Abstract
Managing atrial fibrillation (AF) patients with multiple comorbidities and complex medications is challenging. This study aimed to identify different patient profiles of AF based on comorbidities and medicines combinations and to explore their associations with the risk of adverse outcomes. This cohort study included patients with AF from the anticoagulant outpatient clinic at a hospital in Italy, undergoing follow-ups every 6 months. Comorbidity and medication patterns were identified using latent class analysis. Cox regression was used to explore associations with thromboembolism, major bleeding, falls, and death-separately and composite. A total of 633 patients with AF (mean age 80.5 ± 6.9 years, 52.5% women) treated with direct oral anticoagulants were followed for a median of 24.2 (IQR 12.1-35.5) months. Four patterns were identified: unspecific pattern (39.0%), diabetes and liver pattern (14.8%), neurocognitive and psychiatric pattern (14.1%), and musculoskeletal, immunologic and dermatologic pattern (32.1%). After adjustments, the neurocognitive and psychiatric pattern was associated with a higher risk of the composite outcome (hazard ratio [HR] [95% CI]: 1.75 [1.56-3.82]), thromboembolism (HR: 3.04 [1.28-7.22]) and major bleeding (HR: 2.55 [1.05-6.22]) compared to the non-specific pattern. The musculoskeletal, immunologic, and dermatologic pattern was also associated with a higher bleeding risk (HR: 2.21 [1.05-4.65]). Stratified analyses showed that these links were stronger in patients without cancer, and there was significant interaction in bleeding risk based on cancer status (p = 0.014). Anticoagulated AF patients with the neurocognitive and psychiatric profile are at higher risk, emphasizing the need for holistic AF management.