We have reviewed studies examining the efficacy of various psychotropic medications, primarily antidepressant agents, in the treatment of a group of disorders that appear to exhibit some phenomenologic and genetic relationship to major depression. These disorders all appear to benefit (albeit to varying degrees) from antidepressant medications of several different chemical families. This observation has important theoretical and clinical implications. From a theoretical perspective, these results invite the hypothesis that these various disorders may share some particular etiologic "step" in common with major depression-and that the various antidepressant classes benefit these various disorders and major depression via a common action at this hypothetical "step". Although there is an appealing parsimony to this hypothesis, several reservations must be considered. First, it must be recognized that the quality of the available studies varies widely. As noted in the text, these studies used numerous different designs, varying diagnostic criteria for the disorders under study, and diverse methods of rating outcome. Interpretation is further complicated by the fact that many studies included other concomitant medications or therapeutic interventions in addition to the psychotropic drugs administered. Also, the dose of antidepressant medications administered in many of these studies, especially those using TCAs, was often much less than that normally administered in the treatment of major depressive disorder itself. Finally, many of the studies did not systematically evaluate improvement in both the physical and psychological symptoms of a given disorder. For all of these reasons, any theoretic discussion of the results must be tentative. Nevertheless, the overall tally of results strongly favors the hypothesis that antidepressant agents, regardless of their chemical class, are generally useful in the treatment of these disorders. At a minimum, therefore, we can conclude that antidepressant treatment in these disorders deserves aggressive further investigation in studies with modern, rigorous designs. Second, even allowing that multiple antidepressant agents are effective in these various disorders, it still may be premature to conclude that these disorders are related to major depressive disorder. In particular, many of the studies found little correlation between improvement in psychological symptoms and physical symptoms of a given disorder. This observation would seem to argue against a relationship with major depressive disorder. The alternative hypothesis, however, namely, that these disorders do not share a common etiologic "step," seems even less attractive. It would be a remarkable coincidence if, say, fluoxetine possessed an antidepressant property, an independent antimigraine property, and a third, independent, antipremenstrual dysphoric disorder property. And it would be even more peculiar if various other antidepressant medications chemically unrelated to fluoxetine also, by chance alone, benefited all of these same disorders via still other independent mechanisms. Although we cannot, of course, rule out the possibility of multiple mechanisms and multiple causes, the experience of scientific research often has been that the simpler explanation of a phenomenon has proved to be correct. Therefore, the possibility of a link among these various antidepressant-responsive disorders deserves investigation. From a clinical perspective, too, these results are important. They suggest that trials of antidepressant medications should be strongly considered in patients with these disorders. Furthermore, other types of psychotropic medication appear to have a role in the treatment of individual disorders, as discussed in the corresponding sections.(ABSTRACT TRUNCATED)