Abstract
INTRODUCTION: Heart failure (HF) is a disease often preceded by other cardiac disorders, such as myocardial infarction (MI) and atrial fibrillation (AF). We aimed to evaluate whether the plasma protein profiles of three groups of patients with incident HF-HF with preceding MI, HF with preceding AF, and HF with no preceding MI or HF-would be similar, and whether a common protein profile across all three HF groups could be detected. METHODS: We analysed data from 50 765 UK Biobank (UKB) participants with measurements on 2922 plasma proteins. Participants who developed HF after enrolment (12,6 years, median follow-up) were divided in those who, between enrolment and HF diagnosis, experienced a MI (MI-HF group, n = 269), developed AF (AF-HF, n = 519), or were not diagnosed with MI or AF (noMInoAF-HF, n = 1059). We estimate hazard ratios (HR) with 95% confidence interval (CI) for the associations between protein measurements and incident HF using multivariable adjusted Cox models. Proteins associated with the outcome across all three HF groups where further evaluated in relation to magnetic resonance-measured left ventricular ejection fraction (LV-EF) in 5097 UKB participants, and to echocardiography-measured diastolic E/A-ratio in 502 POEM study participants. RESULTS: After correction for multiple testing, 110 proteins were uniquely associated with HF in the noMInoAF-HF group, 21 proteins in the AF-HF group, and only one protein (ADGRG2) in the MI-HF group. Across all three HF groups, the same 60 proteins were significantly associated with HF, 13 of which were related to LV-EF and another 7 were associated with the E/A-ratio. LRRN1 was inversely associated with incident HF in the MI-HF group [HR 0.49 (95% CI, 0.36, 0.67)], the AF-HF group [0.51 (0.41, 0.63)], and in the noMInoAF-HF group [0.58 (0.50, 0.68)]; all remaining proteins were positively associated with HF. ADM was the top association in the MI-HF group [HR 5.19 (3.06, 8.78)] and the AF-HF group [7.78 (5.31, 11.41)]. The most significant enriched pathways for the 60 shared proteins were interleukin-10 signalling, transcription of death receptors, and TNF receptor binding. CONCLUSION: A plasma protein profile associated with heart failure independently of prior MI or AF was identified. This protein profile represents several pathophysiological pathways of interest, and some of these proteins were also related to either LV-EF or the E/A ratio. In addition, a number of proteins were found to be uniquely associated with only one of the three HF traits.