Atrial fibrillation and excess risk of cardiovascular and non-cardiovascular diseases: population-based linked cohort study
L Mu, R Nadarajah, C Gale, J W WuAbstract
Background
Atrial fibrillation (AF) is common and associated with excess risks of stroke, but the long-term prognosis for cardiovascular diseases (CVDs) and non-CVDs after incident AF, particularly when accounting for competing mortality, is not well quantified in routine care.
Purpose
We aimed to estimate the incidence, absolute risks, and competing-risk adjusted associations of 16 cardiovascular and non-cardiovascular outcomes after incident AF and to examine heterogeneity by age, sex, and deprivation.
Methods
Population-based matched cohort study using CPRD primary care linked electronic health records in England (1998–2021). Individuals aged ≥16 years with incident AF were risk-set matched 1:1 to controls without AF on birth year and sex. The primary endpoint for each outcome was the first recorded of this specific outcome in any of the data sources, occurring after the incident AF diagnosis date or the matched index date for controls. We estimated crude and standardised incidence rates, cumulative incidence using Aalen-Johansen estimators, and subdistribution hazard ratios (sHRs) using Fine–Gray models.
Results
We included 173,519 individuals with incident AF and 173,519 matched controls (mean age 75.6 years; 48.7% women). Over a median follow-up of 2.9 (IQR 1.0–6.0) years, mortality was higher in AF than controls (76,412 vs 34,555 deaths; crude mortality rate 11,692 vs 4,691 per 100,000 person-years; IRR 2.49, 95% CI 2.46–2.52). At 10 years after AF, cumulative incidence was 61.68% (95% CI 61.34–62.03) for all-cause mortality, 35.08% (34.73–35.43) for chronic kidney disease, and 29.07% (28.74–29.40) for heart failure. In competing-risk models, the largest excess hazards were for heart failure (sHR 2.37, 95% CI 2.32–2.42), valvular heart disease (2.32, 2.25–2.40) and stroke (1.48, 1.44–1.52); thyroid disease (1.69, 1.62–1.76) and obstructive sleep apnoea (1.74, 1.57–1.92) showed the largest excess risks among non-cardiovascular outcomes. Risks varied strongly by age, sex, and deprivation.
Conclusions
Incident atrial fibrillation is associated with substantial excess mortality and multisystem morbidity. Beyond stroke, heart failure and valvular heart disease confer the largest excess risks after accounting for competing mortality. Following diagnosis, obstructive sleep apnoea and thyroid disease occur more frequently than in matched controls. The patterns of excess risk across cardiovascular and non-cardiovascular conditions differ by age, sex, and socioeconomic deprivation, with important implications for clinical risk assessment, treatment, prevention, and trial design.Incident rate plotHeat map of CVD in AF