Background: Lithium augmentation is a first-line strategy for depressed patients resistant to antidepressive therapy, but little is known about patients' subsequent long-term course or outcome predictors. We investigated long-term outcomes of unipolar depressed patients who had participated in a study on the effects of lithium augmentation on the hypothalamic-pituitary- adrenocortical system using the combined dexamethasone/corticotrophin-releasing hormone (DEX/CRH) test. Methods: Twelve to 28 months (mean 18.6 ± 4.6 months) after lithium augmentation, 23 patients were assessed with a standardized interview, of which 18 patients had complete DEX/CRH test results. Relapse was diagnosed by DSM-IV criteria (Structured Clinical Interview for DSM-IV; SCID I). Results: Only 11 patients (48%) had a favorable follow-up, defined as absence of major depressive episodes during the observation period. Patients with a favorable and an unfavorable course did not differ in clinical or sociodemographic parameters, endocrinological results or continuation of lithium. However, fewer previous depressive episodes tended to correlate (p = 0.09) with a favorable course. Conclusion: Results from studies using the DEX/CRH test to predict relapse in depressed patients treated with antidepressants were not replicated for lithium augmentation. Our finding could reflect the elevation of DEX/CRH results by lithium, independent of clinical course. Limitations of the study are its small sample size, the heterogeneous clinical baseline conditions and the lack of lithium serum levels. The fact that lithium continuation did not predict the course might be related to the difference between the efficacy of lithium in controlled studies and its effectiveness in naturalistic settings.