In November 1982, a U.S. Peace Corps volunteer in Kenya completed pre-exposure rabies prophylaxis with a standard 3 dose intradermal (ID) series of human diploid cell rabies vaccine (HDCV). In May 1983, she was bitten by a dog and died of rabies 3 months later. An initial investigation revealed that the patient, as well as 9 of 11 others immunized at the same time, had no rabies antibody titers (less than 1:5). We therefore instituted investigations into the immunogenicity of pre-exposure HDCV both in the United States and in developing countries. A serosurvey revealed unexpectedly low rabies titers in both Peace Corps volunteers and others immunized in developing countries. Antibody titers measured 2-3 weeks after ID immunization were compared in 9 groups totaling 271 persons in the United States and Kenya. There was no statistically significant difference in antibody titers in the 6 U.S. groups immunized from 1980-1984 (P greater than 0.15); however, groups immunized in the United States had significantly higher titers than a group of Kenyan nationals (P less than or equal to 0.0001), and the Kenyans had significantly higher titers than 2 Peace Corps groups immunized in Kenya (P less than or equal to 0.0001). No single hypothesis proposed (laboratory error, vaccine potency, vaccination technique, or specific immune suppression) accounted for the observed differences. Although we cannot fully explain the poor response to HDCV, it is probably due to multiple factors. We conclude that persons immunized with ID pre-exposure HDCV in developing countries should have rabies antibody titers determined to ensure their seroconversion; for persons immunized in the United States, such titers need not be routinely determined.