Risks associated with bacterial vaginosis in infertility patients: a systematic review and meta-analysis

Abstract

Is bacterial vaginosis (BV) associated with the cause of infertility and does BV impinge on conception rates and early pregnancy loss following IVF? The incidence of BV is significantly higher among patients with tubal infertility when compared with patients with non-tubal infertility. BV does not impinge on conception rates but is significantly associated with preclinical pregnancy loss, though not with first trimester abortion. BV is prevalent in patients with infertility, as evident from studies across the world. This study is a meta-analysis of data on the prevalence of BV in women with infertility, the association between BV and the cause of infertility, and the associations between BV and conception rates and early pregnancy loss following IVF. The meta-analyses of the various topics involved different numbers of studies: prevalence of BV with infertility--12 studies, association with tubal infertility--3 studies and associations with conception rates--6 studies, with early preclinical pregnancy loss--2 studies and with clinical pregnancy loss--4 studies. Systematic literature searches of the electronic databases, PubMed, EMBASE, CINAHL, the Cochrane Library and ISI Web of Knowledge were performed up to September 2012. Studies were included if they reported on, at least, one of the following: prevalence of BV in infertility patients, association between BV and the cause of infertility, association between BV and conception rates with IVF or association between BV and early pregnancy loss. Studies were considered eligible if BV was diagnosed through standardized criteria like Nugent's criteria or Hay-Ison's criteria. In none of the studies, infertility as such was defined, but patients were described as unselected patients undergoing IVF. The estimated prevalence of BV (Nugent score >6) in infertile women is 19% [95% confidence interval (CI): 14-25%]. Abnormal microflora including BV and intermediate microflora (Nugent scores 4-10) occurs in 39% of the infertile patients (95% CI: 26-52%). BV is significantly more prevalent in women with infertility compared with antenatal women in the same population [OR (odds ratio) 3.32, 95% CI 1.53-7.20]. BV is significantly more prevalent in women with tubal infertility compared with women with other causes of infertility (OR 2.77, 95% CI 1.62-4.75). BV is not associated with decreased conception rates (OR 1.03, 95% CI 0.79-1.33). Similarly, none of the studies found an association between abnormal vaginal flora and conception rates following IVF treatment. BV is associated with a significantly elevated risk of preclinical pregnancy loss (OR 2.36, 95% CI: 1.24-4.51). BV is not associated with an increased risk of first trimester miscarriage (OR 1.20, 95%CI: 0.53-2.75). All included studies were centre based. In addition, publication bias cannot be ruled out. Furthermore, all estimates are obtained using an absolute minimum of studies. The standard error on the estimates is so large that it does not allow for any formal statistical conclusions regarding heterogeneity between the effects reported in different studies. It needs to be recognized that most inferences drawn in our study rely on a limited number of studies, potentially, endangering the generalizability of our findings. Moreover, all studies on cause of infertility in relation to BV included had a cross-sectional design and, therefore, do not allow for causal inferences. Still, there is strong circumstantial evidence that supports a causal link between BV and tubal infertility. Studies with a longitudinal design, on the other hand, strongly support a relation between BV and early pregnancy loss. Unfortunately, no study looked beyond first trimester fetal loss, although it is plausible that the high preterm birth rates with IVF are, at least, in part attributable to BV.