Could skeletal health offer clues about retinal vulnerability? A new study suggests that bone mineral density (BMD) may provide a window into the systemic aging processes associated with age-related macular degeneration (AMD).
Researchers from the University of Cambridge (UK), Yale University (US) and Tianjin Eye Hospital (China) investigated a potential “bone–eye axis” using an unusually broad analytical framework. Published in Cyborg and Bionic Systems, the study combined data from three human cohorts with machine learning, Mendelian randomization (MR), proteomic and metabolomic profiling, and an experimental low-BMD animal model.
The analysis included 275,752 participants from UK Biobank, of whom 5,455 had AMD; 3,425 participants from the US National Health and Nutrition Examination Survey (NHANES), including 382 AMD cases; and a hospital-based Tianjin cohort of 1,000 participants, including 436 with AMD. Across all three datasets, participants with AMD were older and had lower BMD than those without the disease.
Crucially, the association persisted after adjustment for demographic, lifestyle and clinical factors. In the UK Biobank, higher baseline BMD was associated with a lower incidence of AMD, with an adjusted hazard ratio of 0.89 in the researchers’ main multivariable model. Comparable inverse relationships emerged in NHANES and Tianjin, although the magnitude differed substantially between cohorts. Quartile analyses also indicated a broadly graded relationship, with higher BMD associated with lower AMD risk.
Machine learning added another layer to the findings. Across the datasets, age and BMD repeatedly emerged among the most important features for distinguishing AMD cases from controls. However, the study authors emphasize that BMD’s incremental clinical value beyond established risk factors remains uncertain. The predictive models were developed separately within heterogeneous cohorts and have not undergone the external validation necessary for clinical use.
The team also investigated whether the association might extend beyond correlation. Two-sample MR provided genetic evidence consistent with higher BMD making a modest contribution to lower AMD risk, with no strong evidence for reverse causality. A two-step MR analysis highlighted three circulating proteins – granzyme A (GZMA), NEL-like protein 1 (NELL1) and collagen type II alpha 1 chain (COL2A1) – as potential molecular intermediates. Proteomic and metabolomic analyses additionally pointed toward shared pathways involving extracellular matrix remodeling, lipid transport, amino acid metabolism and inflammation.
An experimental model provided further, albeit preliminary, support. Rats with glucocorticoid-induced low BMD developed outer retinal thinning, narrower retinal vessels and delayed visual-spatial performance. Importantly, however, the animals did not develop hallmark features of human AMD, including drusen, retinal pigment epithelium dysfunction or choroidal neovascularization.
Taken together, the findings position BMD less as a new standalone AMD biomarker than as a potentially accessible marker of systemic biological aging. The study authors stress that differences in AMD ascertainment and BMD measurement between cohorts, residual confounding and the exploratory nature of the predictive analyses all warrant caution. For now, the bone–eye axis remains an interesting hypothesis, and one that could broaden the search for systemic signals of retinal vulnerability.
