Osteoporosis Drug Denosumab Outperforms Zoledronic Acid for Muscle Strength Over Three Years

Postmenopausal women with osteoporosis who received denosumab gained substantially more muscle strength and physical performance over three years than comparable women treated with zoledronic acid, according to a prospective observational cohort study conducted at Huashan Hospital, Fudan University in Shanghai. The study followed 186 postmenopausal women aged 50 or older with confirmed osteoporosis, 82 of whom initiated denosumab and 104 of whom initiated zoledronic acid as part of routine clinical care between March 2020 and April 2022. Because osteoporosis and sarcopenia frequently coexist in older adults, a combined state known as osteosarcopenia that carries greater risks of falls, fractures, disability and loss of independence than either condition alone, the researchers asked whether the two widely used antiresorptive drugs differed in their effects on muscle as well as bone.

The functional differences were evident within the first year and widened throughout follow-up. At 12 months, grip strength in the dominant hand, measured with a calibrated dynamometer, increased by 1.45 kilograms on average in the denosumab group versus 0.66 kilograms in the zoledronic acid group, a between-group difference of 0.79 kilograms. Usual gait speed over a 10-metre walkway improved by 0.041 metres per second with denosumab versus 0.017 with zoledronic acid, and the time needed to complete five rises from a standard chair fell by 0.82 seconds versus 0.32 seconds. All three between-group differences at 12 months were statistically significant at p < 0.001. By 36 months the gaps had grown: grip strength had improved by 2.52 kilograms with denosumab versus 0.99 kilograms with zoledronic acid, gait speed by 0.070 versus 0.027 metres per second, and chair-stand time had shortened by 1.47 versus 0.63 seconds. Linear mixed-effects models showed significant group-by-time interactions for all three primary functional outcomes, indicating that the trajectories genuinely diverged rather than differing only at isolated visits.

Changes in muscle mass followed a later and more nuanced course. Whole-body dual-energy X-ray absorptiometry, or DXA, showed no significant between-group difference in the appendicular skeletal muscle mass index at 12 months. A difference emerged at 24 months, when the index had risen by 0.161 kilograms per square metre with denosumab versus 0.076 with zoledronic acid, and it persisted at 36 months at 0.192 versus 0.089. Bone mineral density told a mixed story: lumbar spine density increased more with denosumab at every visit, while femoral neck density showed slightly greater gains in the zoledronic acid group at 24 and 36 months. The authors note that functional improvement appeared earlier, whereas separation in DXA-defined muscle mass emerged later, a staged pattern that has practical implications because DXA alone may miss early treatment effects when muscle function improves before measurable changes in appendicular lean mass appear.

Computed tomography added a structural dimension that DXA cannot capture. Using clinically acquired whole-abdomen CT scans and an open-source deep-learning segmentation tool called TotalSegmentator, with all outputs reviewed by a musculoskeletal radiologist blinded to treatment group, the researchers quantified a volumetric T12-L4 skeletal muscle volume index, skeletal muscle density, and indices of subcutaneous, visceral and intermuscular fat, each normalized to height squared. At 12 months, denosumab was associated with a greater increase in skeletal muscle volume index (1.19 versus 0.45), greater increase in muscle density (1.47 versus 0.47) and greater reduction in intermuscular fat volume index (−0.10 versus −0.04), all p < 0.001. By 24 and 36 months these differences remained pronounced, and the reduction in visceral fat also became significant, while subcutaneous fat did not differ between groups at any time point. The coordinated decline in intermuscular fat alongside rising muscle volume and density points to an improvement in muscle quality rather than a nonspecific shift in body composition.

The biologic rationale for the comparison rests on the receptor activator of nuclear factor-kappa B ligand pathway, commonly abbreviated RANKL. Denosumab is a monoclonal antibody that inhibits RANKL, an established antiresorptive mechanism with robust evidence for suppressing bone turnover, increasing bone mineral density and reducing fracture risk. But experimental work has shown that skeletal muscle also expresses components of the RANK/RANKL/osteoprotegerin axis, and that RANK signalling influences calcium handling, sarcoplasmic and endoplasmic reticulum calcium-ATPase activity, oxidative programming and muscle performance. Preclinical studies have further suggested that RANKL inhibition can improve muscle strength and insulin sensitivity, raising the possibility that denosumab has clinically relevant effects beyond fracture prevention.

The study’s biomarker data fit this framework. Denosumab produced a larger reduction in soluble RANKL, a larger increase in osteoprotegerin, and a substantially greater rise in the OPG/sRANKL ratio at every follow-up visit; at 12 months the ratio had increased by a median of 2.503 with denosumab versus 0.807 with zoledronic acid, and by 36 months the figures were 5.677 versus 1.448. Bone turnover markers, serum C-terminal telopeptide of type I collagen and procollagen type I N-terminal propeptide, fell further with denosumab throughout. Downstream, myostatin declined more and irisin rose more in the denosumab group, as did the follistatin-to-myostatin ratio, while inflammatory markers including high-sensitivity C-reactive protein, interleukin-6 and tumour necrosis factor alpha, along with fasting insulin and the HOMA-IR insulin-resistance index, all decreased more with denosumab. Brain-derived neurotrophic factor did not differ significantly in longitudinal change at any visit.

Within the denosumab group, the 12-month change in the OPG/sRANKL ratio correlated with later clinical gains: it was associated with change in grip strength (r = 0.520), gait speed (r = 0.364), chair-stand time (r = −0.470) and the DXA-derived muscle mass index (r = 0.316). A prespecified responder analysis split denosumab-treated participants at the within-group median of the 12-month OPG/sRANKL increase. High responders, who did not differ from low responders in baseline high-sRANKL prevalence, showed greater 36-month gains in grip strength (2.80 versus 2.21 kilograms), CT-defined muscle volume index (2.816 versus 2.199) and DXA-derived muscle mass index, and a markedly higher sarcopenia remission rate at 36 months (91.7 percent versus 55.0 percent). Exploratory mediation analyses, using bootstrap resampling with 5,000 iterations, found that the 12-month OPG/sRANKL change accounted for 37.0 percent of the association between treatment and 36-month grip-strength gain, and 39.9 percent of the association with CT-defined muscle volume gain, with significant residual direct effects in both models.

Sarcopenia outcomes, re-evaluated retrospectively under the updated Asian Working Group for Sarcopenia 2025 framework using prospectively collected data, followed the same staged pattern. Low muscle strength was less common with denosumab at 12 months (61.8 percent versus 79.1 percent), 24 months (38.8 versus 71.2 percent) and 36 months (26.7 versus 69.6 percent), while the prevalence of low muscle mass did not differ significantly between groups at any visit. Overall study-defined sarcopenia, defined by concurrent low muscle strength and low muscle mass, was less prevalent with denosumab at every visit, reaching 20.0 percent versus 46.0 percent at 36 months, and among participants with sarcopenia at baseline, remission was more frequent with denosumab at all time points, reaching 75.0 percent versus 37.0 percent at 36 months. Falls and new fragility fractures, by contrast, did not differ significantly between groups at any time point, with cumulative three-year fracture frequencies of 5.0 percent versus 10.1 percent that were directionally favourable but not statistically significant.

The researchers took several steps to address the inherent limitation that treatment was chosen by physicians rather than randomly assigned. Sensitivity analyses adjusted for age, baseline vitamin D supplementation and baseline variables that differed between groups, and additional models incorporated regular physical activity, resistance exercise and physical therapy as time-varying covariates; the primary functional findings were materially unchanged. A propensity score-weighted analysis using stabilized inverse probability of treatment weighting improved covariate balance, reducing the maximum absolute standardized mean difference from 0.446 to 0.110, and the weighted 36-month between-group differences remained significant for grip strength (1.607 kilograms, 95 percent confidence interval 1.386 to 1.828), gait speed (0.0455 metres per second) and chair-stand time (−0.863 seconds). An exploratory age-stratified analysis found no evidence of treatment-by-age interaction for any functional outcome.

The authors are careful about what the data do and do not establish. The mediation analyses are exploratory and hypothesis-generating, and the OPG/sRANKL response patterns are best interpreted as statistically compatible with partial mediation rather than proof that the pathway directly caused the later functional or structural benefits. Residual confounding from unmeasured clinical or lifestyle factors cannot be excluded in an observational cohort, and the cohort was intentionally enriched for women prescreened for sarcopenia risk, most commonly because of low handgrip strength, so the high baseline prevalence of low muscle strength and sarcopenia, roughly 84 percent in both groups, does not reflect an unselected osteoporosis population. With those caveats, the study describes a coherent sequence in which denosumab was associated first with stronger biologic pathway activation, then earlier gains in muscle strength and physical performance, coordinated CT-defined remodelling of muscle and fat compartments, later gains in DXA-defined muscle mass, and ultimately a lower study-defined sarcopenia burden, supporting further prospective evaluation of RANKL inhibition as a strategy for osteosarcopenic phenotypes.

Subject of Research: Comparative effects of denosumab versus zoledronic acid on muscle strength, physical performance and body composition in postmenopausal osteoporosis

Article Title: Denosumab Versus Zoledronic Acid for Muscle Strength and Physical Performance in Postmenopausal Osteoporosis: A 3‐Year Prospective Study

Article References: Li, X., Yuan, Z., Ma, N., Huang, B., Tao, H., & Chen, S. (2026). Denosumab Versus Zoledronic Acid for Muscle Strength and Physical Performance in Postmenopausal Osteoporosis: A 3‐Year Prospective Study. Journal of Cachexia, Sarcopenia and Muscle, 17(5), Article e70406. https://doi.org/10.1002/jcsm.70406

Image Credits: AI Generated

DOI: 10.1002/jcsm.70406

Keywords: denosumab, zoledronic acid, osteoporosis, sarcopenia, osteosarcopenia, RANKL inhibition, grip strength, physical performance, body composition, DXA, computed tomography, postmenopausal women