Genetic ancestry and risk of atrial fibrillation in individuals of Black ethnicity in the UK Biobank Authors: Chang Liu, Yan V Sun, Linzi Li, Lindsay J Collin, Amit J Shah, Alvaro Alonso Published: 2026 ABSTRACT AbstractBackgroundBlack individuals have a lower incidence of atrial fibrillation (AF) than White individuals, despite a higher burden of traditional risk factors. Prior studies have suggested that European genetic ancestry may contribute to this paradox, but findings have been inconsistent.MethodsWe examined the association between European genetic ancestry and incident AF among 6920 UK Biobank (UKB) participants who self-identified as Black and were free of AF at baseline. European ancestry proportions were estimated by comparing participants to HapMap populations and were analysed both continuously and categorically using Cox proportional hazards models. Non-linear associations were evaluated using flexible spline curves. We then conducted a meta-analysis combining these results with those from the Atherosclerosis Risk in Communities Study, the Cardiovascular Health Study and the Women’s Health Initiative.ResultsDuring a median follow-up of 13.7 years, 205 (3%) participants developed incident AF. Each 10% increase in European ancestry was non-significantly associated with increased AF risk (HR 1.07 (95% CI 0.95 to 1.20)), and individuals in the highest European ancestry category had a higher AF incidence rate compared with all other categories. Random-effects meta-analysis of all four cohorts yielded a pooled relative risk (RR) of 1.09 (95% CI 0.99 to 1.21), with substantial heterogeneity largely driven by the Women’s Health Initiative (WHI) study. Excluding WHI resulted in a pooled estimate (RR 1.14 (95% CI 1.05 to 1.23)) without heterogeneity.ConclusionIn UKB, European ancestry proportion was not significantly linked to incident AF, and pooled cross-cohort results were heterogeneous. Excluding WHI in sensitivity analyses produced a modest positive pooled association with no heterogeneity. WHAT IS ALREADY KNOWN ON THIS TOPIC Black individuals have a lower incidence of atrial fibrillation (AF) than White individuals despite a higher burden of conventional risk factors, and prior admixture studies suggested that greater European genetic ancestry among Black individuals may be associated with higher AF risk, but findings have been inconsistent. Black individuals have a lower incidence of atrial fibrillation (AF) than White individuals despite a higher burden of conventional risk factors, and prior admixture studies suggested that greater European genetic ancestry among Black individuals may be associated with higher AF risk, but findings have been inconsistent. WHAT THIS STUDY ADDS In 6920 UK Biobank participants who self-identified as Black, higher European genetic ancestry was associated with a modest but non-significant increase in incident AF risk, with imprecise estimates at higher ancestry levels; in meta-analysis with Atherosclerosis Risk in Communities, Cardiovascular Health Study and Women’s Health Initiative (WHI), the pooled association was heterogeneous overall and became modestly positive after exclusion of WHI, the main source of heterogeneity. In 6920 UK Biobank participants who self-identified as Black, higher European genetic ancestry was associated with a modest but non-significant increase in incident AF risk, with imprecise estimates at higher ancestry levels; in meta-analysis with Atherosclerosis Risk in Communities, Cardiovascular Health Study and Women’s Health Initiative (WHI), the pooled association was heterogeneous overall and became modestly positive after exclusion of WHI, the main source of heterogeneity. HOW THIS STUDY MIGHT AFFECT RESEARCH, PRACTICE OR POLICY These findings suggest that genetic ancestry may contribute modestly to AF risk variation within admixed populations, but current evidence is not strong enough to support clinical use. These findings suggest that genetic ancestry may contribute modestly to AF risk variation within admixed populations, but current evidence is not strong enough to support clinical use. Introduction Atrial fibrillation (AF) is a common arrhythmia and a well-established risk factor for stroke, heart failure and other adverse outcomes.1 Its risk factors include older age, male sex, high blood pressure, obesity, diabetes and others.2 Epidemiologic studies in the USA have consistently shown that Black and Hispanic individuals have lower rates of AF compared with White individuals, despite having a higher burden of traditional AF risk factors.3 4 A recent analysis of the UK Biobank (UKB) found a similar pattern, with individuals of self-reported White race experiencing a higher incidence of AF compared with those of Black or Asian race groups.5 An analysis of the Atherosclerosis Risk in Communities (ARIC) Study and the Cardiovascular Health Study (CHS) found that among Black participants with genetic admixture, a higher proportion of European genetic ancestry was associated with an increased risk of AF, suggesting that ancestry-related genetic factors may play a role in this racial difference.6 However, a subsequent analysis of the Women’s Health Initiative (WHI) cohort did not replicate these findings.7 As a result, the relationship between genetic ancestry and AF risk remains unclear. To further explore the potential role of genetic ancestry in explaining racial differences in AF risk, we examined the association between European genetic ancestry and AF incidence among UKB participants who self-identified as Black. Additionally, we conducted a meta-analysis incorporating the current results with those from the ARIC, CHS and WHI cohorts to provide a more comprehensive assessment of this relationship. Genetic ancestry reflects shared population history and should not be conflated with race, a social and political construct. We therefore evaluated European genetic ancestry within self-identified Black individuals, an admixed group with meaningful ancestry variation to reduce confounding by race-associated social determinants. Methods The UKB cohort The UKB is an ongoing, large-scale prospective cohort study based in the UK that enrolled over 500 000 participants aged 40–69 years between 2006 and 2010.8 The study collected extensive data through questionnaires, physical measurements, blood biomarker assays, imaging, genome-wide genotyping and longitudinal follow-up.8 The cohort was linked to Hospital Episode Statistics (HES) data, primary care data and the death registry, which include information on death date and both primary and secondary causes. This analysis was restricted to participants who self-identified as ‘Black or Black British’, ‘Caribbean’, ‘African’ or ‘any other Black background’ in response to the standard questionnaire on ethnic background at enrolment. Participants with genomics data, genetically unrelated (greater than third-degree relatedness) and without prevalent AF were included in this study. AF ascertainment The UKB mapped first occurrence of disease diagnosis to International Classification of Diseases, 10th Revision (ICD-10) codes, integrating HES, primary care, death registry and self-reported medical conditions. AF was defined through linkage with these records and by the presence of ICD-10 code I48, with the date of AF onset determined by its first occurrence in any linked dataset. Prevalent AF was defined as AF first recorded on or before enrolment, while incident AF was defined as AF first recorded after enrolment. Follow-up time was defined as time from enrolment to incident AF, lost to follow-up, death or the end of follow-up as of September 2023. Determination of genetic ancestry We performed ancestry estimation using ADMIXTURE,9 a model-based clustering method that estimates individual ancestry proportions from genome-wide genotype data. The analysis was conducted in supervised mode, using predefined ancestral reference populations to guide the estimation process. A total of 770 unrelated individuals from the HapMap phase 3 reference data were used as ancestry reference populations, including Utah residents with Northern and Western European ancestry from the Centre d'Étude du Polymorphisme Humain, (CEPH) collection (CEU); the Han Chinese in Beijing, China (CHB); the Chinese in Metropolitan Denver, Colorado (CHD); Japanese in Tokyo, Japan (JPT); Maasai in Kinyawa, Kenya (MKK); Luhya in Webuye, Kenya (LWK); and Yoruba in Ibadan, Nigeria (YRI). To reduce the number of highly correlated variants, we performed pruning of single nucleotide polymorphisms (SNPs) based on linkage disequilibrium (LD) using PLINK.10 Starting from all genetic variants with minor allele frequency >5%, a sliding window of 50 consecutive SNPs was used, with the window shifted by five SNPs at each step, and SNP pairs with high LD (an r² value greater than 0.1) were pruned to reduce redundancy. A total of 33 individuals self-reported as Black with a genetic ancestry proportion <15% African ancestry (combination of MKK, LWK and YRI ancestry proportions) were excluded from the analysis. This was a quality control restriction to remove extreme ancestry outliers and to ensure that European ancestry proportion reflected variation in admixture within an African ancestry background. A total of 73 693 independent genetic variants were included in this analysis. Additionally, a principal component analysis was performed using the pruned genomics data to visualise the population structure comparing the reference populations and the UKB participants. Covariates Age, sex, education, total household income before tax, current smoking and blood pressure-lowering medication use were self-reported at enrolment. Height, weight, systolic and diastolic blood pressure were measured at the baseline visit. Body mass index (BMI) was calculated as weight in kilograms divided by height in metres squared. Education was categorised as high (college or university degree; National Vocational Qualification (NVQ), Higher National Diploma (HND) or Higher National Certificate (HNC) or equivalent, other professional qualifications (eg, nursing, teaching)), low (Advanced (A) levels/Advanced Subsidiary (AS) levels or equivalent; O levels/ General Certificate of Secondary Education (GCSE) or equivalent; Certificate of Secondary Education (CSE) or equivalent; none of the above) and missingness was treated as a separate category. Household income was categorised as high (≥£31 000), low (≤£30 999) and missing. Baseline prevalence of type 2 diabetes (ICD-10 E11), coronary artery disease (ICD-10 I20–I25) and heart failure (ICD-10 I50) were determined by ICD-10 codes from linkage to HES, primary care, death registry, self-reported medical conditions and comparing the earliest date of onset to the enrolment date. Patient and public involvement Patients and members of the public were not involved in the design, conduct, reporting or dissemination plans of this research. This study was a secondary analysis of existing UKB data, and no patients or public contributors were involved as research partners in this specific study. Statistical analysis Baseline characteristics of the study population were summarised by levels of European genetic ancestry and incident AF status. Differences across groups were compared using one-way analysis of variance, two-sample t-test and Wilcoxon rank-sum test for continuous variables and χ2 tests for categorical variables where appropriate. We evaluated the association of European genetic ancestry with the incidence of AF using a Cox proportional hazards model among eligible participants self-reporting Black ethnicity. European genetic ancestry was modelled as a continuous variable (per 10% higher). Model 1 adjusted for age at enrolment, sex (male vs female), education (high vs low vs missing) and income (high vs low vs missing). Model 2 additionally adjusted for current smoking, BMI, height, systolic blood pressure, diastolic blood pressure, blood pressure-lowering medication use, prevalent type 2 diabetes, prevalent coronary artery disease and prevalent heart failure. We conducted additional analyses categorising European genetic ancestry into three groups using cutoffs of 11.26% (67th percentile) and 50%, including low (0.001%–11.25%), medium (11.26%–49.80%) and high (50.10%–73.43% (maximum)). The 11.26% cut-off was selected based on the overall distribution of European genetic ancestry and its non-linear association with incident AF risk, with a notable shift in risk observed around this threshold. The 50% cut-off was used to define individuals with predominant European ancestry, which is commonly applied in large-scale genetic and admixture studies to stratify populations based on majority ancestry. The groups of medium and high European genetic ancestry were compared with the low European genetic ancestry, using the same covariate adjustment used in Models 1 and 2. Sex-stratified analysis and analysis among only postmenopausal women were also performed. The proportional hazards assumption was assessed using Schoenfeld residuals and inspection of log(−log (survival)) plots. No material violations were observed. We explored dose–response curves modelling European genetic ancestry as a spline term using the software smooth HR,11 12 based on the fully adjusted Model 2. This approach generated flexible HR curves that capture potential non-linear associations between ancestry and AF risk, allowing for the identification of inflection points along the exposure continuum. We synthesised results with prior studies by pooling the fully adjusted relative risk (RR) estimates per 10% increase in European ancestry (ie, the UKB Model 2 HR and the corresponding fully adjusted estimates reported for ARIC, CHS and WHI6 13 using the metan command in Stata.14 We first estimated a fixed-effect model with inverse of variance weighting and then a DerSimonian-Laird random-effects model because design heterogeneity made between-study heterogeneity plausible (ARIC, CHS and UKB only included incident cases, while WHI included prevalent and incident cases; WHI only included women). Heterogeneity was assessed using the I2 statistic and Cochran’s Q test.15 To assess the impact of each individual study on the overall summary estimate, we conducted a leave-one-out analysis by iteratively removing each study and recalculating the pooled risk ratio and heterogeneity statistics. Results The study cohort included 6920 Black participants free of AF at baseline. The workflow of inclusion and exclusion is shown in online supplemental figure 1. The mean age was 51.9 years (SD) 8.0), and 43.5% were male. The median follow-up time was 13.7 years (IQR) 13.2–14.5). During follow-up, a total of 205 individuals (3.0%) developed incident AF. European genetic ancestry proportions ranged from 0.001% to 73.43%, with a median of 6.3% (IQR 3.0%–15.0%). The overall distribution of European genetic ancestry is shown in figure 1. Principal component analysis showed a clear alignment between UKB self-reported Black participants and the HapMap 3 reference populations of MKK, LWK and YRI (online supplemental figure 2). Distribution of European genetic ancestry among the UK Biobank participants self-reported as Black. Baseline characteristics stratified by European ancestry categories are presented in table 1, and by AF incidence status in table 2. Participants with higher proportions of European ancestry were more likely to smoke, have lower BMI and lower blood pressure, and were less likely to use blood pressure-lowering medications. Incident AF cases were older, more likely to be male, less likely to smoke, taller and had a higher prevalence of cardiometabolic risk factors such as elevated BMI, systolic blood pressure, diabetes, and history of coronary artery disease and heart failure. Table 1Baseline characteristics by categories of European genetic ancestry proportionsCharacteristicsOverallN=6920Low*N=4613Medium*N=2198High*N=109P†European genetic ancestry proportions (minimum–maximum)0.001–73.430.001–11.2511.26–49.8050.10–73.43–Age (years)51.9 (8.0)51.6 (8.0)52.3 (8.0)52.9 (8.5) 0.004 Male (%)3007 (43.5)2065 (44.8)888 (40.4)54 (49.5) 0.001 Education‡ Low (%)2177 (31.5)1297 (28.1)829 (37.7)51 (46.8) <0.001 High (%)4510 (65.2)3164 (68.6)1291 (58.7)55 (50.5) Missing (%)233 (3.4)152 (3.3)78 (3.5)3 (2.8)Income ≤£30 999 (%)3374 (48.8)2286 (49.6)1042 (47.4)46 (42.2) 0.001 ≥£31 000 (%)1929 (27.9)1215 (26.3)675 (30.7)39 (35.8) Missing (%)1617 (23.4)1112 (24.1)481 (21.9)24 (22.0)Current smoking (%)840 (12.2)454 (9.9)358 (16.4)28 (25.7) <0.001 Body mass index (kg/m2)29.5 (5.4)29.8 (5.3)28.9 (5.5)27.7 (4.6) <0.001 Height (cm)167 (9)167 (9)168 (9)168 (9)0.26Systolic BP (mm Hg)140 (19)140 (19)138 (19)134 (18) <0.001 Diastolic BP (mm Hg)85 (11)85 (11)84 (11)82 (9) <0.001 BP-lowering medication (%)2111 (31.3)1495 (33.2)593 (27.7)23 (21.9) <0.001 Type 2 diabetes (%)331 (4.8)204 (4.4)121 (5.5)6 (5.5)0.14Prevalent coronary artery disease (%)265 (3.8)186 (4.0)76 (3.5)3 (2.8)0.43Prevalent heart failure (%)22 (0.3)12 (0.3)10 (0.5)0 (0.0)0.34Incident AF (%)205 (3.0)134 (2.9)66 (3.0)5 (4.6)0.59Data are presented as mean (standard deviation) unless otherwise stated.*European genetic ancestry categories: low: 0.001%–11.25%; medium: 11.26%–49.80% and high: 50.10%–73.43%.†P values compare characteristics across low, medium and high European genetic ancestry groups. Continuous variables were compared using one-way analysis of variance and categorical variables were compared using the χ2 test. Bold entries correspond to p values < 0.05‡Education: high—college or university degree; NVQ, HND or HNC or equivalent, other professional qualifications (eg, nursing, teaching); low—A levels/AS levels or equivalent; O levels/GCSEs or equivalent; CSEs or equivalent; none of the above.AF, atrial fibrillation; BP, blood pressure. Baseline characteristics by categories of European genetic ancestry proportions Data are presented as mean (standard deviation) unless otherwise stated. European genetic ancestry categories: low: 0.001%–11.25%; medium: 11.26%–49.80% and high: 50.10%–73.43%. P values compare characteristics across low, medium and high European genetic ancestry groups. Continuous variables were compared using one-way analysis of variance and categorical variables were compared using the χ2 test. Bold entries correspond to p values < 0.05 Education: high—college or university degree; NVQ, HND or HNC or equivalent, other professional qualifications (eg, nursing, teaching); low—A levels/AS levels or equivalent; O levels/GCSEs or equivalent; CSEs or equivalent; none of the above. AF, atrial fibrillation; BP, blood pressure. Table 2Baseline characteristics by AF incidence statusCharacteristicsNo incident AFN=6715(97.0%)Incident AFN=205(3.0%)P*European genetic ancestry percentage, median (IQR)6.3(3.0–14.9)6.0(3.0–17.0)0.54Age (years)51.6 (7.9)59.6 (8.0) <0.001 Male (%)2889 (43.0)118 (57.6) <0.001 Education† Low (%)2103 (31.3)74 (36.1) 0.001 High (%)4394 (65.4)116 (56.6) Missing (%)218 (3.2)15 (7.3)Income ≤£30 999 (%)3274 (48.8)100 (48.8) 0.03 ≥£31 000 (%)1885 (28.1)44 (21.5) Missing (%)1556 (23.2)61 (29.8)Current smoking (%)826 (12.4)14 (7.0) 0.03 Body mass index (kg/m2)29.4 (5.3)30.8 (6.0) <0.001 Height (cm)167 (9)169 (10) 0.01 Systolic BP (mm Hg)139 (19)147 (21) <0.001 Diastolic BP (mm Hg)85 (11)85 (11)0.96BP-lowering medication (%)1985 (30.3)126 (62.4) <0.001 Type 2 diabetes (%)301 (4.5)30 (14.6) <0.001 Prevalent coronary artery disease (%)228 (3.4)37 (18.0) <0.001 Prevalent heart failure (%)18 (0.3)4 (2.0) <0.001 Data are presented as mean (standard deviation) unless otherwise stated.*P values compare baseline characteristics between participants with and without incident AF. Continuous variables were compared using the two-sample t-test, and categorical variables were compared using the χ2 test, except that European genetic ancestry percentage was compared using the Wilcoxon rank-sum test. Bold entries correspond to p values < 0.05†Education: high—college or university degree; NVQ, HND or HNC or equivalent, other professional qualifications (eg, nursing, teaching); low—A levels/AS levels or equivalent; O levels/GCSEs or equivalent; CSEs or equivalent; none of the above.AF, atrial fibrillation; BP, blood pressure. Baseline characteristics by AF incidence status Data are presented as mean (standard deviation) unless otherwise stated. P values compare baseline characteristics between participants with and without incident AF. Continuous variables were compared using the two-sample t-test, and categorical variables were compared using the χ2 test, except that European genetic ancestry percentage was compared using the Wilcoxon rank-sum test. Bold entries correspond to p values < 0.05 Education: high—college or university degree; NVQ, HND or HNC or equivalent, other professional qualifications (eg, nursing, teaching); low—A levels/AS levels or equivalent; O levels/GCSEs or equivalent; CSEs or equivalent; none of the above. AF, atrial fibrillation; BP, blood pressure. When analysed continuously, each 10% increase in European genetic ancestry was not significantly associated with incident AF in either the minimally adjusted Model 1 (HR 1.03 (95% CI 0.92 to 1.15)) or the fully adjusted Model 2 (HR 1.07 (95% CI 0.95 to 1.20)) (table 3). When the European genetic ancestry was treated as a categorical variable, the incidence rate per 1000 person-years was 2.15 in the low group, 2.23 in the middle group and 3.36 in the high group. Compared with the low ancestry group, in Model 1, the middle group was not associated with AF risk (HR 0.99 (95% CI 0.73 to 1.32)), while the high ancestry group showed a suggestive association (HR 1.26 (95% CI 0.51 to 3.09)), but the estimate was imprecise and not statistically significant. In Model 2, both groups showed a suggestive but insignificant association (middle group: HR 1.06 (95% CI 0.78 to 1.44); high group: HR 1.64 (95% CI 0.66 to 4.04)). Similar findings were observed in the sex-stratified analyses (online supplemental table 1). In Model 2, the association between each 10% increase in European genetic ancestry and incident AF was suggestive and not statistically significant in males (HR 1.04 (95% CI 0.89 to 1.22)), females (HR 1.12 (95% CI 0.93 to 1.34)) and postmenopausal females (HR 1.14 (95% CI 0.92 to 1.41)). When European genetic ancestry was treated as a categorical variable, the high ancestry group had slightly elevated HRs in the middle and high groups, with 95% CIs covering the null. The data were sparse in both males and females with high European genetic ancestry. Table 3Association of European genetic ancestry with AF riskEstimatesPer 10% increase in European genetic ancestry percentageLow*Medium*High*Model 1N692046132198109Number of AF205134665Person-years93 39162 33429 5671490Incidence rate per 1000 person-years2.202.152.233.36HR (95% CI)1.03 (0.92 to 1.15)p=0.66Ref0.99 (0.73 to 1.32)p=0.921.26 (0.51 to 3.09)p=0.61Model 2N649143202071100Number of AF196128635Person-years87 57858 36327 8581357Incidence rate per 1000 person-years2.242.192.263.68HR (95% CI)1.07 (0.95 to 1.2)p=0.27Ref1.06 (0.78 to 1.44)p=0.721.64 (0.66 to 4.04)p=0.28Model 1: adjusted for age at enrollment, sex (male vs female), education (high—college or university degree; NVQ, HND or HNC or equivalent, other professional qualifications versus low—A levels/AS levels or equivalent, O levels/GCSEs or equivalent, CSEs or equivalent, none of the above versus missing), income (≤£30 999 vs ≥£31 000 vs missing).Model 2: adjusted for current smoking, body mass index, height, systolic blood pressure, diastolic blood pressure, blood pressure lowering medication use, prevalent type 2 diabetes, prevalent coronary artery disease and prevalent heart failure, in addition to Model 1.*European genetic ancestry categories: low—0.001%–11.25%; medium—11.26%–49.80%; high—50.10%–73.43%.AF, atrial fibrillation. Association of European genetic ancestry with AF risk Model 1: adjusted for age at enrollment, sex (male vs female), education (high—college or university degree; NVQ, HND or HNC or equivalent, other professional qualifications versus low—A levels/AS levels or equivalent, O levels/GCSEs or equivalent, CSEs or equivalent, none of the above versus missing), income (≤£30 999 vs ≥£31 000 vs missing). Model 2: adjusted for current smoking, body mass index, height, systolic blood pressure, diastolic blood pressure, blood pressure lowering medication use, prevalent type 2 diabetes, prevalent coronary artery disease and prevalent heart failure, in addition to Model 1. European genetic ancestry categories: low—0.001%–11.25%; medium—11.26%–49.80%; high—50.10%–73.43%. AF, atrial fibrillation. The potential non-linear association between European ancestry and AF risk is shown in figure 2. The HR curves showed no strong evidence of a monotonic trend, with the HRs remaining close to the null across most of the ancestry distribution. However, there was a slight elevation in risk at higher European genetic ancestry proportions. The CIs were wide and crossed the null at the upper end of the distribution due to the sparse data. Sex-stratified non-linear curves showed a similar pattern (online supplemental figure 3). Flexible non-linear HR curves of the association between European genetic ancestry and incident atrial fibrillation among the UK Biobank participants self-reported as Black. Models adjusted for age at enrolment, sex (male vs female), education (high—college or university degree; NVQ, HND or HNC or equivalent, other professional qualifications vs low—A levels/AS levels or equivalent, O levels/GCSEs or equivalent, CSEs or equivalent, none of the above vs missing), income (≤£30 999 vs ≥£31 000 vs missing), current smoking, body mass index, height, systolic blood pressure, diastolic blood pressure, blood pressure lowering medication use, prevalent type 2 diabetes, prevalent coronary artery disease and prevalent heart failure. Online supplemental table 2 includes general characteristics of the four meta-analysed studies. The fixed-effect meta-analysis of all four cohorts yielded RR 1.03 (95 % CI 0.99 to 1.08) with substantial heterogeneity (I2=72%, p for heterogeneity=0.01) (figure 3A). The random-effects summary was RR 1.09 (95 % CI 0.99 to 1.21) (figure 3B). The leave-one-out analysis confirmed that exclusion of the WHI study materially affected both the summary estimate and between-study heterogeneity (online supplemental figure 4). When WHI was removed, the fixed-effect RR increased to 1.14 (95 % CI 1.05 to 1.23) and I2 decreased to 0% (p for heterogeneity=0.40), indicating homogeneity across the remaining three cohorts. In contrast, omitting any of the other studies produced minimal changes in the pooled estimate and did not materially reduce heterogeneity, suggesting that WHI was the primary source of inconsistency in the meta-analysis. Association between genome-wide European ancestry and atrial fibrillation risk among Black adults across four cohorts. Panel A shows the fixed-effect meta-analysis; Panel B shows the random-effects meta-analysis. RRs are per 10% increase in European ancestry. ARIC, Atherosclerosis Risk in Communities; CHS, Cardiovascular Health Study; RRs, relative risks; WHI, Women’s Health Initiative. Discussion In this large, prospective analysis of UKB participants who self-identified as Black, we found a suggestive association between higher European genetic ancestry and increased risk of incident AF. Our results are consistent with earlier studies from ARIC and CHS, which reported associations between European ancestry and AF among Black individuals. However, they contrast with the findings from the WHI cohort, which did not observe a similar relationship. The primary meta-analysis including all four cohorts did not demonstrate a statistically significant association and showed substantial between-study heterogeneity. Although exclusion of WHI produced a statistically significant pooled estimate with no detectable heterogeneity, this result represents a study-exclusion sensitivity analysis and should not be interpreted as independent confirmation of the association. Rather, it demonstrates the influence of differences across the small number of available studies and underscores the limited robustness of the pooled result. The observed heterogeneity in the association between European genetic ancestry and AF across cohorts, particularly between WHI16 and the other studies (ARIC,17 CHS18 and UKB8) may reflect key differences in study design, population characteristics and AF ascertainment. WHI included only postmenopausal women, whereas ARIC, CHS and UKB included both men and women, potentially introducing sex-related biological differences in AF risk.19 However, the sex-stratified analysis and analysis among only postmenopausal women in UKB did not reveal meaningful differences, which may point to differences in other study-specific factors. For example, in WHI, AF was identified using hospitalisation records adjudicated by the study and diagnosis codes obtained from linked Medicare claims data,13 which likely failed to capture asymptomatic or subclinical AF cases. In contrast, in ARIC and CHS, AF was ascertained through ECGs conducted at study visits, as well as from hospitalisation records and death certificates.3 20 21 In the UKB, AF was ascertained by integrating data from HES, primary care records, death registries and self-reported medical conditions.8 Additionally, WHI included both prevalent and incident AF cases, while the other studies focused exclusively on incident AF, which could introduce survival or selection bias. Taken together, the heterogeneity likely reflects the combined influence of differences in cohort composition, AF case definition and ascertainment rather than a single sex-specific effect. These between-study differences limit direct comparability and support cautious interpretation of the pooled association. It has been established that the genetic architecture of AF differs between African and European ancestry populations, and the allele frequencies at AF-associated loci may vary across these groups.22 23 Additional factors may help explain the observed paradox in AF incidence across racial groups. AF is often asymptomatic and can remain undetected without continuous or systematic monitoring. Prior studies have reported that AF is underdiagnosed in Black individuals compared with White individuals.3 24 Underascertainment in Black individuals may be driven in part by reduced access to medical care and differences in healthcare utilisation, all of which could contribute to the divergence in reported AF incidence.3 24 Further, Black individuals often have a greater burden of comorbid conditions such as hypertension, diabetes and heart failure, which are associated with increased risk of premature cardiovascular morbidity and mortality.25 These conditions can lead to earlier occurrence of adverse events or death from causes other than AF, effectively shortening the period during which AF could be clinically detected. This analysis focused on genetic ancestry (a biological trait) as a risk factor among individuals self-identified as Black (a social and political construct). Addressing this research question raises several important issues. First, it highlights the need to differentiate between genetic ancestry and racial self-identification to avoid biologising race and reinforcing false notions that racial categories are innate and genetically deterministic, which can obscure the role of structural and social conditions as drivers of observed disparities across racial and ethnic groups.26 Second, this distinction directly informed our study design. Restricting the sample to self-identified Black individuals minimises confounding by race-associated social determinants of health and allows for meaningful variation in European genetic ancestry; in contrast, self-identified White individuals have minimal ancestry variability, such that their inclusion would introduce substantial and largely unresolvable confounding.27 Finally, even though genetic ancestry is not a modifiable factor, its evaluation may assist in clarifying the mechanisms underlying the observed racial differences in AF risk and motivate the application of local-ancestry-aware genomic methods specific for populations with genetic admixture.28 These findings do not currently support the use of global genetic ancestry in clinical AF risk prediction. Established AF risk models rely on clinically measurable factors such as age, sex, blood pressure, anthropometric characteristics, smoking, diabetes, heart failure and other cardiovascular conditions. The association with European ancestry observed in the present study was modest, imprecise and heterogeneous across cohorts. Genetic ancestry should also not be used as a substitute for self-identified race or for social and structural determinants of health. Its potential value at present is primarily etiologic, for generating hypotheses about ancestry-related genetic architecture and motivating future local-ancestry or variant-level studies, rather than for individual clinical decision-making. Our analysis has several key strengths, including a large sample of self-identified Black individuals in the UKB with comprehensive genotype data, robust AF ascertainment through multiple linked data sources, and the capacity to model genetic ancestry both as a continuous and categorical variable. The use of spline modelling allowed for flexible detection of non-linear associations, and the meta-analysis approach enabled integration of data across multiple large cohorts to enhance generalisability. However, several limitations should be acknowledged. First, despite the relatively large overall sample, the number of AF events among individuals with high European ancestry was modest, contributing to wide CIs and reduced precision. This estimate should therefore be considered imprecise and hypothesis-generating. These findings should not be interpreted as demonstrating a threshold effect and require confirmation in substantially larger populations with greater representation at the upper end of the ancestry distribution. Second, heterogeneity in AF case definitions, population characteristics and ascertainment methods across the included cohorts may have introduced variability that influenced the pooled estimates. In addition, even with multivariable adjustment, residual confounding cannot be excluded, and any observed association should be interpreted as descriptive and hypothesis-generating rather than causal. Conclusion Our findings suggest a potential weak association between higher European ancestry and increased AF risk among individuals self-identifying as Black. Future studies with larger, diverse cohorts and harmonised methodologies are needed to clarify the role, if any, of genetic ancestry in AF pathogenesis and to better understand the interplay of biological and social factors shaping racial disparities in risk of AF and its complications. Supplementary material