Postoperative trunk shift in Lenke 1C scoliosis: what causes it? How can it be prevented?

A risk factor analysis study.

To identify the causative factors for postoperative trunk shift in Lenke 1C scoliosis and investigate how to prevent it.

When selective thoracic fusion is performed, postoperative trunk shift is a significant problem in the management of Lenke 1C scoliosis. It is often accompanied by unsatisfactory clinical outcomes and a risk of reoperation.

We reviewed all the patients with adolescent idiopathic scoliosis (AIS) surgically treated in our institution from 2002 through 2008. Inclusion criteria were as follows: (1) patients with Lenke 1C curves who were treated with selective thoracic fusion using posterior pedicle screw-only constructs; (2) the lowest instrumented vertebra (LIV) ending at L1 level or above; and (3) 2-year radiographical follow-up. Eighteen radiographical parameters were chosen as potential risk factors. The 18 parameters measured (1) amount of correction obtained by surgery; (2) preoperative position of LIV; (3) magnitude of major thoracic and thoracolumbar/lumbar (MT and TL/L) curves and ratio of MT: TL/L curve; and (4) curve flexibility. Both comparative and correlation analyses were performed. Those parameters that had shown highest correlation with the 2-year thoracic apical vertebra-center sacral vertical line (AV-CSVL) distance were selected to form a linear regression model, by which the correlations were quantified.

Of the 278 patients reviewed, 44 met the inclusion criteria. The parameters that measured the preoperative position of LIV and ratio of MT: TL/L curve showed high correlation with the 2-year thoracic AV-CSVL distance. With regard to the parameters that measured the amount of correction obtained by surgery, only the correction of the thoracic AV-T1 distance showed low correlation. Among the 18 parameters, preoperative lowest instrumented vertebra-lower end vertebra (LIV-LEV) difference and ratio of MT: TL/L Cobb angle were selected to form a formula to help predict postoperative trunk shift. The formula was as follows: 2-year thoracic AV-CSVL distance = -26.6 + 22.7 (ratio of MT: TL/L Cobb angle) - 3.9 (preoperative LIV-LEV difference). The model R2 = 0.55.

Both LIV selection and ratio of MT: TL/L curve were found to be highly correlated with the onset of postoperative trunk shift in Lenke 1C scoliosis. Amount of correction obtained by surgery, however, did not seem to be an independent causative factor. Postoperative trunk shift is less likely to occur when selecting LEV as LIV and the ratio of MT: TL/L Cobb angle of 1.2° or more.