Ethnic and sex inequalities in premature coronary artery disease across disaggregated South Asian and Black subgroups in England: a population-based cohort study

Background

Premature, or early-onset, coronary artery disease (CAD) carries lifelong consequences. Ethnic inequalities in CAD are established in the UK, but risk estimates mostly derive from events after middle age, while ethnicity is aggregated for South Asian and Black populations. Disaggregated, sex-specific risk estimates would permit better recognition and more targeted prevention.

Methods

We used Clinical Practice Research Datalink (CPRD) Aurum, an English primary care database linked to hospital, mortality, and deprivation data. We included adults aged 18–45 of White European, South Asian (Indian, Pakistani, Bangladeshi), or Black (African, Caribbean) ethnicity, followed for up to 20 years. Incident premature CAD (onset ≤45) was a first myocardial infarction or coronary revascularisation. We estimated age-adjusted and fully adjusted incidence rate ratios (IRRs) versus White Europeans by Poisson regression, testing an ethnicity– sex interaction.

Findings

Among 14.8 million adults contributing 80.9 million person-years, 16,001 premature CAD events occurred (77.6% in men). In men, the combined South Asian age-adjusted IRR was 1.87 [95% CI 1.77, 1.99], ranging from 1.39 [1.28, 1.50] in Indian to 2.79 [2.52, 3.07] in Bangladeshi men, persisting after full adjustment and already evident at ages 18-26. The combined Black IRR was 0.64 [0.57, 0.71], lowest in African (0.57[0.50, 0.65]) and highest in Caribbean men (0.82 [0.68, 0.98]). In women, the combined South Asian IRR showed no overall excess (1.10 [0.95, 1.26]), concealing a clear excess in Pakistani women (1.56 [1.29, 1.89]). The ethnicity–sex interaction was significant (p<0.001); the male-to-female ratio was highest in Bangladeshi individuals (8.4:1 versus 3.3:1 in White Europeans).

Interpretation

Aggregated ethnic categories conceal sex-specific subgroups at high risk of premature CAD, a risk already present in early adulthood. Current screening and health check programmes beginning at age 40, start too late to reach these higher-risk, underserved groups.

Funding

Kusuma Trust and NIHR UCLH BRC.

Research in context

We searched MEDLINE (via PubMed) from database inception to 27 July 2026, restricted to English-language articles. We combined subject headings and free-text terms for coronary artery disease, myocardial infarction, and ischaemic heart disease with terms for premature or early-onset disease and terms for ethnicity, race, and specific ethnic groups. The full search strategy is provided in the appendix. It is well established that South Asian populations in the UK and internationally develop coronary disease earlier and at higher rates than White populations. This evidence comes mainly from hospital case series, cross-sectional risk-factor surveys, and mortality data, and uses aggregated ethnic categories such as “South Asian” and “Black”. The few studies that disaggregated these categories, found substantial within-group variation, but no study has estimated premature coronary artery disease incidence across disaggregated ethnic subgroups in England using linked primary care, hospital, and mortality records, and no study has examined how sex modifies ethnic differences within those disaggregated groups.

Added value of this study

In 14.8 million adults aged 18–45 years, the largest UK study of premature CAD to date, we show that ethnic inequalities are already established in early adulthood, within an age band existing risk screening does not reach. Disaggregation reveals heterogeneity that aggregated analyses obscure: Bangladeshi men carry nearly threefold and Pakistani men approximately twofold the coronary risk of White European men, with a smaller but consistent excess in Indian men, while African incidence sits well below the White rate and Caribbean approaches it. A significant ethnicity–sex interaction exposes a clear excess in Pakistani women entirely hidden within a null combined South Asian female estimate, identifying a high-risk group not previously described.

Implications of all the available evidence

• Disaggregated, sex-specific ethnic reporting should be standard practice across CVD and broader clinical research. Aggregated categories obscure the groups at highest risk and can actively misinform prevention, as here, where a null result for South Asian women overall concealed a clear excess in Pakistani women.

• The elevated risk is established in early adulthood years before the National Health Service (NHS) health check at age 40, therefore current programme reaches too late. Earlier, targeted CVD risk assessment in Bangladesh, Pakistani men and Pakistani women, particularly in deprived communities is warranted.