Does BMP-2 on a collagen sponge placed in the interbody space improve clinical outcomes in patients surgically treated with long fusions for adult scoliosis?

Attention: open in a new window. PDFPrintE-mail



rhBMP-2 has been reliable in producing fusion when placed in a tapered cage through an anterior approach for treatment of single level lumbar degenerative disease. Results using BMP for arthrodesis in patients with adult deformity have not been reported.


This is the first study reporting on clinical and radiographic outcomes in adult scoliosis patients undergoing long fusions using rhBMP-2 in the interbody spaces.


Prospective series of consecutive adult scoliosis patients surgically treated single surgeon for adult scoliosis were clinically and radiographically compared to a consecutive series of similar patients treated with allograft or mixed allograft and rib autograft.


Sixty-eight consecutive patients with adult scoliosis: 53 treated with rhBMP -2 and 15 without. There were 7 smokers in the BMP group and none in the control group.


Oswestry and VAS pain scores at 6 months, 12 months, 2 years, and latest follow-up.


In addition to the posterior instrumented correction and fusion (average 9 levels in both groups, range 4 – 17 levels), each patient had one or more interbody fusion (58 ALIF, 10 TLIF). rhBMP-2 was used in at least one interbody space (average 3 levels/pt, total 172 levels) in 53 patients at a dose of 4-12 mg/disk on collagen sponges along with structural tricortical iliac crest allograft (41 patients) or within a PEEK cage (12 patients). Eight were fused to the sacrum, 13 to L5, 13 to L4, and13 to L3. Structural allograft alone was used in 15 patients (average 6 levels/pt, 90 levels total). Fusion was assessed by CT or flex-ext radiographs with >3° motion or screw halos defining nonunion. Deformity correction was also followed.


Follow-up averaged >4 years (24-154 months). No rhBMP-2 related complications occurred. There were 3 infections in the rhBMP-2 group and 1 in the control group, all occurring posteriorly. All infections were successfully treated with irrigation and drainage and IV antibiotics. In the control group, there were significantly more pseudarthrosis; 5/15 patients developed nonunions (33%). One of these patients had a rod fracture. Of the BMP patients, 3/53 (6%) developed nonunions, one at L5-S1 and two at L4-5. All 3 of these nonunions occurred at the bottom of a long lumbar construct starting T11, and all 3 had structural allograft instead of a cage used for anterior column support. All 3 were ALIF patients (there were no TLIF nonunions). Two of these three patients were nonsmokers. Two additional BMP group patients had delayed unions which fused by 1 year. The dose of rhBMP-2 used in the disk (4mg vs. 12mg) did not correlate to nonunion occurrence. Oswestry and VAS pain scores were better in the BMP group and in the patients that fused.

CONCLUSIONS: rhBMP-2 used in the interbody space along with structural support (allograft or cage), whether by ALIF or TLIF, significantly improves the fusion rate and clinical outcomes in patients surgically treated for adult scoliosis.