OBJECTIVE: The aims of this investigation were to explore the feasibility of a cognitive behavior approach to loss of clinical effect during long-term antidepressant therapy and to compare it with dose increase. METHOD: Ten patients with recurrent depression who relapsed while taking antidepressant drugs were randomly assigned to dose increase and clinical management or to cognitive behavior therapy and maintenance of the antidepressant drug at the same dose. RESULTS: Four of five patients responded to a larger dose, but all had relapsed again on that dose by the 1-year follow-up. Four of five patients responded to cognitive behavior therapy, and only one relapsed during follow-up. CONCLUSIONS: The data suggest that application of a cognitive behavior therapy approach is feasible when there is a loss of clinical effects during long-term antidepressant treatment and may carry long-term benefits. The results need to be confirmed with large-scale controlled studies.