Inflammaging Biomarkers Causally Reshape Retinal Vessels and Predict Cardiovascular Fate

Chronic low-grade inflammation that accumulates with age — termed 'inflammaging' — is increasingly recognized as a driver of microvascular deterioration, cardiovascular disease, and shortened lifespan. However, establishing the causal direction of these relationships has been challenging because conventional observational studies cannot separate cause from effect or rule out confounding. This study used two-sample Mendelian randomization (MR), leveraging genome-wide significant genetic variants as instrumental variables, to causally interrogate three questions: do inflammaging plasma biomarkers alter retinal microvascular architecture; does altered retinal vasculature subsequently affect cardiovascular disease risk; and do these vascular changes influence longevity phenotypes?

The study drew on large-scale GWAS summary data and UK Biobank retinal fundus imaging data from approximately 50,000 to 70,000 individuals of European ancestry. The primary retinal phenotype was fractal dimension (Df), a quantitative measure of the complexity and space-filling capacity of the retinal vascular tree derived from automated image analysis. Higher Df indicates a denser, more branched microvascular network associated with healthy vasculature, while lower Df has been epidemiologically linked to cardiovascular risk factors. A curated panel of inflammaging-associated plasma proteins was assembled based on prior literature, including GDF15, IL-18, CXCL10 (IP-10), MCP-1 (CCL2), and several others, with GWAS instruments drawn from large plasma proteomics studies.

The MR analyses revealed that genetically elevated levels of multiple inflammaging proteins significantly reduced retinal vascular fractal dimension. GDF15 showed among the strongest effects, with each standard deviation increase in genetically-predicted GDF15 associated with a meaningful reduction in Df (beta approximately −0.06 to −0.09, FDR-adjusted p < 0.05). IL-18 and CXCL10 showed similar directional effects. Sensitivity analyses using MR-Egger, weighted median, and weighted mode methods were largely consistent with the primary inverse-variance weighted estimates, supporting robustness against potential pleiotropy. Steiger filtering confirmed the expected causal direction from biomarker to vascular phenotype rather than reverse causation.

In the second MR step, reduced retinal Df (instrumented by SNPs for Df) was causally associated with increased risk of atrial fibrillation, coronary artery disease, and heart failure in large cardiovascular GWAS consortia. Effect sizes were modest but statistically significant after multiple testing correction, with odds ratios in the range of 1.10–1.25 per SD reduction in Df for the cardiovascular outcomes tested. Reduced Df was also associated with shorter healthspan (a composite of major age-related disease-free survival) and parental lifespan in MR analyses of longevity GWAS data, connecting microvascular complexity to the broader aging trajectory.

The study also explored the reverse direction — whether cardiovascular risk factors or disease states causally reduce retinal Df — and found bidirectional relationships for some traits, suggesting a feedback loop between systemic vascular health and retinal microvascular structure. Importantly, the inflammaging proteins that most consistently affected Df were also independently associated with cardiovascular outcomes in direct MR analyses, supporting a partially mediated pathway through retinal vasculature but also suggesting direct systemic effects. Colocalization analyses provided additional evidence that shared genetic architecture underlies the inflammaging-Df-cardiovascular axis at specific loci.

The findings position retinal imaging as a potentially powerful, non-invasive biomarker of systemic inflammaging burden and cardiovascular risk stratification. Because fundus photography is inexpensive and increasingly automated with AI-based analysis, incorporating retinal vascular complexity metrics into clinical risk prediction pipelines could offer an early warning system for patients at elevated inflammaging-driven cardiovascular risk. The authors acknowledge that analyses were conducted primarily in European-ancestry populations, and that the causal estimates reflect lifetime genetic exposure rather than short-term modifiable biomarker changes, warranting validation in diverse cohorts and intervention studies.