【摘要】 目的 比较不同生长分化因子15（GDF-15）变化轨迹对老年非ST段抬高型急性冠状动脉综合征（NSTE-ACS）患者随访3年发生不良心血管事件（MACE）风险的预测价值。方法 回顾性队列研究。纳入2015年1月至2020年1月入住丽水市中心医院心内科的1 522例老年NSTE-ACS患者。利用潜在混合模型识别暴露期间（2016—2023年）的GDF-15水平确定了五种不同变化轨迹：低水平-稳定（Q₁组，620～1 324 ng/L,248例）、中低水平-稳定（Q₂组，1 325～1 576 ng/L,366例）、中高水平-稳定（Q₃组，1 577～1 783 ng/L,559例）、高水平-稳定（Q₄组，1 784～1 991 ng/L,274例）和高水平-升高（Q₅组，>1 991 ng/L,75例）。结局指标是MACE,包括死亡、心力衰竭入院或室性心律失常（需要ICD治疗的室性心动过速/心室颤动）的复合事件。比较随访3年的MACE风险，Cox回归分析MACE的危险因素。应用受试者工作特征（ROC）曲线比较基线GDF-15、GDF-15变化轨迹、肌钙蛋白I（TNI）和N末端B型利钠肽原（NT-proBNP）预测MACE的价值。结果 1 522例老年NSTE-ACS患者中，男性805例（52.9%）,年龄60～87岁，平均（72.2±7.5）岁。五组的年龄、饮酒史、吸烟史、体质指数、糖尿病、心房颤动、非ST抬高型心肌梗死的患病率以及血红蛋白、空腹血糖、TNI、NT-proBNP、GDF-15、血尿酸、左室射血分数、收缩末期容积、SYNTAX评分等的差异均有统计学意义（均为P<0.05）。五组的利尿剂应用率和急诊经皮冠状动脉介入治疗比例等也存在统计学差异（均为P<0.05）。中位随访35个月期间，共发生206例（13.5%）MACE,包括104例（6.8%）心力衰竭入院、37例（2.4%）死亡和65例（4.3%）室性心律失常。混杂变量校正后，GDF-15变化轨迹与心力衰竭入院风险（HR=1.594,95%CI:1.035～2.743）和死亡风险（HR=1.152,95%CI:1.007～1.441）均显著正相关。多因素Cox回归分析结果提示，GDF-15变化轨迹（HR=1.484）、年龄（HR=1.136）、吸烟史（HR=1.285）、陈旧心肌梗死（HR=1.247）、NT-proBNP（HR=1.425）、TNI（HR=1.293）、左室质量指数（HR=1.173）和血管紧张素转换酶抑制剂/血管紧张素Ⅱ受体拮抗剂治疗（HR=0.874）是发生MACE的相关因素。校正其他因素后，GDF-15变化轨迹仍与发生MACE显著相关（HR=1.264）。ROC曲线显示GDF-15变化轨迹对MACE的预测价值较高（AUC=0.786）,高于入院时NT-proBNP（AUC=0.771）、入院时GDF-15（AUC=0.755）和入院时TNI（AUC=0.739）。结论 GDF-15变化轨迹与老年NSTE-ACS患者随访发生MACE风险明显相关，提示连续测量GDF-15可优化老年NSTE-ACS患者的风险分层，具有临床应用价值。更多还原

【Abstract】 Objective To compare the predictive value of different growth differentiation factor（GDF）-15 trajectories on the risk of adverse cardiovascular events（MACE） in elderly non-ST-segment elevation acute coronary syndrome（NSTE-ACS） patients who were followed up for 3 years. Methods This was a retrospective study. A total of 1 522 elderly NSTE-ACS patients admitted to Cardiology Department of Lishui Municipal Central Hospital from January 2015 to January 2020 were included. Five different GDF trajectories were identified by using a latent mixed model to identify GDF-15 levels during exposure time（2015—2020）, including the low stable group（Q₁ group, n=248, 620-1 324 ng/L）, the medium low stable group（Q₂ group, n=366, 1 325-1 576 ng/L）, the medium high stable group（Q₃ group, n=559, 1 577-1 783 ng/L）, the high stable group（Q₄ group, n=274, 1 784-1 991 ng/L）, and the high elevated group（Q₅ group, n=75, >1 992 ng/L）. The primary outcome was MACE, which included death, hospitalization for heart failure（HHF）, or ventricular arrhythmias（ventricular tachycardia/fibrillation requiring ICD treatment）. Baseline clinical data and the incidence of MACE of each group were compared and the risk factors for MACE were analyzed by using Cox regression analysis. Receiver operating characteristic（ROC） curve and the area under the curve（AUC） was used to compare the predictive value of baseline GDF-15, GDF-15 trajectories, troponin I（TNI） and N-terminal pro-B-type natriuretic peptide（NT-proBNP） for MACE. Results Among the 1 522 elderly NSTE-ACS patients, there were 805 males（52.9%）. The average age was（72.2 ± 7.5） years with a range of 60-87 years.There were significant differences in age, drinking history, smoking history, body mass index, the ratio of diabetes, atrial fibrillation, non-ST elevation myocardial infarction, levels of hemoglobin, fasting blood glucose, TNI, NT-proBNP, baseline GDF-15, blood uric acid, left ventricular ejection fraction, end-systolic volume, SYNTAX score, among the five groups（all P<0.05）. Additionally, the rate of diuretic use and the proportion of emergency percutaneous coronary intervention were significantly different among groups（all P<0.05）. During a median follow-up of 35 months, 206 cases（13.5%） had MACE, including 104（6.8%） HHF, 37（2.4%） deaths and 65（4.3%） ventricular arrhythmias. After adjusting for confounding variables, there were significant correlations between the GDF-15 trajectory and the risk of HHF（HR=1.594, 95%CI: 1.035-2.743） and death（HR=1.152, 95%CI: 1.007-1.441）. Multivariate Cox regression analysis suggested that the GDF-15 trajectories（HR=1.484）, age（HR=1.136）, smoking history（HR=1.285）, old myocardial infarction（HR=1.247）, NT-proBNP（HR=1.425）, TNI（HR=1.293）, left ventricular mass index（HR=1.173）, and angiotensin converting enzyme inhibitor or angiotensin Ⅱ receptor blocker（HR=0.874） were related factors for MACE. After adjusting for other factors, a significant correlation was still found between the GDF-15 trajectories and the occurrence of MACE（HR=1.264）. The ROC curve showed that the GDF-15 trajectory was a good predictive index for MACE（AUC=0.786）, which was higher than NT-proBNP（AUC=0.771）, baseline GDF-15（AUC=0.755） and TNI（AUC=0.739）. Conclusions GDF-15 trajectory is significantly correlated with the risk of MACE in elderly NSTE-ACS patients. Continuous measurement of GDF-15 levels may help to optimize risk stratification in elderly NSTE-ACS patients and has clinical application value.更多还原