To evaluate the effects of maintenance treatment and durability of gains after treatment discontinuation, responders to either phenelzine (PZ) or cognitive-behavioral group therapy (CBGT) from an acute trial comparing these two treatments as well as pill placebo and a psychotherapy control (educational supportive group therapy) were enrolled into maintenance and treatment-free follow-up phases. Experimental design: Responders to an acute trial contrasting PZ and CBGT entered a six-month maintenance phase. Patients who continued to respond through the maintenance phase entered a six-month treatment free phase. Patients receiving pill placebo or educational supportive group therapy in the acute trial did not enter the long term study. Principal observations: PZ patients entered maintenance more improved than CBGT patients, and nonrelapsing PZ patients maintained their superior gains throughout the study. Relapse during maintenance did not differ between treatments. However, PZ patients showed a trend toward greater relapse during treatment-free follow-up. There was a greater relapse among patients with generalized social phobia with phenelzine. Conclusions: PZ and cognitive-behavioral group therapy may differ in their long term effects. The superiority seen with PZ on some measures in the acute study persisted in patients who maintained their gains over the course of maintenance and treatment-free follow-up. However, CBGT may lead to a greater likelihood of maintaining response after treatment has terminated. Replication with larger samples is needed, as is a study of the acute and long-term efficacy of combined PZ and CBGT. Depression and Anxiety 10:89–98, 1999. © 1999 Wiley-Liss, Inc.