心房颤动导管消融术后晚期复发预测评分的系统评价

A systematic review of the predictive score of late recurrence after catheter ablation of atrial fibrillation

目的 系统评价心房颤动(简称房颤)患者导管消融术后晚期复发的预测评分。方法 检索中国知网、万方数据、PubMed、EMbase、Web of Science数据库发表的房颤患者导管消融术后晚期复发预测评分的相关文献,检索时间为数据库建库至2023年9月。由研究者独立筛选、提取文献中预测评分的基本特征、建模情况、模型性能和预测因子等信息,使用偏倚风险评估工具PROBAST对预测评分进行质量评价。结果 共纳入15项研究,包含15个预测评分。Cox比例风险回归模型及logistic回归模型是研究人员常用的建模方法。年龄、性别、左房内径、房颤类型、房颤病程、肾功能不全、早期复发是模型常纳入的预测因子。有13个预测评分报告了受试者工作曲线下面积(AUC),范围为0.634~0.94,其中有7个预测评分展现了较好的预测性能(AUC>0.7),有5个预测评分进行了模型校准。PROBAST结果显示,纳入的15篇文献,有1篇为低偏倚风险,其余14篇均为高偏倚风险;在模型适用性上,均为高适用性。结论 房颤患者导管消融术后晚期复发的预测评分具有一定的预测价值,但仍需要进一步评估及验证。

Objective To systematically evaluate the predictive scores of patients with atrial fibrillation(AF)for late recurrence after catheter ablation.Methods The relevant literature on the predictive scores of late recurrence after catheter ablation in patients with atrial fibrillation published in CNKI,Wanfang Data,PubMed,EMbase,and Web of Science databases were searched from the database establishment to September 2023.The investigators inde-pendently screened and extracted the basic characteristics,modeling,model performance and predictors of the pre-diction scores in the literature,and used the bias risk assessment tool PROBAST to evaluate the quality of the pre-diction scores.Results Fifteen studies with 15 prediction scores were included.Cox proportional hazards regres-sion model and logistic regression model were commonly used modeling methods by researchers.Age,gender,left atrial diameter,type of atrial fibrillation,course of atrial fibrillation,renal insufficiency,and early recurrence were common predictors included in models.There were 13 prediction scores that reported the area under the reciver op-erating characteristic curve(AUC)in the range of 0.634-0.94,of which 7 predicted scores showed good prediction performance(AUC>0.7),and 5 prediction scores were calibrated by the model.The results of PROBAST showed that of the 15 included studies,1 was at low risk of bias,and the remaining 14 were at high risk of bias.In terms of model applicability,they were all highly applicable.Conclusions The predictive score of late recurrence after catheter ablation in patients with atrial fibrillation has certain predictive value,but further evaluation and validation are still needed.