Trends in mortality in people with heart failure and atrial fibrillation: a population-based cohort study

Background: Atrial fibrillation and heart failure frequently coexist but the relative effect of atrial fibrillation on survival in people with heart failure, and vice versa, remains uncertain. We aimed to report contemporary estimates of mortality among people with atrial fibrillation and heart failure and analyse trends in mortality over time. Methods: We did a retrospective cohort study of adults aged 45 years or older in England, using primary care data from the Clinical Practice Research Datalink GOLD dataset and linked secondary care data (Hospital Episode Statistics and Office for National Statistics datasets), for a total follow-up period from Jan 1, 2000, to Dec 31, 2018. We recorded incident cases of heart failure and atrial fibrillation in primary or secondary care during the study period, as well as pre-existing cases at the study index date. Individuals were categorised as having both heart failure and atrial fibrillation, atrial fibrillation only, heart failure only, or neither condition, with heart failure and atrial fibrillation included in analyses as time-varying covariates. The primary outcome was all-cause mortality, as recorded in primary or secondary care. We report the incidence and hazard ratios for all-cause mortality by diagnosis status, median overall survival following diagnosis, and the cumulative probability of all-cause mortality from 3 months to 10 years of follow-up and by year of diagnosis to assess trends over time. Estimates of median survival and the cumulative probability of overall mortality were restricted to incident diagnoses during the study period, and calculated overall as well as by sex, age, and Index of Multiple Deprivation quintile. Findings: The cohort consisted of 2 381 941 people, including 100 132 initially diagnosed with heart failure only and 155 061 initially diagnosed with atrial fibrillation only by the study index date or during follow-up. By the end of follow-up, 74 470 people had been diagnosed with both conditions. 314 042 people died during follow-up, including 42 427 (57·0%) of those diagnosed with both heart failure and atrial fibrillation. In people diagnosed with both conditions during the study period (n=43 714), median overall survival was 3·15 years (95% CI 3·08–3·21), and the cumulative probability of mortality was 31·8% (95% CI 30·2–33·6) at 1 year, 61·4% (59·4–63·3) at 5 years, and 80·2% (78·3–82·1) at 10 years after both conditions had been diagnosed, representing significantly worse rates than for an initial diagnosis of either condition alone. Similarly, the risk-adjusted hazard of all-cause mortality was highest among people with both heart failure and atrial fibrillation. For the overall population, cumulative mortality probability estimates were unchanged over successive years of diagnosis for people with both heart failure and atrial fibrillation, while showing small improvements for people initially diagnosed with heart failure only (median reduction in 10-year cumulative probability of 3·8% [95% CI 1·4–6·1] between diagnosis years 2000 and 2008) or atrial fibrillation only (median reduction in 1-year cumulative mortality probability of 2·4% [0·5–4·2] between diagnosis years 2000 and 2017) and improvement over the long-term for people diagnosed with both conditions before age 65 years (median reduction in 10-year cumulative mortality probability of 14·5% [95% CI 3·8–25·2] between diagnosis years 2000 and 2008). For people with both conditions, median overall survival was significantly longer in the least deprived quintile (3·46 years [95% CI 3·31–3·59]; n=9275) than in the most deprived quintile (2·67 years [2·51–2·81]; n=6302). Median overall survival in each exposure group was similar between sexes after stratifying for age. Interpretation: Comorbid heart failure and atrial fibrillation was common and prognosis was poor, with no improvement in mortality estimates for diagnoses over time, and the worst survival in socially deprived groups. Funding: Wellcome Trust and the National Institute for Health and Care Research Collaboration for Leadership in Applied Health Research and Care Oxford.