Sex-stratified associations of C-VBQ with paraspinal muscle degeneration in degenerative cervical disease

Abstract Objective To investigate associations between the cervical vertebral bone quality score (C-VBQ) and C3-level paraspinal muscle morphology, including cross-sectional area (CSA) and a threshold-based, T2-weighted MRI-derived fat-infiltration surrogate (FI), in patients undergoing surgery for degenerative cervical disease, and to compare these associations between women and men. Methods This retrospective study included 93 patients (46 women and 47 men) with preoperative cervical MRI. C-VBQ was calculated exclusively on sagittal T1-weighted images as the median signal intensity of the C3–C6 vertebral bodies normalized to cerebrospinal fluid at C2, whereas CSA and threshold-based FI for four functional muscle groups were measured on the mid-C3 axial T2-weighted image. Associations were evaluated using sex-stratified Spearman correlations with Benjamini–Hochberg false discovery rate correction. Formal effect modification was tested in the overall cohort using sex-by-C-VBQ interaction terms in standardized linear models with heteroscedasticity-robust standard errors; adjusted models included age and BMI. An exploratory multivariable model and exploratory ROC analysis were conducted in women. Results Women had higher C-VBQ [3.25 (2.69–3.73) vs. 2.80 (2.05–3.40), P = 0.023] and higher overall FI values [10.20% (7.41%–12.00%) vs. 8.44% (6.48%–10.74%), P = 0.020], whereas men had larger CSA (all P < 0.001). In women, C-VBQ correlated positively with posteromedial FI ( ρ = 0.408, P = 0.005; FDR q = 0.021) and negatively with minimum T-score ( ρ = − 0.452, P = 0.002; q = 0.014) and posterior-group CSA ( ρ = − 0.410, P = 0.005; q = 0.021). No association remained significant in men after FDR correction. After adjustment for age and BMI, the sex-by-C-VBQ interaction was not statistically significant for the posteromedial FI ( P = 0.110), posterior-group CSA ( P = 0.094), or minimum T-score ( P = 0.379). In the exploratory female model, the posteromedial FI ( B = 0.025, P = 0.013) and minimum T-score (B = − 0.273, P = 0.035) were independently associated with C-VBQ. Exploratory ROC analysis yielded an AUC of 0.750 (bootstrap 95% CI 0.583–0.886) for high posteromedial FI values. Conclusions C-VBQ showed sex-stratified associations with DXA-derived bone status and paraspinal muscle degeneration. Although within-group associations were more consistent in women, formal adjusted interaction analyses did not demonstrate statistically significant effect modification by sex for the principal outcomes included in the interaction analysis. These findings are hypothesis-generating and require validation in larger cohorts.