Temporal trends and patterns in cause-specific mortality and hospitalisations after incident atrial fibrillation in England from 2001 to 2017

Abstract

Abstract Background Atrial fibrillation (AF) incidence is increasing, and in high income countries outstrips that of many cancers.(1) Following AF diagnosis reports of temporal trends of mortality are conflicting,(2-7) whilst reports of hospitalisation provide an incomplete picture of drivers of admission.(4,8) Purpose To investigate trends in cause-specific mortality and hospitalisation for patients with incident AF, including by age, sex, socioeconomic status and diagnostic care setting. Methods We used linked primary and secondary care records of 72 412 patients in England aged ≥16 years with incident AF between Jan 1, 2001 and Dec 31, 2017. The primary outcomes were all-cause and cause-specific mortality and hospitalisation at 1 year following diagnosis. Poisson regression estimated rate ratios (RRs) with 95% confidence intervals (CIs) comparing 2001/02 and 2016/17, adjusted for age, sex, region, socioeconomic status and 18 major comorbidities. Results Mean age at diagnosis was 75·6 years (SD 12·4), and 44 762 (61·8%) had ≥3 comorbidities. One-year mortality following incident AF was 20·0% and declined over the study period (RR 2016/17 vs. 2001/02, 0·72; 95% CI, 0·65–0·80) with larger declines for cardiovascular (RR 0·46; 95% CI, 0·37-0·58) and cerebrovascular mortality (RR 0·41; 95% CI, 0·29–0·60) than other causes of death (Figure 1). By 2016/17 dementia (67, 8·0%) accounted for more deaths than myocardial infarction, heart failure and acute stroke combined (56, 6·7%), and for patients aged ≥80 years death from non-cardio/cerebrovascular causes did not decline (RR 1·05; 95% CI 0·88–1·25). Mortality was higher for patients in the most deprived socioeconomic quintile compared with those in the least deprived quintile (RR 1·22; 95% CI 1·15–1·29), and for those diagnosed in hospital compared with primary care (RR 2·58; 95% CI 2·48-2·68), but similar between men and women. Hospitalisation within the first year after AF diagnosis was high (1·72 hospitalisations per patient-year at risk), with infection the most common cause and cardio/cerebrovascular diseases representing less than a fifth of all admissions (Figure 2). Hospitalisations increased (RR 1·17; 95% CI 1·13-1·22), driven by admissions for non-cardio/cerebrovascular causes (RR 1·42; 95% CI 1·39-1·45). Individuals aged ≥80 years experienced a greater increase in hospitalisation compared with younger individuals (RR 1·39, 95% CI 1·30-1·48) particularly for non-cardio/cerebrovascular causes (RR 1·54, 95% CI 1·44-1·65). Patterns for hospital admissions by sex and diagnostic setting were similar to those observed for mortality, but frequency of hospitalisation was similar between most and least deprived individuals. Conclusion For patients newly diagnosed with AF, cardiovascular and cerebrovascular mortality has declined, and hospitalisations for non-cardio/cerebrovascular diseases increased. The changing outcomes of AF require new strategies to reduce its health burden.Figure 1:Temporal trends in mortalityFigure 2:Temporal trends in admission