OBJECTIVE: The authors undertook a review of the literature on the dexamethasone suppression test (DST) as a predictor of course and outcome in major depression, to illustrate clinical and pathophysiological implications of studying the relation between biological measures and course of illness in psychiatry. METHOD: In computerized searches and cross-references, 144 articles were found that related DST results to prediction of treatment response or outcome. Meta-analysis was performed on pooled data from all of the studies and separately on data from selected studies that had used stricter methodology. RESULTS: 1) Baseline DST status did not predict response to antidepressant treatment or outcome after hospital discharge. 2) Non-suppression of cortisol on the baseline DST predicted poorer response to placebo. 3) Persistent nonsuppression of cortisol on the DST after treatment was associated with high risk of early relapse and poor outcome after discharge. CONCLUSIONS: Baseline DST results may be devoid of prognostic value, but posttreatment nonsuppression of cortisol on the DST is strongly associated with poor outcome. Persistent nonsuppression may reflect a partially treated index episode or identify sicker patients who are unresponsive to treatment. The findings reiterate the importance of studying biological measures in relation to multiple validators of psychiatric illness beyond cross- sectional diagnosis.Abstract Teaser