|
| 心脏手术后谵妄的危险因素分析及预测模型构建 |
| Risk factors analysis and prediction model construction of postoperative delirium after cardiac surgery |
| |
| DOI：10.12089/jca.2020.12.012 |
| 中文关键词: 心脏手术 术后谵妄 危险因素 预测模型 |
| 英文关键词: Cardiac surgery Postoperative delirium Risk factors Predictive modeling |
| 基金项目:南京市卫生科技发展项目（YKK18108） |
|
| 摘要点击次数: 2546 |
| 全文下载次数: 659 |
| 中文摘要: |
目的 探讨心脏手术患者术后谵妄(POD)的危险因素，根据危险因素构建预测模型并评估其预测效能。
方法 本研究为回顾性病例-对照研究。收集2016年5月至2019年5月行心脏手术治疗的患者3 397例临床资料，提取性别、年龄、吸烟史、饮酒史、高血压病史、糖尿病史、高脂血症病史、脑血管疾病史、手术时间、心肺转流（CPB）时间、主动脉阻断时间、术中连续有创动脉血压值、术中出血量、术中麻醉药物使用情况、入ICU时Hb、白细胞计数、氧合指数（PaO2/FiO2）、血乳酸值、呼吸机使用时间、血管活性药物使用情况、术后左心室射血指数（LVEF）等临床指标，采用单因素分析评估上述指标与POD的相关性。采用逐步Logistic回归进行临床指标筛选并构建预测模型，绘制模型列线图，计算预测模型的受试者工作特征（ROC）曲线下面积（AUC）以及最佳阈值下的敏感性和特异性。
结果 有186例（5.48%）患者术后出现谵妄。与非谵妄患者比较，谵妄患者年龄、脑血管疾病和高血压病史比例、术中低血压比例、术中血压变异率、术中丙泊酚用量、术中出血量、术后乳酸值、术后白细胞显著增高的比例、术后左心室收缩功能不全（LVEF<50%）的比例以及术后应用肾上腺素和去甲肾上腺素比例明显增加（P<0.05），手术时间和CPB时间、术中低血压时间以及术后机械通气时间明显延长（P<0.05），术后氧合指数和Hb明显降低（P<0.05）。逐步Logistic回归分析显示，老年、脑血管疾病史、手术时间长、术中血压变异率大、术后Hb<100 g/L、机械通气时间延长以及左心室收缩功能不全、应用去甲肾上腺素是心脏手术后谵妄的独立危险因素。应用这些危险因素构建预测模型，AUC为0.932（95%CI 0.897～0.967），敏感性78.2%，特异性93.5%。
结论 老年、脑血管疾病史、手术时间、术中血压变异率、术后Hb<100 g/L、机械通气时间延长以及左心室收缩功能不全、应用去甲肾上腺素是心脏手术后谵妄的独立危险因素，基于这些危险因素构建的模型可较好地预测心脏手术后谵妄的发生，为早期干预治疗提供参考。 |
| 英文摘要: |
Objective To investigate the risk factors of delirium in patients after cardiac surgery and to develop a prediction model based on the risk factors, and to evaluate its predictive efficacy with the purpose of early diagnosis and early treatment.
Methods A retrospective analysis of 3 397 patients who underwent cardiac surgery in Nanjing First Hospital from May 2016 to May 2019 were performed. Clinical variables including age, gender, smoking history, drinking history, hypertension history, diabetes history, history of hyperlipidemia and cerebrovascular disease, operation time, cardiopulmonary bypass (CPB) time, aortic occlusion time, intraoperative continuous invasive arterial blood pressure, blood loss and anesthetic agent usage, postoperative ICU hemoglobin, white blood cell count, oxygenation index (PaO2/FiO2), blood lactate, use of vasoactive medicine, and postoperative left ventricular ejection fraction(LVEF) were collected and their correlation with postoperative delirium were analyzed. Stepwise Logistic regression was used to screen clinical indicators and develop a prediction model which was visualized by nomogram. The area under the receiver operating characteristic curve (AUC) of the prediction model and the sensitivity and specificity under the optimal threshold were calculated for model evaluation.
Results A total of 186 (5.48%) patients had postoperative delirium. Univariate analysis revealed that, compared with the non-delirium patients, the delirium patients were significantly older, and had a higher proportion of cerebrovascular disease, hypertension history, intraoperative hypotension, the application of epinephrine or noradrenaline, and left ventricular systolic dysfunction (LVEF < 50%) after surgery (P < 0.05). The delirium patients also had larger intraoperative blood pressure variance and lactic acid value, more intraoperative blood loss and dosage of propofol, lower postoperative Hb and oxygenation index, greater percentage of postoperative white blood cells growth, longer operation time and CPB time, and longer intraoperative hypotension time and postoperative mechanical ventilation time (P < 0.05). Stepwise Logistic regression showed that the elderly, history of cerebrovascular disease, operation time, intraoperative blood pressure variance, postoperative hemoglobin < 100 g/L, prolonged mechanical ventilation time, left ventricular systolic dysfunction, and norepinephrine application are independent risk factors for delirium after cardiac surgery. The AUC of the predictive model constructed using these risk factors was 0.932 (95% CI 0.897 - 0.967), the sensitivity and the specificity was 78.2% and 93.5%, respectively.
Conclusion Elderly, history of cerebrovascular disease, operation time, intraoperative blood pressure variance, postoperative hemoglobin <100 g/L, prolonged mechanical ventilation time, left ventricular systolic dysfunction, and norepinephrine application are independent risk factors for delirium after cardiac surgery. Predictive model constructed with those risk factors could predict the occurrence of delirium after cardiac surgery and provide the possibility for early intervention. |
|
查看全文
查看/发表评论 下载PDF阅读器 |
| 关闭 |
|
|
|
