Scholarlink

Introduction and Hypothesis <p>Preoperative lower-extremity venous thrombosis (LEVT) is frequently overlooked in women undergoing pelvic organ prolapse (POP) surgery, despite its potential impact on perioperative outcomes. Conventional risk tools, including the Caprini score and D-dimer, show limited performance in this population, and the true prevalence of occult preoperative LEVT remains uncertain. This study aimed to determine the prevalence of occult preoperative LEVT in women undergoing POP surgery and to develop and validate a clinical prediction model for this condition.</p> Methods <p>This retrospective cross-sectional study included consecutive women scheduled for POP surgery between August 2019 and August 2025. All patients underwent routine preoperative compression ultrasonography of lower extremities. Demographic, clinical, and laboratory variables were collected. Patients were randomly divided into training (<i>n</i> = 878) and testing (<i>n</i> = 375) sets. Independent predictors were identified by multivariable logistic regression. Five predictive models—logistic regression, random forest, extreme gradient boosting (XGBoost), adaptive boosting (AdaBoost), and support vector machine (SVM)—were developed. Model performance was assessed using discrimination, calibration, and clinical utility, with comparisons to the Caprini score and D-dimer. A nomogram and web-based calculator were derived from the final model.</p> Results <p>Among 1253 women, 117 (9.3%) had preoperative LEVT. Independent predictors of preoperative LEVT included advanced age, lower-extremity varicose veins, prior venous thromboembolism, elevated cholesterol, and increased D-dimer. Logistic regression achieved the best performance for predicting preoperative LEVT (AUC 0.827; 95% CI 0.770–0.883), significantly exceeding the Caprini score and D-dimer. At the optimal threshold, sensitivity was 0.971 and negative predictive value was 0.995.</p> Conclusions <p>Routine screening revealed a higher-than-expected prevalence of preoperative LEVT in POP surgery. A simple logistic model provided clinically useful risk stratification, supporting more selective use of preoperative ultrasonography and improved perioperative thrombosis management.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Hidden Preoperative Lower-Extremity Venous Thrombosis in Pelvic Organ Prolapse Surgery: Prevalence and Development of a Clinical Prediction Model

  • 期刊论文

  • 发表日期：

    2026年5月16日

查看全文
  • Qi Wang,
  • Xiaoxiang Jiang,
  • Xiaoyan Li,
  • Chaoqin Lin

摘要

Introduction and Hypothesis

Preoperative lower-extremity venous thrombosis (LEVT) is frequently overlooked in women undergoing pelvic organ prolapse (POP) surgery, despite its potential impact on perioperative outcomes. Conventional risk tools, including the Caprini score and D-dimer, show limited performance in this population, and the true prevalence of occult preoperative LEVT remains uncertain. This study aimed to determine the prevalence of occult preoperative LEVT in women undergoing POP surgery and to develop and validate a clinical prediction model for this condition.

Methods

This retrospective cross-sectional study included consecutive women scheduled for POP surgery between August 2019 and August 2025. All patients underwent routine preoperative compression ultrasonography of lower extremities. Demographic, clinical, and laboratory variables were collected. Patients were randomly divided into training (n = 878) and testing (n = 375) sets. Independent predictors were identified by multivariable logistic regression. Five predictive models—logistic regression, random forest, extreme gradient boosting (XGBoost), adaptive boosting (AdaBoost), and support vector machine (SVM)—were developed. Model performance was assessed using discrimination, calibration, and clinical utility, with comparisons to the Caprini score and D-dimer. A nomogram and web-based calculator were derived from the final model.

Results

Among 1253 women, 117 (9.3%) had preoperative LEVT. Independent predictors of preoperative LEVT included advanced age, lower-extremity varicose veins, prior venous thromboembolism, elevated cholesterol, and increased D-dimer. Logistic regression achieved the best performance for predicting preoperative LEVT (AUC 0.827; 95% CI 0.770–0.883), significantly exceeding the Caprini score and D-dimer. At the optimal threshold, sensitivity was 0.971 and negative predictive value was 0.995.

Conclusions

Routine screening revealed a higher-than-expected prevalence of preoperative LEVT in POP surgery. A simple logistic model provided clinically useful risk stratification, supporting more selective use of preoperative ultrasonography and improved perioperative thrombosis management.