The osteoclast paradox in type 2 diabetes: low turnover, skeletal fragility, and implications for osteoclast-targeted therapy

Type 2 diabetes mellitus (T2DM) is increasingly recognized as a skeletal-fragility condition characterized at the population level by lower circulating markers of bone formation and resorption, frequently preserved or increased DXA-derived areal bone mineral density (aBMD), and increased fracture risk. These observations arise from different measurement levels and have not been demonstrated to constitute a single biological phenotype within the same individuals. In this Review, we use the term “osteoclast paradox” as an author-proposed heuristic framework—not a diagnostic entity or an established mechanistic syndrome—to organize apparent dissociations among circulating bone-turnover markers, osteoclast-lineage biology, tissue-level remodeling, matrix properties, and treatment response in T2DM. Direct human evidence linking these levels remains limited. A small exploratory human study suggests a possible dissociation between osteoclast precursor phenotype and mature function, whereas histomorphometry indicates that remodeling can remain coupled despite delayed progression; neither observation establishes generalized preservation or enhancement of mature osteoclast resorptive function. Experimental studies further show context-dependent effects of hyperglycemia, advanced glycation end products, oxidative stress, insulin signaling, inflammation, and the marrow environment on osteoclast differentiation and function. The observed skeletal phenotype is also modified by obesity and body composition, renal and microvascular disease, and glucose-lowering or weight-loss therapies; these overlapping exposures should not be attributed automatically to T2DM itself. Matrix glycation, altered material behavior, osteocyte dysfunction, osteoblast-lineage impairment, and altered basic multicellular unit kinetics may contribute to skeletal fragility despite low circulating turnover markers. Pooled randomized-trial analyses suggest that antiresorptive treatment responses are not obviously attenuated by T2DM, although diabetes-specific fracture estimates are unevenly powered and treatment response should not be interpreted as mechanistic validation of the proposed framework. Osteoanabolic studies demonstrate aBMD, trabecular bone score, and early microarchitectural responses, but diabetes-specific fracture endpoints remain limited. The framework is therefore intended to separate non-equivalent biological and clinical endpoints, identify evidence gaps, and generate testable hypotheses for osteoclast-targeted research. Current treatment should remain fracture risk–based rather than guided by an unvalidated turnover phenotype.