BACKGROUND In-vitro research has suggested that bacterial vaginosis may increase the survival of HIV-1 in the genital tract. Therefore, we investigated the association of HIV-1 infection with vaginal flora abnormalities, including bacterial vaginosis and depletion of lactobacilli, after adjustment for sexual activity and the presence of other sexually transmitted diseases (STDs). METHODS During the initial survey round of our community-based trial of STD control for HIV-1 prevention in rural Rakai District, southwestern Uganda, we selected 4718 women aged 15-59 years. They provided interview information, blood for HIV-1 and syphilis serology, urine for detection of Chlamydia trachomatis and Neisseria gonorrhoeae, and two self-administered vaginal swabs for culture of Trichomonas vaginalis and gram-stain detection of vaginal flora, classified by standardised, quantitative, morphological scoring. Scores 0-3 were normal vaginal flora (predominant lactobacilli). Higher scores suggested replacement of lactobacilli by gram-negative, anaerobic microorganisms (4-6 intermediate; 7-8 and 9-10 moderate and severe bacterial vaginosis). RESULTS HIV-1 frequency was 14.2% among women with normal vaginal flora and 26.7% among those with severe bacterial vaginosis (p < 0.0001). We found an association between bacterial vaginosis and increased HIV-1 infection among younger women, but not among women older than 40 years; the association could not be explained by differences in sexual activity or concurrent infection with other STDs. The frequency of bacterial vaginosis was similar among HIV-1-infected women with symptoms (55.0%) and without symptoms (55.7%). The adjusted odds ratio of HIV-1 infection associated with any vaginal flora abnormality (scores 4-10) was 1.52 (95% CI 1.22-1.90), for moderate bacterial vaginosis (scores 7-8) it was 1.50 (1.18-1.89), and for severe bacterial vaginosis (scores 9-10) it was 2.08 (1.48-2.94). CONCLUSIONS This cross-sectional study cannot show whether disturbed vaginal flora increases susceptibility to HIV-1 infection. Nevertheless, the increased frequency of HIV-1 associated with abnormal flora among younger women, for whom HIV-1 acquisition is likely to be recent, but not among older women, in whom HIV-1 is likely to have been acquired earlier, suggests that loss of lactobacilli or presence of bacterial vaginosis may increase susceptibility to HIV-1 acquisition. If this inference is correct, control of bacterial vaginosis could reduce HIV-1 transmission.