OBJECTIVESTo investigate the incidence of thoracic curve decompensation or proximal adding-on phenomenon after anterior selective fusion of thoracolumbar or lumbar curve in Lenke type 5 adolescent idiopathic scoliosis (AIS), and to identify its risk factors.

METHODSFrom June 2001 to December 2008, 130 Lenke type 5 AIS patients with a Cobb angle of 40° - 73° treated with anterior selective thoracolumbar or lumbar fusion, and with a minimum 2-year postoperative regular follow-up were recruited in this study. The average age, Cobb angle and Risser sign of all patients was 14.8 ± 1.6 years, 46° ± 6° of major thoracolumbar or lumbar curve (TL or L), 25° ± 7° of proximal thoracic curve and 0-5, respectively. The patients were grouped according to the relationship between the upper instrumented vertebrae (UIV) and the upper end vertebrae (UEV), the patients' Risser sign and the relationship between UIV and C(7) plumb line (C(7)PL). The radiographic data of the patients were compared between patients with and without proximal adding-on by using t test, and the incidence of proximal adding-on was analyzed in terms of determination of UIV and Risser sign to identify the risk factors of this phenomenon by using Fisher's exact test.

RESULTSEleven patients were identified with proximal adding-on, thus the incidence of it was 8.5%. At last follow-up postoperatively, the average Cobb angle of TL or L and proximal thoracic curve was 9° ± 4° and 11° ± 5°, respectively. Moreover, the incidence of adding-on in Rissex sign grade 0 to 1 (3/8) was higher than that of grade 2 to 3 (12.1%) and grade 4 to 5 (4.5%). In addition, the incidence of adding-on in UIV lower than UEV group (20.6%) was obviously higher than that of UIV higher than or equal to UEV group (4.2%). The incidence of adding-on for patients with C(7)PL falls away from UIV (19.5%) were obviously higher than that of patients with C(7)PL falls between the pedicle and lateral margin of UIV (3.6%) and between bilateral pedicles of UIV (3.0%). Each group showed significant difference for the incidence of adding-on by Fisher's exact test (P < 0.05).

CONCLUSIONSThere exists the risk of proximal thoracic curve decompensation, with a not low rate, after anterior selective fusion for major TL or L curve AIS. The determination of UIV relative to UEV and the skeletal maturity of the patient are the two factors closely associated with the presence of such a phenomenon.