Lumbo-sacral arthrodesis after laminectomy.

The necessity for arthrodesis of the lumbo-sacral spine after laminectomy has been performed arises mainly in the treatment of spondylolisthesis and some lesions of the intervertebral discs. In most cases of spondylolisthesis the lamina is loose and unattached to its body, so it can play no useful part in the fusion operation. Apart from this, it should in all cases be removed to allow a thorough decompression of the cord and nerve roots. In the operation for exploration of a disc it is impossible to get adequate exposure of the lower two disc spaces unless the spinous process and at least one half of the lamina of the fifth lumbar vertebra are removed. Sometimes, especially in the L.4-5 segment, it is necessary also to remove the medial facet of the posterior intervertebral joint, and this is not infrequently removed inadvertently. I do not believe that arthrodesis is necessary in the routine treatment of all cases of disc prolapse, but in the cases in which it is considered necessary there will usually be too little of the neural arch left for the performance of a satisfactory posterior arthrodesis. The commonest indication for fusion is the so-called scarred disc, in which severe backache is associated with some sciatic radiation. Although a prolapsed nucleus will not usually be found in these cases it is most important that a thorough exploration of both the affected and adjacent discs should be made to exclude its presence. For the same reason, although disc lesions probably play an insignificant part in the common low back pain syndrome, until the etiology of this condition is better understood it is unwise to perform a posterior arthrodesis without first exploring the disc spaces. In both cases the exploration inevitably removes the basis of a sound posterior fusion. In the past this problem has usually been tackled by using a long graft to bridge the laminectomised vertebra. An attempt is thus made to fuse two intervals, most commonly the fourth lumbar to sacrum. The choice lies between an onlay cortical graft as used by Albee (1911) or one of its many modifications, perhaps a graft fixed to spinous processes by bolts and plates as described by Wilson and Straub (1947), or the H-graft impacted between spinous processes as used by Bosworth (1942). All are open to the same criticisms: the gap to be bridged is too large; the lever-like action of a long graft greatly exaggerates the slight movements which are so hard to eliminate in this region; and the sacral spinous process is so often absent or ill-developed that adequate fixation of the sacral end is impossible. All published results show a greatly increased rate of failed fusion when more than one interval is spanned. Cleveland, Bosworth and Thompson (1948), in a review of 594 cases of grafts of various kinds, quoted an increase from 35 per cent for the lumbo-sacral interval only to 173 per cent for the fourth lumbar to sacral interval, and a figure as high as 35 per cent for spondylolisthesis cases. For reasons to be mentioned later, I believe that even these results may represent a too optimistic view of this problem. My own experience of these types of graft is small, but none the less unfortunate. Excluding fusions done for fracture or tuberculosis, I have been able to study the late result in eight cases, of which at least four are certain failures. Many writers have mentioned the need for improved methods of arthrodesing this region of the spine, especially in cases of spondylolisthesis, and during the past six years I have devoted much time and thought to attempts, mostly unsuccessful, at solving this problem. A bone graft inserted between the vertebral bodies through the same posterior approach and at the same time as the exploratory operation seemed to offer much promise.