Blood protein age estimates linked to eye aging and disease
An observational study of 55,028 people in three UK and Chinese cohorts linked protein-based age estimates to eye aging and future disease.
55,000 participants, independent replication
Limit: The study assessed disease links, not whether using protein estimates improves patient care
Edited by Boris Djordjevic · AgingResearch.News
Researchers analyzed blood proteins from 55,028 people in an observational study of three UK and Chinese cohorts. The UK group included 53,014 participants recruited at ages 40–69; 53.9% were female. The two Guangzhou groups included 1,584 and 430 participants, from cohorts recruiting at ages 35–85 and 18–85, respectively.
They trained computer models to estimate age from protein patterns and compared those estimates with computer-analyzed eye images and disease diagnoses. Older-than-expected protein profiles were associated with image-based eye aging and with existing and future age-related eye diseases, beyond actual age. These included cataracts, which cloud the eye’s lens, and glaucoma, which damages the nerve carrying visual signals to the brain. A simpler protein model retained predictive performance, and its disease associations were replicated in the Guangzhou Diabetic Eye Study.
Why it matters
The authors propose combining protein-based age estimates with artificial intelligence as a scalable way to track eye health and investigate aging pathways shared by several eye diseases.
Caveats
The observational design cannot establish cause and effect or whether using these estimates improves care. The UK group was predominantly White, and external validation came from Guangzhou.
The paper
Proteomic clocks combined with deep learning phenotypes track eye aging and diseases
Sun Yat-sen University
NPJ Digital Medicine, Jun 6, 2026, Open access
