Molecular detection of urogenital microorganisms in women with microscopy-negative vaginitis: a retrospective case–control study

A substantial proportion of women with vaginitis symptoms have no explanatory findings after routine microscopy, which may lead to empirical treatment and persistent symptoms. This study aimed to compare urogenital microbial DNA detection patterns between microscopy-negative symptomatic women and asymptomatic controls. This retrospective single-center case–control study included 309 women evaluated at Minhang Hospital, Fudan University, Shanghai, China, between June and September 2025. Of these, 155 were symptomatic women with clinician-diagnosed vaginitis, no explanatory findings on wet-mount microscopy or Gram staining, and Nugent scores of 0–3; 154 were asymptomatic controls with a Nugent score of 0 and no microscopic abnormalities. Eleven targets were tested using commercial multiplex and single-target real-time PCR assays, and Ureaplasma DNA loads were normalized to human cell equivalents. The secondary composite endpoint was detection of at least one non- Ureaplasma target. Associations for ten estimable targets were estimated using age-adjusted Firth penalized logistic regression, with Benjamini–Hochberg correction across these comparisons. For the secondary composite endpoint, at least one non- Ureaplasma target was detected in 61 of 155 symptomatic participants (39.4%) and 23 of 154 controls (14.9%; age-adjusted odds ratio [aOR], 3.60; 95% confidence interval [CI], 2.12–6.30; P < 0.001). Trichomonas vaginalis was detected in 36 symptomatic participants and no controls; the resulting complete separation yielded a large but imprecise Firth estimate (aOR, 110.16; 95% CI, 15.08–14046.89; P < 0.001). Ureaplasma parvum showed an inverse nominal association. Positive nominal associations for Chlamydia trachomatis (11 vs. 3 detections) and Candida glabrata (5 vs. 0 detections) were based on small numbers of positive detections. Only T. vaginalis remained statistically significant after correction for the ten target-specific comparisons. Cell-normalized Ureaplasma DNA loads did not differ significantly between groups. Non- Ureaplasma targets were more frequently detected among microscopy-negative symptomatic women; T. vaginalis was the only target that remained significant after multiplicity correction. Molecular testing could complement routine assessment in selected patients with unexplained symptoms. However, DNA detection alone cannot establish microbial viability, active infection, or symptom causality, and other nominal associations require confirmation in larger prospective studies.