A hypothesis-generation study was conducted to examine risk factors for bacterial vaginosis recurrences among women participating in a randomized controlled trial of male sexual partner treatment. Of the 140 women enrolled in the trial, 72 had a normal vaginal gram-stained slide at the first follow-up visit and were included in this analysis. Nineteen of the 72 (26.2%) women developed a bacterial vaginosis recurrence documented by a vaginal gram-stained slide during the 6-week follow-up period. Of the 19 women, 4 of 6 sponge users (66%) developed a bacterial vaginosis recurrence (RR 2.93, 95% CI: 1.43-6.02). Logistic regression analysis supported the association between sponge use and bacterial vaginosis recurrences (adjusted RR 2.71, 95% CI: 1.06-6.99) but revealed that an apparent protective effect of diaphragm use was due to the confounding effect of sexual partner treatment. Our hypothesis is that bacterial vaginosis recurrences may be due to factors other than sexual transmission, such as sponge use. Further studies are needed to examine risk factors for bacterial vaginosis recurrences. In the interim, physicians might suggest another form of contraception to their sponge-using patient with frequent bacterial vaginosis infections.