Abstract
Introduction: The occurrence of non-union following instrumented scoliosis correction may predispose to pseudarthrosis and subsequent implant failure. Although non-union is often multifactorial, it is widely accepted that bone graft of adequate quality and quantity is fundamental to achieve solid fusion. Conventionally, autologous rib graft or iliac crest harvest has been utilised for endoscopic anterior instrumented scoliosis surgery. However, these techniques increase the operative duration and cause donor site morbidity, both of which may lengthen hospital stay. Alternatives such as allograft bone and bone morphogenetic proteins have gained more widespread use and may improve fusion rates although this remains controversial. The aim of this study was to compare two-year postoperative fusion rates for a series of patients who underwent endoscopic anterior instrumentation for thoracic scoliosis utilising various bone graft types.
Methods: 19 patients who had undergone endoscopic anterior instrumented scoliosis correction using identical instrumentation (4.5mm diameter titanium anterior rod and vertebral body screws, Eclipse, Medtronic) between May 2000 and August 2005 were identified from a surgical database of 132 consecutively treated individuals. All patients received bone graft to supplement thoracic fusion. Discectomy was performed at the levels to be instrumented and intervertebral spaces were packed with autologous rib heads (8 patients), iliac crest (1 patient), or mulched femoral head allograft (10 patients). The quality of thoracic fusion and implant integrity were evaluated two years following scoliosis correction using low-dose CT performed in accordance with local ethical approval. The intervertebral fusion was assessed using a modified Sucato method (1). Each level was graded using a 4-point scale based on calculated percentage of fusion across the disc space. 0 points indicated no fusion; 1 point, fusion < 25%; 2 points, fusion between 25 and 50%; 3 points, fusion between 50 and 75%; 4 points > 75% or complete fusion. The fusion was considered solid with a score of 3 points or more. Data was analysed with non-parametric tests using a significance level of 0.05.
Results: Of the cohort, nine had evidence of implant failure with rod fracture. All implant failures occurred in the group who received either rib head or iliac crest graft. No rod fractures were identified in the femoral allograft group. The mean fusion grade in the autologous bone graft group was 1.91 whereas in the allograft group this was 3.30 (95% confidence intervals 1.38–2.44 and 2.99–3.61 respectively) with a statistically significant difference in fusion rates between these two groups (p=0.001).
Discussion: This study demonstrated significantly better rates of thoracic fusion in endoscopic anterior instrumented scoliosis correction using mulched femoral allograft compared with autologous rib heads and iliac crest graft. This could be partly explained by the difficulty obtaining sufficient quantities of autologous graft. The lower fusion rate seen in the autologous graft group appears to predispose to rod fracture although the longer-term clinical consequence of implant failure in this group is not clear and warrants further study.
Correspondence should be addressed to Dr Owen Williamson, Editorial Secretary, Spine Society of Australia, 25 Erin Street, Richmond, Victoria 3121, Australia.
