Dou Xian-Ling, Zhang Ke-Ke
Department of Obstetrics and Gynecology, Xi'an People's Hospital (Fourth Hospital of Xi'an), Xi'an, Shaanxi, China.
Department of Obstetrics, Xi'an International Medical Center Hospital, Xi'an, Shaanxi, China.
Med Sci Monit. 2025 Jul 2;31:e947803. doi: 10.12659/MSM.947803.
BACKGROUND Postoperative infections following cesarean sections contribute to increased maternal morbidity, prolonged hospital stays, and elevated healthcare costs. Identifying risk factors and developing predictive models are essential for targeted prevention. MATERIAL AND METHODS A retrospective study of 685 cesarean section patients from January 2021 to December 2023 categorized them into infection (n=33) and non-infection (n=652) groups. Risk factors were identified using multivariable logistic regression. A nomogram was developed and validated using receiver operating characteristic (ROC) curve analysis, calibration plots, and decision curve analysis (DCA). RESULTS Comparative analysis showed diabetes mellitus (39.4% vs 20.0%, P<0.001) and Group B Streptococcus (GBS) colonization (9.1% vs 2.4%, P=0.024) were more common in the infection group. Membrane rupture (57.6% vs 23.8%, P<0.001), complete cervical dilation (6.1% vs 0.9%, P=0.007), and >5 vaginal examinations (30.3% vs 10.0%, P<0.001) increased infection risk. The nomogram showed an AUC of 0.786 (95% CI: 0.681-0.856), sensitivity of 79.7%, and specificity of 76.8%. Internal validation confirmed a corrected C-index of 0.716 and excellent calibration (mean absolute error=0.008, Hosmer-Lemeshow χ²=2.915, P=0.921). Decision curve analysis demonstrated superior net benefit over no or universal intervention. CONCLUSIONS Key risk factors for postoperative infections include excessive vaginal examinations, membrane rupture, cervical dilation, diabetes mellitus, and GBS colonization. The nomogram offers strong predictive accuracy and clinical utility, aiding clinicians in stratifying infection risk and implementing targeted prevention.
剖宫产术后感染会导致产妇发病率增加、住院时间延长以及医疗费用升高。识别风险因素并建立预测模型对于针对性预防至关重要。材料与方法：对2021年1月至2023年12月期间的685例剖宫产患者进行回顾性研究，将其分为感染组（n = 33）和非感染组（n = 652）。使用多变量逻辑回归识别风险因素。通过受试者操作特征（ROC）曲线分析、校准图和决策曲线分析（DCA）建立并验证列线图。结果：对比分析显示，糖尿病（39.4% vs 20.0%，P < 0.001）和B族链球菌（GBS）定植（9.1% vs 2.4%，P = 0.024）在感染组中更为常见。胎膜破裂（57.6% vs 23.8%，P < 0.001）、宫颈完全扩张（6.1% vs 0.9%，P = 0.007）以及>5次阴道检查（30.3% vs 10.0%，P < 0.001）会增加感染风险。列线图的曲线下面积（AUC）为0.786（95%可信区间：0.681 - 0.856），灵敏度为79.7%，特异度为76.8%。内部验证确认校正后的C指数为0.716，校准良好（平均绝对误差 = 0.008，Hosmer - Lemeshow χ² = 2.915，P = 0.921）。决策曲线分析表明，与不干预或普遍干预相比，净效益更高。结论：术后感染的关键风险因素包括过多的阴道检查、胎膜破裂、宫颈扩张、糖尿病和GBS定植。列线图具有较高的预测准确性和临床实用性，有助于临床医生对感染风险进行分层并实施针对性预防。
