Purpose: To identify patients in whom anterior scoliosis correction was not possible and to determine pre-operative factors that may predict such an outcome.

Methods: From 1999–2005, 257 patients were listed for anterior correction with the Kaneda Anterior Spine System (KASS). Of these 246 were completed successfully. However in 11 cases it was not possible to complete the procedure.

We performed a retrospective review of case notes and X-rays. A control group of 22 patients, in whom anterior surgery was completed, matched to age, sex and type of curve, was used.

Results: Two reasons for abandoning anterior instrumentation were identified; loss of cord signal (7) and failure to achieve adequate correction after anterior release and reduction (4).

Of the seven patients with lost signal three were syndromic and four were associated with syrinx. In all seven, loss of signal occurred on clamping of segmental vessels. All seven had no residual neurological deficit post-operatively and had uncomplicated posterior correction the following week.

All four patients in whom inadequate correction was achieved after anterior release and repositioning had idiopathic curves. Of these two were thoracic and two were thoracolumbar. Mean pre-operative Cobb angle was 67 (range 59–85) compared to a mean of 56 (range 42–68) in the control group. Mean pre-operative stiffness index was 91% (range 85%–100%) compared to a mean stiffness index of 65% (range 53–80) in the control population.

Conclusion: Whilst a successful outcome is achieved in a majority of KASS instrumentations we have identified two reasons why anterior surgery has to be abandoned. Whilst one often cannot pre-operatively anticipate intra-operative loss of cord signal, we found that in cases with an underlying syrinx there is a particular risk of this occurrence. Our experience has shown particularly stiff curves (Stiffness index ≥ 85%) may not be suitable for stand-alone anterior surgery.

Nitric oxide (NO) is a free radical labile gas which has important physiological functions and is synthesised by the action of a group of enzymes called nitric oxide synthases (NOS) on L- arginine. We have shown that nitric oxide modulates fracture healing¹. Bone morphogenic proteins (BMP) are potent differentiating factors that augment the process of new bone formation. Recombinant human BMP-2 (rhBMP-2) enhances spinal fusion². With progression of fusion there is a remodelling of the fusion mass bone accompanied with a decrease in the fusion mass size. It is not known whether nitric oxide has a role in spinal fusion or rhBMP-2 enhanced spinal fusion.

We studied this in a novel rat intertransverse fusion model using a defined volume of bone graft (7 caudal vertebrae) along with 157 mm³ of absorbable Type-1 collagen sponge (Helistat®) carrier, which was compacted and delivered using a custom jig for achieving a similar graft density from sample to sample. The control groups consisted of a sham operated group (S, n=20), an autograft + carrier group (AC, n=28) and a group consisting of 43μg of rhBMP-2 (Genetics Institute, Andover, MA) mixed with autograft + carrier (ACB, n=28). Two experimental groups received a nitric oxide syn-thase (NOS) inhibitor, N^G-nitro L-arginine methyl ester (L-NAME, Sigma Chemicals, St Louis, MO) in a dose of 1mg/ml ad lib in the drinking water (ACL, n=28) and one of these experimental groups had rhBMP-2 added to the graft mixture at the time of surgery (ACLB, n=28). Rats were sacrificed at 22 days and 44 days, spinal columns dissected and subjected to high density radiology (faxitron) and decalcified histology. The faxitrons were subjected to image analysis (MetaMorph).

On a radiographic score (0–4) indicating progressive maturation of bone fusion mass, no difference was found between the AC and ACL groups, however, there was a significant enhancement of fusion when rhBMP-2 was added (ACB group,3.3±0.2) when compared to the AC group (1±0) (p< .001). However, on day 44, the ACLB group (3.3±0.2) showed significantly less fusion progression when compared to the ACB group (4±0) (p< 0.01). There was a 25% (p< 0.05) more fusion-mass-area in day 44 of ACLB group (297±26 mm³) when compared to day 44 of the ACB group (225±16 mm³) indicating that NOS inhibition delayed the remodelling of the fusion mass. Undecalcified histology demonstrated that there was a delay in graft incorporation whenever NOS was inhibited (ACL and ACLB groups).

Our results show that the biology of autograft spinal fusion and rhBMP-2 enhanced spinal fusion can be potentially manipulated by nitric oxide pathways.