To the Editor: The observation noted by Dr. Benazzi is indeed puzzling: Why should the beneficial effect demonstrated in the group receiving continuation fluoxetine plus olanzapine not be seen in the group receiving olanzapine and placebo? The rationale is that since fluoxetine had been recently discontinued in the latter group, sustained plasma levels of fluoxetine and norfluoxetine would have been expected because of the long half-life of the two compounds. The point made by Dr. Benazzi that the differences could have occurred by chance clearly is a possibility, and we agree strongly that a larger clinical trial is needed to establish any treatment effect. However, are there other possible explanations? After discontinuation, the plasma levels of fluoxetine and norfluoxetine would have begun to decline progressively and might have fallen below the critical threshold required to produce the effect. However, a transient effect in the olanzapine group is suggested by the response seen in the group receiving olanzapine and placebo in the third week of treatment (Figure 1). Or the combination effect could be a result of a heretofore unknown mechanism of action. Clearly, more research is needed to test the effectiveness of this combination and any possible pharmacological effects of treatment.