This study evaluates whether patients with cocaine-alcohol dependence might benefit from naltrexone (NTX) pharmacotherapy when delivered in conjunction with psychotherapy. Eighty outpatients meeting DSM-IV criteria for alcohol and cocaine dependence were randomly assigned to receive NTX (placebo or 50 mg/d) combined with psychotherapy (Relapse Prevention [RP] or Drug Counseling [DC]) for twelve weeks. It was hypothesised that the skills training focus of RP therapy, in combination with NTX 50 mg/d, would produce greater reductions in cocaine and alcohol use. Outcome measures included self- and objective reports of substance use, treatment retention, medication compliance, and adverse effects. During the first four weeks of treatment, the percentage of cocaine-positive urine screens was significantly lower for those receiving RP therapy (22%) than those receiving DC (47%); however, this difference subsequently diminished. No medication effects were found. All groups reported less alcohol use at the end of treatment. Treatment retention was the same among the groups, with about 33% of the subjects completing all twelve weeks of treatment. The active medication group showed better medication compliance, while the number of adverse events was low overall and not significantly different by group. In conclusion, NTX at 50 mg/d did not reduce cocaine or alcohol use. These findings stand in contrast to previously reported positive findings for NTX and RP in patients with a single diagnosis of cocaine dependence.