基于衰弱和围手术期临床资料构建老年髋部骨折手术神经认知障碍的预测模型

    construction of the prediction model of neurocognitive impairment in elderly patients with hip fracture surgery based on frailty and perioperative clinical data

    • 摘要:
      目的: 基于衰弱和围手术期临床资料构建老年髋部骨折手术围术期神经认知障碍(PND)预测模型,并探讨其应用价值。
      方法: 选取231例老年髋部骨折手术病人,根据是否发生PND分为PND组和非PND组。比较2组临床资料,通过logistic分析老年髋部骨折手术病人发生PND的影响因素,根据影响因素构建老年髋部骨折手术病人发生PND的Nomogram预测模型,通过ROC曲线、DCA评价模型的预测效能和临床效用。
      结果: 231例老年髋部骨折手术病人中37例发生PND,PND发生率为16.01%;年龄(OR = 3.329,95%CI = 1.315 ~ 8.426)、术前合并症数量(OR = 2.801,95%CI = 1.197 ~ 6.554)、MMSE评分(OR = 0.375,95%CI = 0.231 ~ 0.609)、白蛋白(OR = 0.327,95%CI = 0.185 ~ 0.577)、CRP水平(OR = 3.339,95%CI = 1.357 ~ 8.216)、术前衰弱(OR = 6.104,95%CI = 6.104 ~ 2.375)、手术方式(OR = 3.182,95%CI = 1.259 ~ 8.042)、麻醉方式(OR = 4.456,95%CI = 1.884 ~ 10.537)、术中镇静深度(OR = 4.805,95%CI = 1.921 ~ 12.018)、术后使用镇痛泵(OR = 3.971,95%CI = 1.536 ~ 10.264)均为老年髋部骨折手术病人发生PND的影响因素;ROC曲线、DCA结果显示该预测模型的AUC为0.917,具有良好预测效能,且具有明显的正向净收益,在预测PND发生风险方面拥有良好临床效用。
      结论: 年龄、术前合并症数量、MMSE评分、白蛋白、CRP水平、术前衰弱、手术方式、麻醉方式、术中镇静深度、术后使用镇痛泵均为老年髋部骨折手术病人发生PND的影响因素,临床可据此制定针对性防治措施,进一步改善预后。

       

      Abstract:
      Objective To construct the prediction model of perioperative neurocognitive disorders (PND) in elderly patients with hip fracture surgery based on frailty and perioperative clinical data, and explore its application value.
      Methods A total of 231 elderly patients with hip fractures were divided into the PND group and non-PND group according to whether they had PND. The clinical data were compared between two groups, the influencing factors of PND in elderly patients treated with hip fracture surgery were analyzed using logistic regression. Based on the influencing factors, a Nomogram prediction model for PND in elderly patients treated with hip fracture surgery was constructed. The predictive performance and clinical utility of the model were evaluated using ROC curves and DCA.
      Results Among 231 elderly patients with hip fractures treated with surgery, 37 cases developed PND, with a PND incidence of 16.01%. The age (OR = 3.329, 95%CI = 1.315–8.426), number of preoperative comorbidities (R = 2.801, 95%CI = 1.197–6.554), MMSE score (OR = 0.375, 95%CI = 0.231–0.609), albumin (OR = 0.327, 95%CI = 0.185–0.577), CRP level (OR = 3.339, 95%CI = 1.357–8.216), preoperative weakness (OR = 6.104, 95%CI = 6.104–2.375), surgical approach (OR = 3.182, 95%CI = 1.259–8.042), anesthesia method (OR = 4.456, 95%CI = 1.884–10.537), intraoperative sedation depth (OR = 4.805, 95%CI = 1.921–12.018) and postoperative use of analgesia pump (OR = 3.971, 95%CI = 1.536–10.264) were the influencing factors of the occurrence of PND in elderly patients treated with hip fracture surgery. The results of ROC curve and DCA showed that the AUC of prediction model was 0.917, the predictive model has good predictive efficiency and significant positive net benefit. It had good clinical utility in predicting the risk of PND occurrence.
      Conclusions The age, number of preoperative comorbidities, MMSE score, albumin, CRP level, preoperative weakness, surgical approach, anesthesia method, intraoperative sedation depth and postoperative use of analgesia pump are the influencing factors of the occurrence of PND in elderly patients treated with hip fracture surgery. Clinical measures can be developed to further improve the prognosis.

       

    /

    返回文章
    返回