Abstract
Threatened abortion is a common complication in pregnancy. Studies have indicated an association between reproductive tract infections, vaginal microbiota imbalance, and adverse pregnancy outcomes, but evidence linking routine clinical vaginal microecological parameters to miscarriage risk remains insufficient. This retrospective study aims to systematically compare routine microecological indicators and specific pathogen infections between pregnant women at risk of miscarriage and a healthy pregnancy control group. A retrospective analysis was conducted among 136 patients with threatened abortion (threatened abortion group) and 69 healthy pregnant women (control group) who underwent prenatal examination at Shandong Maternal and Child Health Care Hospital between January 2024 and July 2025. We compared the vaginal microecological indicators between the 2 groups, including microbial density, diversity, hydrogen peroxide (H2O2), sialidase, leukocyte esterase, pH value, Nugent score, aerobic vaginitis (AV) score, as well as the presence of pathogenic microorganisms such as Ureaplasma urealyticum (UU), Candida spp., Gardnerella vaginalis/Prevotella spp., and Trichomonas vaginalis. The results showed no significant differences in microbial density, diversity, H2O2, sialidase, leukocyte esterase, pH, Nugent score, AV score, or detection rates of bacterial vaginosis, vulvovaginal candidiasis, AV or Trichomonas vaginitis between the 2 groups (all P > .05). However, the UU infection was significantly more prevalent in the threatened abortion group than in the control group (53.7% vs 24.6%, P < .001). After adjusting for maternal age and gestational age, UU infection was significantly associated with threatened abortion (adjusted odds ratio = 3.50, 95% confidence interval: 1.82-6.71, P < .001). UU infection was significantly more prevalent with threatened abortion. Whereas conventional vaginal microecological indicators were not significantly associated with threatened abortion, the limitations of this study primarily lie in the insufficient control for potential confounding factors and the constraints inherent in the laboratory methodologies employed. Future prospective studies employing more comprehensive detection techniques are needed to validate these findings.