The idiopathic hypereosinophilic syndrome is empirically defined as the presence of prolonged eosinophilia without identifiable underlying cause, and with evidence of end-organ dysfunction. Virtually any organ system may be involved, most frequently the heart, the central and peripheral nervous system, the lungs and the skin. We report two cases where the diagnosis of hypereosinophilic syndrome was proposed although the classic criteria were not met. In the first case total peripheral eosinophil counts were relatively low, but pathological evidence clearly showed infiltration of eosinophils in the damaged tissues. An hypothesis to explain this discrepancy is formulated. The second case did not fulfil the first feature either, although the clinical presentation and disease course corresponded well with other cases reported in the literature. The delay in diagnosis was caused by early institution of corticosteroids, clearing all evidence of eosinophil involvement in the observed tissue damage.