Predicting stress urinary incontinence during pregnancy: combination of pelvic floor ultrasound parameters and clinical factors.

INTRODUCTION: The aim of this study was to develop and validate a predictive tool that combines pelvic floor ultrasound parameters and clinical factors for stress urinary incontinence during pregnancy.

MATERIAL AND METHODS: A total of 535 women in the first or second trimester of pregnancy were included for an interview and transperineal ultrasound assessment from two hospitals. Imaging data sets were analyzed offline to assess for bladder neck vertical position, urethra angles (α, β and γ angles), hiatal area and bladder neck funneling. All significant continuous variables at univariable analysis were analyzed by receiver operating characteristics. Three multivariable logistic models were built on clinical factors, and combined with ultrasound parameters. The final predictive model with best performance and fewest variables was selected to establish a nomogram. Internal and external validation of the nomogram was performed by both discrimination represented by C-index and calibration measured by Hosmer-Lemeshow test. A decision curve analysis was conducted to determine the clinical utility of the nomogram.

RESULTS: After excluding 14 women with invalid data, 521 women were analyzed. β angle, γ angle and hiatal area had limited predictive value for stress urinary incontinence during pregnancy, with area under curves of 0.558-0.648. The final predictive model included body mass index gain since pregnancy, constipation, previous delivery mode, β angle at rest, and bladder neck funneling. The nomogram based on the final model showed good discrimination with a C-index of 0.789 and satisfactory calibration (p = 0.828), both of which were supported by external validation. Decision curve analysis showed that the nomogram was clinically useful.

CONCLUSIONS: The nomogram incorporating both the pelvic floor ultrasound parameters and clinical factors has been validated to show good discrimination and calibration, and could be an important tool for stress urinary incontinence risk prediction at an early stage of pregnancy.