Drug-induced agranulocytosis is relatively rare. It is a heterogeneous disorder in pathogenetic terms, not surprisingly in view of its idosyncratic nature. Indeed, one drug might cause agranulocytosis by different mechanisms in different patients. Recent investigations suggest that there are at least three mechanisms by which it can be produced, namely differences in drug pharmacokinetics, abnormal sensitivity of myeloid precursors, and adverse immune responses to drug administration. Genetic factors are important and could act via any of the above. Examples have been reported where one drug has been shown to cause damage at different levels of neutrophil development in one patient, and others where the drug acted at different sites in different patients. Any drug is potentially capable of toxicity and should be viewed with suspicion in the appropriate context. Drug-induced agranulocytosis is usually a self-limiting condition (provided toxic drugs are withdrawn) with complete resolution within two weeks. However, the mortality rate during the acute phase is high, and therefore prompt supportive therapy with isolation and broad-spectrum antibiotics for infection are mandatory during periods of severe neutropenia. The place of granulocyte transfusions in this disorder has not yet been established, and needs to be decided in each individual case.