Elevated fracture risk in people with type 2 diabetes is well documented, yet the cellular and structural mechanisms behind it remain poorly understood. If bone loss rates are similar across glycemic categories, the fracture paradox must arise from qualitative changes in bone tissue — not from the rate of density decline — which has significant implications for how clinicians screen and intervene in this population.
This longitudinal analysis from the Framingham Study tracked 769 adults (mean age 67, 59% women) using high-resolution peripheral quantitative computed tomography (HR-pQCT) at two skeletal sites — the tibia and radius — across a mean 8-year interval between 2012–2016 and 2021–2023. Participants were stratified into normoglycemia (57%), prediabetes (31%), and type 2 diabetes (12%). After adjustment for age, sex, weight, and height, changes in cortical volumetric bone mineral density, trabecular microarchitecture, and estimated bone strength did not differ significantly across glycemic groups. At the radius, cortical density declined by approximately 1.5% in diabetes versus 2.4% in normoglycemia — numerically less loss in the diabetic group, though statistically equivalent.
This finding reframes the clinical problem considerably. The assumption that higher glucose exposure accelerates bone resorption or impairs formation at a rate detectable by standard densitometric metrics appears unsupported over this timeframe. Instead, the literature increasingly points to altered collagen cross-linking driven by advanced glycation end-products (AGEs), impaired osteocyte mechanosensing, and cortical porosity as the structural culprits behind diabetic skeletal fragility — none of which are captured well by density measurements alone. HR-pQCT is more sensitive than DXA, yet even this advanced modality may not fully reveal the qualitative tissue-level deterioration occurring within seemingly stable bone compartments. A key limitation is the 12% diabetes prevalence, which may limit statistical power to detect subtle between-group differences. That the cohort is predominantly older, white, and community-dwelling also restricts generalizability. Overall, this is a confirmatory, methodologically rigorous study that strengthens the case for developing fracture-risk tools in diabetes that go beyond bone density metrics.