Identificador: TDX:4500
Autores: Moltó Balado, Pedro
Resumen:
INTRODUCTION: Atrial fibrillation (AF) is a cardiac arrhythmia with increasing prevalence and is associated with an increased risk of major adverse cardiovascular events (MACE), posing challenges for early identification and treatment. OBJECTIVES: a) To identify the incidence of AF and episodes of MACE b) To determine the characteristics of patients at risk of developing AF c) To define predictors of MACE in patients with newly diagnosed AF using IA. METHODOLOGY: Multicentre, observational, retrospective, community-based, multicentre study of a cohort (n=40,297) aged 65-95 years, between 1 January 2015 and 31 December 2021, with no previous diagnosis of AF or MACE. Five ML models were developed to determine predictors of MACE in AF patients. RESULTS: A total of 2,574 persons (6.39%) developed a first episode of AF with an overall incidence of 8.9/1,000 person-years (95%CI: 8.6-9.2). The incidence of MACE among AF patients was 75.1/1,000 person-years (95%CI 70.8-79.5), compared to 20.6/1,000 person-years (95%CI 20.2-21.1) without AF with a rate ratio of 3.65 (95%CI 3.43-3.88). Higher malnutrition was detected among patients with MACE (49.7% vs. 26.6%). AdaBoost performed best (accuracy: 0.9999; recall: 1; F1 score: 0.9997). The model estimated ICC, cancer, diabetes mellitus, COPD, cognitive impairment, vascular disease, CHA2DS2-VASc and Wells as the main predictors of MACE. CONCLUSION: Patients in the 4th quartile (Q4) at risk of developing AF had a higher cardiovascular risk prior to AF diagnosis, especially in CKD, ischaemic heart disease and peripheral arterial disease. A diagnosis of AF increases the incidence of heart failure fourfold and MACE eightfold. AdaBoost offers a more accurate predictive approach for early identification of MACE risk in the evaluation of AF patients.
Repositori URV