Medication was discontinued under a placebo-controlled, double-blind, six-month protocol with 17 chronically depressed patients who had been taking an average daily dose of 138 mg amitriptyline (AMI) for an average of 3.7 years. Only one of nine patients became depressed on active medication, while of the 15 patients receiving a placebo trial, 11 had a depressive recurrence at an average time of 9.3 weeks. These 11 were subsequently restarted on AMI, and responded similarly to the way in which acutely depressed patients respond, although the patients showed either a need for less AMI or decreased symptoms, compared to entry. Tolerance did not develop to anticholinergic side-effects during long-term medication. Twelve of the 15 patients on placebo showed a withdrawal reaction during the first few weeks of tapered AMI discontinuation which could be distinguished from recurrence of depression. This study suggests that the majority of patients on long-term antidepressant will suffer a recurrence of depressive symptoms when the medication is discontinued.