Risk factor analysis and predictive model for the alleviation of postpartum stress urinary incontinence: a retrospective cohort study.

Background: Postpartum stress urinary incontinence (SUI) can substantially affect women's daily functioning, psychosocial well-being, and rehabilitation needs. Predictive models for persistent postpartum SUI after basic pelvic floor muscle training (PFMT) remain limited. This study aimed to identify factors associated with persistent SUI and to develop a predictive nomogram for early risk stratification. Methods: This retrospective cohort study included 1,538 women diagnosed with SUI at one month postpartum who underwent basic PFMT and completed six-month follow-up at Wuhan Central Hospital between January 2020 and December 2024. Persistent SUI at six months postpartum was defined according to International Continence Society criteria as self-reported involuntary urine leakage during effort, exertion, sneezing, or coughing, confirmed during outpatient follow-up by clinical symptom assessment. Women without leakage symptoms at six months were classified into the Recovery Group, whereas women with ongoing leakage symptoms were classified into the Persistent SUI Group. Multivariable logistic regression was used to identify factors associated with persistent SUI. Model discrimination was evaluated using receiver operating characteristic (ROC) curve analysis. The optimal cutoff value was determined using the Youden index, and sensitivity, specificity, and accuracy were calculated. Internal validation was performed using 1,000 bootstrap resamples to estimate the apparent and optimism-corrected C-indexes. Calibration was assessed using a calibration curve and calibration slope, and clinical utility was evaluated using decision curve analysis (DCA). Nonlinear associations between postpartum BMI loss and persistent SUI were explored using restricted cubic splines and threshold analysis. Results: Among 1,538 women, 826 achieved symptom resolution and 722 had persistent SUI at six months postpartum. Multivariable analysis showed that older age, higher BMI, gestational diabetes mellitus, parity >=2, perineal tear, cystocele, uterine prolapse, rectocele, and levator ani hiatus defect were associated with higher odds of persistent SUI. Cesarean delivery and greater postpartum BMI loss were associated with lower odds of persistent SUI. The nomogram demonstrated good discrimination, with an AUC of 0.813 (95% CI: 0.792-0.834). At the optimal cutoff value of 0.549, sensitivity, specificity, and accuracy were 0.644, 0.828, and 0.742, respectively. Bootstrap validation yielded an apparent C-index of 0.813 and an optimism-corrected C-index of 0.808. The calibration slope was 0.974, and DCA suggested potential clinical net benefit across a range of threshold probabilities. A nonlinear U-shaped association was observed for postpartum BMI loss, with lower odds of persistent SUI below approximately 5 kg/m Conclusion: Older age, higher BMI, gestational diabetes mellitus, higher parity, perineal trauma, pelvic organ prolapse, and levator ani hiatus defect were associated with persistent postpartum SUI after basic PFMT. Cesarean delivery and moderate postpartum BMI loss were associated with lower odds of persistent SUI. The nomogram showed favorable internal performance but requires external validation before routine clinical application. The nonlinear association between BMI loss and persistent SUI should be interpreted cautiously because residual confounding and unmeasured postpartum factors may have influenced this finding.