内镜下治疗结直肠腺瘤术后出血的危险因素及预测模型的建立

Risk Factors and Prediction Model Establishment for Postoperative Hemorrhage of Colorectal Adenoma under Endoscopy

摘要:

〔摘 要〕 目的：探讨内镜下治疗结直肠腺瘤术后出血的危险因素，以此建立预测模型，并评估模型的预测能力。 方法：收集 2015 年 5 月至 2018 年 12 月在福建省立医院南院消化内镜中心行内镜下结直肠腺瘤手术患者的临床资料， 选取术后出血的 100 例患者作为观察组，术后未出血的 100 例患者作为对照组。对临床中观察到的潜在出血危险因素 （年龄、性别、手术部位、腺瘤大小、手术方式）进行单因素和多因素 logistic 回归分析，探索结直肠腺瘤术后出血 的危险因素，并建立预测模型，以受试者工作特征曲线（ROC）评估模型的预测能力。 结果：多因素 logistc 回归显示， 年龄（OR = 0.436，P = 0.021）、性别（OR = 0.283，P = 0.002）、手术部位（OR = 4.993，P = 0.003）、手术方式 （OR = 8.632，P = 0.000）、腺瘤大小（OR = 6.163，P = 0.011）均为内镜下结直肠腺瘤术后出血的危险因素。预测 模型的 ROC 曲线下面积（AUC）为 0.815，P ＜ 0.001，95 % CI（0.775，0.874），最佳临界点为 0.515，灵敏度为 0.72， 特异度为 0.79。 结论：患者年龄、性别、手术部位、腺瘤大小、手术方式均是术后出血的危险因素，基于危险因素构 建的预测模型可以有效识别出术后出血的高危患者。

Abstract:

〔Abstract〕 Objective To investigate the risk factors of hemorrhage after endoscopic treatment of colorectal adenoma, so as to establish a predictive model and evaluate the predictive ability of the model. Methods The clinical data of patients who underwent endoscopic colorectal adenoma surgery in the Digestive Endoscopy Center of Fujian Provincial Hospital South Branch from May 2015 to December 2018 were collected. 100 patients with postoperative hemorrhage were selected as the observation group, and 100 patients without postoperative hemorrhage were selected as the control group. Observed in the clinical potential bleeding risk factors (age, gender, operative site, adenoma size, operation mode) on the single factor and multiple factors logistic regression analysis, to explore the risk factors of colorectal adenomas postoperative bleeding, and establish the forecast model, with the receiver operating characteristic curve (ROC) evaluation model of ability to predict. Results Multi-factor logistic regression showed that age (OR = 0.436, P = 0.021), gender (OR = 0.283, P = 0.002), operative site (OR = 4.993, P = 0.003), and surgical method (OR = 8.632, P = 0.003). P = 0.000) and adenoma size (OR = 6.163, P = 0.011) were risk factors for bleeding after endoscopic colorectal adenoma surgery. The area under ROC area under curve (AUC) of the prediction model was 0.815, P < 0.001, 95% CI (0.775, 0.874), the best critical point was 0.515, the sensitivity was 0.72, and the specificity was 0.79. Conclusion Age, gender, operative site, adenoma size and surgical method in patients are all risk factors for postoperative hemorrhage. The prediction model based on risk factors can effectively identify high-risk patients with postoperative hemorrhage.