OBJECTIVES: Bacterial vaginosis (BV) increases women's susceptibility to sexually transmitted infections (STIs) and HIV and may partly explain the high incidence of STI/HIV among girls and young women in East and southern Africa. The objectives of this study were to investigate the association between BV and sexual debut, to investigate other potential risk factors of BV and to estimate associations between BV and STIs. METHODS: Secondary school girls in Mwanza, aged 17 and 18 years, were invited to join a cross-sectional study. Consenting participants were interviewed and samples were obtained for STI and BV testing. Factors associated with prevalent BV were analysed using multivariable logistic regression. Y-chromosome was tested as a biomarker for unprotected penile-vaginal sex. RESULTS: Of the 386 girls who were enrolled, 163 (42%) reported having ever had penile-vaginal sex. Ninety-five (25%) girls had BV. The prevalence of BV was 33% and 19% among girls who reported or did not report having ever had penile-vaginal sex, respectively. BV was weakly associated with having ever had one sex partner (adjusted odds ratio (aOR) 1.59;95% CI 0.93 to 2.71) and strongly associated with two or more partners (aOR = 3.67; 95% CI 1.75 to 7.72), receptive oral sex (aOR 6.38; 95% CI 1.22 to 33.4) and having prevalent human papillomavirus infection (aOR = 1.73; 95% CI 1.02 to 2.95). Of the 223 girls who reported no penile-vaginal sex, 12 (5%) tested positive for an STI and 7 (3%) tested positive for Y-chromosome. Reclassifying these positive participants as having ever had sex did not change the key results. CONCLUSIONS: Tanzanian girls attending school had a high prevalence of BV. Increasing number of sex partner was associated with BV; however, 19% of girls who reported no penile-vaginal sex had BV. This suggests that penile-vaginal sexual exposure may not be a prerequisite for BV. There was evidence of under-reporting of sexual debut.