Risk of Bacterial Vaginosis Is Elevated for Women Who Douche, Wheter or Not They Have Had Symptoms

D. Hollander

First published online:

Demographic, behavioral and hormonal factors all appear to play a role in the occurrence of bacterial vaginosis, a condition caused by an overabundance of certain types of bacteria normally present in the vagina.1 In a clinic-based study conducted in Michigan, black women with no more than 13 years of education and women who douched had significantly elevated odds of the condition, while those who had recently used hormonal contraceptives had reduced odds. Douching was associated with both symptomatic and asymptomatic disease, and the association was equally strong whether women douched because they had symptoms or for other reasons.

The analyses are based on 298 women receiving services at a county health department clinic in 1998. Women were eligible to participate if they were scheduled to have a vaginal examination. After providing basic information about their demographic characteristics and reasons for visiting the clinic, participants completed a detailed, self-administered questionnaire that covered vaginal symptoms and gynecologic, reproductive and lifestyle factors that may affect the odds of bacterial vaginosis. Vaginal smears taken during their examination were assessed for bacterial vaginosis.

Women attending the public clinic were predominantly 20 or older (89%) and unmarried (77%). The majority had more than 13 years of schooling (57%), and about half said that at least one of their parents had gone beyond high school (53%). Sixty-five percent were white, 25% black and 11% members of other racial or ethnic groups. One in five had Medicaid coverage.

Overall, 30% of participants had bacterial vaginosis. Results of chi-square testing indicated that women with 13 or fewer years of schooling had significantly higher prevalence levels (38-48%) than those with more education (23%), and Medicaid recipients had a higher prevalence (42%) than those without such coverage (28%). Black women's prevalence was almost twice that of whites (42% vs. 25%). The racial difference was more pronounced among women with 13 or fewer years of education than among those with more schooling: At the lower educational level, 59% of black women and 35% of white women had bacterial vaginosis, while at the higher level, the proportions were 28% and 17%, respectively.

The prevalence of bacterial vaginosis was also related to several gynecologic and reproductive factors. It increased significantly as a woman's parity rose--from 19% among those who had never had a live birth to 56% among those who had had more than three--and was lower among women who had used hormonal contraceptives within the past six months (24%) than among nonusers of these methods (41%). Two-thirds of the sample had ever douched, and these women had a higher prevalence than women who had never done so (36% vs. 20%); prevalence was also significantly elevated among women who had first douched in their teens, those who had douched at least monthly within the past year and those who had douched within the past two months.

Finally, the prevalence of infection declined from 59% to 28% as the frequency with which a woman showered increased from zero to four or more times a week, but it rose from 24% to 50% as the frequency of bathing increased over that same range. It was lower among women who took vitamins or nutritional supplements (21%) than among those who did not (34%).

Results of logistic regression analysis revealed that two behavioral factors had independent effects on the odds of bacterial vaginosis. When all factors that were significantly related to prevalence or that were considered clinically relevant were controlled for, hormonal contraceptive use continued to have a protective effect (odds ratio, 0.5), and having douched within the past two months remained a risk factor (2.9). Of the background factors examined, only the interaction between race and education was significantly associated with the odds of bacterial vaginosis: Compared with white women who had more than 13 years of schooling, less-educated black women had significantly and sharply higher odds (odds ratio, 5.5).

Because of the elevated prevalence of douching among women with bacterial vaginosis, the investigators explored the association between douching and bacterial vaginosis among women who reported vaginal symptoms (discharge, change in color of discharge, odor or itching, and odor after intercourse) and those who did not. They found that the prevalence of douching was similar in these groups (29% and 23%, respectively), and in both groups, women who douched had elevated odds of bacterial vaginosis (odds ratios, 5.8 and 2.6, respectively). Furthermore, whether women douched because they had symptoms or for other reasons, their odds of bacterial vaginosis were about three times those of women who did not douche (odds ratios, 3.1-3.4).

At one time, the authors observe, the motivation for treating bacterial vaginosis was to eliminate the unpleasant odor and discharge that it often causes. However, a growing body of evidence has linked the condition to numerous adverse health outcomes, and some research has suggested that it facilitates the transmission of HIV. They conclude, therefore, that "the relatively high prevalence of regular vaginal douching and the modifiable nature of this behavior argue for more comprehensive studies on (1) the impact of vaginal douching...and (2) the personal and culture-based motivations for vaginal douching." They also stress the importance of exploring the "natural history" of bacterial vaginosis and understanding the interactions of all potential factors.--D. Hollander


1. Holzman C et al., Factors linked to bacterial vaginosis in nonpregnant women, American Journal of Public Health, 2001, 91(10):1664-1670.