The failure of subjects to complete clinical trials is a common problem with important implications for the interpretation of study results. Although a substantial literature exists on the high prevalence of premature termination from psychiatric and substance abuse treatment setting, there has been little attention paid to early discontinuation in clinical trials. There is evidence that the presence of substance abuse predicts higher rates of early discontinuation. This, combined with a recent increase in efforts to develop medications for treatment of substance use disorders, led us to conduct a literature review to determine whether pharmacotherapeutic trials for patients with these disorders have higher rates of premature discontinuation than comparable studies of patients with other psychiatric disorders. Of 267 articles that were initially identified, 83 met predetermined criteria for inclusion in the analysis. As hypothesized, after controlling for a number of potential contributing variables, treatment trials with substance abuse patients showed a significantly poorer retention rate than those of patients with other psychiatric diagnoses. The difference in retention rate was also evident when studies specific to alcohol dependence, the largest subgroup of substance use disorders, were evaluated separately. Although the retrospective nature of the study design limits the conclusions that can be drawn, the results suggest that, in pharmacotherapy trials with alcoholics or other substance abuse patients, particular attention should be paid to enhancing treatment retention.