Scholarlink

<p>Surgical site infection (SSI) is a serious complication after proximal humeral fracture (PHF) surgery in the elderly, yet nutritional status is rarely incorporated into risk prediction. This exploratory study aimed to evaluate nutritional assessment tools for SSI prediction and to integrate the most informative marker with clinical predictors into a preliminary nomogram. We retrospectively reviewed elderly patients who underwent surgery for closed PHFs between 2017 and 2022. Five nutritional assessment tools were calculated from admission blood samples, and SSI was defined per CDC criteria. ROC analysis compared their predictive performance. Restricted cubic spline (RCS) regression characterized dose-response trends and identified an exploratory reference-based stratification value for the selected tool. Predictors with <i>P</i> &lt; 0.05 were entered into multivariable logistic regression to construct a nomogram. Model performance was evaluated by discrimination, calibration, and decision curve analysis (DCA). Sensitivity analyses using Firth penalized logistic regression and a model retaining triglycerides, total cholesterol, and body weight index (TCBI) as a continuous predictor were performed. Temporal validation was conducted in a later 2023–2024 cohort from the same institution. Among 734 patients in the development cohort (median age 72 years; 45% male), 28 (3.81%) developed SSI. Among the tested nutritional tools, TCBI had the highest, although modest, standalone AUC (0.654). RCS regression revealed a linear, positive association with SSI risk (P-overall &lt; 0.05). The exact exploratory reference-based stratification value was 1036.07, which was rounded to 1036 for use in the model. Six independent predictors were identified: older age, surgical delay ≥ 6 days, combined dislocation, longer operative time, elevated fasting blood glucose, and TCBI ≥ 1036. The nomogram showed an apparent AUC of 0.814. After bootstrap validation, the optimism-corrected C-index was 0.776 and the optimism-corrected calibration slope was 0.841. Firth penalized logistic regression yielded broadly consistent associations, and when TCBI was modeled as a continuous variable, the overall pattern of associations remained unchanged. The temporal validation cohort yielded an AUC of 0.809 (95% CI, 0.613–0.962). TCBI, together with five clinical predictors, may help stratify SSI risk after PHF surgery in elderly patients. Given the low number of events and the limited same-institution temporal validation, the model should be considered preliminary and hypothesis-generating; independent prospective multicenter validation is required.</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

An exploratory nomogram using nutritional and clinical predictors was temporally evaluated for surgical site infection after proximal humerus fracture surgery

  • 期刊论文

  • 出版模式：

    开放获取

  • 发表日期：

    2026年7月28日

查看全文
  • Zhihao Ren,
  • Yongle Wei,
  • Xuebin Zhang,
  • Erdong Zhang,
  • Ziping Li,
  • Zhiyong Hou,
  • Lin Jin

摘要

Surgical site infection (SSI) is a serious complication after proximal humeral fracture (PHF) surgery in the elderly, yet nutritional status is rarely incorporated into risk prediction. This exploratory study aimed to evaluate nutritional assessment tools for SSI prediction and to integrate the most informative marker with clinical predictors into a preliminary nomogram. We retrospectively reviewed elderly patients who underwent surgery for closed PHFs between 2017 and 2022. Five nutritional assessment tools were calculated from admission blood samples, and SSI was defined per CDC criteria. ROC analysis compared their predictive performance. Restricted cubic spline (RCS) regression characterized dose-response trends and identified an exploratory reference-based stratification value for the selected tool. Predictors with P < 0.05 were entered into multivariable logistic regression to construct a nomogram. Model performance was evaluated by discrimination, calibration, and decision curve analysis (DCA). Sensitivity analyses using Firth penalized logistic regression and a model retaining triglycerides, total cholesterol, and body weight index (TCBI) as a continuous predictor were performed. Temporal validation was conducted in a later 2023–2024 cohort from the same institution. Among 734 patients in the development cohort (median age 72 years; 45% male), 28 (3.81%) developed SSI. Among the tested nutritional tools, TCBI had the highest, although modest, standalone AUC (0.654). RCS regression revealed a linear, positive association with SSI risk (P-overall < 0.05). The exact exploratory reference-based stratification value was 1036.07, which was rounded to 1036 for use in the model. Six independent predictors were identified: older age, surgical delay ≥ 6 days, combined dislocation, longer operative time, elevated fasting blood glucose, and TCBI ≥ 1036. The nomogram showed an apparent AUC of 0.814. After bootstrap validation, the optimism-corrected C-index was 0.776 and the optimism-corrected calibration slope was 0.841. Firth penalized logistic regression yielded broadly consistent associations, and when TCBI was modeled as a continuous variable, the overall pattern of associations remained unchanged. The temporal validation cohort yielded an AUC of 0.809 (95% CI, 0.613–0.962). TCBI, together with five clinical predictors, may help stratify SSI risk after PHF surgery in elderly patients. Given the low number of events and the limited same-institution temporal validation, the model should be considered preliminary and hypothesis-generating; independent prospective multicenter validation is required.