A clinic-hospital-based leptospirosis surveillance program was conducted to determine the morbidity and risk factors in nonepidemic settings. The study was conducted on two islands, Kauai and Hawaii (Big Island), in the state of Hawaii for one year during 1988 and 1989. An active, more comprehensive case detection system was used on the Big Island that enabled us to determine the incidence of clinical disease. Subjects from both islands were used to conduct a case-control study for risk factors. One hundred seventytwo subjects from the Big Island (who presented with any two of the following symptoms: fever, headache, myalgia, or nausea/vomiting) were enrolled in the study. Twenty cases were diagnosed by culture, serology, or fluorescent antibody tissue staining at autopsy. Six cases required hospitalization and two succumbed to fatal infections. We estimated that these cases represented an annual incidence rate of 128 per 100,000 person-years in our target population. For 33 cases, 77 controls were matched for island, age, sex, and time of onset of illness. Interviews were conducted retrospectively in a double-blinded fashion with cases and controls and evaluated approximately 30 risk factors. Factors that were associated most strongly with development of leptospirosis were household use of rainwater catchment systems (P = 0.003), presence of skin cuts during the incubation period (P = 0.008), contact with cattle or the urine of cattle (P = 0.05 and P = 0.03, respectively), and handling of animal tissues (P = 0.005). Suspected risk factors that did not show an association with the disease were contact with mud or freshwater, swimming in freshwater, handling of garbage, and keeping of pets. Our study showed that leptospirosis is a surprisingly common disease, accounting for a high proportion of flu-like illness. Furthermore, a case-control study is relatively simple to conduct and useful for identifying risk factors specific to each population and setting.