Diagonal paraspinal muscle degeneration and osteoporosis were associated with rotatory subluxation in degenerative lumbar scoliosis.

purposeThis study aimed to investigate the potential association between paraspinal muscle (PSM) degeneration, osteoporosis, and vertebral rotatory subluxation (VRS) in patients with degenerative lumbar scoliosis (DLS).

methodsThis retrospective study analyzed standing anteroposterior radiographs to assess the coronal (Cobb angle, coronal balance distance, and lateral translation) and sagittal parameters (thoracic kyphosis, lumbar lordosis, and sagittal vertical axis). Patients were categorized into rotatory subluxation (RS, lateral translation ≥ 5 mm) and non-RS groups, with the RS group further subdivided into single- and double-level subgroups based on the frequency of RS occurrence. The bilateral PSM cross-sectional area (CSA) and fat infiltration rate (FI%) at L1-S1 were evaluated using lumbar magnetic resonance imaging. Vertebral rotation angles and Hounsfield units (HU) were quantified using reconstructed axial computed tomography images, with osteoporosis defined as L1 HU ≤ 110. The ratios of convex-to-concave measurements (Ro FI%, Ro CSA, and Ro HU) were calculated. Spearman's correlation and logistic regression analyses were used to explore the associations between PSM degeneration, osteoporosis, and VRS. Intraoperative multifidus samples from patients with RS-1 (upper RS level) and RS + 1 (below RS level) DLS were analyzed histologically to assess regional fat infiltration and muscle atrophy.

resultsPrecisely 166 patients were included in this study: 90 (54.2%) with and 76 without VRS. The apex vertebrae were predominantly at the RS-1 level (67%). The single- and double-level RS groups showed significantly higher FI% of erector spinae and multifidus than the non-RS group at various levels (P < 0.05). Patients with VRS generally exhibited osteoporosis (64.4% vs. 47.4%, P = 0.027). The HU values for both single- and double-level RS patients were significantly lower than those in the non-RS group at multiple levels (P < 0.05). Notably, the FI% of the multifidus and HU measurements in the RS group showed more severe asymmetry at multiple levels than those in the non-RS group (P < 0.05). Ro HU and Ro FI% of the multifidus were generally < 1 above and > 1 below the RS level, suggesting that asymmetric PSM degeneration and osteoporosis above and below the RS level were reversed. Logistic regression analysis showed that VRS was significantly associated with multifidus Ro FI% above the RS (odds ratio, 0.021; 95% confidence interval [CI], 0.001 - 0.147, area under the curve [AUC] = 0.766; P < 0.001) and multifidus Ro FI% below the RS (odds ratio, 1.956; 95%CI, 0.930 - 4.114, AUC = 0.664; P = 0.007). Youden's index analysis showed cut off value of Ro FI% was 0.7 or 1.3. The Cobb angle and HU value at L1 (odds ratio, 0.982; 95%CI, 0.971 - 0.992, AUC = 0.712; P = 0.005) were also independently associated with VRS.

conclusionsVRS demonstrates a significant association with PSM degeneration and osteoporosis in patients with DLS. The low HU of L1 and diagonal fatty degeneration of multifidus (Ro FI%<0.7 or > 1.3) across the VRS level suggests a biomechanical mechanism contributing to VRS progression. These findings highlight the importance of evaluating PSM and bone density to identify potential VRS and predict the natural history of DLS, thereby informing targeted interventions.