Yang Feng, Xiaoqian Zhang
Purpose: Bacterial vaginosis (BV) is a common condition resulting from vaginal microbiota dysbiosis. The red cell distribution width-coefficient of variation (RDW-CV) is a hematological parameter derived from routine blood tests, indicating the variation in red blood cell size. This study aimed to investigate the correlation between RDW-CV and BV. Patients and Methods: Data were derived from NHANES (2001-2004), a U.S. national survey. Propensity score matching balanced group differences. Associations were evaluated using conditional logistic regression and restricted cubic splines. XGBoost identified key predictors, and mediation analysis explored serum uric acid's role. For external validation, consistency analysis used MIMIC-IV, where the outcome was "vaginitis" per ICD codes due to data limitations. Results: In the NHANES study, a positive correlation was found between elevated RDW-CV and BV. Individuals in the highest RDW-CV quartile had significantly higher odds of BV compared to those in the lowest quartile (OR=1.565, 95% CI: 1.128-2.173, P=0.007), after adjusting for confounders. Findings were consistent across subgroups and supported by RCS. Mediation analysis revealed that serum uric acid, a marker linked to inflammation and oxidative stress, mediated 5.81% of this relationship. In the MIMIC cohort, RDW-CV was positively associated with vaginitis (OR=4.265, 95% CI: 1.129-16.117, P=0.0325), but caution is advised due to the wide confidence interval resulting from the limited number of vaginitis cases (n=87). Conclusion: This study revealed a significant positive association between elevated RDW-CV and BV, partially corroborated by external validation. The consistent directional trends observed across cohorts reinforce this association. However, the broader outcome definition in the MIMIC database presents an inherent limitation. Causality cannot be established due to the cross-sectional design, underscoring the need for prospective studies.