Back

Impact of Life's Essential 8 Cardiovascular Health Score and Hypertension on Retinal Biological Aging

Squirrell, D.; Nielsen, C.; Vaghefi, E.; An, S.; Moghadam, S.; Xie, L.; McConnell, M. V.

2025-10-18 cardiovascular medicine
10.1101/2025.10.16.25338198 medRxiv
Show abstract

BackgroundThe American Heart Associations Lifes Essential 8 (LE8) score was designed to quantify an individuals cardiovascular health (CVH). Previously we developed a novel biomarker (retinal BioAge) to estimate biological aging from a deep-learning analysis of retinal images. This study investigates the association between retinal BioAge and CVH, as quantified by LE8, and identifies the key health factors driving this relationship. MethodsThis retrospective study included 4,887 participants (aged 40-70 years) from the UK Biobank. The retinal BioAgeGap (retinal BioAge minus chronological age) was calculated from retinal images. A CVH score was computed for each participant. ResultsParticipants in the highest CVH quartile were shown to have a negative retinal BioAgeGap while the lowest CVH quartile had a positive retinal BioAgeGap (-0.36 vs. +0.30 years; difference = - 0.65 years, P<0.001). Ablation analysis identified blood pressure as the most significant driver of this association, accounting for 66.3% of the total change in effect size (|{Delta}d|=0.11, P<0.001). Among individuals with high CVH scores, a negative retinal BioAge was no longer found among those with hypertension ([&ge;]130/80mmHg) compared to those without (+0.08 vs -0.52 years, P<0.001). ConclusionsBetter CVH is strongly associated with a "younger" retinal BioAge. Blood pressure was the predominant contributor to this association and an accelerated retinal BioAge was noted even in those individuals who had an otherwise favorable CVH score but were hypertensive. This study highlights retinal BioAges potential as a scalable, non-invasive tool for refining cardiovascular risk assessment.

Matching journals

The top 14 journals account for 50% of the predicted probability mass.

50% of probability mass above

"Similar papers" are the closest papers from that journal in the model's embedding space. They show what the match is built on, but the ranking comes mostly from a classifier over the whole training set, not from these examples alone.