The purpose of this study was to compare the conventional radioimmunoassay (RIA) to an enzyme-linked immunoassay (EIA) for the measurement of Histoplasma antigen in banked urine specimens. A correlation between the two methods would allow the EIA to be used as a nonradioactive alternative to the established 125I RIA. The study used stored urine from patients diagnosed with histoplasmosis during an outbreak in Indianapolis which began in 1988. Control specimens from healthy adults, patients with other fungal infections, urinary tract infections, or nonfungal pneumonia were also tested. Both the RIA and EIA were run concurrently. The RIA system measured antigen levels of 0.4 to 27.0 RIA units, while the EIA measured antigen levels of 0.6 to 20.1 units. Both the EIA and RIA detected measurable antigen levels in urine from 50 of 56 patients (89%) with disseminated disease and 11 of 30 patients (37%) with self-limiting disease. One of 96 control specimens, from a patient with paracoccidioidomycosis, was positive with both systems. Antigen levels measured by EIA correlated well with those measured by the established RIA method (correlation coefficient, 0.974). The EIA is an acceptable alternative to the RIA for measuring Histoplasma antigen levels in urine specimens.