To summarize the epidemiologic, diagnostic, and clinical features of the 32 laboratory-confirmed cases of human rabies diagnosed in the United States from 1980 to 1996. Data were obtained from case reports of human rabies submitted to the Centers for Disease Control and Prevention by state or local health authorities. All cases of human rabies reported in the United States from 1980 to 1996 in which infection with rabies virus was confirmed by laboratory studies. Patients were reviewed for demographic characteristics, exposure history, rabies prophylaxis, clinical presentation, treatment, clinical course, diagnostic laboratory tests, identification of rabies virus variants, and the number of medical personnel or family members who required postexposure prophylaxis after coming in contact with an exposed person. 32 cases of human rabies were reported from 20 states. Patients ranged in age from 4 to 82 years and were predominantly male (63%). Most patients (25 of 32) had no definite history of an animal bite or other event associated with rabies virus transmission. Of the 32 cases, 17 (53%) were associated with rabies virus variants found in insectivorous bats, 12 (38%) with variants found in domestic dogs outside the United States, 2 (6%) with variants found in indigenous domestic dogs, and 1 (3%) with a variant found in indigenous skunks. Among the 7 patients with a definite exposure history, 6 cases were attributable to dog bites received in foreign countries and 1 was attributable to a bat bite received in the United States. In 12 of the 32 patients (38%), rabies was not clinically suspected and was diagnosed after death. In the remaining 20 cases (63%), the diagnosis of rabies was considered before death and samples were obtained specifically for laboratory confirmation a median of 7 days (range, 3 to 17 days) after the onset of clinical signs. Of the clinical differences between patients in whom rabies was diagnosed before death and those in whom it was diagnosed after death, the presence of hydrophobia or aerophobia was significantly associated with antemortem diagnosis (odds ratio, 11.0 [95% CI, 1.05 to 273.34]). The median number of medical personnel or familial contacts of the patients who received postexposure prophylaxis was 54 per patient (range, 4 to 179). None of the 32 patients with rabies received postexposure prophylaxis before the onset of clinical disease. In the United States, human rabies is rare but probably underdiagnosed. Rabies should be included in the differential diagnosis of any case of acute, rapidly progressing encephalitis, even if the patient does not recall being bitten by an animal. In addition to situations involving an animal bite, a scratch from an animal, or contact of mucous membranes with infectious saliva, post-exposure prophylaxis should be considered if the history indicates that a bat was physically present, even if the person is unable to reliably report contact that could have resulted in a bite. Such a situation may arise when a bat bite causes an insignificant wound or the circumstances do not allow recognition of contact, such as when a bat is found in the room of a sleeping person or near a previously unattended child.