Several case reports have suggested that treatment with the benzodiazepine alprazolam can result in behavioral disinhibition. To address this question, the authors reviewed the medical records (blinded to all pharmacologic treatments the patients received) of 323 psychiatric inpatients treated with alprazolam (108 patients), clonazepam (111 patients), or no benzodiazepine (104 patients) between January 1989 and June 1990. During benzodiazepine treatment, there were no significant differences among the three groups on the following measures: (1) acts of self-injury (alprazolam, 1.9%; clonazepam, 1.8%; no benzodiazepine, 2.9%); (2) assaults on staff or other patients (alprazolam, 0%; clonazepam, 0.9%; no benzodiazepine, 1.0%); (3) need for seclusion or restraints (alprazolam, 3.7%; clonazepam, 6.3%; no benzodiazepine, 4.8%); (4) increased need for observation by hospital staff (alprazolam, 8.3%; clonazepam, 7.2%; no benzodiazepine, 6.7%); and (5) decrease in patient privileges (alprazolam, 11.1%; clonazepam, 12.6%; no benzodiazepine, 11.5%). The results indicate that in an inpatient psychiatric population, the frequency of behavioral disturbances with alprazolam, clonazepam, or no benzodiazepine does not differ. This suggests that alprazolam does not possess unique disinhibitory activity. Second, these data suggest that disinhibition may not be an important clinical problem associated with benzodiazepine use. The design of the study does not allow one to establish a relationship between the prescription of the benzodiazepine and worsening behaviors, and the findings need to be interpreted conservatively because it was a retrospective review of a heterogeneous population. However, it is noteworthy that the incidence of adverse events was low even in this high-risk population, and because the patients were in the hospital and under constant observation, the objective assessment of so-called paradoxical reactions was undertaken in a controlled setting.