Leptospirosis has rarely been reported in Puerto Rico, although in the period from 1948 to 1952, 208 cases of leptospirosis and an island-wide seroprevalence of antibody to Leptospira of 14% were documented. In Puerto Rico in October 1996, following rainfall and a period of flooding generated by Hurricane Hortense, serum specimens of 4 patients with suspected dengue fever that were negative for dengue tested positive for Leptospira-specific IgM antibodies in a dipstick assay. Subsequently, we used an island-wide dengue laboratory-based surveillance system to determine the increase in leptospirosis after hurricane-generated floods. All anti-dengue IgM-negative patients (n = 142) with disease onset from August 8 to October 6, 1996 from prehurricane and posthurricane groups were investigated for leptospirosis. Laboratory-confirmed leptospirosis cases were defined as microscopic agglutination test titers > or = 1 :400 to 1 or more serovars, or positive immunohistochemistry in autopsy tissues. Four (6%) of 72 prehurricane and 17 (24%) of 70 posthurricane patients had laboratory-confirmed cases of leptospirosis (relative risk [RR] = 4.4, 95% confidence interval [CI] = 1.6-12.4). The mean age of case-patients was 34 years (range = 13-64). Eighteen (86%) of 21 confirmed case-patients were males, including one patient who died (31 years old). Patients were located in 18 (38%) of 48 municipalities that submitted serum samples. Clinical features significantly associated with leptospirosis were eye pain (RR = 1.5, 95% CI = 1.3-1.9), joint pain (RR = 1.4, 95% CI = 1.1-1.6), diarrhea (RR = 1.7, 95% CI = 1.2-2.5), and jaundice (RR = 3.3, 95% CI = 1.5-7.2). This study demonstrates the utility of a dengue laboratory-based surveillance system for the detection of an increase of leptospirosis, which most likely would have gone unrecognized. Leptospirosis is treatable with antibacterial agents; knowledge of this diagnosis may significantly reduce morbidity and mortality.