Morphology-defined bacterial vaginosis and HPV-related cervical screening abnormalities: a two-year real-world study...

Bacterial vaginosis (BV) is characterized by reduced Lactobacillus dominance and enrichment of anaerobic bacteria. Although BV has been associated with human papillomavirus (HPV) infection, its relationship with cytologic abnormalities and biopsy-confirmed cervical lesions remains incompletely defined in real-world laboratory settings.

We conducted a retrospective real-world study integrating vaginal fluorescence microscopy, 21-genotype HPV genotyping, thin-prep cytology (TCT), and cervical histopathology records from January 2024 through December 2025 at a tertiary hospital in China. Morphology-defined BV was defined by clue cells or Gardnerella-like anaerobic bacteria on vaginal fluorescence microscopy and was not equivalent to Nugent scoring, Amsel criteria, culture-based diagnosis, or molecular microbiome profiling. HPV and TCT records were matched within a prespecified 30-day window after patient-level deduplication. Multivariable logistic regression adjusted for year, age, and vaginal microecological covariates. A 90-day sensitivity analysis was performed. The histopathology cohort was clinically selected and was analyzed as a correlation subgroup rather than as a random sample of the screening cohort.

The main 30-day analysis included 4,492 unique patients, of whom 587 (13.07%) had morphology-defined BV. After multivariable adjustment, morphology-defined BV remained associated with overall HPV positivity (adjusted odds ratio [aOR] 1.560; 95% confidence interval [CI] 1.293-1.883), high-risk HPV positivity (aOR 1.611; 95% CI 1.327-1.957), HPV multiple infection (aOR 1.976; 95% CI 1.526-2.559), TCT abnormality (aOR 1.465; 95% CI 1.125-1.907), and concurrent high-risk HPV positivity plus TCT abnormality (aOR 1.558; 95% CI 1.174-2.066). After additional adjustment for high-risk HPV, the BV-TCT association was attenuated and non-significant. Among 825 patients with cervical histopathology records, BV was not independently associated with any histologic lesion grade, whereas high-risk HPV, HPV16/18, and TCT abnormality were the principal predictors of histopathologic disease.

Morphology-defined BV was associated with HPV infection and HPV-related cytologic abnormalities, but not with histologic cervical lesions in the clinically selected biopsy subgroup. These findings suggest that routine morphologic evidence of BV marks an HPV-related screening-positive phenotype rather than an independent histopathologic lesion predictor.