Socioeconomic disadvantage is associated with excess cardiovascular disease (CVD), but the extent to which this gradient operates through modifiable biological pathways has not been fully quantified. This study used genetic approaches to estimate how much of the association between genetically proxied educational attainment (EA) and CVD is mediated by conventional cardiometabolic risk factors, and to search for shared genomic architecture linking EA and CVD endpoints.
The primary goals were to (1) quantify associations between genetically proxied EA and multiple CVD outcomes using Mendelian randomisation (MR), (2) assess independence from other socioeconomic traits, (3) estimate mediation through a broad set of cardiometabolic risk factors, and (4) identify loci where EA and CVD might share causal variants using proteome-wide cis-pQTL MR and colocalisation.
The study implemented two-sample MR to examine associations between genetically proxied EA and seven CVD endpoints. Multivariable MR (MVMR) was applied to evaluate whether EA effects were independent of related socioeconomic traits, including intelligence, income, occupational status and cognitive function.
Mediation was assessed using a two-step MR approach with a product-of-coefficients method to quantify the contribution of 22 individual cardiometabolic risk factors. A joint MVMR model estimated the combined attenuation when multiple mediators were included simultaneously.
To identify shared genetic architecture, the authors performed proteome-wide cis-pQTL MR and colocalisation analyses to detect loci where genetic signals for EA and CVD endpoints may reflect the same causal variant.
Higher genetically proxied EA was associated with lower risk of several cardiovascular and cardiometabolic endpoints. Significant protective associations were observed for coronary artery disease (CAD), myocardial infarction (MI), heart failure (HF), atrial fibrillation (AF), ischaemic stroke (IS), and type 2 diabetes (T2DM). The reported odds ratio range across these endpoints was 0.61–0.78, with associated P values ≤ 1.21 × 10⁻¹¹. The association with chronic kidney disease was reported as weaker.
When accounting for other socioeconomic traits in multivariable MR models, genetically proxied EA retained an independent effect on CVD outcomes. This suggests the associations observed are not fully explained by correlated socioeconomic variables such as intelligence, income, occupational status or measures of cognitive function.
Individual two-step MR analyses evaluated mediation through 22 cardiometabolic risk factors. In joint MVMR, accounting for multiple cardiometabolic mediators simultaneously resulted in substantial attenuation of EA’s protective associations for several outcomes.
Specifically, the combined set of cardiometabolic risk factors explained between 63% and 82% of the protective effect of EA on CAD, HF and T2DM. For AF, joint mediation analysis indicated full mediation, with the direct effect of EA on AF estimated as null when mediators were included. For ischaemic stroke (IS), mediation accounted for 63% of the EA effect, but a residual direct effect remained; confidence limits for mediation estimates extended to or above 100% for upper bounds, reflecting uncertainty in estimate precision.
These results indicate that a substantial portion of the protective association between higher EA and several CVD outcomes operates via modifiable metabolic and haemodynamic pathways encompassed by the assessed risk factors.
Proteome-wide cis-pQTL MR and colocalisation analyses identified four protein loci—LMOD1, DAG1, CD40, and MEGF9—as loci showing hypothesis-generating evidence of shared genetic architecture between EA and CVD endpoints. The authors present these loci as signals warranting further investigation, rather than definitive causal mechanisms.
The study’s findings suggest that much of the cardiovascular burden associated with lower educational attainment is mediated through modifiable cardiometabolic pathways. From a clinical and public-health perspective, intensified management of cardiometabolic risk factors in socioeconomically disadvantaged populations may substantially reduce education-related cardiovascular inequalities.
These results support targeted prevention and risk-factor control strategies (metabolic and haemodynamic management) as potential routes to attenuate the excess CVD risk observed with lower EA.
The article reports that the datasets analysed are available within the article and its Supplementary Material, and further information is available from the corresponding author on request. The authors declared no competing interests and confirm that appropriate ethical approvals and participant consents were obtained or documented as per journal requirements.
Readers should note that the proteome-wide loci identified are described as hypothesis-generating and require further validation. Details about instrument selection, specific mediator effect sizes, and statistical sensitivity analyses are provided in the full article and supplementary materials.