老年患者行无痛胃肠镜检查时发生不良事件的风险因素分析及列线图模型的构建

Analysis of risk factors for adverse events in painless gastroenteroscopy and construction of a nomogram model

摘要:

目的 探讨老年患者行无痛胃肠镜检查时发生不良事件的危险因素，并构建预测不良事件发生风险的列线图模型。方法 选取2023年9月－2024年3月于该院接受无痛胃肠镜检查的老年患者302例，按照7∶3的比例随机分为训练集（211例）和验证集（91例）。训练集根据是否发生不良事件，分为不良事件组（64例）和非不良事件组（147例）。采用多因素Logistic回归模型，分析老年患者行无痛胃肠镜检查时发生不良事件的危险因素；选用R 3.4.3软件，构建列线图模型，并采用验证集进行外部验证。通过绘制受试者操作特征曲线（ROC curve）评价列线图模型区分度，并绘制校准曲线评价列线图模型的校准度。结果 单因素分析结果显示，不良事件组与非不良事件组中的超重或肥胖、吸烟、高血压、冠心病、文化程度、检查前睡眠时间和运动习惯比较，差异均有统计学意义（P < 0.05）；多因素Logistic回归分析显示，超重或肥胖（OR＾ = 4.821，95%CI：1.052～11.651）、吸烟（OR＾ = 1.056，95%CI：1.313～3.109）、高血压（OR＾ = 1.356，95%CI：1.175～2.677）和冠心病（OR＾ = 1.385，95%CI：1.168～2.765）是老年患者行无痛胃肠镜检查时发生不良事件的危险因素（P < 0.05）。绘制ROC curve，训练集和验证集的曲线下面积分别为0.921和0.795，敏感度为90.62%（95%CI：0.846～0.965）和82.14%（95%CI：0.689～0.856），特异度为74.83%（95%CI：1.056～2.939）和76.19%（95%CI：1.245～4.161），说明列线图模型具有较好的区分度；Hosmer-Lemeshow拟合优度检验结果显示，列线图模型具有较好的拟合度（P > 0.05）；训练集和验证集校准曲线均显示，列线图模型中，不良事件实际发生率与预测发生率具有较高的一致性（P > 0.05）。结论 超重或肥胖、高血压、心脏病和吸烟是老年患者行无痛胃肠镜检查时发生不良事件的独立危险因素，根据以上因素构建的无痛胃肠镜检查时发生不良事件风险的列线图预测模型，具有良好的预测效能。

Abstract:

Objective To investigate the risk factors of adverse events in elderly patients with painless gastroenteroscopy, and construct a nomogram model for predicting the risk of adverse events.Methods 302 patients who underwent painless gastroenteroscopy from September 2023 to March 2024 were retrospectively enrolled and randomly divided into training set (211 cases) and validation set (91 cases) according to the 7∶3 ratio. The training set was divided into adverse event group (64 cases) and non-adverse event group (147 cases) according to whether adverse events occurred. Multivariate Logistic regression analysis was used to screen the risk factors for adverse events in elderly painless gastroenteroscopy patients, R 3.4.3 software was used to construct a nomogram model, and the validation set was used for external verification. The nomogram model discrimination was evaluated by drawing the receiver operator characteristic curve (ROC curve), and the calibration curve was drawn to evaluate the calibration degree of the nomogram model.Results Univariate analysis showed that, the comparison of overweight or obesity, smoking, hypertension, coronary heart disease, education level, pre examination sleep time, and exercise habits between the adverse event group and the non-adverse event group were statistically significant (P < 0.05); Multivariate Logistic regression analysis showed that overweight or obesity (OR＾ = 4.821, 95%CI: 1.052 ~ 11.651), smoking (OR＾ = 1.056, 95%CI: 1.313 ~ 3.109), hypertension (OR＾ = 1.356, 95%CI: 1.175 ~ 2.677), and coronary heart disease (OR＾ = 1.385, 95%CI: 1.168 ~ 2.765) were risk factors for adverse events during painless gastroenteroscopy in elderly patients (P < 0.05); The areas under the ROC curves for the training and validation sets were 0.921 and 0.795, respectively. The sensitivity was 90.62% (95%CI: 0.846 ~ 0.965) and 82.14% (95%CI: 0.689 ~ 0.856), and the specificity was 74.83% (95%CI: 1.056 ~ 2.939) and 76.19% (95%CI: 1.245 ~ 4.161), indicating that the nomogram model had good discriminability; The Hosmer-Lemeshow goodness of fit test results showed that the nomogram model had a good fit (P > 0.05), and the calibration curves of the training and validation sets both showed that the nomogram model had a good consistency between the actual and predicting incidence of adverse events (P > 0.05).Conclusion Overweight or obesity, hypertension, cornary heart disease and smoking are independently correlated with adverse events in painless gastroenteroscopy, and the nomogram model constructed based on the above factors to predict the risk of adverse events in painless gastroenteroscopy has good predictive efficacy.