Characteristics of the pelvic axial rotation in adolescent idiopathic scoliosis: a comparison between major thoracic curve and major thoracolumbar/lumbar curve

Background context As the biomechanical foundation of the spine, the pelvis was found to display rotation in the transverse plane in adolescent idiopathic scoliosis (AIS). However, the possible factors influencing the pelvic axial rotation (PAR) and its mechanism in patients with AIS remain unclear. Purpose To characterize the PAR in AIS patients with right major thoracic (MT) or major left thoracolumbar/lumbar (TL/L) curve and to explore the associated influencing factors and probable mechanism of compensation by analyzing the association between PAR and other spinal radiographic parameters. Study design Retrospective study. Patient sample All patients with the primary diagnosis of AIS from January 2008 to November 2009 were retrieved from our scoliosis database. Outcome measures Age, Cobb angle, and apex rotation of the main curve and the compensatory curve, curve flexibility, and PAR. Methods One-hundred thirty-two patients with right MT (14.7±1.8 years, 48±6.9°) and 83 patients with left major TL/L (14.7±1.5 years, 46±6.6°) were retrospectively reviewed. On standing full-spine posteroanterior radiographs, the PAR was quantified by the left/right hemi-pelvis width ratio (L/R ratio); then the subjects in both MT and major TL/L groups were divided into two subgroups: L/R ratio ≤1 (pelvis rotated to the right, R-PAR group); and L/R ratio >1 (pelvis rotated to the left, L-PAR group). Comparisons of all variables were performed between the L- and R-PAR subgroups; correlation and regression analysis were carried out to identify the influencing factors of PAR. Results The majority of the MT and major TL/L patients (75.8% vs. 60.2%) displayed right pelvic rotation, which was in the same direction as the thoracic curve. The incidence of R-PAR was greater in the MT patients than the major TL/L ones (p=.016). Lumbar flexibility in MT patients with R-PAR was greater than in MT patients with L-PAR (0.96±0.27 vs. 0.81±0.33, p=.038), which was contrary to the findings in the major TL/L patients (L-PAR>R-PAR, 0.79±0.15 vs. 0.70±0.22, p=.024). In the MT patients with R-PAR, the L/R ratio showed significant positive correlations with the lumbar Cobb angle (r=0.424) as well as with the apex rotation (r=0.488), which was further identified as an influencing factor (R=0.418) of the PAR. Significant positive correlations between L/R ratio and thoracic apex rotation (r=0.361) also were detected. Conclusion The majority of AIS patients with right MT or left major TL/L curves were found to have PAR to the right, in the same direction as the thoracic curve. The lumbar flexibility and apex rotation significantly influenced the PAR direction and magnitude. Moreover, the pelvis might be involved in compensation for the MT deformity through its connection with the lumbar spine.