目的 分析内窥镜鼻窦手术（ESS）治疗慢性鼻窦炎伴鼻息肉（CRSwNP）患者术后复发的危险因素，并构建预测模型。 方法 回顾性地选取2022年3月至2023年2月在医院进行ESS术的203例CRSwNP患者作为研究对象，按照术后是否复发分为复发组（n = 43）和未复发组（n = 160）。收集临床资料，采用单因素分析差异项，使用Lasso回归分析筛选潜在影响因素，采用多因素logistic回归分析危险因素，列线图构建术后复发模型，受试者工作特征（ROC）曲线及校准曲线对模型进行评估。 结果 单因素分析结果显示：ESS术后复发组病程、Lund-Mackay CT评分、鼻腔鼻窦结局测试-22（SNOT-22）、嗜酸性粒细胞（EOS）、嗜酸性粒细胞阳性蛋白（ECP）、总免疫球蛋白E（IgE）、C反应蛋白（CRP）、白细胞介素-5（IL-5）、IL-1β高于ESS术后未复发组（P < 0.05），宾夕法尼亚大学的嗅觉测试（UPSIT）评分、CD3⁺、CD4⁺、CD8⁺、CD4⁺/CD8⁺低于ESS术后未复发组（P < 0.05）；ESS术后复发组合并变应性鼻炎、哮喘、未术前糖皮质激素治疗、手术时间≥ 2 h占比高于ESS术后未复发组（P < 0.05）。多因素logistic回归分析模型结果显示，病程（OR = 1.389，95% CI：1.094 ~ 1.763，P = 0.007）、合并哮喘（OR = 2.997，95% CI：1.065 ~ 8.432，P = 0.038）、Lund-Mackay CT评分（OR = 1.156，95% CI：1.027 ~ 1.301，P = 0.016）、EOS（OR = 1.540，95% CI：1.249 ~ 1.898，P < 0.001）、总IgE（OR = 1.005，95% CI：1.000 ~ 1.009，P = 0.041）、IL-5（OR = 1.165，95% CI：1.078 ~ 1.260，P < 0.001）为ESS术后复发的危险因素。基于多因素logistic回归分析的列线图，其ROC曲线下面积（AUC）为0.9057、灵敏度为76.74%、特异度为87.5%，校准曲线平均绝对误差（MAE）为0.03、均方误差（MSE）为0.00157，0.9分位点的绝对误差为0.065。 结论 病程、合并哮喘、Lund-Mackay CT评分、EOS、总IgE、IL-5均是CRSwNP患者ESS术后复发危险因素，且相关列线图模型可作为CRSwNP术后复发风险评估的可靠工具。

Objective To analyze the risk factors for postoperative recurrence in patients with chronic sinusitis and nasal polyps （CRSwNP） treated by endoscopic sinus surgery （ESS）， and to construct a predictive model. Methods A retrospective study was conducted on 203 patients with CRSwNP who underwent ESS in the hospital from March 2022 to February 2023. These patients were divided into a recurrence group （n = 43） and an non-recurrence group （n = 160） based on whether they experienced recurrence after surgery. Clinical data were collected and analyzed using univariate analysis to identify significant differences. Lasso regression was used to screen potential influencing factors. Multivariate logistic regression was employed to analyze the risk factors. A nomogram was constructed for postoperative recurrence model， and the receiver operating characteristic （ROC） curve and calibration curve were used to evaluate the model. Results Single-factor analysis showed that disease duration， Lund-Mackay CT score， SNOT-22 score， EOS， ECP， total IgE， CRP， IL-5， and IL-1β in the recurrence group after ESS were higher than those in the non-recurrence group （P < 0.05）， while the UPSIT score， CD3⁺， CD4⁺， CD8⁺， and CD4⁺/CD8⁺ were lower （P < 0.05）； the proportion of patients with allergic rhinitis， asthma， no preoperative glucocorticoid treatment， and surgery time ≥ 2 hours in the recurrence group after ESS was higher than that in the non-recurrence group （P < 0.05）. Multivariate logistic regression analysis model showed that disease duration （OR = 1.389， 95% CI： 1.094 ~ 1.763， P = 0.007）， combined asthma （OR = 2.997， 95% CI： 1.065 ~ 8.432， P = 0.038）， Lund-Mackay CT score （OR = 1.156， 95% CI： 1.027 ~ 1.301， P = 0.016）， EOS （OR = 1.540， 95% CI： 1.249 ~ 1.898， P < 0.001）， total IgE （OR = 1.005， 95% CI： 1.000 ~ 1.009， P = 0.041）， and IL-5 （OR = 1.165， 95% CI： 1.078 ~ 1.260， P < 0.001） were risk factors for ESS recurrence. Based on multivariate logistic regression analysis nomogram， the area under the ROC curve （AUC） was 0.9057； the sensitivity 76.74%； the specificity 87.5%； the average absolute error （MAE） of the calibration curve 0.03； the mean square error （MSE） 0.00157， and the absolute error at the 0.9 percentile 0.065. Conclusions Disease course， the presence of asthma， Lund-Mackay CT score， EOS， total IgE， and IL-5 are all risk factors for postoperative recurrence in patients with CRSwNP. Moreover， the relevant nomogram model can be used as a reliable tool for assessing the risk of postoperative recurrence in CRSwNP.