急性冠状动脉综合征患者经皮冠状动脉介入治疗术后抗凝治疗与院内血红蛋白下降风险及其影响因素

Risk factors for an in-hospital decrease in hemoglobin concentration in patients with acute coronary syndrome treated with percutaneous coronary intervention

目的分析急性冠状动脉综合征(acute coronary syndrome,ACS)患者接受经皮冠状动脉介入治疗(percutaneous coronary intervention,PCI)术后抗凝治疗对院内血红蛋白(hemoglobin,Hb)下降的风险及其影响因素,建立抗凝治疗相关院内出血预测风险模型。研究对象连续纳入2014年7月至2018年12月国家电子疾病监测系统数据库提取的接受PCI治疗的ACS患者。干预措施是否在PCI术后进行抗凝治疗。观测指标及测量方法纳入人口学、病史和院内治疗相关的临床变量共34个作为协变量,采用多变量校正的logistic回归比较ACS患者PCI术后抗凝治疗与未抗凝治疗组Hb下降≥5 g/L的风险;将住院期间院内用药情况进行倾向性匹配,利用逐步回归混合法筛选出有意义的基线和院前危险因素,建立抗凝相关院内出血预测风险模型。结果在184850例接受PCI治疗的ACS患者中,具备在抗凝前后进行血常规检测,且研究协变量无缺失者共11853例,其中抗凝治疗组6998例,未抗凝治疗组4855例;发生PCI术抗凝治疗后院内Hb下降≥5 g/L患者共99例,其中抗凝治疗组81例(81.1%),未抗凝治疗组18例(18.2%),在进行多因素校正后抗凝治疗组发生Hb下降≥5 g/L的比值比(odds ratio,OR)为2.52[95%置信区间(confidence interval,CI)为1.44~4.41,P=0.001]。在将院内用药进行倾向性匹配的基础上,利用双向逐步回归共筛选出性别、入院Hb水平、Killip分级、估算肾小球滤过率(estimated glomerular filtration rate,eGFR)、PCI史、院前P2Y12抑制剂、高血压病史及PCI术后抗凝共8个危险因素建立抗凝相关出血模型。抗凝相关院内出血模型与CRUSADE评分比较,该模型受试者工作特征曲线下面积(area under curve,AUC)为0.757(95%CI为0.645~0.895),优于CRUSADE评分(ΔAUC=0.120,P<0.001)。结论ACS患者PCI术后抗凝治疗与院内Hb降低≥5 g/L风险增加1.5倍有关,本研究探索的基于危险因素的抗凝相关院内出血预测风险模型有助于早期识别高危患者,指导抗凝方案的选择。

Objective The aims of this study were to identify the risk factors for an in‑hospital decrease in hemoglobin(Hb)concentration in patients with acute coronary syndrome(ACS)after percutaneous coronary intervention(PCI)with anticoagulation strategies and to establish a predictive risk model for anticoagulation‑related in‑hospital bleeding for the clinical assessment of bleeding risk.Subjects We extracted information for ACS patients treated with PCI from the National Electronic Disease Surveillance System database from July 2014 to December 2018.Interventions Patients with post‑PCI anticoagulation therapy were included in the study.Main Outcomes and Measurements Multivariable logistic regression model was adopted to evaluate the risk of Hb decline(≥5 g/L)between with or without anticoagulant treatment after PCI in ACS patients.There were 34 clinical indicators for demography,medical history and hospital therapy treated as covariates.Based on the propensity score matching for in‑hospital medication,the risk prediction model of anticoagulation‑related in‑hospital bleeding was established using the meaningful risk factors,which screened by stepwise regression.Results Data were extracted from 184850 patients.After excluding those with missing values for key covariates and those without routine blood test results before and after post‑PCI anticoagulation therapy,11853 patients with ACS who underwent PCI were included in the analysis.Ninety‑nine patients,including 81 patients(81.1%)in the anticoagulation group,were found to have a decrease in Hb concentration≥5 g/L during hospitalization.After adjustment for multiple variables,the odds ratio(OR)was 2.52[95%confidence interval(CI):1.44~4.41,P=0.001].We further performed propensity score matching of in‑hospital medications to identify the risk factors associated with the decrease in Hb concentration after anticoagulation therapy.Eight risk factors,including sex,Hb concentration at admission,Estimated glomerular filtration rate(eGFR),Killip classification,a history of PCI,pre‑admission P2Y12 inhibitor treatment,a history of hypertension,and postoperative anticoagulation treatment were screened out using bidirectional stepwise regression.The resulting model for the risk of an in‑hospital decrease in Hb concentration≥5 g/L yielded an area under the curve(AUC)of 0.757(95%CI:0.645-0.895),which was significantly higher than the AUC achieved using the CRUSADE score(ΔAUC=0.120,P<0.001).Conclusions Anticoagulation therapy after PCI in ACS patients was associated with a 1.5-fold increase in the risk of an in‑hospital decrease in Hb concentration≥5 g/L.Risk‑factor‑based anticoagulation‑related in‑hospital bleeding prediction models may facilitate the early identification of high‑risk patients and provide personalized guidance for post‑PCI anticoagulation therapy.