Atrial fibrillation (AF) is the most prevalent cardiac arrhythmia worldwide and is strongly associated with increased risks of stroke, heart failure, and mortality. Traditional methods to predict AF and prognostic its associated risks often fail to capture the full complexity of AF patterns, limiting their predictive accuracy. In spite of the improvements achieved by machine learning (ML) techniques, state-of-the-art AF-focused predictors do not generally incorporate longitudinal data, reducing their capacity to model the dynamic and evolving nature of individual behaviors and physiological indicators over time. The absence of a longitudinal perspective restricts understanding of how AF risk develops and changes across prognostic windows. This study addresses these limitations by developing superior ML models tailored to predict adverse events within a longitudinal Portuguese cohort of individuals with AF. The work targets six clinical endpoints: stroke, all-cause death, cardiovascular death, heart failure hospitalizations, inpatient visits, and acute coronary syndrome. The predictors yielded an AUC of 0.65 for 1- year stroke prediction, outperforming CHA2DS2-VASc (0.59). For all-cause mortality prediction, the models achieved an AUC of 0.78 against the 0.72 reference of GARFIELD-AF. In addition to predictive advances, the study identifies determinants of AF-related risks and introduces a prototype decision-support tool for clinical use.

The authors have declared no competing interest.

This work was financed by national funds from FCT - Fundacao para a Ciencia e a Tecnologia, I.P., in the scope of the project UID/04378/2025 (DOI identifier 10.54499/UID/04378/2025), and UID/PRR/04378/2025 (DOI identifier 10.54499/UID/PRR/04378/2025), of the Research Unit on Applied Molecular Biosciences - UCIBIO and the project LA/P/0140/2020 (DOI identifier 10.54499/LA/P/0140/2020) of the Associate Laboratory Institute for Health and Bioeconomy - i4HB, INESC-ID plurianual (UID/50021/2025); and the project FRAIL (2024.07266.IACDC).

I confirm all relevant ethical guidelines have been followed, and any necessary IRB and/or ethics committee approvals have been obtained.

Yes

The details of the IRB/oversight body that provided approval or exemption for the research described are given below:

Ethical approval granted by both the Ethics Committee of Unidade Local de Saude de Matosinhos (ULSM) and Data Protection Officer at ULSM for the undertaken analysis. The original data was de-identified according to HIPAA Safe Harbour Method.

I confirm that all necessary patient/participant consent has been obtained and the appropriate institutional forms have been archived, and that any patient/participant/sample identifiers included were not known to anyone (e.g., hospital staff, patients or participants themselves) outside the research group so cannot be used to identify individuals.

Yes

I understand that all clinical trials and any other prospective interventional studies must be registered with an ICMJE-approved registry, such as ClinicalTrials.gov. I confirm that any such study reported in the manuscript has been registered and the trial registration ID is provided (note: if posting a prospective study registered retrospectively, please provide a statement in the trial ID field explaining why the study was not registered in advance).

Yes

I have followed all appropriate research reporting guidelines, such as any relevant EQUATOR Network research reporting checklist(s) and other pertinent material, if applicable.

Yes

Access to all data in the present study should be requested to Unidade Local de Saude de Matosinhos (ULSM), together with a description of the target aims and any additional documentation deemed relevant for ethics approval.