrenic and in 25 % of the affective psychosis suicide victims had more often been withdrawn or reduced during the ten days preceding suicide than among controls. No schizophrenic suicide victims but five controls had been on mood stabilisers. The use of antipsychotics (classical and atypical) and a recent change in tricyclic drug or drug dose were more frequent in suicide victims with affective psychosis. Lithium had been given to one patient, but it had also been administered to six controls; this difference is significant. CONCLUSIONS Mood stabilisers, especially lithium, should be considered more often in patients with previous suicide attempt(s). When changing antidepressants in affective psychosis, benzodiazepines might be given more deliberate consideration. Patients in all diagnostic categories should be closely guided by means of intensified psychotherapeutic interventions while undergoing a benzodiazepine reduction. The treatment of patients suffering from schizophrenia with full-dose tricyclic regimens should be considered as possibly enhancing the acute suicide risk in some individuals.