Compartment-specific effects of muscle strength on bone microarchitecture in women at high risk of osteoporosis

Abstract

Background: It is well known that skeletal integrity is influenced by the musculature. Poor muscle strength (i.e. sarcopenia) is considered a major predictor of fragility fractures. While this observation appears particularly relevant for older women with increased risk of osteoporosis, there has been no comprehensive investigation to determine the influence of muscle performance on compartment-specific bone microarchitecture in multiple body regions.

Methods: We retrospectively analysed data from different muscle performance and bone microarchitecture assessments in 230 women (aged 21 to 87 years) at high risk of osteoporosis. Muscle performance tests included grip strength and chair rising test (CRT) combined with mechanography. Balance was determined by Romberg posturography. Areal bone mineral density (BMD) was measured by dual-energy X-ray absorptiometry (DXA) at the hip and lumbar spine. Compartment-specific volumetric BMD, microarchitecture, and geometry were assessed by second-generation high-resolution peripheral quantitative computed tomography (HR-pQCT) at multiple skeletal sites (distal radius, tibia, and fibula). Regression models were applied to test for interactions between muscle and bone parameters. Subgroups were defined to compare women with osteoporosis and osteosarcopenia regarding BMD and microarchitecture.

Results: While osteoporosis was diagnosed in 115/230 (50.0%) women, sarcopenia was detected in 38/230 (16.5%). Positive associations of both grip strength and CRT maximum force with cortical geometric and microarchitectural parameters were detected at all measured sites, with the strongest effect applying to CRT maximum force and tibial parameters (e.g. tibial cortical area R² = 0.36, P < 0.0001, and tibial cortical thickness R² = 0.26, P < 0.0001). Balance parameters showed much weaker or no associations with HR-pQCT parameters. Major associations between muscle strength and trabecular parameters could not be confirmed. Age and body mass index were confirmed as negative and positive predictors for several microarchitectural parameters, respectively. An independent predictive value of grip strength on radial, tibial, and fibular (all P < 0.01) cortical area and of CRT maximum relative force on cortical thickness (all P < 0.05) was revealed. Women with osteosarcopenia showed significantly reduced cortical HR-pQCT parameters but no differences in DXA values compared with women with osteoporosis but no sarcopenia. Stratification by fracture and treatment status revealed that vertebral fractures and denosumab treatment altered the muscle-bone interaction.

Conclusions: A systemic interaction between muscle strength and bone microarchitecture was demonstrated, and this interaction appears to be primarily with the cortical bone compartment. The value of muscle assessments in fracture risk evaluation may be partly mediated by their effects on bone microarchitecture.

Keywords: HR-pQCT; Mechanography; Microarchitecture; Muscle performance; Osteoporosis; Sarcopenia.

Figure 1

Associations of muscle performance and balance measures with HR‐pQCT parameters assessed at multiple sites. Influence of (A) grip strength, (B) chair rising test (CRT) maximum force, (C) CRT time per repetition, (D) Romberg path length eyes open (EO), and (E) Romberg path length eyes closed (EC) on HR‐pQCT parameters at the distal radius, tibia, and fibula. Colour‐coded charts represent the coefficients of determination R ² determined by linear regression analyses. *P < 0.05, **P < 0.001.

Figure 2

Influence of grip strength on volumetric BMD and cortical parameters at different skeletal sites. Visualization of the associations between grip strength and (A) total BMD (Tt.BMD), (B) cortical thickness (Ct.Th), and (C) cortical area (Ct.Ar) at the distal radius, tibia, and fibula. Linear regression analyses were performed and confidence intervals (CIs) of the respective regression slopes, the coefficients of determination R ², and the P‐values were calculated. Numbers in bold indicate statistical significance (P < 0.05).

Figure 3

Differences in HR‐pQCT parameters between women with osteoporosis (Opo) and women with osteosarcopenia (Osa) at different skeletal sites. Comparison of (A) total BMD (Tt.BMD), (B) bone volume to tissue volume (BV/TV), and (C) cortical thickness (Ct.Th) at the distal radius, tibia, and fibula. The Shapiro–Wilk test was used to evaluate the normal distribution of the data. Then, the unpaired two‐sided t test was used for normally distributed data and the Mann–Whitney U test was used for non‐parametric data. The dashed lines of violin plots represent the median and the quartiles. Exact P‐values are displayed.

Figure 4

Associations between different muscle performance measures and HR‐pQCT parameters stratified by fracture status. (A) Comparison of the associations between grip strength and HR‐pQCT parameters at the distal radius, tibia, and fibula in women without fracture, with vertebral fracture, and with peripheral fracture. (B) Comparison of the associations between chair rising test (CRT) maximum force and HR‐pQCT parameters at the distal radius, tibia, and fibula in women without fracture, with vertebral fracture, and with peripheral fracture. Colour‐coded charts represent the coefficients of determination R ² determined by linear regression analyses. *P < 0.05, **P < 0.001.

Figure 5

Associations between different muscle performance measures and HR‐pQCT parameters stratified by treatment status. (A) Comparison of the associations between grip strength and HR‐pQCT parameters at the distal radius, tibia, and fibula in women without specific treatment, treated with bisphosphonates, and treated with denosumab. (B) Comparison of the associations between chair rising test (CRT) maximum force and HR‐pQCT parameters at the distal radius, tibia, and fibula in women without specific treatment, treated with bisphosphonates, and treated with denosumab. Colour‐coded charts represent the coefficients of determination R ² determined by linear regression analyses. *P < 0.05, **P < 0.001.