Cross-classification of bone mass status and possible sarcopenia delineates a gradient in fall risk: parallel concordance of BMD severity, fragility-fracture burden, and bone turnover biomarker profiles

Background Falls and related fragility fractures are common and consequential in community-dwelling older adults, yet bone mineral density (BMD) alone does not fully capture fall-related vulnerability.As skeletal and functional impairment may progress asynchronously, we performed a cross-classification based on bone mass status and possible sarcopenia to evaluate whether it identifies a gradient in fall risk and demonstrates parallel trends in BMD severity, fragility-fracture burden, and bone turnover biomarkers. Methods Baseline data were analyzed cross-sectionally in 1786 participants with dual-energy X-ray absorptiometry (DXA)-derived lumbar spine and total hip measures, available physical-function measures (handgrip strength and/or chair-stand testing), and fall history in the prior year. Outcomes were any fall (≥ 1) and recurrent falls (≥ 2) in the prior year. Stepwise-adjusted regression models estimated associations with 95% confidence intervals and derived marginal standardized risks and risk differences; trends, interactions, and restricted cubic splines were evaluated. Parallel outcomes included continuous T-scores, prior fracture history, and biomarkers of bone turnover (β-CTx, PINP, osteocalcin, alkaline phosphatase) and total 25-hydroxyvitamin D. Results The prevalence of any fall (≥1) increased stepwise across strata, from 12.8% in participants with normal bone mass without possible sarcopenia to 27.1% in those with osteoporosis and possible sarcopenia. In fully adjusted analyses, standardized risk was 14.8% versus 22.2% ( P for trend = 0.032), and the odds ratio for any fall was 1.68 (95% CI 1.01–2.79). Recurrent-fall estimates were directionally consistent but less precise. Lumbar spine and hip T-scores declined monotonically across strata; within osteoporosis, possible sarcopenia was associated with a further reduction in hip T-score (adjusted mean difference −0.24, 95% CI −0.39 to −0.09). Prior fragility fractures accumulated across strata. Bone turnover markers shifted upward, while total 25-hydroxyvitamin D trended modestly downward. Dose–response patterns for grip strength, chair-stand time, and T-scores were supported by spline analyses, and sensitivity analyses were consistent with the primary findings. Conclusions A joint “bone–function” stratification yields a reproducible fall-risk gradient with parallel clinical and metabolic correlates, providing a testable framework for longitudinal validation and risk phenotyping.