Treatment-associated changes in the cervicovaginal microbiome and microbial correlates of HPV clearance and cervical disease outcomes: a systematic review

Abstract Background Treatment for high-risk human papillomavirus (HR-HPV), cervical intraepithelial neoplasia, and cervical cancer is effective, but HPV may persist or recur. We synthesised evidence on post-treatment cervicovaginal microbiota (CVM) changes and pre-treatment CVM correlates of outcomes. Methods Five databases were searched to August 20, 2026 for studies assessing the CVM molecularly before, or before and after, treatment. Risk of bias was assessed using RoB-2, ROBINS-I, and the Newcastle–Ottawa Scale. Findings were synthesised narratively under Synthesis Without Meta-analysis guidance and stratified by population, treatment, sampling site, and outcome. Meta-analysis and formal publication-bias assessment were not performed. Results Twenty-one studies were included: 16 cohorts, three non-randomised studies, and two randomised trials. Twenty reported the number of women enrolled ( n = 1,345), whereas one reported 125 vaginal samples without a corresponding number of women. Most excision or probiotic studies reported lower alpha diversity and greater Lactobacillus dominance after treatment; after loop excision, Lactobacillus-depleted communities fell from 32% to 10% among women who cleared HR-HPV in one study. Only five studies included a comparison group. Eight of 11 studies found no baseline alpha-diversity difference by outcome. Baseline Lactobacillus iners was higher with HPV persistence (62% vs 16%; 60% vs 45%); the only adjusted analysis associated it with non-clearance after non-operative management (aOR 3.94, 95% CI 1.06–14.70) but yielded an imprecise estimate after excision (aOR 1.86, 95% CI 0.12–29.42). Individual studies linked several anaerobic taxa to HPV persistence or recurrence and Lacticaseibacillus casei, Peptostreptococcus anaerobius, and Anaerococcus lactolyticus to clearance, but none of these associations was externally validated and multiple-testing correction was inconsistently reported. Conclusions After treatment, the CVM often shifts towards lower-diversity, Lactobacillus-dominant communities, but causality is unproven. Baseline L. iners dominance is a candidate correlate of unfavourable outcomes requiring prospective confirmation.