Comorbidity patterns and mortality in atrial fibrillation: a latent class analysis of the EURopean study of Older Subjects with Atrial Fibrillation (EUROSAF)

• Published in: Annals of medicine (Helsinki) Vol. 57; no. 1; p. 2454330
• Main Authors: Ng, Huah Shin, Woodman, Richard, Veronese, Nicola, Pilotto, Alberto, Mangoni, Arduino A.
• Format: Journal Article
• Language: English
• Published: England Taylor & Francis 01.12.2025; Taylor & Francis Group
• Subjects: Aged; Aged, 80 and over; anticoagulants; Anticoagulants - therapeutic use; Atrial fibrillation; Atrial Fibrillation - drug therapy; Atrial Fibrillation - epidemiology; Atrial Fibrillation - mortality; comorbidities; Comorbidity; Europe - epidemiology; Female; Geriatrics; Humans; Latent Class Analysis; Male; mortality; older adults; Phenotype; Prognosis; Proportional Hazards Models; Risk Factors; Europe; frailty; phenotypes; comorbidities; Atrial fibrillation; anticoagulants; mortality; older adults; latent class analysis; multidimensional prognostic index
• ISSN: 0785-3890; 1365-2060; 1365-2060
• DOI: 10.1080/07853890.2025.2454330

Most older patients with atrial fibrillation (AF) have comorbidities. However, it is unclear whether specific comorbidity patterns are associated with adverse outcomes. We identified comorbidity patterns and their association with mortality in multimorbid older AF patients with different multidimensional frailty. Hospitalised adults aged ≥65 years with non-valvular AF were followed for 12 months in the multicentre EURopean study of Older Subjects with Atrial Fibrillation (EUROSAF). Demographic characteristics, coexisting medical conditions, use of medications including anticoagulants, and the Multidimensional Prognostic Index (MPI) were captured on discharge. We used latent class analysis (LCA) to identify comorbidity phenotypes and Cox regression to determine associations between identified phenotypes and 12-month mortality. Amongst  = 2,019 AF patients (mean ± SD age 82.9 ± 7.5 years), a 3-class LCA solution was considered optimal for phenotyping. The model identified phenotype 1 (hypertensive, other circulatory conditions, metabolic diseases; 33%), phenotype 2 (digestive diseases, infection, injury, non-specific clinical and laboratory abnormalities; 26%), and phenotype 3 (heart failure, respiratory diseases; 41%). Overall, 512 patients (25%) died within 12 months. Compared to phenotype 1, after adjusting for age, sex, use of anticoagulants, cardiovascular medications, and proton pump inhibitors, and individual MPI domains, phenotype 3 had a significantly higher risk of mortality (adjusted hazard ratio = 1.27, 95% CI = 1.01 to 1.60). In contrast, the risk of mortality in phenotype 2 was not different to phenotype 1. We observed an association between comorbidity phenotypes identified using LCA and mortality in older AF patients. Further research is warranted to identify the mechanisms underpinning such associations.