Large Magnitude Adolescent Idiopathic Scoliosis: Anterior vs. Posterior Surgical Stabilization

Introduction: Large magnitude (>70°) adolescent idiopathic scoliosis (AIS) has traditionally been treated operatively. Posterior instrumentation with or without an anterior release is presently the standard technique. As anterior instrumentation gains favor concern has been raised over its ability to produce similarly good results in these larger curves.

Purpose: This study was undertaken to evaluate the radiographic results of both anterior and posterior instrumentation in the treatment of AIS with primary curves equal to or greater than 70 degrees.

Methods: Patients with AIS and primary thoracic curves of greater than or equal to 70 degrees were selected from a multi–center AIS study group population compiled over the years from 1995– 1998. Preoperative, bending, immediate postoperative and 1 and 2 year follow–up radiographs were reviewed. Primary and secondary curve Cobb angles, flexibility of the curves, apical translation and rotation, end instrumented vertebral angulation, overall coronal and sagittal balance, and sagittal Cobb angles were measured. Statistical analysis was performed using ANOVA within the each group and Student’s T–test for comparison between groups.

Results: Twenty–one patients met criteria for this study with a mean age of 14.1 years. There were 17 females and 4 males with a minimum two–year follow–up. Nine procedures were performed from a posterior approach and 12 from an anterior approach. Three of the 9 posterior procedures also had an anterior release while only 2 of 12 anterior procedures had a concomitant posterior release. Flexibility of the curves in each group was not significantly different based on preoperative bending radiographs. The mean primary preop curve angle was 75 degrees for the anterior group (70–96°) and 77 degrees for the posterior group (70–97°) while immediate postoperative primary curves measured 32 and 24 degrees respectively (p=0.07). At 1 year curve correction had been well maintained in both groups but at 2 years the primary curves in the posterior group had a greater loss of correction and were equivalent to the anterior group (34 degrees and 36 degrees; p=0.8). The average anterior instrumentation was 2.3 levels shorter for comparable curves than the posterior group. Apical translation showed a statistically significant improvement in the posterior group compared to the anterior group but this difference had been lost by the 2–year follow–up. The compensatory lumbar curve was not significantly different between the two groups preoperatively or at any time postoperatively. However, there was a statistically significant improvement postoperatively within each group (p<0.001 anterior; p<0.001 posterior) which continued to improve throughout the follow–up period. There was a greater percentage change in end instrumented vertebral angulation (EIV–A) at each postoperative time point for the posteriorly instrumented group. Neither T1–rib angle nor global coronal balance showed any difference between the groups throughout the study. The preoperative sagittal profile demonstrated no significant differences between the two groups. However, the anterior group showed a postoperative trend toward increasing overall thoracic kyphosis from 37 degrees to 46 degrees (T2–T12; p=0.3) and from 27 degrees to 37 degrees (T5–T12; p=0.2). The T5–T12 Cobb angle was significantly greater at each postoperative measurement in the anterior group (27, 32 and 37 degrees at immediate, 1 year and 2 year respectively) compared to the posterior group (16, 23 and 21 degrees; p=0.04, 0.05 and 0.05).

Conclusions: Initial correction of large curves was better accomplished through a posterior approach, however, by two years greater loss of correction in both the coronal and sagittal planes had occurred in the posteriorly instrumented group rendering essentially equivalent results. Correction of end instrumented vertebral angulation is better accomplished through the posterior approach possibly due to the greater number of segments fused. Anterior correction and stabilization engenders greater thoracic kyphosis. Anterior instrumentation and fusion for large magnitude AIS curves is equally as effective as a posterior instrumentation and fusion with the advantage of requiring arthrodesis in an average of 2.3 fewer motion segments.