Convergent evidence indicates that HPA-axis hyperactivity is a risk factor for suicide in major depressive disorder, and seven independent reports have shown that patients with abnormal dexamethasone suppression test (DST) results have significantly higher rates of eventual suicide. The identification of interactions between DST results and other clinical predictors would enhance risk assessment, but modest sample sizes have limited such analyses in earlier cohorts. Subjects with major depressive disorder who participated in research protocols at the University of Michigan between 1980 and 1991, who had fully structured diagnostic interviews, and who underwent a 1-mg DST while actively depressed were screened with the National Death Index for a mean (S.D.) follow-up period of 18.0 (5.5) years. Of 334 subjects, 69 (20.7%) were identified as having died. Of these, 13 (18.8%) had died by suicide and 32 (46.4%) from cardiovascular causes. Baseline DST results did not significantly predict death from suicide or from cardiovascular disease for the sample as a whole. Significant relationships between DST results and later suicide did exist for inpatients, for patients with manifest suicidality and, in particular, for inpatients with manifest suicidality. Because nearly all previous reports of DST results and suicide described depressed inpatients, it is possible that the DST is a useful predictor only within this population.