Bone Density Linked to Vision Loss Risk in Landmark Study

Researchers at the University of Cambridge have identified a biological link between bone strength and vision loss that could reshape how doctors identify people at risk for age-related macular degeneration, one of the leading causes of irreversible blindness in older adults. The finding, published September 14, 2026, in Cyborg and Bionic Systems, emerged from an unusually comprehensive analysis that wove together data from three separate populations—the UK Biobank, the U.S. National Health and Nutrition Examination Survey, and a hospital-based cohort in Tianjin, China—and tested the connection using genetic analysis, molecular profiling, machine learning, and animal experiments.

Age-related macular degeneration, or AMD, destroys central vision gradually and irreversibly. It progresses through stages marked by fatty deposits in the retina, deterioration of the pigmented layer beneath the retina, and eventually either abnormal blood vessel growth or tissue atrophy. The disease robs people of the ability to read, recognize faces, and move independently. Yet identifying who will develop it remains difficult. Xuehao Cui, the lead researcher, explained that previous studies hinting at a connection between bone density and AMD had been small, limited to single populations, or conducted only at one point in time—leaving fundamental questions unanswered about whether the link was real, whether it applied across different groups, and what biological mechanisms might explain it.

The team's approach was methodical. They analyzed data from thousands of participants across the three cohorts, using statistical models to assess whether lower bone mineral density predicted higher AMD risk. They applied machine learning algorithms—including random forests and gradient boosting—to identify which factors mattered most for predicting AMD. They performed genetic analysis to test whether the association reflected causation or mere correlation. They examined blood proteins and metabolites to find shared molecular signatures. And they created a rat model with artificially weakened bones to observe whether retinal changes followed.

The results pointed consistently in one direction: people with lower bone mineral density had higher AMD risk, across all three populations and regardless of how researchers measured or analyzed the data. Participants with AMD tended to be older, had weaker bones, and showed more signs of vascular disease, metabolic dysfunction, and inflammation. When researchers built predictive models, bone density and age emerged repeatedly as important contributors. The genetic analysis suggested that higher bone density may causally reduce AMD risk, and identified three genes—GZMA, NELL1, and COL2A1—as potential intermediaries. The molecular analyses revealed that bone and eye may share vulnerability through common pathways: inflammation, problems with how cells handle lipids, disrupted amino acid metabolism, and remodeling of the structural proteins that hold tissues together.

In the rat experiments, animals with artificially induced low bone mass showed thinning of the outer retina, narrowing of retinal blood vessels, and slower performance on spatial memory tests. The researchers were careful to note that these changes reflected retinal degeneration under low-bone-mass conditions but did not directly replicate human AMD pathology.

The significance lies in what bone density might become: an accessible, measurable marker of systemic biological aging that could flag people at higher risk for vision loss before symptoms appear. Bone density is routinely measured in clinical practice, particularly in older women screened for osteoporosis. If validated, the same scan could serve a dual purpose. Yet the researchers acknowledged substantial limitations. The study cannot prove that low bone density causes AMD—only that the two are linked. AMD diagnosis and bone density measurement varied across the three cohorts. The machine learning models performed differently in each population, suggesting their predictive value may not transfer universally. And the animal model, while suggestive, cannot fully capture human disease.

Cui and colleagues outlined the next steps: standardizing how AMD is classified and measured across studies, using unified bone density protocols, conducting long-term imaging follow-up to track retinal changes over time, gathering molecular data from eye tissue itself, and running larger animal experiments. Only then will it become clear whether bone density truly belongs in the clinic as a tool for identifying people who need closer eye monitoring, or whether it remains a fascinating biological signal without practical clinical weight.