Incidence rates of aplastic anemia (AA) are rare among defined populations. Since June, 1984, a cooperative group, including 83 University medical centers throughout metropolitan France, prospectively recorded new cases of AA and followed them up. Inclusion criteria were: at least two depressed blood cell lineages (hemoglobin less than or equal to 10 g/100 mL and reticulocytes less than or equal to 50 x 10(9)/L, granulocytes less than or equal to 1.5 x 10(9)/L, platelets less than or equal to 100 x 10(9)/L) and a bone marrow biopsy compatible with the disease. Between May, 1984, and April, 1987, 292 cases were recorded. After exclusion of constitutional disease, 27 patients did not satisfy the inclusion criteria with relation to either bone marrow or blood evaluations and seven patients were initially misdiagnosed (shown in the follow-up), leaving 250 confirmed AA cases in the register. The annual incidence in France appeared to be about 1.5 per million inhabitants. The sex ratio of AA cases was similar to that of the population. In men, two peaks of incidence were observed: one between 15 and 30 years and one after 60 years. In women, the only peak was observed after 60 years. An excess of cases was observed in small towns but not in rural areas. About two of every three cases had severe AA, with a possible excess in younger cases. Based on a minimum follow-up of 1 year for 238 patients, the fatality rate was estimated at 17% at 3 months after diagnosis and at 34% at 1 year. Among 243 suspected etiologies reported by the physicians, 74% were declared idiopathic, 13% presumably associated to drug toxicity, and 5% related to hepatitis. AA appears to be a rare and severe disease in metropolitan France, often of unknown origin, a fact that emphasizes the necessity of controlled etiologic studies.