Aripiprazole blocks psychostimulant seeking in a rat model of relapse. However, in humans, it may increase ongoing use. We tested aripiprazole specifically for relapse prevention. Methadone-maintained outpatients who were abstinent from cocaine in weeks 11-12 were randomized to double-blind aripiprazole (15 mg daily) or placebo in weeks 13-27 after 12 weeks of contingency management. Participants reported craving through ecological momentary assessment. We stopped the trial because very few (18/41) participants fulfilled the abstinence criterion. The results suggested that aripiprazole delayed lapse [hazard ratio (HR)=0.45, 95% confidence interval (CI)=0.14-1.42, P=0.17] and relapse (HR=0.31, 95% CI=0.07-1.27, P=0.10), but the effects did not reach statistical significance. Unexpectedly, the proportion of participants reporting cocaine craving was higher in the aripiprazole group (Fisher's exact P=0.026), although the frequency of craving was similar in the aripiprazole and placebo groups (1.89 vs. 1.16%, reffect=0.43, 95% CI=-0.08-0.76). The results suggest that in recently abstinent cocaine users, aripiprazole might delay relapse, but might also slightly increase craving. Difficulty in trial implementation underscores the fact that initial abstinence from cocaine is not a trivial hurdle.