This prospective observational study aims to advance our knowledge of the natural history and abnormal alteration in vaginal microbial flora, and to identify risk factors that are associated with these changes. A total of 3600 healthy, non-pregnant women of reproductive age have been enrolled and were followed in Birmingham, Alabama for one year (baseline visit plus once every three months afterwards, totaling five visits). Detailed information on sociodemographic status, medical history, hygiene practice, sexual behavior and psychosocial stress was collected through in-person interviews. Routine gynecological and dental examinations were carried out during the study visits. Samples of vaginal secretion and blood were collected. Laboratory tests to quantify vaginal microbial flora and to identify infection of bacterial vaginosis as well as common sexually transmitted diseases were conducted. Sub-studies evaluate the day-to-day change in vaginal flora, and the effect of cessation of douching.? ? The final study visits were completed in June 2003, and various analyses are underway. In results published or accepted for publication during this reporting period, we reported on nutritional factors associated with BV prevalence. We found that total fat intake was positively associated with BV- the odds ratio for the highest quartile of fat intake was 1.5. Saturated fat intake was associated with severe BV, with an odds ratio >2 for the highest quartile. Total energy intake was only weakly associated with BV. Folate, vitamin E and calcium intake were inversely associated with BV.? ? A second publication reported on a case-crossover analysis of factors that prompt a change in douching behavior. In this design, womens characteristics in intervals when they did not douche are compared to characteristics during intervals when she douched. The case-crossover design eliminates confounding by factors that differ between women, since each woman serves as her own control. We found that women were more likely to douche during intervals in which they noticed a fishy vaginal odor, or in which they reported vaginal irritation. Douching was more common during the summer months and during times when a woman had more sex partners. Treatment for BV or trichomoniasis, as well as absence of menstrual periods, were associated with reduced douching. These results provide some evidence that douching may be a result, not a cause of BV. Interestingly, nearly all the estimates from the within-woman analysis were further from the null than those of the conventional (between woman) analysis.? ? A third publication evaluated the longitudinal and cross-sectional association between BV and contraceptive use. BV was found to be less prevalent among women who used hormonal contraception. This was the case for both oral and injectable contraceptives. Women who had tubal ligation were at increased risk of BV. BV incidence was also increased among with with tubal ligations, while BV remission was increased among women using hormonal contraceptives.? ? A fourth publication evaluated factors associated with recurrent BV among adolescents attending a sexually transmitted disease clinic.? ? A fifth publication used marginal structural models to evaluate the association between douching and bacterial vaginosis. These models are a state-of-the-art way to address the bias that can occur when an intermediate variable in the causal pathway is also a confounder. In this case, vaginal symptoms in the past might cause a woman to douche, which then causes bacterial vaginosis, which itself then causes vaginal symptoms. Marginal structural models employ inverse probability weighting to control for past symptoms without over-controlling for past bacterial vaginosis.? ? A sixth publication updated the US Preventive Services Task Force recommendations regarding screening asymptomatic women for BV. It was concluded that screening women without a history of preterm birth in a past pregnancy is not indicated, but the evidence regarding screening women who have such a history is too conflicting to make a recommendation.
Showing the most recent 10 out of 26 publications
