Lack of association between testosterone and suicide attempts.

OBJECTIVE To determine whether testosterone levels differ in male suicide attempters versus healthy controls and to explore the associations between testosterone levels and time of blood collection, and between testosterone levels and characteristics of suicide attempts. METHOD A sample of 112 male suicide attempters was studied. Thirty-seven male blood donors were recruited as controls. RESULTS The mean testosterone levels were 5.1 ± 2.9 ng/ml in male attempters and 4.6 ± 1.6 ng/ml in controls. Group differences in testosterone levels were not significant when we studied the interaction with time of extraction (F = 0.37; d.f. = 2; p = 0.70) or when matched by age and time of extraction (t = -0.74; d.f. = 26; p = 0.47). When partial correlations were performed correcting for the effect of time of extraction, significant partial correlations were found in testosterone levels with history of aggressive behavior and lethality of the attempt. CONCLUSIONS When circadian variation and age were considered, we found no support for the putative role of testosterone as a biological marker of suicidal behavior. Further research should consider: (1) testosterone and neurosteroids; (2) serial determinations with a minimal time gap between the attempt and the blood extraction; (3) controls within the same time periods, and (4) other variables that may affect testosterone levels, such as body mass index, physical activity and sleep disturbances.