We conducted a double-blind, randomized clinical trial of mazindol (n = 37) for the prevention of relapse to cocaine abuse in methadone-maintained patients who were in the "action" stage of change, i.e., had a history of cocaine dependence but who had been abstinent for at least 2 weeks prior to entry into the study. Eight-one percent of subjects completed the 12-week course of treatment. Overall, cocaine use during the study was comparatively low-17% of the urine screens submitted were positive for cocaine metabolite. Differences between the mazindol and placebo groups of rates of relapse, number of days to relapse, and cocaine use did not reach statistical significance, but were in the direction of a treatment effect. Results suggest that stage of abstinence initiation may be a potentially useful category to employ as an independent variable in future pharmacotherapy trials for the treatment of cocaine addiction in this patient population.