In view of the fact that controlled prospective studies on the benefits of dose escalation of the selective serotonin re-uptake inhibitor (SSRI) paroxetine are lacking, we conducted a double-blind, randomized, parallel-group multicentre study designed to compare the possible benefits of dose escalation of paroxetine and maprotiline in patients suffering from major or minor depression according to modified Research Diagnostic Criteria (RDC) with inadequate treatment response. The study sample consisted of 544 out-patients with different degrees of severity of depression. Patients received either 20 mg paroxetine (n = 271) or 100 mg maprotiline (n = 273) for the first 3 weeks in a double-blind manner. Response after 3 weeks was defined using explicit operationalized criteria. Patients with inadequate treatment response (paroxetine group, n = 86; maprotiline group, n = 88) were again randomized to either continuation of the previous dosage (paroxetine, n = 36; maprotiline, n = 48) or increased doses, i.e. 40 mg paroxetine (n = 50) or 150 mg maprotiline (n = 40), respectively. Intention-to-treat and completer analyses were performed. Defining response as a reduction in Hamilton Depression Rating Scale (17-item version) (HAMD-17) score of at least 50% from baseline, no significant benefits of dose escalation were found for either paroxetine or maprotiline. Stratification according to baseline severity of depression also revealed no significant benefits of dose escalation. After dose escalation, new adverse events that had not been present during treatment with lower doses rarely occurred. Our results support the view that a dose of 20 mg paroxetine is optimal for the acute treatment of depression in the majority of patients.