20-year trends in cause-specific heart failure outcomes by sex, socioeconomic status, and place of diagnosis: a population-based study

Abstract

Background: Heart failure is an important public health issue affecting about 1 million people in the UK, but contemporary trends in cause-specific outcomes among different population groups are unknown.

Methods: In this retrospective, population-based study, we used the UK Clinical Practice Research Datalink and Hospital Episodes Statistics databases to identify a cohort of patients who had a diagnosis of incident heart failure between Jan 1, 1998, and July 31, 2017. Patients were eligible for inclusion if they were aged 30 years or older with a first code for heart failure in their primary care or hospital record during the study period. We assessed cause-specific admission to hospital (ie, hospitalisation) and mortality, by age, sex, socioeconomic status, and place of diagnosis (ie, hospital vs community diagnosis). We calculated outcome rates separately for the first year (first-year rates) and for the second-year onwards (subsequent-year rates). Patients were followed up until death or study end. This study is registered with Clinical Practice Research Datalink Independent Scientific Advisory Committee, protocol number 18_037R.

Findings: We identified 88 416 individuals with incident heart failure over the study period, of whom 43 461 (49%) were female. The mean age was 77·8 years (SD 11·3) and median follow-up was 2·4 years (IQR 0·5 to 5·7). Age-adjusted first-year rates of hospitalisation increased by 28% for all-cause admissions, from 97·1 (95% CI 94·3 to 99·9) to 124·2 (120·9 to 127·5) per 100 person-years; by 28% for heart failure-specific admissions, from 17·2 (16·2 to 18·2) to 22·1 (20·9 to 23·2) per 100 person-years; and by 42% for non-cardiovascular admissions, from 59·2 (57·2 to 61·2) to 83·9 (81·3 to 86·5) per 100 person-years. 167 641 (73%) of 228 113 hospitalisations were for non-cardiovascular causes and annual rate increases were higher for women (3·9%, 95% CI 2·8 to 4·9) than for men (1·4%, 0·6 to 2·1; p<0·0001); and for patients diagnosed with heart failure in hospital (2·4%, 1·4 to 3·3) than those diagnosed in the community (1·2%, 0·3 to 2·2). Annual increases in hospitalisation due to heart failure were 2·6% (1·9 to 3·4) for women compared with stable rates in men (0·6%, -0·9 to 2·1), and 1·6% (0·6 to 2·6) for the most deprived group compared with stable rates for the most affluent group (1·2%, -0·3 to 2·8). A significantly higher risk of all-cause hospitalisation was found for the most deprived than for the most affluent (incident rate ratio 1·34, 95% CI 1·32 to 1·35) and for the hospital-diagnosed group than for the community-diagnosed group (1·76, 1·73 to 1·80). Age-adjusted first-year rates of all-cause mortality decreased by 6% from 24·5 (95% CI 23·4 to 39·2) to 23·0 (22·0 to 24·1) per 100 person-years. Annual change in mortality was -1·4% (95% CI -2·3 to -0·5) in men but was stable for women (0·3%, -0·5 to 1·1), and -2·7% (-3·2 to -2·2) for the community-diagnosed group compared with -1·1% (-1·8 to -0·4) in the hospital-diagnosed group (p<0·0001). A significantly higher risk of all-cause mortality was seen in the most deprived group than in the most affluent group (hazard ratio 1·08, 95% CI 1·05 to 1·11) and in the hospital-diagnosed group than in the community-diagnosed group (1·55, 1·53 to 1·58).

Interpretation: Tailored management strategies and specialist care for patients with heart failure are needed to address persisting and increasing inequalities for men, the most deprived, and for those who are diagnosed with heart failure in hospital, and to address the worrying trends in women.

Funding: Wellcome Trust.

Copyright © 2019 The Author(s). Published by Elsevier Ltd. This is an Open Access article under the CC BY 4.0 license. Published by Elsevier Ltd.. All rights reserved.

Figures

Figure 1

Predicted rates of cause-specific hospitalisation during the first year after heart failure diagnosis, by population group Data are predicted hospitalisation rates, estimated at the mean population age (77·8 years), per 100 person-years by cause, between 1998 and 2015. Whiskers are 95% CIs. Hospitalisations included all non-elective admissions with at least one overnight stay and occurring after but not including the date of heart failure diagnosis. Follow-up was until death or study end.

Figure 2

First-year rates of mortality, by population group and calendar year of diagnosis Data are predicted rates of mortality per 100 person-years at mean population age (77·8 years) per 100 person-years between 1998 and 2015. Whiskers are 95% CIs. Rates were calculated in survivors of the first month after heart failure diagnosis.