Association between LIV + 1 bone mineral density asymmetry and distal adding-on phenomenon in adolescent idiopathic...

Distal adding-on phenomenon remains a common complication following corrective surgery for adolescent idiopathic scoliosis (AIS). This study aimed to characterize the spatial distribution of vertebral bone mineral density (BMD) across the scoliotic spine and to examine whether preoperative BMD asymmetry at the vertebra immediately distal to the lowest instrumented vertebra (LIV + 1) may serve as a complementary biomechanical risk marker for postoperative distal adding-on within existing LIV-selection frameworks.

This retrospective cohort study included 90 patients with AIS who underwent posterior spinal fusion and had preoperative full-spine computed tomography scans. Patient-specific three-dimensional finite element analysis (3D-FEA) models were constructed to quantify bilateral vertebral BMD along simulated pedicle-screw trajectories. Patients were grouped by Cobb-angle strata and by postoperative adding-on status. ROC analysis was used to evaluate the predictive performance of LIV + 1 BMD asymmetry. Multivariable and Firth penalized logistic regression analyses were performed to assess its association with distal adding-on.

The mean whole-spine BMD was 0.85 ± 0.14 g/cm². No significant difference was observed between the overall concave and convex sides (p = 0.510), although region-specific asymmetry patterns were identified. Distal adding-on occurred in 15 patients (16.7%). Patients with adding-on showed greater LIV + 1 BMD asymmetry than those without adding-on (0.14 ± 0.06 vs. 0.11 ± 0.03 g/cm², p = 0.006). LIV + 1 BMD asymmetry demonstrated good discrimination (AUC = 0.865, 95% CI 0.77-0.96), with an optimal cutoff value of 0.112 g/cm². Based on the optimal cutoff, adding-on occurred in 44.8% of patients above this cutoff and 3.3% of patients at or below this cutoff (Fisher's exact OR = 24.0, p < 0.001); the negative predictive value was 96.7%. The between-group effect size was large (Cohen's d = 0.85; Hedges' g = 0.85). In Firth penalized logistic regression, LIV + 1 BMD asymmetry remained associated with adding-on (OR = 1.338 per 0.01 g/cm² increase, 95% CI 1.077-1.664, p = 0.009).

Vertebral BMD in AIS exhibits a region-specific asymmetric distribution pattern. Increased preoperative LIV + 1 BMD asymmetry is associated with postoperative distal adding-on and may provide complementary, hypothesis-generating biomechanical information for risk stratification and distal fusion-level planning. These findings should be validated in larger cohorts before BMD asymmetry is used as a stand-alone determinant of LIV selection.