We examined the ability of several baseline variables to predict treatment outcome in a pharmacotherapy trial that included 164 participants who were both cocaine- and alcohol-dependent and were selected for a randomized, double-blind, placebo-controlled study. Predictor variables included results from the baseline Addiction Severity Index (ASI), initial Urine Drug Screen results, cocaine and alcohol craving and cocaine and alcohol withdrawal symptoms at the start of treatment. Successful treatment was defined as four continuous weeks of self-reported cocaine abstinence verified by urine drug screens. In respect to demographic characteristics, there were no significant differences between patients who achieved four weeks of abstinence from cocaine and those who did not. Baseline variables that most consistently predicted cocaine abstinence included initial urine drug screen (UDS) results, the initial Cocaine Selective Severity Assessment (CSSA) scores, and initial self-reported cocaine use in past 30 days, whereas cocaine craving, cocaine composite scores, alcohol craving, alcohol withdrawal symptoms, and alcohol composite scores did not. The results of this study suggest that cocaine dependence severity in general, and initial UDS results, the CSSA scores and frequency of recent cocaine use in particular, have a significant impact on treatment outcome in the treatment of cocaine-dependent patients with comorbid alcoholism. Initial UDS results and CSSA scores are very useful predictors of treatment outcome and could be used as stratifying variables in outpatient cocaine and alcohol medication trials.