Rheumatoid arthritis and cardiovascular disease associations in the UK Biobank

BACKGROUND: This study evaluated observational and causal relationships between rheumatoid arthritis (RA) and cardiovascular disease and imaging phenotypes in the UK Biobank. METHODS: RA was defined using linked hospital records, self-reported diagnostics, and medication data. Controls were participants without a record of RA. Cardiovascular diseases (CVDs) were defined using linked hospital records over an average of 14 years of prospective follow-up, including: ischaemic heart diseases (IHD), acute myocardial infarction (AMI), atrial fibrillation, any arrhythmia, non-ischaemic cardiomyopathies, pericardial disease, stroke, peripheral vascular disease, and venous thromboembolism. For participants with cardiovascular magnetic resonance (CMR) available as part of the UK Biobank Imaging Study, we considered measures of cardiac structure and function extracted using automated pipelines. Associations of RA with prevalent and incident CVDs were calculated using logistic and Cox regression. Linear regression was used to examine associations with CMR metrics. Models were adjusted for demographic, lifestyle, and clinical confounders. Causal associations were assessed using two-sample Mendelian randomisation. Genetic instruments for RA (22,350 cases and 74,823 controls), nine CVDs (FinnGen, n = 224,737), and 11 CMR phenotypes (UK Biobank) were extracted and associations assessed using inverse-variance weighting with pleiotropy adjustments and multiple testing corrections. RESULTS: The analysis included 1,436 RA cases (mean age 59.9 years; 70.6% female) and 476,975 controls (mean age 56.5 years; 54.3% female). Participants with RA lived in more socioeconomically deprived areas (as per the Townsend Deprivation Index), had lower physical activity levels, were more likely to smoke, and had a higher baseline prevalence of CVDs. In fully adjusted models, participants with RA had a significantly higher hazard of multiple incident CVDs, with the greatest risks related to pericardial disease (HR 2.63 (1.85, 3.74)), heart failure (HR 1.68 (1.42, 1.99)), and AMI (HR 1.53 (1.20, 1.96)). Mendelian Randomisation analyses supported causal links between RA and AMI (OR 1.07 (1.02, 1.09), p = 0.009), arrhythmias (OR 1.05 (1.02, 1.06), p = 0.0007), and IHD (OR 1.05 (1.01, 1.06), p = 0.036). No significant associations were identified between RA and CMR phenotypes. CONCLUSIONS: People with RA have a heightened risk of multiple prevalent and incident CVDs, independent of shared risk factors, with suggestions of causal links with IHD, AMI, and arrhythmias.