Ethnic differences in new-onset atrial fibrillation using multidimensional data

Background Atrial fibrillation (AF) is the most common sustained arrhythmia and a growing public health challenge, significantly increasing the risk of stroke and heart failure (HF). Although AF and arrhythmia phenotypes have been examined in the UK Biobank, a focused analysis of ethnic differences in new-onset AF and their potential mediators is lacking. Purpose In this UK Biobank study, we aimed to (1) quantify ethnic differences in new-onset AF; (2) estimate the proportion of these differences explained by cardiac magnetic resonance (CMR) structural and functional measures; and (3) evaluate genetic susceptibility using polygenic risk scores (PRS). Methods Among 502,412 UK Biobank participants, new-onset AF was identified using hospital episode statistics and electrocardiograms. Multivariable logistic regression assessed associations between self-reported ethnicity and new-onset AF, adjusting for age, sex, body mass index, vascular risk factors and the CHA2DS2-VASc score. In an imaging sub-study (n=24,464), mediation analysis estimated the proportion of ethnic differences explained by CMR measures. Genetic contributions were evaluated by assessing associations of AF-PRS and HF-PRS with new-onset AF using logistic regression, comparing effect estimates between European and non-European ancestry groups. To quantify the attenuation of the European ancestry association, nested models (age, sex, genetic principal components) were fit with and without PRS. Results 16,122 (3.2%) had prevalent AF and 12,403 (2.5%) developed new-onset AF, as shown in Table 1. Compared to White Europeans, the odds of new-onset AF were significantly lower in South Asian (Odds Ratio [OR] 0.68; 95% Confidence Interval [CI] 0.57–0.80), Black (OR 0.48; 95% CI 0.38–0.60), and Chinese (OR 0.56; 95% CI 0.30–0.94) participants (Figure 1). In the imaging sub-study, participants with AF demonstrated larger atrial size, reduced ventricular function, and impaired myocardial strain. However, mediation analysis indicated that CMR-derived structural and functional parameters explained only a small, non-significant proportion of the ethnic differences in new-onset AF risk. The AF-PRS was strongly associated with incident AF, with stronger effect estimates in European versus non-European ancestry groups (European OR 1.68 vs non-European OR 1.25). The HF-PRS showed a modest association (European OR 1.30 vs non-European OR 1.08). Inclusion of PRS attenuated the association with European ancestry by ~11%, leaving substantial residual ethnic differences. Conclusions Significant ethnic differences in new-onset AF persist after adjustment for clinical factors. CMR-derived phenotypes and polygenic burden explain only a limited proportion of this variation. This highlights the need to investigate ancestry-specific genetic architecture, electrophysiological mechanisms, and social or environmental determinants.