Autologous bone graft harvested from remote locations has been used to fuse the human spine for the treatment of disease, degeneration, or injury for over a century. For the latter half of the twentieth century, autologous bone harvested from the iliac crest was the “gold standard” bone graft used for spinal fusion. Donor-site morbidity, particularly pain lasting more than a year, has been reported in >30% of patients who have undergone spinal fusion with autologous iliac crest bone graft (ICBG). This morbidity led many investigators to search for alternatives to ICBG. Research began to focus on recombinant forms of a variety of bone morphogenetic proteins (BMPs) as a potential bone-graft substitute. BMPs, discovered by Marshall Urist in the early 1960s, are a subclass of a somewhat inhomogeneous group of proteins known as cytokines that have cell signaling as a common feature. BMPs were found to be osteoinductive; that is, they induce osteoblastic activity in the formation of bone. BMPs, in their natural form, are very difficult to generate in quantity. Recombinant forms that could be produced in quantity were developed, but manufacturing them was expensive and required substantial investment by industry. This investment appeared to show fruit as animal studies demonstrated efficacy of BMPs as a bone-graft substitute. These studies were followed by human clinical trials, in which two classes of BMPs (recombinant human [rh] BMP-7 and rhBMP-2) demonstrated non-inferiority compared with ICBG for achieving spinal fusion and one of them (rhBMP-2) showed potential superiority to ICBG.

Vaccaro et al.reported the results more than three years after the use of rhBMP-7 putty (known as osteogenic protein-1, or OP-1) in patients treated with decompression and uninstrumented posterolateral fusion for degenerative spondylolisthesis. Initially, 335 patients were randomized 2:1 to receive OP-1 Putty (Stryker) or ICBG, and 202 of them (144 in the OP-1 group and fifty-eight in the ICBG group) were followed for more than thirty-six months. New bone formation bridging across the transverse processes, angulation of ≤5°, and translation of ≤3 mm, as judged by a panel of independent spine surgeons, were all required for a patient to be considered to have radiographic evidence of fusion in this study. At the time of final follow-up, more than thirty-six months after the fusion, 69.3% of the OP-1 group and 68.4% of the ICBG group (p = 1.0) met the angulation criteria for solid fusion and 74.5% of the OP-1 group and 75.7% of the ICBG group (p = 1.0) met the translation criteria for solid fusion. Computed tomography (CT) scans were obtained for 196 (97%) of the 202 patients followed for more than thirty-six months, and they showed bridging bone in 74.8% of the OP-1 group and 77.4% of the ICBG group (p = 0.852). Other results, noted only at the twenty-four-month time point, included rates of success according to the Oswestry Disability Index (ODI) of 74.5% in the OP-1 group and 75.7% in the ICBG group (p = 0.839); an absence of serious treatment-related adverse events in 85.6% and 84.7% (p = 0.863), respectively; and neurologic success rates of 92.1% and 84.1% (p = 0.057), respectively. Although the twenty-four-month rates of successful fusion, as judged by the radiographic angulation and translation criteria noted above, were similar to those noted at more than thirty-six months, the percentage of patients with bridging bone was lower in the OP-1 group (61.7%) than in the ICBG group (83.1%) (p < 0.001) at twenty-four months. As pointed out by Delawi et al., the U.S. Food and Drug Administration (FDA) ultimately granted OP-1 only a Humanitarian Device Exemption approval for revision of posterolateral lumbar fusion in compromised patients.

The methodology that Delawi et al. used in their study was similar, but not identical, to that used by Vaccaro et al. Delawi et al. included patients with either isthmic or degenerative spondylolisthesis treated with decompression and instrumented posterolateral fusion; one group received local autograft combined with ICBG, and the other received local autograft combined with Osigraft (Stryker). Osigraft is an OP-1 formulation available in Europe that contains the same amount of OP-1 protein, in the same carrier, as is found in the OP-1 Putty that has been used in the U.S. but without the carboxymethylcellulose additive that is present in OP-1 Putty.

Delawi et al. enrolled 134 patients, 119 of whom underwent the surgery, had adequate imaging at the time of follow-up, and had been randomized 1:1 at the time of bone-grafting during the procedure (fifty-nine in the ICBG group and sixty in the OP-1 group). The primary outcome measures at one year, including a combination of clinical outcomes and evidence of fusion as determined on CT scans, were analyzed for 113 of these patients (fifty-six in the ICBG group and fifty-seven in the OP-1 group). Non-inferiority was not achieved in the OP-1 group (40% success rate compared with 54% in the ICBG group) primarily because of the lower rate of radiographic evidence of fusion, which was seen in 54% in the OP-1 group versus 74% in the ICBG group (p = 0.03). No adverse events were attributed to the use of OP-1.

The strengths of this study include its randomized design, the use of CT as a measure of fusion, and the relative homogeneity of the study population. Its primary weakness is a one-year end point for the determination of clinical outcomes and fusion status. Also, there was a significantly larger percentage of smokers in the OP-1 group (48%) than in the ICBG group (31%) (p = 0.04). This difference may account for some, if not all, of the difference in the fusion rates between the two groups.

The study by Delawi et al. seems to be another bit of evidence against the use of rhBMP-7 as a substitute for ICBG in posterolateral spinal fusion. The other available form of BMP, rhBMP-2, has been embroiled in a controversy that has included allegations of methodological biases and underreporting of adverse events in industry-sponsored publications. The controversy has been partially quelled by the unprecedented Yale University Open Data Access (YODA) studies. The YODA studies have indicated that the efficacy of rhBMP-2 is at least equivalent, if not superior, to that of autologous bone graft in achieving spinal fusion and reducing pain levels at up to twenty-four months, possibly at the expense of increased pain in the immediate postoperative period and (questionably) an increased risk of cancer. The controversy surrounding the use of BMPs, particularly rhBMP-2, has certainly reduced the use of this class of proteins as an adjunct for spinal fusion, especially its “off-label” usage (which had been its primary clinical usage before the controversy erupted). While all of this does not necessarily mean the death knell of BMP usage in spinal fusion, it certainly does appear that the search for the ideal ICBG substitute will go on.