论文部分内容阅读
目的:建立非心脏手术老年患者术后谵妄(POD)的预测模型并进行验证。方法:选择2019年3月至2019年8月全身麻醉下行非心脏手术的老年患者531例，采用简单随机抽样方法，按照7∶3的比例将患者分为建模组(n n＝386)和验证组(n n＝145)。收集患者术前一般情况各指标、术中和术后各指标。术后1～3 d采用意识模糊评估法评估POD发生情况。建模组采用logistic回归分析，筛选POD的独立危险因素，利用列线图法对各风险因素进行综合并建模，最后利用验证组对于得到的预测模型进行验证。n 结果:logistic回归分析发现年龄、麻醉时间、气管拔管时间、入ICU、简易精神状态检查量表评分(MMSE)、查尔森合并症指数(CCI)、术后中性粒细胞与淋巴细胞比率是非心脏手术老年患者POD的独立危险因素(预测值＝0.053×年龄＋0.004×麻醉时间＋0.025×气管拔管时间＋0.815×入ICU-0.115×MMSE＋0.274×CCI)，建立列线图预测模型。该预测模型经Hosmer-Lemshow检验，n P ＝0.492。利用本模型中的预测值拐点值作为诊断指标预测验证组患者POD发生的受试者工作特征曲线下面积为0.805，敏感度为61.92%，特异度为99.21%。n 结论:本试验成功建立了非心脏手术老年患者POD预测模型，该模型可有效评估POD的风险程度。“,”Objective:To establish and verify the postoperative delirium (POD) prediction model in elderly patients undergoing non-cardiac surgery.Methods:A total of 531 elderly patients who underwent non-cardiac surgery under general anesthesia from March 2019 to August 2019 were selected and divided into model group (n n=386) and verification group (n n=145) by using a simple random sampling method.The indexes of patient′s baseline characteristics before operation and indexes during and after operation were collected.The occurrence of POD was evaluated by Confusion Assessment Method within 1-3 days after surgery.Logistic regression analysis was used to analyze the data to stratify the independent risk factors of POD in model group, and the nomogram method was used to synthesize various risk factors and establish the model, and the prediction model established was finally verified by using the data obtained in verification group.n Results:Multivariate logistic regression analysis showed that age, anesthesia time, extubation time, ICU admission, Mini-mental State Examination (MMSE) score, Charlson Comorbidity Index, and postoperative neutrophil to lymphocyte ratio (NLR) were independent risk factors for POD (the predicted value=0.053×age+ 0.004×anesthesia time+ 0.025×extubation time+ 0.815× ICU admission-0.115×MMSE+ 0.274×CCI). The nomogram prediction model was established.The nomogram model to predict POD was verified by Hosmer-Lemshow, n P =0.492.The cut-off point value of the predictive value in this model was used as the diagnostic index to predict the occurrence of POD in verification group, and the area under the receiver operating characteristic curve was 0.805, with a sensitivity of 61.92% and a specificity of 99.21%.n Conclusion:This study successfully establishes a POD prediction model for elderly patients undergoing non-cardiac surgery, which can effectively evaluate the risk degree of POD.