Selective Thoracic Fusion in Adolescent Idiopathic Scoliosis

Abstract

Analysis of multicenter, prospectively collected data. To determine how selection of the lowest instrumented vertebra (LIV) relative to the stable vertebra (SV) and the end vertebra (EV) effects correction of the main thoracic curve, compensatory lumbar curve, and incidence of coronal decompensation after selective thoracic fusion. Traditionally, in Lenke type 1B and 1C curves, the LIV is selected as the SV; however, selecting the LIV continues to be controversial. Inclusion criteria were patients with adolescent idiopathic scoliosis (AIS) with Lenke type 1B, 1C, or 3C curves that had a selective thoracic fusion with the LIV from T11 to L1 (n=172). The patients were divided into three curve patterns on the basis of the relative position of SV and EV. Group SBE (stable below end) (n=93) had SV below EV, group SAE (stable at end) (n=66) had SV at the EV, and group EBS (end below stable) (n=13) has EV below SV. In addition, each group was divided into six subgroups based on the selected LIV: LIV above SV, at the SV, below SV, above EV, at the EV, and below EV. Each was compared for preoperative and 2-year postoperative radiographic parameters and clinical data. In group SBE, the 2-year postoperative thoracic curve correction rate when the LIV was below the EV (64%+16%) was significantly greater than when the LIV was at the EV (54%+13%; P<0.001). The 2-year postoperative spontaneous lumbar curve correction (SLCC) rate similarly correlated with the LIV selection subgroups, 52%+20% and 43%+19%, respectively (P=0.03). In group SAE, the 2-year postoperative thoracic curve correction rate when the LIV was below the EV/SV (64%+14%) was significantly greater than when the LIV was at the EV/SV (52%+14%; P=0.004). The 2-year postoperative SLCC rate for group SAE similarly correlated with the LIV selection subgroup, 56%+16% and 38%+21%, respectively (P<0.01). In group EBS, the 2-year postoperative thoracic curve correction and SLCC rates were not significantly different among the LIV selection subgroups; however, the incidence of decompensation was 38%. When performing a selective thoracic fusion of Lenke type 1B, 1C, and 3C AIS curves in which the SV was at/or below the EV, the greatest correction of the main thoracic and compensatory lumbar curves occurred when the LIV was at/or at least one level distal to the SV. This more distal LIV did not result in an increased rate of truncal imbalance.