Accurate Prediction of Height Loss in Adolescent Idiopathic Scoliosis: Cobb Angle Alone is Insufficient

Benlong Shi,Saihu Mao,Leilei Xu,Xu Sun,Zhen Liu,Jack C Y Cheng,Zezhang Zhu,Yong Qiu
DOI: https://doi.org/10.1007/s00586-016-4530-4
2016-01-01
Abstract:PURPOSE:Spinal height loss due to scoliosis was mostly calculated by equations with Cobb angle as the unique independent variable in adolescent idiopathic scoliosis (AIS), with their accuracy being seriously doubted. The purpose of this study was to compare and correlate the measured loss in spinal height (ΔSHm) with the calculated loss in spinal height (ΔSHc), and to identify the key factors associated with height loss in AIS. METHODS:This retrospective study included two stages. In stage I, 277 surgically treated AIS patients were reviewed, and divided into single curve and double curve groups. The accuracy and limitation of each correction equation was testified by comparing the data sets of ΔSHc with the ΔSHm, as well as comparing the pre- and post-op spinal length for each correction equation. In stage II, 235 curves within 100 AIS patients were selected to explore the relationship between loss in vertical curve height (△CH) and Cobb angle, curve length (CL), number of vertebrae within the curve (NVC), tilt angle of each curve (α) through partial correlation analysis. RESULTS:In stage I, the △SHm averaged 3.5 cm in single curve group and 4.8 cm in double curve group. The ΔSHc was only comparable with ΔSHm using the Bjure's equation. Pre- and post-op spinal length was only comparable in Bjure's corrected group with Cobb angle <60° (p > 0.05). In stage II, the partial correlation coefficients of △CH with Cobb angle, CL, α, and NVC were 0.889, 0.493, -0.723 and -0.175 (p < 0.01), respectively. The △CH could be calculated by the following formula: △CH (mm) = 29.305 + 0.506Cobb + 0.083CL - 0.592α - 0.796NVC. CONCLUSIONS:Previous height correction equations employing Cobb angle as the unique independent variable are inaccurate. Cobb angle, curve length, curve inclination, and number of vertebrae within the curve are all strong determinants responsible for the height loss in AIS.
What problem does this paper attempt to address?