AIM: To determine whether raised prolactin concentrations could be identified using postmortem venous blood and whether the level of prolactin correlated with antemortem stress. METHODS: Blood was obtained from the right femoral vein of 100 random adult necropsy cases, all of whom had been dead less than four days. Prolactin was determined in the samples by microparticle immunoassay. The levels of prolactin obtained were then analysed in relation to sex and cause of death, with particular emphasis on a history of antemortem stress and drug use. RESULTS: Prolactin in all cases of trauma was in the normal range (up to 500 mU/l). In cases of sudden unexpected deaths the mean concentration was 533 mU/l (95% confidence interval (CI), 372 to 694 mU/l). Postoperative deaths or cases with chronic disease had a mean value of 1027 mU/l (95% CI, 735 to 1319 mU/l). Cases of suicide had a mean value of 1398 mU/l. Analysis of the suicides by sex showed a significant difference, the mean in female cases being 2072 mU/l compared with 692 mU/l in male cases. In three of the four female suicides with the highest prolactin, the hyperprolactinaemia might have been attributable to a drug effect, but one case still had unexplained hyperprolactinaemia. CONCLUSIONS: It is possible to detect prolactin reliably in postmortem venous blood samples. Prolactin values at necropsy differ according to the cause of death, with markedly higher values in postoperative deaths and in the chronically ill. Hyperprolactinaemia in cases of suicide is likely to result from the effects of the drugs used, but the levels were higher than previously reported.