OBJECTIVES: To determine the etiology of genital ulcer disease (GUD) among patients attending sexually transmitted disease (STD) clinics in Pune, India, and to examine the relationship to HIV infection and compare the clinical diagnosis of GUD with the results of a multiplex polymerase chain reaction (M-PCR) assay for Treponema pallidum, herpes simplex virus (HSV), and Hemophilus ducreyi infection. METHODS: Between June 20, 1994, and September 26, 1994, 302 patients with a genital ulcer were evaluated. Clinical etiology of GUD was based on physical appearance and microbiologic evaluations which included darkfield microscopy and serology for syphilis. Swabs of each genital ulcer were tested for HSV antigen by enzyme immunoassay (Herpchek; Dupont, Wilmington, DE) and processed in a multiplex PCR assay (M-PCR; Roche, Branchburg, NJ) for simultaneous detection of HSV, Treponema pallidum, and Hemophilus ducreyi. RESULTS: Two hundred seventy-seven men and 25 women with a median age of 25 were evaluated. The seroprevalence of HIV was 22.2%. The etiology of GUD as determined by M-PCR was HSV (26%), H. ducreyi (23%), T. pallidum (10%), and multiple infections (7%); no etiology was identified in 34%. HIV seroprevalence was higher among those patients positive for HSV compared with other etiologies (OR = 2.1, CI: 1.2-3.7; p = 0.01). When compared with M-PCR, the Herpchek test was 68.5% sensitive and 99.5% specific. Darkfield detection for T. pallidum was 39% sensitive and 82% specific, in contrast to rapid plasma reagin and fluorescent treponemal antibody absorption test, which was 66% sensitive and 90% specific. Clinical diagnosis alone or in combination with basic laboratory tests showed poor agreement with M- PCR.