Predictors of the vaginal microflora

Objective: Our purpose was to define influences on the patterns of the vaginal microflora. Study Design: We enrolled 617 African American and Mexican American women in a 1-year longitudinal study of sexual behaviors and the vaginal microflora on the basis of the presence of gonorrhea, chlamydial infection, trichomoniasis, or syphilis at the initial visit. The patients were assigned randomly to a behavioral intervention or standard counseling regarding sexually transmitted disease. We reevaluated 508 (82%) and 549 (89%) women at 6 and 12 months, respectively. A comprehensive survey of lower genital tract organisms was conducted at baseline and at 6 and 12 months. Behavioral and microbiologic associations were screened by bivariate analysis. All variables associated with an organism at P ≤ .15 were included in a multivariate analysis. Associations between behavior and the genital tract microflora were identified by logistic regression coefficients with P< .05. Results: African American race had a consistent association with vaginal microflora, specifically, Mycoplasma hominis, Trichomonas vaginalis, bacterial vaginosis, group B streptococci, Neisseria gonorrhoeae, and Chlamydia trachomatis. Various behaviors had a less consistent effect, including multiple partners, douching, frequency of coitus >3 times a week, and cunnilingus, fellatio, and anal intercourse at the last sexual encounter. M hominis (but not Ureaplasma urealyticum), Gardnerella vaginalis, and Lactobacillus species were associated with bacterial vaginosis. Lactobacillus species appeared to protect against bacterial vaginosis and infection with G vaginalis. Sexually transmitted diseases (caused by M hominis, N gonorrhoeae, C trachomatis, and T vaginalis) were associated with each other. In contrast, hormonal status, vaginal blood, and foreign bodies had little effect. Conclusion: The presence of other microorganisms and race have a more consistent association with the presence or absence of a cervical-vaginal organism than sexual behavior, hormonal status, vaginal devices, or the presence of abnormal vaginal bleeding. (Am J Obstet Gynecol 2001;184:845-55.)