Dysregulation of the hypothalamic–pituitary–adrenal (HPA) axis function is associated with suicidal behavior. In suicide attempters with mood disorder, the non-suppressor status in the dexamethasone suppression test (DST) is associated with suicide indicating that HPA-axis hyperactivity is a biological risk factor for suicide and may be a useful predictor. The threshold of 5 μg/dl for cortisol levels measured at 08:00 a.m. or 4:00 p.m. following dexamethasone at 11:00 p.m. to define the DTS nonsuppression was derived as being optimal for the separation of melancholia and nonmelancholic conditions rather than the prediction of suicide. A different threshold may offer a better identification of suicide. The aim of this study was to find the optimal threshold level of post DST plasma cortisol at 4 p.m. for suicide prediction using receiver operating characteristics (ROC) analysis. A cohort of 106 depressed inpatients with an index suicide attempt admitted to the department of Psychiatry at the Karolinska University Hospital between 1980 and 2000, were submitted to DST and followed up for causes of death. During the follow-up (mean 17 years), 25 suicides (24%) were identified. The ROC analysis revealed that a lower threshold of 3.3 μg/dl for the nonsuppressor status predicted 17 of 25 suicides (sensitivity of 68%) compared with 15 of 25 suicides (sensitivity 60%) with a conventional threshold of 5 μg/dl at 4:00 p.m. In male suicide attempters the lower threshold for pathological DST result (3.3 μg/dl) changed the Odds ratio from 6.7 till 18. In female suicide attempters a higher threshold (7.3 μg/dl) optimised the value of DST as a biological test for suicide prediction indicating a gender difference.