Objective. To estimate rates of lower extremity amputations (LEAs) in persons with peripheral vascular disease, diabetes mellitus, trauma, neoplasm, osteomyelitis, or emphysematous gangrene. Methods. Regional amputee registries were used to estimate the rate of lower extremity amputations with the capture-recapture (CR) technique. Data were extracted from three amputee registries in Rio de Janeiro: source 1, with 1 191 cases from 23 hospitals; source 2, with 157 cases from a limb-fitting center; and source 3, with 34 cases from a rehabilitation center. Amputee death certificates from source 1 identified 257 deaths from 1992 to 1994. Three CR models were evaluated using sources 2 and 3. In order to avoid an overestimation of the rate of LEAs, two models were applied for the data analysis: in one case, deceased patients listed in source 1 were excluded from the model, and in the other case, deceased patients were included as well. Results. Excluding the 257 deaths, the estimated number of amputations in the municipality of Rio de Janeiro from 1992 to 1994 was 3 954, for a mean annual incidence rate of 13.9 per 100 000 inhabitants. Among persons with diabetes, the annual incidence rate of lower extremity amputations was substantially higher (180.6 per 100 000 persons per year), representing 13 times the risk of individuals without diabetes. The yearly rate of LEAs according to the routine surveillance system was estimated at 5.4 and 96.9 per 100 000 in the general population and in diabetics, respectively. If data from the three registries are added, 1 382 patients with LEAs were identified, with the reasons for the amputations distributed as follows: peripheral vascular disease = 804 (58.1%); diabetes mellitus = 379 (27.4%); trauma = 103 (7.4%); osteomyelitis = 44 (3.1%); gangrene = 36 (2.6%), and neoplasm = 16 (1.1%). Conclusions. These findings show a high incidence of LEAs in Brazil, when compared to countries such as Spain, that is attributable mainly to peripheral vascular disease and diabetes mellitus.