Objective: To systematically review the evidence for pharmacologic management of outwardly directed aggressive behavior in general adult psychiatry.

Data Sources: Literature searches in PubMed, EMBASE, PsycINFO, and Cochrane libraries from 1966 through March 2005 were used to identify relevant studies. The keywords aggression, violence, anger, and hostility combined with drug therapy, psychotropic drugs, adrenergic beta-antagonists, anticonvulsants, antidepressants, antipsychotic agents, benzodiazepines, and lithium were searched. Furthermore, the retrieved publications were searched for additional references.

Study Selection: All randomized controlled trials addressing pharmacotherapy for aggression or aggression-related symptoms were included, except studies addressing the "emergency situation" and studies conducted in specialized psychiatric or nonpsychiatric settings.

Data Extraction: Evidence synthesis was performed using the "best-evidence principle." Two authors independently adjudicated methodological quality and generalizability to daily clinical practice.

Data Synthesis: Thirty-five randomized controlled trials met the inclusion criteria and were evaluated. On the basis of a best-evidence synthesis model, weak evidence for antiaggressive effects of antipsychotics, antidepressants, anticonvulsants, and beta-adrenergic-blocking drugs was found. Atypical antipsychotics appeared superior to typical antipsychotics. The use of various outcome measures and insufficient data reporting in the individual studies hampered the quantitative assessment of efficacy across studies. Further limitations of the available randomized controlled trials included small sample sizes, short study duration, and poor generalizability to daily clinical practice setting.

Conclusions: Whereas pharmacotherapy is frequently applied in aggressive patients, only weak evidence of efficacy of various drug classes was found. Consensus about the use of aggression measurement scales in clinical trials is necessary for future research. Furthermore, large-scale trials with more naturalistic designs, as opposed to classical randomized controlled trials with strict inclusion and exclusion criteria, may be advisable in order to obtain results that are more generalizable to daily clinical practice.

J Clin Psychiatry 2006;67(7):1013-1024