Knowing the negative effect of depression on lymphocyte number and activity in humans, we investigated the effect of antidepressant therapy on various lymphocyte subgroups. Cancer patients receiving treatment for at least 6 months were asked to take the antidepressant, fluvoxamine, for 28 days. Before and at the end of the study, physical and psychiatric examinations were performed, and the severity of depression was assessed by the Hamilton Scale for Depression (HAM-D). In addition, a sample of blood was withdrawn from the patients to quantify the following parameters: total leukocyte and lymphocyte counts, T4, T8, and Natural Killer (NK) cells, and lymphocyte response to the mitogens phytohemagglutinin (PHA) and pokeweed (PWM). Ten adult patients completed the study. Five of the 10 responded favorably to fluvoxamine treatment. Mean improvement was 50% from the score on day one. There was a significant correlation between the change in the HAM-D score of the "responders" and the change in NK cell counts (p = 0.02). The mean increment in NK cell number was 53%. In 4 of the 5 "non-responders", NK cell number dropped by 65% (mean). No correlation between the change in HAM-D score and any other immunological parameters was detected. Fluvoxamine increases NK cell counts in cancer patients, probably by its antidepressant effect.