Medication discontinuation due to intolerable side-effects remains a significant clinical problem in the treatment of depression. We were unable to locate studies which found predictors of medication cessation due to side-effects. We posited that an identifiable subgroup of medically healthy, depressed adults who discontinued medication because of adverse events would have higher pre-treatment somatic symptoms than patients who completed a course of treatment. The sample (n =940) was drawn from a series of double-blind, placebo-controlled studies of antidepressants (imipramine, phenelelzine, L-deprenyl, mianserin and desipramine). Within the medication group, side-effect dropouts had more somatic symptoms than study completers and those who discontinued treatment for miscellaneous reasons. Within the placebo-treated group, the small number of subject who discontinued treated because of side-effects precluded valid statistical analyses, but the findings were in the same direction as those in the medication group. Clearly, further research is required to determine whether these results, obtained from a series of university-based clinical trials with healthy subjects, are generalizable to patients with significant comorbid medical and/or psychiatric disorders, treated with newer antidepressants agents in a general clinical practice setting.