营养炎症风险评分及其列线图模型对老年心力衰竭患者30天死亡率的预测价值

Construction and validation of a nomogram model for predicting heart failure in coronary heart disease patients

目的:心力衰竭(heart failure,HF)是住院患者死亡的重要原因。本研究旨在探究营养炎症风险评分(nutrition-inflammation risk score,NIRS)及其列线图模型对老年心力衰竭患者30d死亡率的预测价值。方法:本研究回顾性分析2018年1月至2020年6月,青岛市胶州中心医院收治的老年HF患者。根据患者30d的存活情况,分为死亡组和存活组。采用单因素分析和多因素Logistic回归构建NIRS和死亡预测模型。采用ROC分析、校准曲线和决策曲线评估老年HF患者死亡预测列线图模型的预测能力、校准能力和临床净获益。结果:本研究共纳入797例老年HF患者,164例(20.6%)老年HF患者在30d内死亡。NIRS由预后营养指数(prognosis nutrition index,PNI)、HGB-白蛋白-淋巴细胞-血小板(hemoglobin,albumin,lymphocyte,platelet,HALP)评分、单核细胞-高密度脂蛋白比值(monocyte to high density lipoprotein ratio,MHR)和CRP-白蛋白比值(C-reaction protein to albumin ratio,CAR)组成。多因素Logstic回归结果表明,NIRS(OR=11.867,95%CI:7.681~18.333,P<0.001)、高血压(OR=1.935,95%CI:1.18~3.175,P<0.001)和慢性阻塞性肺疾病(OR=4.306,95%CI:2.611~7.1,P<0.001)是老年HF患者30d死亡的危险因素。此外,老年HF患者30d死亡列线图模型ACU为0.855;校准曲线显示该模型预测概率与实际概率基本吻合;决策曲线显示该模型净获益良好。结论:INRS是老年HF患者30d死亡的独立预测因素。此外,老年HF患者30d死亡列线图模型可个体化预测老年HF患者30d内的死亡风险,帮助临床医生早期识别死亡高风险个体。

Objective:Heart failure(HF)is an important cause of death in hospitalized patients.This study aims to explore the predictive value of the nutritional inflammation risk score(NIRS)and its nomogram model for 30-day mortality in elderly patients with heart failure.Methods:Elderly HF patients were admitted to Qingdao Jiaozhou Central Hospital from January 2018 to June 2020.According to the patient's30day survival status,they are divided into a death group and a survival group.Using univariate analysis and multivariate Logistic regression to construct NIRS and death prediction models.Using ROC analysis,calibration curves,and decision curves to evaluate the predictive ability,calibration ability,and clinical net benefit of the mortality prediction column chart model for elderly HF patients.Results:This study included a total of 797elderly HF patients,of which 164(20.6%)died within 30days.NIRS is composed of prognosis nutrition index(PNI),hemoglobin-albumin-lymphocytes-platelets(HALP)score,monocyte to high density lipoprotein ratio(MHR),and C-reaction protein to albumin ratio(CAR)The multiple Logistic regression results showed that NIRS(OR=11.867,95%CI:7.681-18.333,P<0.001),hypertension(OR=1.935,95%CI:1.18-3.175,P<0.001),and chronic obstructive pulmonary disease(OR=4.306,95%CI:2.611-7.1,P<0.001)were independent risk factors for30day mortality in elderly HF patients.In addition,the ACU of the 30d mortality column chart model for elderly HF patients was 0.855;The calibration curve shows that the predicted probability of the model is basically consistent with the actual probability;The decision curve shows that the model has a good net benefit.Conclusions:INRS is an independent predictor of 30d mortality in elderly HF patients.In addition,the 30d mortality column chart model for elderly HF patients can personalize the preadiction of mortality risk within3odays,helping clinicians identifay high-risk individuals for death early on.