Introduction Circumferential fusion is becoming increasingly popular and has been advocated by many authors to improve the fusion rates and clinical outcomes of the degenerative lumbosacral spine. Anterior lumbar interbody fusion (ALIF) with posterolateral fusion does provide direct access to the disc via a separate incision/ approach but has inherent neurovascular risks. Posterior lumbar interbody fusion (PLIF) with posterolateral fusion mandates bilateral exposure with significant retraction of neural elements with higher incidence of postoperative radiculitis. PLIF also reduces surface area for fusion and disrupts the posterior tension band. TLIF allows for a circumferential fusion through a single posterior incision with only slight retraction of the thecal sac and nerve roots, with much less morbidity and costs as compared to traditional PLIF and ALIF techniques. To our knowledge, there are no studies that report radiographic and clinical results of using recombinant human bone morphogenic protein (rhBMP-2) and allograft in a TLIF setting. The purpose of this study was to assess clinical & radiographic outcomes of patients treated with one or two level instrumented transforaminal lumbar interbody fusions (TLIF) performed with allograft and rh-BMP2 for treatment of symptomatic spondylolisthesis or degenerative disc disease.

Methods During a consecutive 13 month period, 77 patients underwent TLIF procedures utilizing rhBMP-2 by one spine surgeon for lumbosacral degenerative and deformity conditions with simultaneous posterolateral fusions with allograft. Pedicle screw instrumentation (Monarch, DePuy Spine) provided distraction and a carbon-fiber curvilinear cage (Leopard, Depuy Spine) packed with rhBMP-2 (Large II Kit, total graft volume 8ml onto 77.4 sq. cm collagen sponge; Infuse: Medtronic Sofamor Danek) was placed into the disk space after hemifacetectomy and discectomy. The patients were followed at two weeks and three, six, 12 and 24 months after surgery patients were followed with several functional parameters such as the visual analog scale (VAS), SF-36 and Oswestry Disability Index (ODI) questionnaires. Fusion was assessed by static and dynamic radiographs at 6, 12, and 24 months as well as CT scans at 24 months.

Results 71 patients were available for follow-up (92%) evaluation (mean 16 months; range 6–24 months). At 24 months, 85 percent and 81 percent of patients had improvement over preoperative ODI and SF-36 measures respectively. At 24 months, 70% of patients had good to excellent outcomes by both ODI and SF-36. We achieved 94% fusion rate with only four pseudarthroses. There was one wound infection treated with hardware removal and intravenous antibiotics. One patient had excessive bone growth into the foramen, necessitating surgical decompression with subsequent excellent clinical outcome. Ten patients had paresthesias on the side of the TLIF, all of which resolved by three weeks. There were no permanent neurologic deficits.

Discussion The use of rhBMP-2 inside the cage, in combination with posterolateral allograft, can provide a high fusion rate and good clinical outcomes in a TLIF setting. The morbidity associated with iliac crest bone graft is avoided, with fusion rates approaching that of a true anterior/posterior circumferential fusion. Complications were few, with no significant neurologic sequelae from the placement of a structural graft into the anterior column through a posterior approach. Overgrowth of bone into the neural foramen, likely related to the residue of rhBMP-2 at the TLIF entry site, can occur. Care must be taken to place the TLIF cage device and the contained rhBMP-2 into the anterior half of the disk space to minimize the risk of this complication.