Human papillomavirus infection and cervical dysplasia in HIV-positive women: potential role of the vaginal microbiota.

OBJECTIVES: To assess the associations between microbiological markers of vaginal dysbiosis and incident/cleared/type-swap/persistent high-risk human papillomavirus (hrHPV) infection; and incident/cured/cleared/persistent high-grade cervical intraepithelial neoplasia (CIN2+) while controlling for persistent hrHPV infection. DESIGN: Two nested case-control studies (N = 304 and 236) within a prospective cohort of HIV-positive women in Johannesburg, South Africa. METHODS: Participants were examined for hrHPV type (INNO-LiPA), cervical dysplasia (histology), and vaginal microbiota (VMB) composition (V3-V4 Illumina HiSeq 2x300 bp) at baseline and endline, a median of 16 months later. RESULTS: Women with incident hrHPV compared to those who remained hrHPV-negative were less likely to have an optimal Lactobacillus crispatus or jensenii-dominated VMB type at end-line [relative risk ratio (RRR) 0.125, P = 0.019], but not at baseline. Having different hrHPV types at both visits was associated with multiple anaerobic dysbiosis markers at baseline (e.g. increased bacterial vaginosis-associated anaerobes relative abundance: RRR 3.246, P = 0.026). Compared to women without CIN2+, but with hrHPV at both visits, women with incident CIN2+ had increased Simpson diversity (RRR 7.352, P = 0.028) and nonsignificant trends in other anaerobic dysbiosis markers at end-line but not baseline. These associations persisted after controlling for age, hormonal contraception, and CD4 cell count. Current hormonal contraceptive use (predominantly progestin-only injectables) was associated with increased CIN2+ risk over-and-above persistent hrHPV infection and independent of VMB composition. CONCLUSIONS: hrHPV infection (and/or increased sexual risk-taking) may cause anaerobic vaginal dysbiosis, but a bidirectional relationship is also possible. In this population, dysbiosis did not increase CIN2+ risk, but CIN2+ increased dysbiosis risk. The CIN2+ risk associated with progestin-only injectable use requires further evaluation.