Prediction of early urinary continence after radical prostatectomy based on preoperative pelvic floor parameters: a retrospective study.

Background: Postoperative urinary incontinence is one of the most significant complications following radical prostate cancer surgery. We aimed to predict the risk of urinary incontinence within one month of radical prostatectomy (RP) using preoperative physiological parameters of the pelvic floor muscles. Methods: This is a retrospective study with a convenience sample. A cohort of 188 patients with prostate cancer was recruited from March to December 2023 from a single urology department at Peking University First Hospital. The cohort was divided into a training set of 132 patients and a validation set of 56 patients at a 7:3 ratio. This study used multivariate logistic regression analysis to predict urinary incontinence and calculated the area under the receiver operating characteristic curve (AUC) for model validation. Nomograms and calibration plots were generated for training sets. Preoperative and operative parameters were collected, including age, body mass index (BMI), International Prostate Symptom Score (IPSS), prostate-specific antigen (PSA) level, Gleason score, surgical method, urethral reconstruction, lymph node dissection, nerve-sparing status, catheterization duration, D'Amico risk classification, American Society of Anesthesiologists (ASA) score, Charlson Comorbidity Index, postoperative duration, prostate volume, and pelvic floor muscle parameters. Results: The incidence of urinary incontinence within a month after RP was 78.7%. Advanced age and low fast-twitch muscle scores have emerged as independent risk factors for urinary incontinence. Patients older than 70 had a 6.283-fold higher risk of incontinence compared to those younger than 60 (95% CI: 1.47-26.95). The fast-twitch muscle score was significantly associated with the risk of incontinence (OR = 1.25; 95% CI: 1.05-1.49). The AUC was 0.764 (95% CI: 0.675-0.854) for the training set and 0.776 (95% CI: 0.644-0.908) for the validation set, with calibration plots indicating high model accuracy. Conclusions: Advanced age and low fast-twitch muscle scores in functional level were significant risk factors for RP. This risk predictive model enables healthcare professionals to perform accurate preoperative risk assessments and predictions based on patients' individualized indicators, and provide tailored postoperative rehabilitation strategies.