Answers are grounded in this paper and cite their sources.
Bacterial vaginosis (BV) is an important cause of vaginal symptoms among women in the United States. BV is likely an important contributor to preterm delivery1–3 as well as to susceptibility to sexually transmitted infection (STI).4–10 Estimates from the National Health and Nutrition Survey indicate that as many as 1 in 3 women of reproductive age are affected at any given time.11 Strong racial disparities in the occurrence of BV have been documented; black and Mexican American women have significantly higher odds of BV with prevalence rates of 51.6% and 32.1%, respectively, than white, non-Hispanic women (23.2%). These disparities persist after adjusting for age, sociodemographic characteristics, sexual history, and recent douching, common risk factors for BV. In this issue of Sexually Transmitted Disease, Klebanoff and colleagues report on data from the Longitudinal Study of Vaginal Flora, a prospective study of nonpregnant women between the ages of 15 and 44 years who were followed quarterly for 1 year to identify new occurrences of BV. Acquisition of BV was defined as a change from BV negative (Nugent score <7) to BV-positive (Nugent score ≥7) in consecutive study visits with reliability of Nugent scoring verified by an independent laboratory. Klebanoff et al examine 2 important questions in the field: (1) can bacterial vaginosis be considered aSTI? and (2) how does race of male sexual partners influence BV risk? First, is BV a sexually transmitted infection? The authors present a concise overview of the arguments in support and opposition and suggest that it may be sexually transmitted. The risk profile of BV strongly resembles that of other STIs. BV is associated with increased risk for both viral12–14 and nonviral STIs4,9 in both cross-sectional and prospective studies. BV is more common among those with a new sexual partner, frequent intercourse and/or more recent or lifetime sexual partners.4,5,15,16 And studies in monogamous lesbian sexual partners have found concordance between BV status and the prevalence of BV-associated bacteria.17,18 Similarly, consistent condom use likely decreases the risk of incident BV,16,19–21 but as the authors point out, findings are mixed and include an earlier report from the same cohort which found no association between condom use and BV incidence, prevalence or remission.22 The current study provides further support for sexual transmission. Specifically, Klebanoff et al reported that BV was present in 12.8% of sexually active cycles compared with 5.1% cycles in which no sexual activity was reported and that male partner demographic characteristics and condom use significantly impact the risk of BV acquisition. It may be useful to borrow a few concepts from the field of epidemiology to evaluate the role of sexual transmission in BV acquisition. Rothman supplies a framework that links “metaphysical notions of cause with basic epidemiologic parameters.”23 Cumulative disease risk is comprised of the sum of population attributable risks (or etiologic fractions) for the constellation of pathways that lead to the development of a particular disease. Causes may be necessary, that is, present in all pathways that lead to disease, therefore without this cause there is no disease. But they may also be sufficient-the presence of a causal factor produces disease, but this factor is not present in all possible pathways. Sexual transmission may be a sufficient, even predominant cause of BV, yet may not be a necessary cause. Among women reporting no history of sexual intercourse in the National Health and Nutrition Survey sample, almost 15% had BV.11 Absence of a history of sexual intercourse was correlated with young age, suggesting that misclassification of sexual history alone did not explain the findings. Additional evidence comes from a study of women entering the military, which found that 18% of self-identified sexually inexperienced women were positive for BV. While misclassification of sexual activity-whether due to inaccurate self-report, variation in interpretation of intercourse, nonconsensual contact, etc.-may partially explain some of this association, these studies suggest some portion of transmission could be due to the existence of other pathways, such as colonization from the intestinal tract. Future research focusing on estimation of population attributable risks will enable us to quantify the magnitude of the effect of sexual transmission on incident BV. The second question Klebanoff et al examined was the impact of male partner race on BV acquisition. Male partner race was not associated with recent sexual activity (number of partners, frequency of intercourse, receptive oral sex, or anal sex), having a new sexual partner, douching or condom use in the last month, yet, it was associated with BV incidence. The probability of acquiring BV among white women who had black partners was more than twice that of white women with white male partners (RRadj = 2.2; 95% CI, 1.5-3.4). The observed effect of partner race was modified by condom use. Among those using condoms consistently there was no increase in BV incidence in women with black partners. Only among women not using condoms or using condoms inconsistently was there an association with male partner race. Understanding the meaning of this finding, that black male partners are associated with higher incidence of BV, is challenging. Does this mean that there are factors of the male partners that increase BV risk? Conversely, does it mean that the male partners are coincidental, and are the true factors influencing BV risk instead associated with the woman and sexual partner selection? There has been active debate in public health literature on the varied and complex meanings of race and the best approach to quantify race-related effects.24,25 Klebanoff et al correctly state that race is a social construct26 and as such may represent the disparate influences of socioeconomic position, individual behaviors or sexual networks, to name a few. In the study of racial disparities, Sherman illustrates that race is a “catch-all” term which reflects host, environmental, and behavioral factors.27 Host factors may include stress, health history, and current medical status as well as biologic differences including genotype. Environmental factors include both economic (education, insurance, poverty, discrimination) and sociophysical (social and sexual networks, residential segregation, neighborhood condition or density) forces. Behavioral factors may include sexual activity, contraceptive choice, or smoking. These factors may act individually or in concert under the rubric of race to influence disease risk which may, in turn, be attenuated or exacerbated by health care access and utilization. Finally, could male partner race simply be an intermediary, that is, does it reflect the influence of an independent variable? While speculative, it is theoretically possible that the association with male partner race reflects the likelihood of having had a prior partner with BV. That is, if black men have a higher probability of having had prior black female partners who, in turn, have higher risks of BV, does this increase the risk of BV in future partners of these men? It is essential for future prevention efforts that the components of race be drawn out to ascertain the individual modifiable factors that alter rates of BV incidence. An important limitation of the Flora study is that the findings may not be generalizable to nonwhite women. While the study included black women, the number of intervals where black women reported exclusively white male partners was exceedingly small (0.5% of sexually active intervals). Of note, a similar evaluation among pregnant women found an association among white women with black partners and no association among black women with white male partners.28 This highlights the social complexities underlying sexual partner selection. Many factors including social networks and cultural norms influence the partner selection of women. Among white women those with black male partners did not differ in terms of age, education or sexual behavior from those with white male partners, but did have higher BMI and lower rates of smoking. Given the small number of black women with white male partners, this study was not able to make a similar comparison for black women. Understanding how the factors underlying sexual partner selection influence BV risk are not understood, yet, they are important for future study. An explicit perspective on “race,” whether of the individual woman or her partner(s), is essential to prevent further stigmatization of this already disadvantaged population. The potential for misapplication of scientific findings or sensational headlines in the lay press is real. Research that disentangles the complex notion of race into its component causes will allow us to segregate risk factors into modifiable versus nonmodifiable, to develop and evaluate interventions to reduce racial disparities, and to more accurately quantify and attribute individual risks for the acquisition of BV.
No takes yet. Share an insight, caveat, or question.
Jenifer E. Allsworth (2010) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: