Betz, Randal R. MD*; Lavelle, William F. MD†; Samdani, Amer F. MD*
Part I. Bone Grafting Options in Patients With Idiopathic Scoliosis
Decisions about bone grafting options must be made for almost every spinal deformity surgery in children and adolescents. The purpose of this article is to provide the evidence-based information from the literature to aid in making these choices.
The first question to be addressed: Is autogenous iliac crest bone graft necessary in posterior spinal fusions for patients with adolescent idiopathic scoliosis (AIS)? As additional bone graft substitute options have become available, the iliac crest as a source for bone graft has been scrutinized. An ongoing controversy exists with regards to iliac crest site morbidity in AIS. Kager et al reported that the morbidity associated with the use of autograft in 71 patients was minimal.1 However, a review by Skaggs et al reported frequent complications, including persistent pain at the donor site in 31% of patients at 2 years postoperatively and 24% of patients at 4 years after surgery, with 15% of patients having functional deficits.2 Goulet et al reported 19% of patients having pain at 2 years after surgery.3
In addition, more significant complications have also been reported, including hematoma, seroma, superficial wound dehiscence, nerve and artery injuries, and ilium fractures. The morbidity appears to increase with increasing bone volume harvested. Specifically, Kessler et al found that all of the bone graft-related complications seen in their study were associated with volumes of harvested bone graft exceeding 17 cm3.4 In the early 1980s, autogenous rib graft was promoted by Howard Steel,5 especially when associated with a thoracoplasty for cosmesis. Several reports, however, have reported that there is a substantial decline in percent predicted pulmonary function tests averaging 28% following rib resection surgery.6,7
Allograft supplementation of local autogenous bone graft has been demonstrated in the literature to be as effective as autogenous iliac crest bone grafting in contributing to a successful posterior spinal fusion.8–12 However, all of these reports have essentially been uncontrolled Level IV studies. Unfortunately, an additional concern that is associated with the use of allograft is both the actual and perceived risk of disease transmission, especially the risk of viral or prion contamination that may not be susceptible to sterilization or even testable through sterility evaluations.13–15 In a recent study, Mroz et al16 performed a retrospective review of allograft recall data by tissue type, reason, and year during the period from January 1994 to June 30, 2007 which included a search for disease transmission reports in the Centers for Disease Control and Prevention database as well as the literature through Medline. These authors found that there has been only one known case of human immunodeficiency virus infection transmission to a patient undergoing spine surgery, from a fresh femoral head in 1988. This is the only case of viral transmission ever reported. There are no reports of bacterial disease transmission from the use of allograft bone to patients undergoing spine surgery.
In 2006, Betz et al reported that a posterior spinal fusion with multisegmented hook-screw-rod systems is effective without the use of autogenous bone graft.17 This was based on a randomized, controlled trial where the surgeon was blinded up to the point in surgery where the bone graft was requested. At that point, the surgeon was given either a bowl of allograft cortical cancellous chips or an empty bowl. This study reviewed 91 patients with AIS who were less than 20 years of age and scheduled for a posterior spinal fusion with multisegmented hook-screw-rod systems. Criteria included a Risser score ≥2 to minimize the risk of crankshaft. All of the curves were <80°. About 83% of the patients potentially having 2 years follow-up were reviewed for this article. The mean patient age in both groups was 15 years. A standard posterior spinal fusion was performed, including removing the cartilage of the lumbar and thoracic facet joints and decortication of the posterior elements. “Definite pseudarthrosis” (direct visualization of defect during surgical exploration of the fusion or broken instrumentation seen on radiograph) occurred in no patients (0%) in the no-graft group and 1 patient (2.7%) in the allograft group, P = 0.98. Two patients in each group met the criteria for “possible pseudarthrosis” (persistent midline, moderate to severe back pain, a defect in fusion mass in the unfused facet visible on radiograph, or curve progression >10° from initial erect postoperative radiograph [5.1% vs. 5.4%, P = 0.65]). Delayed infection occurred in 1 patient in the allograft group and superficial infection occurred in 1 patient in the no-graft group.
In an editorial concerning this article,18 Skaggs commented that if patients do well with intermediate follow-up under extreme conditions of no bone graft placement, allograft bone combined with local autograft (spinous processes) should be sufficient in most patients with idiopathic scoliosis provided adequate facet destruction is performed at every level.
The senior author (R.R.B.) still uses an extender/enhancer with local autogenous bone graft in AIS. Five years is still considered to be intermediate follow-up. Although 20% of the patients in the article just discussed had 5-year follow-up, it is not long enough that one can be confident in not adding supplemental bone graft. In addition, a thicker fusion may be present with the addition of an extender/enhancer, but there is no evidence from exploration to support this or evidence as to whether a thick fusion is better or not.
Definitions of Graft Extenders, Enhancers, and Substitutes
Louis-Ugbo et al19 defined graft extenders, enhancers, and substitutes as follows: When added to autogenous bone, graft extenders enable the fusion of more levels or the use of smaller amounts of autogenous bone and should produce a successful fusion equal to that of autogenous bone graft alone. When added to the usual amount or a decreased amount of autogenous bone, graft enhancers produce a higher rate of successful fusions than autogenous bone graft alone. Graft substitutes completely replace autogenous bone and enable a comparable or increased rate of successful fusion compared with autograft alone. Allograft in the freeze-dried form of cancellous cortical chips presents the advantage of decreased immunogenicity and disease transmission, but because of the processing, it is purely osteoconductive with no osteoinductivity.19–23
The need for a product that is both osteoinductive and osteoconductive led to the development of demineralized bone matrix (DBM). In general, DBM products are prepared by decalcification of cortical bone, exposing the extracellular matrix which contains small amounts of osteoinductive growth factors, including minute fractions of bone morphogenetic proteins (BMPs). Different formulations have been developed to improve its handling characteristics. The primary difference in the various DBM products on the market has to do with the decalcification and preparation process. These preparation processes plus particle size, sterilization, etc. relate to the osteoconductive differences of the products and are usually patent protected.24,25 Although different formulations of the DBM have been shown to be osteoinductive in rat ectopic bone models, few products have demonstrated efficacy in higher species, especially in the more challenging posterolateral spine fusion areas in rabbits and rhesus monkeys. Additionally, the various commercially available DBM products have been found to have differing osteoinductive potentials when tested in an athymic rat model.26
A study by Martin et al24 showed that newer formulations of DBM were a more effective graft alternative in a rabbit posterolateral lumbar spine arthrodesis as compared to prior generation formulations. Specifically in this study, the putty form and the flex form (both having shaved fibers) of the DBM product performed better as compared to the gel form (bone particles) used to make the DBM formulations (Table 1). In this same article, they reported that DBM formulations could actually function as an autograft substitute in this rabbit intertransverse process arthrodesis model. They showed that there was 73% fusion in the group with autograft alone and a fusion rate, 83% and 100% in the putty and flex group (with fibers), respectively. In those with bone particles as a formulation, the fusion rate was only 58% (Table 2).
Next, Louis-Ugbo et al19 tested 2 formulations of the DBM: Grafton DBM Flex and Matrix (Osteotech, Eatontown, NJ), both fiber-based products, for their potential efficacy in a rhesus monkey model, which is difficult to fuse. The posterolateral spine model is quite challenging, as evidenced by the rate of only 21% successful lumbar spine fusions when autogenous iliac crest bone graft is used alone.27 The results show that when combined with local autogenous bone, the Flex form resulted in fusion in 2 of 4 monkeys (50%) and the Matrix in 3 of 4 monkeys (75%) as compared with 21% with autograft alone. This study suggests that the use of Flex or Matrix formulations of the DBM may play a role as a bone graft enhancer (yielding a higher rate of successful fusions than autogenous bone graft alone), not merely an extender.
On the basis of the results of this preclinical work on the Matrix DBM product, the senior author of this article (R.R.B.) retrospectively reviewed his own clinical series of the use of Matrix DBM strips in posterior spinal fusions for AIS (Pahys et al, unpublished data, July 2008). Between 2002 and 2005, 30 patients underwent PSF for AIS at our institution who met the inclusion criteria of a stand alone PSF using pedicle screws or a hook/screw hybrid construct augmented with either allograft cancellous cortical chips (Red Cross) or DBM Matrix strips (Osteotech, Eatontown, NJ) and had a minimum 2-year follow-up. Allograft alone was used in 13 patients during the earlier years of the study and DBM alone was used in 17 in the latter part of the study period. Patients were clinically and radiographically followed at regular intervals to assess fusion using a previously developed radiographic grading system (1 = fusion to 4 = no fusion).28 There were no cases of definitive or probable pseudarthrosis in either the Allograft or DBM group (Figures 1 A–D) as assessed by plain radiographs. No computed tomography scans were done, and the authors acknowledge that there could be pseudarthrosis not detectable radiographically.
Radiographic grading of the fusion mass was evaluated, and all the patients in both groups attained Grade 1 fusion. The timing of achieving fusion was then evaluated. The Allograft group and the DBM group demonstrated a Grade I definitive fusion in an average of 19.5 months (range, 7–36 months) and 13.8 months (range, 5–25 months), respectively (P < 0.05). Using a mixed model ANOVA test, the DBM group demonstrated 63% fusion in 4.23 months compared to 10.57 months in the Allograft group (P < 0.05). Thus, the fusion rate was 60% faster in the DBM group compared to the Allograft group. At 12 months, the average fusion grades for the DBM and Allograft groups were 1.17 and 1.66, respectively (P < 0.05) (Figure 2). This conclusion suggests that both DBM Matrix strips and corticocancellous allograft chips are viable extenders (and possible enhancers) for local bone graft for PSF in the AIS population. However, there appeared to be a faster time to fusion in the DBM matrix strip group.
In addition, the senior author (R.R.B.) explored the in situ posterior spinal fusion of a 7-year-old boy with congenital scoliosis who had been treated with casting and no instrumentation. (It is the senior author's standard practice to explore all noninstrumented in situ fusions at 6 to 12 months for confirmation of a solid fusion or repair of a pseudarthrosis as taught by R. Winter [personal communication]). The posterior spinal fusion had been performed from C6 to T6 with Matrix strips and no autogenous bone graft. At 11 months following original fusion, on exploration, a solid bone mass was seen. Bone biopsies were obtained under an IRB approved protocol. Both specimens consisted of mature bone with no evidence of residual Matrix DBM material evident in either biopsy (Figures 3 A–C).
Another bone graft extender/enhancer is Healos (DePuy-Spine, Raynham, MA), which is based on a bovine type 1 fibular collagen (70% hydroxyapatite coating and 30% proprietary cross-linked matrix). This composition is reported to closely mimic the early phase of natural bone formation. It is always recommended to be used with bone mineral aspirate as a primary source of osteoprogenitor cells.28
Preclinical studies with Healos have been contradictory. In a posterior spinal fusion rabbit model, Tay et al30 reported that the fusion rate at 8 weeks for Healos with bone mineral aspirate (from the tibia) was better than with autograft. Both the Healos with BMA and the autograft group were better than Healos alone (Table 3). However, a study by Kraiwattanapong et al31 of the Boden posterolateral fusion rabbit model showed no fusion in any animal at 8 weeks in a group with Healos and bone mineral aspirate (from the iliac crest). The difference in these 2 studies may be in the source of BMA cells and their osteoprogenitor capability. In addition, the assessment of fusion is not the same. Further preclinical and clinical study is needed, but what is clear is that Healos needs a source of additional cells to be effective.
In a comparison of 2 noncontrolled clinical studies of Healos, both Grosse et al32 and Kitchel33 found that the fusion rate in patients treated with a one-level posterior lumbar spine fusion and using either autograft or Healos and BMA were similar. However, it is important to point out that in the Kitchel study,33 all patients had an interbody fusion and did not achieve these high levels of fusion with just a posterolateral lumbar spinal fusion. In addition, the number of cases in each group may not be significant. A fusion rate of 84% (21/25) was achieved for the autologous bone grafts and 80% (20/25) for the bone graft substitutes. As reported by McLain et al,34 it is possible to obtain adequate mesenchymal stem cells from direct vertebral body aspiration during pedicle screw placement as compared to iliac crest aspiration.
Another in the group of extenders and enhancers is synthetic bone graft. A common example is beta tricalcium phosphate (β-TCP). There are several generic forms of the product, but the one the senior author (R.R.B.) is most familiar with carries the proprietary name Vitoss (Orthovita, Malvern, PA). Like DBM products, not all generic forms of β-TCP will perform the same, as there are proprietary differences such as porosity that can significantly affect results. The study by Fleming et al14 shows that Vitoss mixed with BMA from iliac crest results in a composite containing a significant number of osteoconductive growth factors and osteogenic precursor bone cells. Epstein35 conducted a prospective study on 40 patients undergoing posterolateral lumbar fusion with 53 levels instrumented with pedicle screws (27 single level, and 13 two level). Vitoss foam strips with bone mineral aspirate from the iliac crest or pedicle plus local bone was used in approximately a 50:50 mix. There was no control group. The results showed a successful fusion for 26 of 27 single-level fusions (with 1 pseudarthrosis) and 11 of 13 two-level fusions (with 2 pseudarthroses) as assessed by dynamic plain radiographs and computed tomography scan. However, postoperative follow-up in this series was only 12 months.
In a randomized study by Lerner et al,36 20 patients using β-TCP (Vitoss) as an extender for local bone graft were compared with 20 patients using iliac crest bone graft. There were no significant differences in the preoperative or follow-up Cobb angles or in the percent correction and maintenance of correction. Follow-up averaged 46 months in the β-TCP group and 48 months in the autograft group. The postoperative visual analog score for iliac crest site showed considerable pain in 14 of 20 patients who experienced a mean of 4.1 (range, 1–8) at discharge, which improved to 4 of 20 patients experiencing pain with a mean of 3.5 (range, 2–5) at last follow-up. No patients underwent reoperation in the iliac crest bone graft group, while 2 patients did in the β-TCP group. One case was a technical failure with only 1 pedicle screw in the lower instrumented level on 1 side, and the other was revision of a pedicle screw, unrelated to the issue being studied.
Part II: Bone Grafting in Children With Neuromuscular Deformity
There is a paucity of information regarding bone grafting in patients with neuromuscular deformity. Montgomery et al37 compared 2 groups, 18 patients having autogenous bone grafting versus 12 patients who had freeze dried allograft cancellous and cortical bone as an extender to local autogenous bone. Loss of correction was 46% and 38%, respectively. This study took place before the advent of modern fixation techniques and therefore the failure rate itself is irrelevant, but the comparison between allograft and autograft is important. In the allograft group, anesthesia time decreased from 344 to 281 minutes (P < 0.050), and intraoperative blood loss decreased from an average 2730 to 1740 mL (P < 0.250) as no iliac crest graft was harvested. While the time and amount of blood loss may seem excessive for a degenerative surgery case, it is assumed that the authors were trying to harvest a large amount of ICBG. Yazici and Asher38 reported on the use of freeze-dried allograft for patients with neuromuscular spinal deformities undergoing posterior spinal fusion. In this uncontrolled cohort of 32 patients, the pseudarthrosis rate was 3.1%. There were no deep wound infections; however, there was 1 patient with a delayed wound infection which was presumed to be unrelated to the bone grafting. The level of evidence in these articles is quite poor (level IV) but, in fact, there is no literature of quality published for bone grafting of surgeries in patients with neuromuscular disease.
While confirmed pseudarthrosis may be low in the few published articles, certainly the loss of correction and probable pseudarthrosis rate associated with the loss in correction is of concern. This has led to consideration of the use of BMP in patients with neuromuscular deformity.
BMP-2, under the proprietary name INFUSE (Medtronic Sofamor Danek, Memphis, TN), is indicated for spinal fusion procedures in skeletally mature patients with degenerative disc disease at 1 level from L2 to S1. According to the package insert, it is indicated for use in an anterior interbody cage and is contraindicated in patients with known hypersensitivity, near a resected tumor, in patients who are skeletally immature (<18 years of age), pregnant women, and in those with an active infection. Although these are the indications and contraindications that have been approved by the FDA, it is important to note that this does not necessarily dictate the surgeon's practice. If appropriately justified, the surgeon can use the BMP in other situations.
Mulconrey and Lenke et al (“Safety and efficacy of bone morphogenetic protein [rhBMP-2] in a complex pediatric spinal deformity at a minimum 2-year follow-up.” Presented at International Meeting on Advanced Spine Techniques, July 8–11, 2008, Hong Kong) reported the safety and efficacy of BMP in a series of patients treated for complex pediatric spine deformity and minimum 2-year follow-up. In this retrospective review, Dr. Lenke used BMP in 20 patients with neuromuscular scoliosis, of which 9/20 were revisions for pseudarthroses. Diagnoses included cerebral palsy, muscular dystrophy, myelomeningocele, and Conradi-Hunerman Syndrome. Of the 20, 6 were anterior spinal fusions and 14 were posterior spinal fusions. In the anterior spine fusions, the fusion rate was 100% using 6 mg/level of BMP in a total of 20 vertebral levels. In the posterior group, a total of 118 levels using 5.9 mg/level of BMP had a 94% fusion rate. There was 1 infection but no other complications. One of their conclusions was that a lesser amount of BMP is necessary to develop a posterior spinal fusion in the pediatric population compared to the reported adult concentrations of 20 mg/level.
In patients treated for adolescent idiopathic scoliosis, the authors propose that iliac crest graft may not be needed for obtaining a solid posterior spinal fusion. Local bone graft with an allograft or synthetic extender of local bone graft is recommended. DBM Matrix strips may promote an earlier spinal fusion because of the osteoinductive properties in addition to the osteoconductive properties of all the extenders. It is important to note that these are recommendations for children and cannot be extrapolated to adults. For patients with neuromuscular scoliosis, more study is needed; however, the use of BMP appears to be promising for enhancing the fusion or as a substitute for bone graft.
* Up to 31% of patients have persistent pain at 2 years postsurgery when autogenous iliac crest bone graft is harvested.
* Allograft supplementation of local autograft has been demonstrated in the literature to be as effective as autogenous iliac crest bone grafting in contributing to a successful posterior spinal fusion.
* Both demineralized bone matrix (DBM) and bone morphogenetic protein (BMP) appear to contribute to a successful posterior spinal fusion.
1. Kager AN, Marks M, Bastrom T, et al. Morbidity of iliac crest bone graft harvesting in adolescent deformity surgery. J Pediatr Orthop 2006;26:132–4.
2. Skaggs DL, Samuelson MA, Hale JM, et al. Complications of posterior iliac crest bone grafting in spine surgery in children. Spine 2000;25:2400–2.
3. Goulet JA, Senunas LE, DeSilva GL, et al. Autogenous iliac crest bone graft. Complications and functional assessment. Clin Orthop Relat Res 1997;76–81.
4. Kessler P, Thorwarth M, Bloch-Birkholz A, et al. Harvesting of bone from the iliac crest: comparison of the anterior and posterior sites. Br J Oral Maxillofac Surg 2005;43:51–6.
5. Steel HH. Rib resection and spine fusion in correction of convex deformity in scoliosis. J Bone Joint Surg Am 1983;65:920–5.
6. Faro FD, Marks MC, Newton PO, et al. Perioperative changes in pulmonary function after anterior scoliosis instrumentation: thoracoscopic versus open approaches. Spine 2005;30:1058–63.
7. Kim YG, Lenke LG, Bridwell KH, et al. Pulmonary function in adolescent idiopathic scoliosis relative to the surgical procedure. J Bone Joint Surg Am 2005;87:1534–41.
8. Aurori BF, Weirman RJ, Lowell HA, et al. Pseudarthrosis after spinal fusion for scoliosis: a comparison of autogenic and allogenic bone grafts. Clin Orthop Relat Res 1985;199:153–8.
9. Bridwell KH, O'Brien MF, Lenke LG, et al. Posterior spinal fusion supplemented with only allograft bone in paralytic scoliosis. Spine 1994;19:2658–66.
10. Dodd CAF, Ferguson CM, Freedman L, et al. Allograft versus autograft in scoliosis surgery. J Bone Joint Surg Br 1988;70:431–4.
11. McCarty RE, Peek RD, Morrissy RT, et al. Allograft bone in spinal fusion for paralytic scoliosis. J Bone Joint Surg Am 1986;68:370–5.
12. Price CT, Connolly JF, Carantzas AC, et al. Comparison of bone grafts for posterior spinal fusion in adolescent idiopathic scoliosis. Spine 2003;28:793–8.
13. Boyce T, Edwards J, Scarborough N. Allograft bone: the influence of processing on safety and performance. Orthop Clin North Am 1999;30:571–81.
14. Fleming JE Jr, Cornell CN, Muschler GF. Bone cells and matrices in orthopedic tissue engineering. Orthop Clin North Am 2000;31:357–74.
15. Ehrler DJ, Vaccaro AR. The use of allograft bone in lumbar spine surgery. Clin Orthop Relat Res 2000;371:38–45.
16. Mroz TE, Joyce MJ, Lieberman IH, et al. The use of allograft bone in spine surgery: is it safe? Spine J 2009;9:303–8.
17. Betz RR, Petrizzo AM, Kerner PJ, et al. Allograft versus no graft with a posterior multisegmented hook system for the treatment of idiopathic scoliosis. Spine 2006;31:121–7.
18. Skaggs DL. Posterior spinal fusion was not improved by supplemental bone graft in adolescent idiopathic scoliosis. J Bone Joint Surg Am 2006;88:2313.
19. Louis-Ugbo J, Murakami H, Kim HS, et al. Evidence of osteoinduction by Grafton demineralized bone matrix in nonhuman primate spinal fusion. Spine 2004;29:360–6.
20. Betz RR, Harms J, Clements DH, et al. Comparison of anterior and posterior instrumentation for correction of adolescent thoracic idiopathic scoliosis. Spine 1999;24:225–39.
21. Lee KJ, Roper JG, Wang JC. Demineralized bone matrix and spinal arthrodesis. Spine J 2005;5:217S–23S.
22. Asselmeier MA, Caspari RB, Bottenfield S. A review of allograft processing and sterilization techniques and their role in transmission of the human immunodeficiency virus. Am J Sports Med 1993;21:170–5.
23. Vaccaro AR, Chiba K, Heller JG, et al. Bone grafting alternatives in spinal surgery. Spine J 2002;2:206–15.
24. Martin GJ, Boden SD, Titus L, et al. New formulations of demineralized bone matrix as a more effective graft alternative in experimental posterolateral lumbar spine arthrodesis. Spine 1999;24:637–45.
25. Morone MA, Boden SD. Experimental posterolateral lumbar spinal fusion wtih a demineralized bone matrix gel. Spine 1998;23:159–67.
26. Peterson B, Whang PG, Iglesias R, et al. Osteoinductivity of commercially available demineralized bone matrix: preparations in a spine fusion model. J Bone Joint Surg Am 2004;86:2243–50.
27. Boden SD, Schimandle JH, Hutton WC. Evaluation of a bovine-derived osteoinductive bone protein in a non-human primate model of lumbar spinal fusion. Trans Orthop Res Soc 1996;21:118.
28. Burwell RG. Studies in the transplantation of bone: part VII. The fresh composite homograft-autograft of cancellous bone; an analysis of factors leading to osteogenesis in marrow transplants and in marrow-containing bone grafts. J Bone Joint Surg Br 1964;46:110–40.
29. Bauer TW. An overview of the histology of skeletal substitute materials. Arch Pathol Lab Med 2007;131:217–24.
30. Tay BK, Le AX, Heilman M, et al. Use of a collagen-hydroxyapatite matrix in spinal fusion: a rabbit model. Spine 1998;23:2276–81.
31. Kraiwattanapong C, Boden SD, Louis-Ugbo J, et al. Comparison of Healos/bone marrow to INFUSE(rhBMP-2/ACS) with a collagen-ceramic sponge bulking agent as graft substitutes for lumbar spine fusion. Spine 2005;30:1001–7.
32. Grosse S, Argenson C, Tavna A, et al. “Mineralized collagen as a replacement for autogenous bone in posterolateral lumbar spinal fusion.” Presented at: EuroSpine; September 7–11, 1999; Munich, Germany.
33. Kitchel SH. A preliminary comparative study of radiographic results using mineralized collagen and bone marrow aspirate versus autologous bone in the same patients undergoing posterior lumbar interbody fusion with instrumented posterolateral lumbar fusion. Spine J 2006;6:405–11.
34. McLain RF, Fleming JE, Boehm CA, et al. Aspiration of osteoprogenitor cells for augmenting spinal fusion: comparison of progenitor cell concentrations from the vertebral body and iliac crest. J Bone Joint Surg Am 2005;87:2655–61.
35. Epstein NE. A preliminary study of the efficacy of Beta Tricalcium Phosphate as a bone expander for instrumented posterolateral lumbar fusions. J Spinal Disord Tech 2006;19:424–9.
36. Lerner T, Bullmann V, Schulte TL, et al. A level-1 pilot study to evaluate of ultraporous beta-tricalcium phosphate as a graft extender in the posterior correction of adolescent idiopathic scoliosis. Eur Spine J 2009;18:170–9.
37. Montgomery DM, Aronson DD, Lee CL, et al. Posterior spinal fusions: allograft versus autograft bone. J Spinal Disord 1990;3:370–5.
38. Yazici M, Asher MA. Freeze-dried allograft for posterior spinal fusion in patients with neuromuscular spinal deformities. Spine 1997;22:1467–71.
© 2010 Lippincott Williams & Wilkins, Inc.