Scholarlink

<p>We developed and internally validated a prediction model based exclusively on routinely available preoperative (pre-) parameters to identify patients at high risk of postoperative (post-) sepsis following upper urinary tract calculi surgery. A retrospective cohort of 260 patients treated between May 2024 and April 2025 was analyzed. Candidate predictors were selected using least absolute shrinkage and selection operator (LASSO) regression with the one-standard-error rule (λ.1se), followed by multivariable logistic regression modeling. Model performance was evaluated in terms of discrimination, calibration, and clinical utility. Among the included patients, 43 (16.5%) developed postoperative sepsis. The final model incorporated four predictors: stone count ≥ 2 (OR = 3.25, <i>P =</i> 0.006), perirenal fat stranding (OR = 3.08, <i>P =</i> 0.006), pre-positive urine nitrite (OR = 4.29, <i>P =</i> 0.003), and pre-positive urine culture (OR = 4.53, <i>P &lt;</i> 0.001). The model demonstrated good discrimination (AUC = 0.878, 95% CI: 0.832–0.924). Calibration analysis using bootstrap resampling showed good agreement between predicted and observed outcomes (calibration slope = 0.923; intercept = − 0.075). Decision curve analysis indicated a favorable net clinical benefit across a wide range of threshold probabilities. This parsimonious model may facilitate early risk stratification and support perioperative decision-making; however, external validation is warranted.</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

A parsimonious preoperative prediction model for postoperative sepsis following upper urinary tract calculi surgery

  • 期刊论文

  • 发表日期：

    2026年5月26日

查看全文
  • Xin Tian,
  • Ruimin Liu,
  • Fei Qin,
  • Qiuping Zhang

摘要

We developed and internally validated a prediction model based exclusively on routinely available preoperative (pre-) parameters to identify patients at high risk of postoperative (post-) sepsis following upper urinary tract calculi surgery. A retrospective cohort of 260 patients treated between May 2024 and April 2025 was analyzed. Candidate predictors were selected using least absolute shrinkage and selection operator (LASSO) regression with the one-standard-error rule (λ.1se), followed by multivariable logistic regression modeling. Model performance was evaluated in terms of discrimination, calibration, and clinical utility. Among the included patients, 43 (16.5%) developed postoperative sepsis. The final model incorporated four predictors: stone count ≥ 2 (OR = 3.25, P = 0.006), perirenal fat stranding (OR = 3.08, P = 0.006), pre-positive urine nitrite (OR = 4.29, P = 0.003), and pre-positive urine culture (OR = 4.53, P < 0.001). The model demonstrated good discrimination (AUC = 0.878, 95% CI: 0.832–0.924). Calibration analysis using bootstrap resampling showed good agreement between predicted and observed outcomes (calibration slope = 0.923; intercept = − 0.075). Decision curve analysis indicated a favorable net clinical benefit across a wide range of threshold probabilities. This parsimonious model may facilitate early risk stratification and support perioperative decision-making; however, external validation is warranted.