Iliac Crest Bone Graft

Iliac crest bone graft (ICBG) harvest has historically been the gold standard for spine fusion for many reasons. ICBG leads to high arthrodesis rates and allows for harvesting large amounts of bone graft either through a small separate incision or through the existing posterior spine surgical incision. ICBG also allows a surgeon to harvest either structural or cancellous graft and can be obtained much cheaper compared to commercially available grafts. Finally, there is no risk of disease transmission or host compatibility with ICBG.

Despite the successes of ICBG, this procedure has been associated with a number of complications. Harvesting complications are not only problematic in the acute surgical period, but also several years later from scar irritation, numbness, and pain that affects one’s ability to perform activities of daily living. For example, Kim and colleagues showed that at 12 months from ICBG harvest, 11.3% of patients reported bothersome harvest area numbness, 3.9% bothersome scar appearance, and 16.5% harvest site pain and variable difficulties with activities of daily living. Other studies report chronic surgical site pain of 29%. The incidence of a surgical site infection associated with ICBG is 1.2% to 1.7% as a result of postoperative hematomas and wound dehiscence. Catastrophic complications from the harvest procedure such as superior gluteal artery injury and spinopelvic dissociation from an acetabular reamer have also been described. Although the overall rate of complications with ICBG harvest ranges from 15% to 48%, there are data to support that these complications are slightly reduced with posterior compared to anterior approaches.

Many surgeons believe that with modern-day surgical techniques, including cortical bone window formation, inner cortex preservation, and appropriate wound closure, the incidence of these reported complications can be significantly decreased. Also, harvesting techniques have evolved over time to utilize smaller incisions, use of power trephines for optimizing bone extraction, and potentially back-filling with graft to help fill in the bone defect. It is still unclear as to the exact clinical benefit for each of these techniques. Furthermore, the complications that have been associated with ICBG as part of industry-sponsored implant trials may have been overstated in the literature secondary to a flawed study design. Despite all of the potential problems with ICBG use, it is an effective method to achieve high rates of bony fusion in all areas of the spine.

Local Bone Graft

Local bone autograft can be harvested from the spinous processes, lamina, and facets during spinal decompression. Historically, surgeons have believed that local is inferior to ICBG in terms of achieving a successful bony arthrodesis. However, in vitro studies suggest that the cancellous osteoblast concentration is higher in laminar bone compared to iliac crest. Furthermore, several clinical studies have shown that for a single-level posterior lumbar fusion, local bone graft is equivalent to ICBG in fusion rate. A systematic review of ICBG and local bone graft fusion rates from clinical studies demonstrates similar rates of 79% and 89%, respectively. However, the threshold of bone needed per fusion level is not available in the local environment for many spine fusion procedures. Additionally, the increased use of minimally invasive spine surgery does not allow sufficient access to locally harvested autograft bone. The use of local bone graft leads to fewer harvest-related complications such as graft site pain and sensory disturbances. Surgical time is also decreased significantly.

Because the success of local bone graft is likely dependent on a threshold volume of bone available per level fused, multilevel fusions may have a different outcome. In a clinical study of patients requiring a three-level posterolateral spinal fusion, the authors concluded that significantly lower rates of arthrodesis were seen if the amount of local bone graft was divided bilaterally, as opposed to utilizing all of the graft unilaterally (34% in the bilateral fusion group versus 86% in the unilateral fusion group). In these clinical situations, bone graft extenders can be helpful to augment the volume of local bone graft.

Allograft

Depending on the anatomic area, allograft or processed cadaveric bone can be used either in conjunction with autograft or as a stand-alone bone-graft substitute. The advantage of allograft is access to a large supply of bone, avoidance of a ICBG harvest procedure, and cost-effectiveness compared to other substitutes such as recombinant bone morphogenetic protein-2. Allograft can be used in different structural forms such as powder, strips, bone chips, or cage-type formulations. The variability in processing allograft can lead to differences in the structural properties and immunogenicity of these grafts. Freeze-drying preparation can cause some loss of osteogenic potential, making them primarily osteoconductive grafts that provide a scaffold for cellular and bony ingrowth. On the other hand, fresh frozen allograft has higher immunogenicity but retains its mechanical properties. Whereas stand-alone posterolateral lumbar fusion rates for allograft are unacceptable (52%), other areas such as for anterior cervical fusions have enjoyed success with allograft. An and colleagues demonstrated that bone fusion density in patients undergoing posterolateral fusion was greatest in autograft fusions followed by an autograft/allograft mixture, frozen allograft, and freeze-dried allograft. Risks with allograft use include potential disease transmission and bacterial contamination of the graft at the time of implantation .

Demineralized Bone Matrix

Demineralized bone matrix (DBM) is derived from human allograft through the removal of the mineralized component of bone through acid extraction. Although there are variabilities in the processing mechanism, all preserve the type 1 collagen and osteoinductive properties of the graft substance. However, there is significant variability in the concentration of the osteoinductive properties of each product shown through in vitro extraction of growth factors and in vivo animal fusion models, suggesting that the clinical efficacy can be unpredictable as well.

There are more than 50 DBM products available on the market, but few clinical studies to support their specific use. Cammisa and associates performed a prospective study that showed equivalent fusion rates between DBM and local autograft composite compared to ICBG in posterolateral fusion. Kang and colleagues also demonstrated, through a prospective randomized multicenter trial, that DBM and local autograft composite compared to ICBG had similar fusion rates of 86% and 92%, respectively. As a stand-alone product in spinal fusion, DBM has not yet been supported by the literature. Failure of DBM as a stand-alone biologic may be attributable to the fact that DBM contains no viable cells, thus necessitating autograft in the fusion bed. Spinal arthrodesis is facilitated through the stimulation of osteoprogenitor cells by DBM and its osteoinductive factors. Although the osteoinductive properties of DBM are variable, it can serve as a reasonable bone graft extender when used in combination with autograft.

Ceramics

Ceramic-based bone grafts constitute another category that can serve as an osteoconductive scaffold for bone growth. These materials vary depending on porosity, structure, and manufacturing. Different components that make up the ceramic family include tricalcium phosphate, calcium phosphate, collagen, calcium sulfate, and hydroxyapatite. Calcium sulfate resorbs over weeks, beta-tricalcium phosphate resorbs over months, and hydroxyapatite can be present in the body for years. The optimal porosity size is considered to be 150 to 500 µm which allows for cell adhesion, proliferation, and differentiation into osteoblasts to facilitate bone in-growth. Additional benefits of ceramics are that they are inert and therefore nonimmunogenic and have a relatively low cost. Ceramics can also be sterilized without loss of structural integrity. The biodegradability of ceramic products can potentially create a problem for appropriate fusion to take place. If the implant dissolves and subsides prior to fusion, then instability and neurologic symptoms may occur. There are many clinical studies demonstrating comparable efficacy and outcomes of ceramics with local bone/autograft compared to ICBG. A systematic review found the fusion rate for ceramics and posterolateral fusion to be 85.6% and for circumferential fusion to be 88.8%. The overall fusion rate of ceramics as bone graft extenders in lumbar fusion was 86.4% ( Table 32-1 ). Ceramic-based scaffolds allow for adequate spinal fusion without the problems associated with obtaining adequate autograft or potential risks of allograft.

Mesenchymal Stem Cells

Mesenchymal stem cells (MSCs) are pluripotent cells that can differentiate into multiple cell lineages and can be isolated from adipose, bone marrow, and muscle tissue. When induced in the proper biochemical environment, MSCs can produce osteoprogenitor cells. MSCs from bone marrow aspiration (BMA) can yield a subpopulation of osteogenic osteoprogenitor cells, which can be delivered with an appropriate carrier. In addition to the iliac crest, this procedure can also be performed through a transpedicular route during pedicle cannulation prior to pedicle screw placement. BMA of the vertebral body has demonstrated 71% greater concentrations of stem cells than iliac crest aspiration. Limitations include the extraction of more than 2 ml from a single location, which can decrease the MSC concentration by dilution of blood. Although centrifuging processes can double the stem cell concentration, the exact dose-response applicability to clinical use has not been established.

Allogeneic MSCs are derived from a host donor and used in a different patient. There are only a few published clinical trials regarding allogeneic MSCs, and they are predominantly level IV studies with questionable methodology. These products are regulated by the U.S. Food and Drug Administration (FDA) throughout the human cells, tissues, and the cellular and tissue-based products (HCT/P) pathway, thereby needing minimal data prior to coming to market. Despite the lack of data to support allogeneic MSCs, use as a bone graft extender has been growing. A subpopulation of cells called mesenchymal progenitor cells, which have STRO+ cell markers, have gained some promise as potential targets for bony fusions. Because of its prohibitive cost, stem cell utilization in spinal fusion is not routinely recommended as there are limited data to support its use as an isolated bone graft substitute.

Recombinant Bone Morphogenetic Protein–2 (rhBMP-2)

In the late 1960s, Urist identified bone morphogenetic proteins (BMPs) as factors of bone healing that can initiate the osteoblastic signaling cascade. RhBMP-2 is a growth factor of the transforming growth factor family that initiates a biochemical pathway that leads to increased expression of genes, which potentiates mesenchymal stem cells into osteogenic differentiation. In 2002, the FDA approved the use of rhBMP-2 to enhance fusion for anterior lumbar interbody fusion procedures using a lumbar tapered titanium cage. From 2002 to 2011, the adjusted number of procedures performed annually with rhBMP-2 increased from 1116 to 79,294 with the majority being off-label use. Its success in achieving spinal fusion led to widespread applications in all areas of the spine, including posterolateral fusion, posterior lumbar interbody fusion, and cervical surgery.

Multicenter, randomized, prospective data demonstrated an increase of fusion rates for one-level anterior lumbar interbody fusion from 88.7% with ICBG to 94.5% with rhBMP-2, with 32% of the patients having ICBG morbidity at 24 months postoperatively. Another multicenter, randomized, prospective study found a 94% fusion rate with rhBMP-2 in patients undergoing one- or two-level instrumented posterolateral lumbar fusion. In other anatomic areas, there are multiple studies that show greater than 90% fusion rates for rhBMP-2, including posterior and direct lateral interbody fusions.

RhBMP-2 has also been associated with a number of complications including inflammatory responses/radiculitis, osteolysis, increased infection and reoperation rates, graft subsidence, cage migration, wound/seroma complications ( Fig. 32-1 ), airway compromise, prevertebral swelling, reintubation, retrograde ejaculation, ectopic bone formation, and potential increased cancer rates. Many factors have been associated with these adverse events, such as the carrier, dose used, and surgical technique. RhBMP-2 exposure in a host can lead to increased expression of angiogenic, osteoclastic, and proinflammatory cytokines, which may help to explain some of the complications, including inflammatory, osteolytic, and seroma formation.

T1/T2 axial MRI images of a patient 9 weeks from L4-L5 TLIF using off-label rhBMP-2 with posterior instrumentation that caused an inflammatory/seromatous response. The right-sided fluid collection was compressing the L4 nerve root causing radicular symptoms. Lateral lumbar x-ray shows stable hardware and polyetheretherketone interbody.

One major concerning potential complication of BMP is induction/metastasis of new cancers. Because of its role as a growth factor and the fact that a number of human cancer lines are known to utilize BMP receptors, the potential mechanism for carcinogenesis exists. On the other hand, rhBMP-2, as a member of the transforming growth factor (TGF)-beta family, also contains tumor suppressive properties depending on the cell involved. In June 2011, The Spine Journal focus issue raised the notion of cancer and retrograde ejaculation with rhBMP-2 and reported significantly higher rates of new cancers among patients exposed to rhBMP-2 with industry-sponsored data (3.8% versus .89%). The issue of whether rhBMP-2 causes cancer is a highly complex topic that involves a myriad of factors. For example, in this study, all cancer types and stages were grouped together and reported as a collective incidence as opposed to defining an annual rate. It is important to separate the potential associations of a stage 4 pancreatic cancer patient found 2 weeks after rhBMP-2 use compared to a local sarcoma at the site of rh-BMP2 use years after exposure. Furthermore, a critical delineation must be made between mutagenesis and metastasis, as the role of rhBMP-2 in these settings might vary.

Cooper and associates looked at 26 cancer types as identified on the National Cancer Institute’s Surveillance Epidemiology and End Results (SEER) program and found no increased rate of cancer with rhBMP-2 use in 146,278 patients aged 67 years and older who underwent surgery and were followed for an average of 4.7 years postoperatively. Anderson and colleagues retrospectively analyzed 467,916 patients and concluded there was no increased risk of cancer at a mean of 2.9 years following surgery with rhBMP-2. Other authors have opined that there is “no definitive association between BMPs and promotion of tumorigenesis or metastasis.” Regardless, many surgeons do not use rhBMP-2 when treating metastatic spinal disease or when there is an oncologic patient history.

In an industry-commissioned independent review of industry-sponsored patient clinical data regarding rhBMP-2 use and spinal arthrodesis, individual research groups from the University of York (UOK) and Oregon Health & Science University (OHSU) released their unbiased assessment. In the lumbar spine, they found higher fusion rates with rhBMP-2; however, there were no differences between ICBG and rhBMP-2 groups in clinical outcomes and incidence of postoperative adverse events. Both research groups also concluded that rhBMP-2 use in the anterior cervical spine was not safe. Whereas the UOK group found no significant differences in cancer incidence, the OHSU group found a statistically significant increase in cancer with rhBMP-2 use when compared to ICBG. Although both groups had access to the same data bank, the OHSU group included one randomized controlled clinical study that the UOK group did not use, which accounted for the differences. The authors ultimately concluded that even though there may be some increased risk of tumor with rhBMP-2 use, the absolute risk is minimal.

The technique for rhBMP-2 use has evolved since the FDA first approved use in 2002. Although the absorbable collagen sponge (ACS) is included with INFUSE, many clinical studies have suggested the concomitant use of a ceramic “bulking agent” when utilized in the posterolateral lumbar space. Other important technical factors involve not using an excessive amount of rhBMP-2 and to allow for sufficient binding times for rhBMP-2 and the ACS prior to use. In the lumbar interbody space, surgeons have proposed the use of a hydrogel such as DuraSeal to act as a barrier, helping to prevent rhBMP-2 from extravasating into an undesirable location. Some surgeons place an rhBMP-2 sponge anterior to their interbody implant and then place additional bone graft posterior to the interbody to seal off the rhBMP-2 and prevent/decrease rhBMP-2 leakage.

Using rhBMP-2 does initially result in a higher hospital cost ranging between 11% and 41% depending on the type of fusion surgery performed, of which the highest inpatient hospital costs were associated with anterior cervical fusion. Due to higher upfront costs, rhBMP-2 use is not cost-effective from a short-term payer perspective; however, from a more global societal view, it may prove to be cost-effective by decreasing lost productivity.

Surgeons must formulate their own evidence-based, cost-effective algorithms for rhBMP-2 use, as there are certainly high-risk patients who could benefit from such technology. Because one-level lumbar fusion rates in healthy patients are successful with local bone, any potential complication and cost associated with rhBMP-2 may not be reasonable. On the other hand, with high nonunion risks or those with previous ICBG harvest, any small risk of carcinogenesis with use of rhBMP-2 pales in comparison to the potential benefits of avoiding a reoperation. Even though rhBMP-2 has an overall fusion rate of 94% in the posterolateral lumbar spine, its use is not necessary in every spinal arthrodesis procedure. There are significant risks with the use of rhBMP-2 in the anterior cervical spine, so it should not be used there. Being aware of the potential complications that can occur with rhBMP-2 allows the surgeon to rapidly identify and appropriately treat them.