OBJECTIVE AND METHOD: To compare the response of a dipstick assay (DSA) detecting Leptospira-specific immunoglobulin M (IgM) antibodies with that of an enzyme-linked immunosorbent assay (ELISA), an indirect haemagglutination assay (IHA), the microagglutination test (MAT) and a polymerase chain reaction assay (PCR) in patients with leptospirosis confirmed by MAT alone or by MAT and/or PCR (MAT/PCR). RESULT: In 75 patients with acute leptospirosis diagnosed by MAT (respectively, 90 patients diagnosed by MAT/PCR), the response in paired early and convalescent sera was positive in 78.9% (67.9%) by DSA, 76.0% (67.8%) by ELISA, 58.7% (55.6%) by IHA, 44.0% (53.3%) by PCR, and 100% (90.0%) by MAT. In early serum only, the response in patients diagnosed by MAT (respectively by MAT/PCR) was positive in 36.0% (38.9%) by DSA, 36.0% (37.8%) by ELISA, 14.7% (18.9%) by IHA, 39.2% (48.3%) by PCR, and 53.3% (58.9%) by MAT titre > or =1:100. DSA detected the main serogroups implicated in human leptospirosis in Seychelles and demonstrated sensitivity comparable to ELISA. In 124 single sera from control subjects without overt disease, the response was positive in 4.8% by DSA, 3.2% by ELISA, 3.2% by IHA, 13.8% by PCR, 37.9% by MAT titre > or =1:100, and 2.4% by MAT titre > or =1:800, giving evidence of the frequency of both past and current subclinical infection in Seychelles and that DSA was less sensitive than MAT to detect moderate levels of leptospiral antibodies. CONCLUSION: DSA is a simple and reproducible assay well adapted to field conditions and could usefully contribute to the evaluation of leptospirosis in areas devoid of serological laboratory facilities.