Objective To systematicly evaluate the risk prediction models for recurrence after radiofrequency catheter ablation in patients with atrial fibrillation and to provide a basis for the clinical treatment of AF. Methods We retrieved the relevant literature on “Risk prediction models for recurrence in AF patients after RFCA” from the CNKI, CSPD, CCD, CBM, PubMed, CINAHL, Web of Science, Embase, and Cochrane Library databases. The search period was from the establishment of the database to May 11,2...更多
Objective To systematicly evaluate the risk prediction models for recurrence after radiofrequency catheter ablation in patients with atrial fibrillation and to provide a basis for the clinical treatment of AF. Methods We retrieved the relevant literature on “Risk prediction models for recurrence in AF patients after RFCA” from the CNKI, CSPD, CCD, CBM, PubMed, CINAHL, Web of Science, Embase, and Cochrane Library databases. The search period was from the establishment of the database to May 11,2024. We extracted the literature data and used predictive model bias risk assessment tools to evaluate the bias risk and applicability of the included models. Results Nineteen articles were included, which were published between 2021 and 2024, and all originated from China. Eleven articles were on model establishment, and 8 articles were on model establishment and validation. Totally 19 models were included, with a total sample size of 136-6,127 cases and a recurrence rate of 8. 70%-48. 57%. Thirteen articles constructed models using Logistic regression analysis,3 articles constructed models using Cox regression, and 1 article constructed models using Lasso regression algorithm and random forest algorithm. Nineteen literature reported the area under the curve, ranging from 0. 723 to 0. 938. Left atrial diameter, age, type of atrial fibrillation, duration of atrial fibrillation, sex, and BMI were the top 6 predictors of repeated reporting of the models. One article presented as“unclear” in the risk of bias assessment while the remaining 18 articles were “high risk of bias”. The applicability of 5 articles was “unclear”, while the applicability risk of the remaining 14 articles is relatively low. Conclusions The research on recurrence risk prediction models for AF patients after RFCA started relatively late, and most modeling methods were based on Logistic regression analysis to construct nomogram models. The 19 included models had a relatively high area under the curve, and the predictive factors could be screened from clinical data, imaging data, and electrophysiological examinations. The overall risk of bias was high.收起
