Abstract:

To analysis the risk factors of intraoperative blood glucose abnormalities（dysglycemia） in elderly patients without diabetes，and to develop a nomogram risk prediction model. Methods Convenience sampling was conducted in elderly patients aged ≥ 65 years without diabetes who received surgical treatment. Univariate analysis and binary logistic regression analysis were used to determine the risk factors of intraoperative blood glucose abnormalities，establish a risk prediction model and draw a nomogram. Receiver Operating Characteristic（ROC） and Hosmer-Lemeshow tests were used to verify the predictive effect of the model，and Bootstrap method was used for internal validation. Results The logistic regression analysis showed that type of anesthesia，duration of surgery，baseline operating room blood glucose level，and age were independent predictors of intraoperative dysglycemia. The predictive formula for intraoperative dysglycemia was established as follows：Logit P=-12.810+0.066 × age +1.966 × baseline operating room blood glucose level +0.008 × duration of surgery -2.778 × type of anesthesia. The area under the ROC curve was 0.815，and the optimal critical value boundary was 0.765. The sensitivity and specificity were 83.00% and 67.00%，respectively. The result of Hosmer-Lemeshow test was χ²=5.557（P=0.697）. The prediction curve fit well with the ideal curve，suggesting that the model has good predictive ability. External validation showed that the sensitivity of the model was 77.01%，the specificity was 75.36%，and the overall accuracy was 76.56%. Conclusion The prediction model constructed in this study has a good effect，which can provide a reference for clinical evaluation of the risk of abnormal blood glucose in elderly patients without diabetes.

Key words: Aged, Intraoperative, Dysglycemia, Nomogram, Anesthesia, Duration of Surgery, Periope-rative Nursing

摘要：

目的 分析非糖尿病老年患者术中血糖异常的影响因素，构建列线图风险预测模型。方法 采用便利抽样法，选取2021年3月—11月在北京市某三级甲等医院住院的607例非糖尿病老年手术患者作为调查对象，采用单因素分析和Logistic回归分析确定术中血糖异常的危险因素，建立风险预测模型并绘制列线图。采用受试者操作特征曲线（receiver operating characteristic，ROC）和Hosmer-Lemeshow检验验证模型预测效果，采用Bootstrap法对模型进行内部验证。结果 Logistic回归分析显示，麻醉类型、手术时间、进入手术室血糖、年龄是术中血糖异常的独立预测因素。建立术中血糖异常的预测公式如下：Logit P=-12.810+0.066 × 年龄+1.966 × 进入手术室血糖+0.008 × 手术时间-2.778 × 麻醉类型。ROC曲线下面积为0.815，最佳临界值为0.765。灵敏度和特异度分别为83.00%和67.00%。Hosmer-Lemeshow检验的结果为χ²=5.557（P=0.697），模型预测曲线与理想曲线拟合良好，表明模型具有良好的预测能力。外部验证表明，该模型的灵敏度为77.01%，特异度为75.36%，总体准确性为76.56%。结论 该研究构建的预测模型预测效果较好，可为临床有效评估非糖尿病老年患者术中血糖异常发生风险提供参考依据。

关键词: 老年人, 术中, 血糖异常, 列线图, 麻醉, 手术时间, 围手术期护理