Properly harvested iliac crest bone autograft applied to a meticulously prepared fusion bed produces a consistently high rate of fusion with a low incidence of donor site morbidity. Some reports advocate substituting bone morphogenic protein (BMP) for iliac crest bone autograft, but in posterolateral lumbar spinal fusion, BMP appears better suited to facilitate iliac crest bone autograft maturation than to substitute for it.
In this single-center, nonrandomized, prospective study (minimum 2-year follow-up), cancellous-only iliac crest bone autograft was harvested for use in posterolateral lumbar spinal fusion. Reviewers blinded to graft condition and age assigned fusion scores to the random radiographs of 31 consecutive patients who underwent 1- to 3-level posterolateral lumbar spinal fusion using iliac crest bone autograft supplemented with either an implanted spinal fusion stimulator or BMP.
There was no significant immediate or remote iliac crest bone autograft harvest morbidity, and there was a significant reduction in pain scores postoperatively (P<.001). at="" 12="" months,="" bmp="" radiographs="" were="" more="" likely="" than="" spinal="" fusion="" stimulator="" radiographs="" to="" be="" rated="" as="" fused="">P<.019). all="" bmp="" patients="" were="" deemed="" fused="" at="" 12="" months="" and="" all="" spinal="" fusion="" stimulator="" patients="" at="" 24="" months.="">
In this study, iliac crest bone autograft supplemented with either BMP or spinal fusion stimulator resulted in a solid contiguous fusion without significant iliac crest bone autograft harvest-related morbidity. Bone morphogenic protein-supplemented iliac crest bone autograft fused at a faster rate, producing the more mature-appearing, trabeculated, robust fusion.
Successful instrumented posterolateral lumbar spinal fusion requires not only proper pedicle and graft bed preparation, but also a sufficient supply of readily available grafting material capable of producing a solid fusion mass. Ideally, graft material should impart low risk and morbidity to the patient, and we believe that iliac crest bone autograft when properly harvested and implanted fulfills these criteria.
The literature regarding iliac crest bone autograft harvest morbidity and substitution products is intertwined. Studies describing iliac crest bone autograft harvest site morbidity often contradict one another. Some suggest that harvest site morbidity is significant, resulting in prolonged postoperative pain in 25% to 60% of patients,1-3 while others report only the anticipated immediate postoperative pain, which resolves without disabling harvest site pain or deformity.4-6 Iliac crest bone autograft harvest morbidity is often reported without regard to differences in graft harvest site, measures of morbidity, or surgical technique.
Claims of significant iliac crest bone autograft harvest-related morbidity, insufficient yield of graft material from iliac crest bone autograft, and unacceptable fusion rates have fueled the development of a host of products marketed as iliac crest bone autograft augmenters, extenders, or even substitutes, the most recent being biologics such as bone morphogenic protein (BMP).7
Bone morphogenic protein is commercially available as recombinant bone morphogenic protein-2 (rhBMP-2), marketed as INFUSE (Sofamor Danek, Memphis, Tennessee), and rhBMP-7, marketed as OP-1 (Stryker Biotech, Hopkinton, Massachusetts). Originally, these BMP products were approved for graft augmentation in anterior lumbar interbody fusion. More recently, however, some researchers have promoted their use in posterolateral lumbar spinal fusion as well, claiming that a host of iliac crest bone autograft harvest-related costs and morbidities are thereby eliminated.1,8-10
Bone morphogenic protein is an exciting technique for lumbar spinal arthrodesis. Its well-documented osteoinductive ability suggests a potential to accelerate iliac crest bone autograft maturation and enhance fusion integrity in posterolateral lumbar spinal fusion.11,12 However, we have not experienced the degree of iliac crest bone autograft-associated morbidity reported in the BMP-for-iliac crest bone autograft substitution literature and believe that the reports of iliac crest bone autograft harvest morbidity and associated increased costs used to justify BMP-for-iliac crest bone autograft substitution are misstated.
In this study, we hypothesized that (1) adequate amounts of cancellous-only iliac crest bone autograft could be harvested from the posterior superior iliac spine without significant harvest site morbidity and that (2) iliac crest bone autograft supplemented with BMP would fuse at a faster rate and produce a more robust-appearing fusion mass than iliac crest bone autograft augmented with an implantable battery-powered spinal fusion stimulator.13
Materials and Methods
This single-center, nonrandomized, prospective study used multiple dependent variables including blinded radiograph reviews in a between-groups comparison of BMP-supplemented to spinal fusion stimulator-supplemented posterolateral lumbar spinal fusion using iliac crest bone autograft. All patients were required to be tobacco- and nonsteroidal anti-inflammatory drug (NSAID)-free pre- and postoperatively. The a priori value for statistical significance was set at P<.05.>
Between March 2004 and October 2005, 15 consecutive patients with lumbar degenerative disk disease underwent posterolateral lumbar spinal fusion with iliac crest bone autograft using the SpineLink II segmental fixation system (EBI Biomet, Parsippany, New Jersey) supplemented with an implantable battery-powered spinal fusion stimulator (SpF; EBI Biomet) (spinal fusion stimulator group).
Between November 2005 and October 2006, 15 consecutive patients with lumbar disk disease underwent posterolateral lumbar spinal fusion with iliac crest bone autograft using SpineLink II, and 1 patient diagnosed with grade II+ spondylolisthesis underwent posterolateral lumbar spinal fusion with iliac crest bone autograft using the SpineLink I segmental fixation system (EBI Biomet) (BMP group). All grafts in the BMP group were supplemented with recombinant human bone morphogenic protein-2 (rhBMP-2) using a single INFUSE kit containing 12 mg of rhBMP-2 (1.5 mg/mL) and collagen sponge, regardless of the number of levels fused. The same 2-orthopedic surgeon team (A.R., C.R.) performed all surgeries, and the same surgical techniques were used to perform each posterolateral lumbar spinal fusion and each iliac crest bone autograft harvest.
Iliac Crest Bone Autograft Harvest Technique and Application
After exposing the posterior-most prominence of the posterior superior iliac spine through a separate paramedial incision, an osteotome was used to ostectomize a 2×3-cm cap off the prominence, allowing access to the underlying cancellous bone. Using various gouges, curettes, and rongeurs, cancellous-only autograft was harvested, taking care not to violate the sacroiliac joint or broach the cortical margins. After completing the harvest, the ilium cavitary defect was backfilled with morselized allograft chips, which were impacted into position with bone impactors. Then the dorsal fascia and subcutaneous tissue were sutured, and the skin was closed with staples. At closure, 0.5% marcaine was infiltrated into the subcutaneous tissue and fascia overlying the graft site.
Fusion Site Preparation
The transverse processes, pars region, and facet joints were decorticated prior to either spinal fusion stimulator or BMP graft application. For spinal fusion stimulator implantation, the harvested morselized iliac crest bone autograft was placed onto the intertransverse fascial sling on each side of the lateral gutter spanning the transverse processes and/or sacral ala, each side receiving a quarter of the graft material. The cathode wire mesh and a second quarter of the iliac crest bone autograft were then overlaid, filling the lateral gutter and creating an iliac crest bone autograft/spinal fusion stimulator/iliac crest bone autograft graft sandwich with one-half of the graft. The anode was placed along the spinous process in-pocket proximally.
For BMP application, INFUSE was reconstituted according to the manufacturer’s protocol and placed on the collagen sponge, which was cut into 4 strips. A sponge strip was then placed in each lateral gutter spanning the transverse processes and/or sacral ala, and a quarter of the morselized iliac crest bone autograft was placed over it. This process was repeated using a second strip of sponge followed by another quarter of the iliac crest bone autograft to form a lateral gutter BMP/iliac crest bone autograft sandwich with one-half of the graft. Iliac crest bone autograft was apportioned equally between each lateral gutter for both spinal fusion stimulator and BMP patients so that graft material was not overlying the exposed dura or the exited nerve roots. After graft application, attention was directed to placement of the pedicle screws at the previously prepared pedicles and to assembling the spinal fixation device.
For both groups, graft status was assessed at 1.5, 3, 6, 12, and 24 months postoperatively with standard lumbar anteroposterior (AP) and lateral radiographs. To minimize viewer bias, BMP radiographs were made using an actual spinal fusion stimulator with battery and lead wires taped externally to the patient’s back to simulate the radiographic presence of an implanted spinal fusion stimulator. At 2 years postoperatively, 8 BMP patients and 1 spinal fusion stimulator patient underwent thin-slice computed tomography (CT) through the fusion site.
Reviewers and presenters were blinded to graft site age and type by removing patient names and dates from all films and replacing them with an identifying code. Radiographs were presented separately to each reviewer in a sequence randomized for graft site age and graft type. The reviewers were the 2 orthopedic surgeons who had performed all 31 surgeries. The orthopedic surgeons and 1 radiologist reviewed the CT scans. Reviewers assigned a numeric score to each set of radiographs using a forced choice method, with 1=no fusion, 2=partial unilateral fusion, 3=partial bilateral fusion, 4=solid unilateral fusion, and 5=solid bilateral fusion. Grades 1 and 2 were defined as a nonunion, grade 3 as immature/indeterminate, and grades 4 and 5 as a solid, contiguous, mature fusion.
Each patient completed a pain diagram as well as a 10-point visual analog pain scale preoperatively and at each postoperative follow-up. Orthopedic examination at each follow-up routinely included visual inspection and palpation of the midline lumbar and graft harvest site incisions.
Patient and Surgery Characteristics
No statistically significant difference was found between the BMP and spinal fusion stimulator group means for intraoperative blood loss, surgery time, patient age at time of surgery, preoperative patient pain scores, or postoperative length of hospitalization (Table 1). Intraoperative blood loss and surgery time were a function of the number of levels fused and whether or not surgery was performed on a virgin back or was a revision or extension of a previous surgery. Types of fusion surgeries were equally distributed between the BMP and spinal fusion stimulator groups (Table 2), although more patients in the BMP group had previously undergone a laminotomy or laminectomy.
Iliac Crest Bone Autograft
Single-crest harvest provided cancellous bone adequate to complete all posterolateral lumbar spinal fusions, including 1- to 3-level posterolateral lumbar spinal fusion, redo of previous failed posterolateral lumbar spinal fusion, or extension of a previous posterolateral lumbar spinal fusion to adjacent levels. No cortical margin breaches or nerve or arterial injuries were encountered during iliac crest bone autograft harvest. There were no intensive care unit admissions or blood transfusions and no harvest or primary incision site postoperative infections. Bone graft site pain never exceeded patients’ reported midline incision pain level as determined by visual analog scores and pain on palpation during physical examination. No Tinel’s signs were present over the graft site or midline incisions at 6, 12, or 24 months postoperatively.
When compared to preoperative scores, 3-month postoperative pain scores decreased significantly (P<.0001, two-tailed="">t test), and scores continued to decrease progressively over the 24-month postoperative period for both groups (P<.001, analysis="" of="" variance).="" interval="" means="" and="" ranges="" are="" listed="" in="" table="" 3.="">
When rating radiographs for fusion quality, the 2 surgeons matched rating scores exactly 64% of the time (correlation coefficient=0.67; R2=0.45), differed by 1 point or none 91% of the time, and differed by 2 points 8.6% of the time.
Radiograph Scores and Graft Maturation
As expected, both groups demonstrated significantly higher radiographic scores at 24 months postoperatively than at 6 weeks postoperatively (analysis of variance); however, all BMP group patients were deemed fused by 12 months postoperatively, but all spinal fusion stimulator group patients were not deemed fused until 24 months postoperatively. Specifically, 12-month radiographs for the BMP group were more likely to be rated either 4 or 5, signifying a unilateral or bilateral fusion, than radiographs for the spinal fusion stimulator group (P<.019, fisher’s="" exact="" test,="" two-tailed).="" this="" was="" the="" only="" postoperative="" time="" interval="" at="" which="" between-group="" fusion="" ratings="" were="" significantly="" different="" statistically.="" fusion="" scores="" for="" the="" 2="" groups="" converged="" at="" 24="" months.="" the="" figure="" illustrates="" fusion="" progression="" as="" determined="" by="" radiographic="" scores.="" the="" means="" for="" radiographic="" fusion="" scores="" are="" listed="" in="" table="" 3.="">
Figure: Bone morphogenic protein- and spinal fusion stimulator-facilitated iliac crest bone autograft maturation as demonstrated by radiographic fusion grades: 1 & 2=nonunion; 3=indeterminate/immature; 4 & 5=solid contiguous mature fusion. Abbreviations: BMP, bone morphogenic protein; SFS, spinal fusion stimulator.
For the BMP group, significant increases in radiographic fusion scores occurred between the 1.5- and 3-month intervals and the 6- and 12-month intervals (two-tailed t test, P<.03 and="" .02,="" respectively),="" consistent="" with="" an="" interval="" increase="" in="" graft="" maturation="" and="" consolidation="" that="" did="" not="" occur="" in="" the="" comparable="" spinal="" fusion="" stimulator="" group.="">
One patient in the spinal fusion stimulator group who resumed smoking during the immediate postoperative period demonstrated radiographic evidence of a nonunion at 12 months, and this was confirmed surgically. This patient underwent a global lumbar fusion at the previously operated level, and at 24 months after repeat surgery was deemed fused. There were no nonunions in the BMP group, and even a patient with significant osteopenia received fusion ratings of 4 and 5 at 6, 12, and 24 months postoperatively. This patient’s fusion was confirmed by thin-cut CT at 24 months. The radiologist and surgeons agreed that thin-cut CT studies performed through the fusion site at 24 months postoperatively and including views of the facets demonstrated a solid intertransverse process and facet fusion with intact fixation and no implant loosening for 9 of 9 patients who underwent CT.
One patient who had previously undergone posterolateral lumbar spinal fusion from L3-S1 using a spinal fusion stimulator underwent a posterolateral lumbar spinal fusion extension to the adjacent L2 level using BMP. This patient’s coronal CT was striking. The segments fused using BMP exhibit a distinctly more robust, better trabeculated fusion mass than the lower segments previously fused with a spinal fusion stimulator. The spinal fusion stimulator/BMP fusion margin is well delineated. When the surgeons (blinded to condition) were presented with BMP/spinal fusion stimulator radiograph pairs (matched for postoperative ages of 6 or 12 months) and were asked to select the one that appeared more mature/robust, they selected the BMP radiograph 80% of the time.
Iliac Crest Bone Autograft Morbidity
In this study, the iliac crest bone autograft harvest site surgical morbidity, prolonged and/or severe postoperative site pain and disability, prolonged hospital stay, and insufficient graft material referenced by others did not occur.1,2,4,7 There was a 30% reduction in average patient pain scores by 3 months postoperatively with a continuous, gradual decrease in average pain scores throughout the remainder of the 24-month postoperative period. At 24 months postoperatively, average patient pain scores were 45% lower than preoperative levels. Only 19% of patients were working preoperatively, while 48% were working at 24 months postoperatively. Those working preoperatively resumed work postoperatively. No donor blood transfusions were required, there were no intensive care unit admissions, average postoperative hospitalization was 2.5 days, and graft site pain never exceeded lumbar incision pain.
The wide variation in iliac crest bone autograft morbidity reported in the literature reflects the differences in morbidity associated with different sites of harvest and different harvesting techniques. Iliac crest bone autograft harvests performed in the 1980s and 1990s were more likely to result in significant and enduring graft site morbidity than the techniques commonly used today. In fact, more recent studies suggest that iliac crest bone autograft harvest site pain5,14 and harvest site operative and postoperative complications4,15 may be overestimated and patient satisfaction underestimated.14 A combination of experience- and evidence-based medicine has improved our harvesting techniques. Knowing that iliac crest bone autograft harvested from the anterior iliac spine results in increased intensity and duration of donor site pain as well as a greater risk for major complications relative to posterior iliac crest harvest,4 we harvest from the posterior crest as previously described. Because of increased morbidity associated with corticocancellous iliac crest bone autograft harvest,16 we perform cancellous-only iliac crest bone autograft. Backfilling the harvest site cavitary defect17 and instilling anesthetic into the site18 further decrease iliac crest bone autograft harvest morbidity.
Estimated operating time related to iliac crest bone autograft harvest is in the 15- to 25-minute range, with Glassman et al19 describing an average increased operating time of 22 minutes for posterolateral lumbar spinal fusion with iliac crest bone autograft and graft extenders compared to posterolateral lumbar spinal fusion without iliac crest bone autograft, and Putzier et al15 reporting 17 minutes on average to harvest cancellous-only iliac crest bone autograft. Glassman et al19 reported an average hospital stay of 6 days for posterolateral lumbar spinal fusion–iliac crest bone autograft patients and 5 days for posterolateral lumbar spinal fusion–BMP patients, 2.5 to 3.5 days longer than encountered in the present study.
Measures of post-iliac crest bone autograft harvest pain and intensity vary and include retrospective and prospective patient satisfaction/pain surveys and chart reviews,3,5,20 as well as direct visual and physical inspection of the site over different postoperative intervals.1 Following the immediate postoperative period, patients reported a progressive improvement in harvest site pain and function.2,6,8,14 The question then becomes what constitutes “significant” morbid postoperative harvest site pain? Heary et al,20 using a patient survey at a “time remote from [iliac crest bone autograft] surgery,” found that although 34% reported some degree of pain, 31% reported only occasional or mild pain, and only 3% reported unacceptable pain.
Fusion Quality and Integrity
One nonunion occurred in the iliac crest bone autograft–spinal fusion stimulator supplemented group in our study: a patient who resumed cigarette use soon after surgery. All patients in the iliac crest bone autograft–BMP supplemented group (including 1 moderately osteopenic patient) were deemed fused. The calculated nonunion rate was 6.6% for the spinal fusion stimulator group, 0% for the BMP group, and 3.2% for all iliac crest bone autografts combined.
In a meta-analysis of the efficacy of iliac crest bone autograft and BMP in posterolateral fusion of the lumbar spine, Papakostidis et al21 noted that some studies used 20 mg of rhBMP-2 per side to achieve a fusion without iliac crest bone autograft. These amounts are >3 times that used in the present study, and it is questionable if BMP holds any monetary advantage over iliac crest bone autograft when used in these amounts or if such high doses will result in adverse effects at the application site12 or in BMP migration into areas in which it is not desired. By performing posterolateral lumbar spinal fusions with 6 mg rhBMP-2 per side per level in combination with iliac crest bone autograft, Singh et al11 obtained significantly better fusion rates than in posterolateral lumbar spinal fusions performed with iliac crest bone autograft alone.
The high rates of iliac crest bone autograft harvest morbidity reported in the literature were not seen in our study, and posterolateral lumbar spinal fusion using iliac crest bone autograft augmented with BPM produced more rapid graft maturation and a more robust, trabeculated solid fusion when compared to posterolateral lumbar spinal fusion using iliac crest bone autograft augmented with a spinal fusion stimulator. Properly performed posterolateral lumbar spinal fusion using iliac crest bone autograft augmented with low-dose BMP produces a robust, rapidly maturing solid fusion without significant iliac crest bone autograft harvest-related enduring pain or surgical morbidity.
- Vaccaro AR, Whang PG, Patel T, et al. The safety and efficacy of OP-1 (rhBMP-7) as a replacement for iliac crest autograft for posterolateral lumbar arthrodesis: minimum 4-year follow-up of a pilot study. Spine J. 2008; 8(3):457-465.
- Dimar JR II, Glassman SD, Burkus JK, Pryor PW, Hardacker JW, Carreon LY. Two-year outcomes in 224 patients treated with a single-level posterolateral fusion with ICBG. Spine J. In press.
- Sasso RC, LeHuec JC, Shaffrey C; Spine Interbody Research Group. Iliac crest bone graft donor site pain after anterior lumbar interbody fusion: a prospective patient satisfaction outcome assessment. J Spinal Disord Tech. 2005; (18 Suppl):S77-81.
- Ahlmann E, Patzakis M, Roidis N, Shepherd L, Holtom P. Comparison of anterior and posterior iliac crest bone grafts in terms of harvest-site morbidity and functional outcomes. J Bone Joint Surg Am. 2002; 84(5):716-720.
- Delawi D, Dhert WJ, Castelein RM, Verbout AJ, Oner FC. The incidence of donor site pain after bone graft harvesting from the posterior iliac crest may be overestimated: a study on spine fracture patients. Spine (Phila Pa 1976). 2007; 32(17):1865-1868.
- Dai LY, Jiang LS. Single-level instrumented posterolateral fusion of lumbar spine with beta-tricalcium phosphate versus autograft: a prospective, randomized study with 3-year follow-up. Spine (Phila Pa 1976). 2008; 33(12):1299-1304.
- Gupta MC, Maitra S. Bone grafts and bone morphogenic proteins in spine fusion. Cell Tissue Bank. 2003; 3(4):255-267.
- Burkus JK, Sandhu HS, Gornet MF, Longley MC. Use of rhBMP-2 in combination with structural cortical allografts: clinical and radiographic outcomes in anterior lumbar spinal surgery. J Bone Joint Surg Am. 2005; 87(6): 1205-1212.
- Glassman SD, Carreon L, Djurasovic M, et al. Posterolateral lumbar spine fusion with INFUSE bone graft. Spine J. 2007; 7(1):44-49.
- Glassman SD, Dimar JR III, Burkus K, et al. The efficacy of rhBMP-2 for posterolateral lumbar fusion in smokers. Spine (Phila Pa 1976). 2007; 32(15):1693-1698.
- Singh K, Smucker JD, Gill S, Boden SD. Use of recombinant human bone morphogenetic protein-2 as an adjunct in posterolateral lumbar spine fusion: a prospective CT-scan analysis at one and two years. J Spinal Disord Tech. 2006; 19(6):416-423.
- Bridwell KH, Anderson PA, Boden SD, Vaccaro AR, Wang JC. What’s new in spine surgery. J Bone Joint Surg Am. 2007; 89(7):1654-1663.
- Rogozinski A, Rogozinski C. Efficacy of implanted bone growth stimulation in instrumented lumbosacral spinal fusion. Spine (Phila Pa 1976). 1996; 21(21):2479-2483.
- Robertson PA, Wray AC. Natural history of posterior iliac crest bone graft donation for spinal surgery: a prospective analysis of morbidity. Spine (Phila Pa 1976). 2001; 26(13):1473-1476.
- Putzier M, Strube P, Funk J, Gross C, Perka C. Periosteal cells compared with autologous cancellous bone in lumbar segmental fusion. J Neurosurg Spine. 2008; 8(6):536-543.
- Varga E, Hu R, Hearn TC, Woodside T, Yang JP. Biomechanical analysis of hemipelvic deformation after corticospongious bone graft harvest from the posterior iliac crest. Spine (Phila Pa 1976). 1996; 21(13):1494-1499.
- Bojescul JA, Polly DW Jr, Kuklo TR, Allen TW, Wieand KE. Backfill for iliac-crest donor sites: a prospective, randomized study of coralline hydroxyapatite. Am J Orthop. 2005; 34(8):377-382.
- Singh K, Phillips FM, Kuo E, Campbell M. A prospective, randomized, double-blind study of the efficacy of postoperative continuous local anesthetic infusion at the iliac crest bone graft site after posterior spinal arthrodesis: a minimum of 4-year follow-up. Spine (Phila Pa 1976). 2007; 32(25):2790-2796.
- Glassman SD, Carreon LY, Campbell MJ, et al. The perioperative cost of Infuse bone graft in posterolateral lumbar spine fusion. Spine J. 2008; 8(3):443-448.
- Heary RF, Schlenk RP, Sacchieri TA, Barone D, Brotea C. Persistent iliac crest donor site pain: independent outcome assessment. Neurosurgery. 2002; 50(3):510-516.
- Papakostidis C, Kontakis G, Bhandari M, Giannoudis PV. Efficacy of autologous iliac crest bone graft and bone morphogenetic proteins for posterolateral fusion of lumbar spine: a meta-analysis of the results. Spine (Phila Pa 1976). 2008; 33(19):E680-692.
Drs Rogozinski (Abraham) and Rogozinski (Chaim) and Mr Cloud are from the Rogozinski Orthopedic Clinic, Jacksonville, Florida.
Drs Rogozinski (Abraham) and Rogozinski (Chaim) receive royalties from EBI Biomet. Mr Cloud has no relevant financial relationships to disclose.
Correspondence should be addressed to: Gregory Cloud, PA-C, MHA, Rogozinski Orthopedic Clinic, 3716 University Blvd S #3, Jacksonville, FL 32216 (firstname.lastname@example.org).