Objectives

To evaluate the association of hormonal contraceptive methods with having an intermediate or bacterial vaginosis (BV) microflora pattern.

Methods

650 asymptomatic women were followed bi-monthly for up to 18 months in a randomized vaccine trial. Follow-up was available for 5597 visits over 860.8 total woman years. Demographics, contraceptive methods, antibiotic use, and sexual behavior were collected by interview. Rectal swabs were collected for culture detection of lactobacilli and tested for hydrogen peroxide (H2O2) production. Vaginal smears were evaluated by the Nugent criteria. Multinomial logistic regression was used to evaluate risk factors for having 1) intermediate vaginal flora (Nugent score 4-6) and 2) BV (Nugent score 7-10) with 3) having normal flora (Nugent score 0-3, the referent group).

Methods

650 asymptomatic women were followed bi-monthly for up to 18 months in a randomized vaccine trial. Follow-up was available for 5597 visits over 860.8 total woman years. Demographics, contraceptive methods, antibiotic use, and sexual behavior were collected by interview. Rectal swabs were collected for culture detection of lactobacilli and tested for hydrogen peroxide (H2O2) production. Vaginal smears were evaluated by the Nugent criteria. Multinomial logistic regression was used to evaluate risk factors for having 1) intermediate vaginal flora (Nugent score 4-6) and 2) BV (Nugent score 7-10) with 3) having normal flora (Nugent score 0-3, the referent group).

Results

Women had intermediate flora at 829 (14.8%) and BV at 1406 (25.1%) of follow-up visits. Women reported using hormonal contraceptive methods (HC), including oral contraceptive tablets (OC), vaginal rings, intrauterine devices containing levonorgestrel (LNG IUD), transdermal patches, and implants, at 2706 (48.3%) and depot medroxyprogesterone acetate (DMPA) at 612 (10.9%) of visits. After adjusting for vaccine arm, race, and education, women who used HC or DMPA had decreased risks of both intermediate and BV flora (see Table).

Rectal colonization by H2O2-producing lactobacilli at the prior visit was associated with a 36% reduced risk of intermediate flora and a 64% reduced risk of BV (95% CI: 0.53-0.76 and 0.27-0.41, respectively). Women using β-lactam antibiotics since the last visit were more likely to have intermediate flora (relative risk=1.92, 95% CI: 1.34-2.78) but no more likely to have BV (p=0.49) compared to women who denied β-lactam exposure. New or number of partners, frequency of vaginal intercourse, or engaging in rectal intercourse was not associated with vaginal flora patterns.