Clozapine is associated with a high incidence of agranulocytosis. We had previously found that it is oxidized by granulocytes, or simply HOCl, to a reactive metabolite that irreversibly binds to the cells, and we proposed that this reactive metabolite is responsible for clozapine-induced agranulocytosis. The reactive metabolite appeared to be a nitrenium ion formed by chlorination of the nitrogen bridge between the two aromatic rings. If this is correct, analogs that contain this structural feature should also be oxidized to a reactive intermediate while those not possessing this feature would, at least, not form the same type of reactive intermediate and, therefore, may not induce agranulocytosis. We tested the first part of this hypothesis with three clozapine analogs that do contain a nitrogen bridge and three that do not. Consistent with the hypothesis, the three analogs that do contain the nitrogen bridge formed reactive intermediates that could be trapped with glutathione when oxidized by HOCl, myeloperoxidase or activated neutrophils. In contrast, we found no evidence of a reactive intermediate on oxidation of analogs that contained an oxygen or sulfur bridge rather than a nitrogen bridge. If such reactive metabolites are responsible for drug-induced agranulocytosis, it should be possible to use such a simple screening method to test drugs at an early stage in their development for the potential to induce agranulocytosis. (C) 1997 Elsevier Science Ireland Ltd.