The efficacy of vaginal tablets (Gynoflor) containing 50 mg of a lyophilisate of viable, H2O2-producing Lactobacillus acidophilus (at least 10(7) colony forming units/tablet) and 0.03 mg estriol (CAS 50-27-1) for the treatment of bacterial vaginosis (BV) was tested in a multicentric, randomised, placebo-controlled clinical trial with parallel-group design. 32 non-menopausal women with positive diagnoses for BV, including intermediate cases, participated in the trial. Patients were diagnosed using the classical clinical parameters of BV according to Amsel and using microscopic analysis of the Gram-stained vaginal smear. A positive clinical diagnosis of BV required at least 2 of the following 4 clinical criteria to be positive; greyish-white, homogeneous leukorrhea; vaginal pH > 4.5; KOH test for volatile amines; presence of clue cells. Microscopic diagnosis of BV, on the other hand, was obtained if examination of the Gram-stained vaginal smear showed less than 6 lactobacilli per field of view (1000 x magnification). This corresponds to another definition of BV as "lactobacilli deficiency syndrome". The efficacy of the 6-day therapy with 1-2 vaginal tablets daily was evaluated using both clinical and microscopic analysis. Using Amsel's classical clinical parameters of BV, the cure rate (defined as < or = 1 of the 4 clinical criteria positive) two weeks after the start of therapy was 77% in the verum group and 25% in the placebo group. Four weeks after the start of therapy, the cure rate was 88% in the verum group and 22% in the placebo group. At both control examinations, the cure rate for the test group was significantly higher than that for the placebo group (p < 0.05, Fisher's exact test, 2-sided, significance level 0.05). In addition, the trial showed that after 6 days of treatment with the test preparation, the lactobacilli were capable of recolonising the vagina. A significant increase in the number of lactobacilli was observed in the Gram-stained vaginal smear for the patient group treated with the test preparation compared to the placebo patient group (p < 0.05, Fisher's exact test, 2-sided, significance level 0.05), two and four weeks after the start of the 6-day treatment.