衰弱评估工具在老年心血管病患者5年全因死亡率预测中的应用价值

A study on the application value of frailty assessment tools in predicting 5-year all-cause mortality rate in elderly cardiovascular disease patients

目的比较Fried衰弱表型与基于老年综合评估的衰弱指数(CGA-FI)预测老年心血管病患者5年全因死亡率的价值,并探索优化预测方法。方法采用前瞻性队列研究设计,纳入2018年9月至2019年2月于北京医院心内科住院治疗的老年心血管病患者,通过随访记录主要不良事件。应用Fried衰弱表型与CGA-FI进行衰弱评估;利用受试者工作特征曲线(ROC)评价不同评估方法的预测效能;通过LASSO回归筛选预测因子,并进一步构建logistic回归预测模型;应用Bootstrap法进行内部验证。结果研究最终纳入420例患者,年龄为73.8(68.9,79.1)岁,男性223例(53.0%)。随访时间为1879(1833,1931)d,73例(17.4%)患者死亡。Fried衰弱表型和CGA-FI评估为衰弱的患者分别为107例(25.5%)和113例(26.9%)。Fried衰弱表型与CGA-FI评估为衰弱的患者5年全因死亡率显著高于非衰弱患者(Log-rank P<0.001)。Fried衰弱表型的ROC曲线下面积(AUC)为0.735(95%CI:0.667~0.804),CGA-FI的AUC为0.777(95%CI:0.713~0.840)。通过LASSO回归从CGA-FI中筛选出6个预测因子,构建的预测模型AUC达到0.882(95%CI:0.835~0.929)。经1000次Bootstrap法验证的AUC平均值为0.871(95%CI:0.866~0.877)。结论衰弱是老年心血管病患者全因死亡的独立预测因子。Fried衰弱表型和CGA-FI均具备预测价值,而新构建的预测模型能更准确地预测5年全因死亡风险。

Objective To compare the prognostic value of the Fried frailty phenotype and the frailty index based on the Comprehensive Geriatric Assessment(CGA-FI)in predicting 5-year all-cause mortality among elderly patients with cardiovascular disease,as well as to explore methods for optimizing predictive models.Methods A prospective cohort study design was employed,involving elderly patients with cardiovascular disease admitted to the cardiology department of Beijing Hospital from September 2018 to February 2019,with follow-up to record major adverse events.Frailty assessments were conducted using the Fried frailty phenotype and CGA-FI.Receiver operating characteristic(ROC)curves were utilized to evaluate the predictive performance of different assessment methods.Predictive factors were selected using LASSO regression,and logistic regression was subsequently employed to construct predictive models.Applying Bootstrap method for internal validation.Results A total of 420 patients were included in the study,with a median age of 73.8(68.9,79.1)years,of which 223(53.0%)were male.The median follow-up duration was 1879(1833,1931)days,during which 73 patients(17.4%)died.Of these,107(25.5%)and 113(26.9%)patients were identified as frail according to the Fried frailty phenotype and CGA-FI,respectively.Frail patients identified by both methods had significantly higher 5-year all-cause mortality rates compared to non-frail patients(Log-rank P<0.001).The area under the curve(AUC)for predicting mortality was 0.735(95%CI:0.667-0.804)for the Fried frailty phenotype and 0.777(95%CI:0.713-0.840)for CGA-FI.A predictive model constructed using LASSO regression with 6 selected factors achieved an AUC of 0.882(95%CI:0.835-0.929).The average AUC validated by Bootstrap method is 0.871(95%CI:0.866-0.877).Conclusions Frailty is an independent predictor of all-cause mortality in elderly patients with cardiovascular disease.Both the Fried frailty phenotype and CGA-FI all have predictive value,and the newly constructed predictive model provides more accurate prediction of 5-year all-cause mortality risk.