Subjective complaints, including those traditionally considered tricyclic antidepressant side effects, were studied in 43 depressed inpatients during a three-week trial of desipramine hydrochloride. Multiple regression analysis was employed to examine the independent relationship of pretreatment symptoms, concurrent depression, and plasma drug concentrations to subjective complaints reported during treatment. As a group, subjective complaints were positively associated with pretreatment symptoms and the concurrent severity of depression, but not with plasma desipramine concentration. Of the 23 individual complaints studied, three increased during treatment and nine improved. Only two complaints, tremors and light-headedness, were significantly associated with plasma drug concentration. The data indicate that during initial treatment of severe depression with desipramine, subjective complaints are more likely to be symptoms of depression than side effects of the drug and that plasma desipramine determinations would not be useful for predicting or avoiding these complaints. The best management of most symptoms studied was adequate treatment of the depression.