Tricyclic antidepressant (TCA) medications are among the most widely used pharmacologic treatments for depression. However, a substantial number of patients fail to respond to these agents. Few standardized trials of pharmacologic treatments for TCA nonresponders are available. Bupropion has an apparently different mechanism of action than TCAs and represents a possible treatment for the TCA nonresponder. Based on positive results from pilot studies, a standardized study evaluating bupropion in TCA nonresponders was conducted. Forty-one depressed outpatients who had failed to respond to adequate documented TCA treatment, defined by specific criteria of dosage, duration, and plasma concentrations, entered a 1-week single-blind placebo phase, followed by an open 8-week bupropion treatment phase utilizing doses of up to the maximum daily dose of 450 mg. Response was measured by change on the Hamilton Depression Rating Scale (HAM-D), Clinical Global Impressions-Severity (CGI-S) and Improvement (CGI-I) Scales, and the Hamilton Anxiety Rating Scale (HAM-A). Bupropion treatment resulted in an improvement in depression on all outcome measures. Forty-nine percent were considered "responders," in that they achieved > or = 50% improvement on the 28-item HAM-D. Fifty-four percent were classified as responders based on a CGI-I rating of much or very much improved. Statistically significant improvement in depressive symptoms, measured by mean scores for the HAM-D, CGI-S, and HAM-A, was achieved at the end of the study (as compared to baseline). Twelve patients (32%) had a HAM-D score of < or = 10 at termination.(ABSTRACT TRUNCATED AT 250 WORDS)