2022 年第 3 期 第 17 卷

术前血清N末端B型脑钠肽前体和心肌肌钙蛋白I水平对老年重症心脏瓣膜病患者术后发生低心排血量综合征的预测价值

Predictive value of preoperative N-terminal pro-brain natriuretie peptide and cardiac troponin I in the occurrence of postoperative low cardiac output syndrome in elderly patients with severe heart valve diseases

作者：宋跃曹剑刘冬

英文作者：Song Yue Cao Jian Liu Dong

单位：首都医科大学附属北京安贞医院心脏外科，北京100029

英文单位：Department of Cardiac Surgery Beijing Anzhen Hospital Capital Medical University Beijing 100029 China

关键词：老年重症心脏瓣膜病；低心排血量综合征；N末端B型脑钠肽前体；心肌肌钙蛋白I

英文关键词：Senilesevereheartvalvediseases；Lowcardiacoutputsyndrome；N-terminalpro-brainnatriuretiepeptide；CardiactroponinI

• 摘要：

• 目的 探讨术前血清N末端B型脑钠肽前体(NT-proBNP)和心肌肌钙蛋白I(cTnI)水平对老年重症心脏瓣膜病患者术后发生低心排血量综合征（LCOS）的预测价值。方法  选取2018年2月至2020年1月于首都医科大学附属北京安贞医院行手术治疗的老年重症心脏瓣膜病患者202例。收集患者一般临床资料、心脏生化指标。出院后对患者进行3年随访，按照是否发生LCOS，将患者分为发病组（47例）和对照组（155例）。采用Logistic多因素回归方法分析患者术后发生LCOS的独立危险因素，通过受试者工作特征（ROC）曲线确定术前NT-proBNP和cTnI水平的最佳截断点，依据患者术后的独立危险因素建立列线图预测模型并进行验证。结果 多因素Logistic 回归分析结果显示，术前心律失常、既往心肌梗死、术前NT-proBNP水平高、术前cTnI水平高、体外循环时间长、重症监护时间长、术后NT-proBNP水平高、术后cTnI水平高为老年重症心脏瓣膜病患者术后发生LCOS的独立危险因素（均P＜0.05）。术前NT-proBNP水平预测老年重症心脏瓣膜病患者术后发生LCOS的ROC曲线下面积为0.742，截断点为7 117.80 ng/L，敏感度为72.60%，特异度为74.30%；术前cTnI水平预测老年重症心脏瓣膜病患者术后发生LCOS的ROC曲线下面积为0.808，截断点为0.73 μg/L，敏感度为81.00%，特异度为78.40%。列线图模型调查显示，术前心律失常、既往心肌梗死、术前NT-proBNP水平高、术前cTnI水平高、体外循环时间长、重症监护时间长、术后NT-proBNP水平高、术后cTnI水平高是术后发生LCOS的独立危险因素，评分分别为31、25、21、11、18、16、30、16分，总评分为168分，对应LCOS术后发生概率为33.6%。该模型一致性指数为0.773，ROC曲线下面积为0.851，具有较好的区分度；校准曲线评价结果显示该模型具有较高的一致性。结论  术前血清NT-proBNP和cTnI水平高是老年重症心脏瓣膜病患者术后发生LCOS的独立危险因素。

• Objective   To explore the predictive value of preoperative N-terminal pro-brain natriuretie peptide (NT-proBNP) and cardiac troponin I (cTnI) for postoperative low cardiac output syndrome (LCOS) in elderly patients with severe heart valve diseases. Methods   A total of 202 elderly patients with severe heart valve diseases who were treated in Beijing Anzhen Hospital, Capital Medical University from February 2018 to January 2020 were selected. General clinical data and cardiac biochemical indicators of the patients were collected. After discharged from the hospital, patients were followed-up for 3 years, and were divided into onset group (n=47) and control group (n=155) according to whether or not LCOS occurred after operation. Multivariate Logistic regression analysis was used to analyze the independent risk factors for postoperative LCOS, and the receiver operating characteristic curve (ROC) was used to determine the optimal cut-off point of NT-proBNP and cTnI levels before surgery. Based on the patient′s independent risk factors after surgery, a nomogram prediction model for risk factors was established and verified. Results  Multivariate Logistic regression analysis showed that preo-perative arrhythmia, previous myocardial infarction, high preoperative NT-proBNP level, high preoperative cTnI content, long cardiopulmonary bypass time, long intensive care time, high postoperative NT-proBNP level and high postoperative cTnI level were independent risk factors for postoperative LCOS in elderly patients with severe heart valve diseases（all P＜0.05）. The area under the ROC curve for preoperative NT-proBNP predicting LCOS in elderly patients with heart valve diseases after surgery was 0.742, the cut-off point was 7 117.80 ng/L, the sensitivity was 72.60%, and the specificity was 74.30%; the area under the ROC curve for preoperative cTnI predicting LCOS in elderly patients with heart valve diseases after the operation was 0.808, the cut-off point was 0.73 μg/L, the sensitivity was 81.00%, and the specificity was 78.40%. The nomogram model survey showed that preoperative arrhythmia, previous myocardial infarction, high preoperative NT-proBNP level, high preoperative cTnI level, long cardiopulmonary bypass time, long intensive care time, high postoperative NT-proBNP level and postoperative cTnI level were independent risk factors for postoperative LCOS, and the scores were 31, 25, 21, 11, 18, 16, 30 and 16 respectively, with a total score of 168, the postoperative probability of corresponding LCOS as 33.6%. The consistency index of the model was 0.773 and the area under the ROC curve was 0.851, which had good discrimination. The calibration curve evaluation show that the model had high consistency. Conclusion Preoperative serum NT-proBNP and cTnI are independent risk factors for postoperative LCOS in elderly patients with severe heart valve diseases.