OBJECTIVE: To test the efficacy of treatment with clindamycin of a partner on the recurrence rate of bacterial vaginosis in women within 3 months from diagnosis. SUBJECTS: Eligible for the study were sexually active women with one current sexual partner, who were aged 18-45 years, with a clinical diagnosis of bacterial vaginosis and whose partner agreed to be treated. METHODS: A double blind, randomised, controlled trial was conducted comparing the effect of treating the partner with either clindamycin capsules or placebo on the reduction of the recurrence rate of bacterial vaginosis. Women were treated with clindamycin 2% vaginal cream, administered intravaginally once daily at bedtime for 7 consecutive days. The partners were randomly allocated to clindamycin hydrochloride capsules, 150 g by mouth four times daily for 7 consecutive days, or a placebo. A total of 139 couples were randomised--69 were treated with clindamycin vaginal cream group and 70 with placebo. One, 4, and 12 weeks after the end of treatment the patients and their partners were examined; vaginal discharges were examined to check for clue cells, vaginal pH was determined, and a KOH test carried out. RESULTS: Overall, 131 women out of the 139 who entered the study were cured (94.2%, lower 95% confidence interval 79.8, based on Poisson's approximation). There was no difference in the cure rate among women whose partner received clindamycin or placebo (chi(2) p = not significant). A total of 55 couples (26 in the clindamycin and 29 in the placebo group) withdrew from the study during the follow up period. Of the 69 women whose partner received clindamycin, 22 (31.9%) reported "recurrence" or persistence. The corresponding number was 21 (30%) of the 70 women whose partner received placebo (chi(2) p = not significant). Of the 84 couples in which the woman was cured by the first week's visit and who completed the study; there were five recurrences (11.6%) among the 43 women whose partner received clindamycin and nine (22.0%) of the 41 whose partner received placebo (chi(2) p = not significant). CONCLUSION: This study indicates that vaginal clindamycin is effective and safe in the treatment of bacterial vaginosis, but it does not support the suggestion that male treatment markedly reduces the short term recurrence rate.