Characteristics and clinical outcomes in atrial fibrillation patients classified using cluster analysis: the Fushimi...

Area of Science:

• Cardiology
• Data Science
• Epidemiology

Background:

• Atrial fibrillation (AF) patient risk is typically assessed using clinical risk scores or individual risk factors.
• Multimorbidity clusters, where risk factors co-occur, may have significant prognostic implications.
• Understanding these complex comorbidity patterns is crucial for accurate risk stratification in AF.

Purpose of the Study:

• To perform cluster analysis on a cohort of AF patients.
• To identify distinct comorbidity cluster phenotypes within the AF population.
• To assess the prognostic implications and outcomes associated with these identified phenotypes.

Main Methods:

• Utilized hierarchical cluster analysis on data from the Fushimi AF Registry, a community-based prospective survey.
• Analyzed 4304 AF patients using 42 baseline clinical characteristics.
• Identified six statistically driven comorbidity clusters based on patient demographics and clinical profiles.

Main Results:

• Six distinct AF patient clusters were identified, ranging from younger individuals with low comorbidity burden to the very elderly with complex conditions.
• These clusters included: (i) young, low-risk (n=209); (ii) elderly, low-risk (n=1301); (iii) high atherosclerotic risk factors, no disease (n=1411); (iv) atherosclerotic comorbidities (n=440); (v) history of stroke (n=681); (vi) very elderly (n=262).
• Significant stratification of all-cause mortality and major adverse cardiovascular or neurological events was observed across the six clusters (P < 0.001).

Conclusions:

• Cluster analysis successfully identified six clinically relevant phenotypes in AF patients.
• These distinct phenotypes are associated with varying comorbidity profiles.
• The identified phenotypes demonstrate significant associations with the incidence of clinical outcomes, including mortality and major adverse events.