We reviewed 20 English-language weight-reduction studies, reported between 1967 and March 1993, of the effect of > or = 6 mo of pharmacologic therapy on weight loss and its maintenance to determine the clinical benefits of extended treatment, propose treatment guidelines, and identify future research needs. Pharmacologic agents included phentermine, mazindol, fenfluramine, dexfenfluramine, and fluoxetine. Study designs varied with respect to blinding, use of a single agent vs a combination, dosing, length of therapy, patient selection, adjunctive therapy, and visit frequency. At endpoint, weight loss varied from study to study but a plateauing of weight loss or weight regain was observed after approximately 6 mo. The benefits of extended treatment appear to outweigh the risks for those patients who are unable to lose sufficient weight without pharmacologic therapy but who maintain adequate weight loss with long-term pharmacologic therapy. Future studies should define and evaluate pharmacologically responsive and unresponsive subgroups.