EARLY COMPLICATIONS ASSOCIATED WITH BMP IN LUMBAR INTERBODY FUSIONS

Abstract

Introduction: Bone morphogenetic protein (BMP) has become widely used in the interbody space as part of a lumbar fusion. Complications can occur but are not completely understood or well documented.

Methods: A prospective review of consecutive lumbar interbody fusions performed by a single surgeon was undertaken over a 2-year period. Early complications (defined as occurring within the first six weeks) were noted. The interbody cages (titanium Syncage, Synthes cages for ALIFs and PEEK Capstone, Medtronic cages for TLIFs) were filled with Infuse BMP-collagen sponge. Until early 2007, the amount of Infuse used was not strictly measured but after that time, only enough to fill the cage was used, with the volume assessed according to the manufacturer’s guidelines. Patients were routinely assessed preoperatively and at the six-week postoperative review using a visual analogue scale and the Oswestry disability score. Plain x-ray and MRI were obtained preoperatively, and plain x-ray was obtained postoperatively. In addition, if early problems developed, MRI scan was obtained. The incidence of complications was compared to that seen in similar procedures but where BMP was not used.

Results: 114 patients, including 78 transforaminal inter-body fusions (TLIFs) and 36 anterior lumbar interbody fusions (ALIFs) were available for review. Early complications were noted in 10 of 114 patients. Two (both with TLIF) were not directly linked with BMP use: in the first, the cage migrated posteriorly and in the second, a deep infection developed. The remaining eight were associated with an exaggerated inflammatory response likely related to BMP use. Severe back pain associated with marked vertebral body inflammation seen on MRI was noted in two ALIF patients. The response occurred within 2 weeks of surgery, and settled with conservative treatment. Severe back pain and recurrence of leg pain developed in six TLIF patients. Fluid cyst formation within the spinal canal was seen on MRI in 4 of these. The cyst extended from the region of the posterior aspect of the cage into the canal and toward the area of the excised facet joint, resulting in compression of the exiting nerve root. In one case, the surgical site was re-explored and the cyst removed. In two cases, the cyst was aspirated under CT guidance and injected with steroid. In the final case a course of oral Prednisone was administered. In the remaining two TLIF cases, there was a diffuse inflammatory response in the region of the posterior aspect of the cage and adjacent epidural space but without discrete cyst formation. In one, oral Prednisone was prescribed. The second was treated expectantly. The majority of these complications were noted in 2007, after the dose of BMP was titrated in line with the manufacturer’s guidelines. In contrast, no such complications were seen when a similar technique but without BMP was undertaken in 33 posterior and 41 anterior interbody fusions.

Discussion: The incidence of an exaggerated inflammatory response with BMP in the lumbar spine may be under-recognised. The majority of complications published to date relate to vertebral osteolysis and bony overgrowth, although a number of adverse responses to BMP reported to the FDA relate to fluid cyst formation or inflammation. With the rapid increase in BMP use, it is important that surgeons are aware of potential complications, and possible strategies to prevent and address them.