A Multicenter Clinical and Radiographic Analysis of Anterior and Posterior Surgical Techniques for Treatment of Thoracolumbar/Lumbar AIS: The Current State of the Art

Purpose: To review the perioperative clinical data and the radiographic results after ASF or PSF techniques for treatment of thoracolumbar/lumbar (TL/L) AIS.

Materials and Methods: The Spinal Deformity Study Group multicenter retrospective AIS database was queried to identify patients with TL/L curves with greater than two-year follow-up. Demographic, clinical and pre/ postoperative radiographic data was collected. A comparison was made between those treated with ASF or PSF. Significant levels were computed using the Mann-Whitney and t-test. The Mann-Whitney results are reported. Significance was accepted at p £ 0.05.

Results: 96 patients were identified with TL/L (Lenke 5C) AIS: ASF (n=59), PSF (n=32), ASF/PSF (n=5). There were 87 females (91%) and 9 males (9%). The mean age was 14+11 years (11+0 to 21+5 years). The mean preoperative main thoracic (MT) Cobbº was 31º (4º to 68º); the TL/L Cobbº was 48º (20º to 83º). At two years postop, ASF achieved 32% "spontaneous correction" of the MT curve, 49% for the PSF (p£ 0.001) and 43% for the ASF/PSF. ASF achieved a mean TL/L curve "surgical correction" of 63%, PSF 60%, and ASF/PSF 70% (Figure 1). The flexibility index (FI) was higher for TL/L curves than for the MT curves. The TL/L curves that were treated with ASF were less flexible than those TL/L curves treated with PSF techniques (Figure 2). The total number of segments within the TL/L curve was the same in the ASF (5.6 segments) and the PSF group (5.7 segments). The mean number of segments fused, however, with an ASF technique was 5.0 (3-7 segments) compared to 10.3 (5-14 segments) for a PSF (p£ 0.001) (Figure 3). Operative time for ASF was 285 min (88-450 min) and for PSF it was 245 min (96-480 min) (p£ 0.001). The complication rate for ASFs (10.2%) and PSFs (12.5%) was similar (Figure 4). Blood loss was 480 mls (143-2000 mls) for the ASF and 1143 mls (200-2500 mls) for the PSF (p£ 0.001).

Discussion: The TL/L curves in the ASF and PSF groups were similar in magnitude. ASF provided slightly better instrumented correction of the TL/L curve but less "spontaneous correction" of the MT at two years postop. The MT curves were of similar flexibility (p>0.6); therefore, the apparent greater "spontaneous correction" of the MT curves may be due to the longer constructs used in the PSF to instrument the nonstructural MT curve: i.e. "surgical" rather than "spontaneous" correction. Despite a lower FITL/L in the ASF group (.66) vs. PSF group (.71), the % correction of the TL/L curve for the ASF group was greater. This is likely due to the beneficial effect that complete discectomy and annulectomy has on facilitating correction of the TL/L spine. Although all of the TL/L curves had the same number of vertebral segments (ASF: 5.6 vs. PSF: 5.7), the ASF group had a mean 5 fewer segments fused than the PSF group (ASF; 5 vs. PSF; 10). Complication rates were similar for the ASF and PSF techniques.

Figure 1. Thoracolumbar/Lumbar (Lenke 5C) AIS:

Figure 2. Thoracolumbar/Lumbar (Lenke 5C) AIS:

MT - Main thoracic curve

TL/L - Thoracolumbar/Lumbar Curve

FI - Flexibility Index = Standing tanding Cobb° - Side Bending ending Cobb°/Standing Cobb°

Figure 3: # of Vertebral Segments within the - Thoracolumbar/Lumbar (Lenke 5C) AIS :

Figure 4. Early & Late Complications for Thoracolumbar/Lumbar (Lenke 5C) AIS: