We hypothesized that patients who had experienced clozapine-induced agranulocytosis would have abnormalities in their free radical scavenging enzyme activity (FRESA) and levels of related trace metals. We therefore measured FRESA profiles and related trace metals in four groups: post-clozapine agranulocytosis (POST CLOZ AGRAN) (N = 9); clozapine no agranulocytosis (CLOZ NO AGRAN) (N = 12); West Coast controls (WC CONTROLS) (N = 14); and Long Island Jewish Medical Center controls (LIJ CONTROLS) (N = 12). Glutathione peroxidase (GSH-Px, P1) levels in plasma were slowest in the POST CLOZ AGRAN group (34.3 +/- 6.9 IU/dl [standard deviation; SD]; p < 0.002); red blood cell glutathione peroxidase (GSH-Px, RBC) was highest in the WC CONTROLS (38.7 +/- 4.7 IU/g hemoglobin [Hgb]; p < 0.008); and selenium (SE) levels in plasma were lower in both the POST CLOZ AGRAN group (111.6 +/- 14.7 ng/ml) and the CLOZ NO AGRAN group(115.0 +/- 17.8) than in the WC CONTROLS (142.5 +/- 18.3; p < 0.0006). SE was also lower in the POST CLOZ AGRAN group than in the LIJ CONTROLS (129.1 +/- 21.6; p < 0.04). The presence of at least one of the following: (1) GSH-Px, P1 < 37.6 IU/dl; (2) GSH-Px, RBC < 31.0 IU/g Hgb; or (3) SE < 112.4 ng/ml, distinguished POST CLOZ AGRAN subjects from the WC CONTROLS, but not from the LIJ CONTROLS. Data from this cross-sectional pilot study suggest that abnormalities in the body's antioxidant defense system may be involved in the pathogenesis of clozapine-associated agranulocytosis. If confirmed in large-scale, prospective studies, these preliminary findings have potential clinical application in the screening and prophylaxis of clozapine agranulocytosis.