Complications associated with single-level transforaminal lumbar interbody fusion

The Spine Journal 9 (2009) 623–629

Clinical Studies

Complications associated with single-level transforaminal lumbar interbody fusion Jeffrey A. Rihn, MDa,*, Ravi Patel, BSb, Junaid Makda, MDa, Joseph Hong, MDa, David G. Anderson, MDa, Alexander R. Vaccaro, MD, PhDa, Alan S. Hilibrand, MDa, Todd J. Albert, MDa a

Department of Orthopaedic Surgery, The Rothman Institute, Thomas Jefferson University Hospital, 925 Chestnut Street, Fifth Floor, Philadelphia, PA 19107, USA b Thomas Jefferson Medical College, 1025 Walnut Street, Philadelphia, PA 19107, USA Received 24 November 2008; accepted 6 April 2009

Abstract

BACKGROUND CONTEXT: The transforaminal lumbar interbody fusion (TLIF) procedure has become an increasingly popular means of obtaining a circumferential fusion while avoiding the morbidity of the anterior approach. Concerns remain, however, regarding the clinical efficacy and safety of its use. PURPOSE: The purpose of this study was to evaluate the complications of the single-level TLIF procedure. The difference in complications observed with the use of iliac crest autograft compared with rhBMP-2 will be assessed. STUDY DESIGN: Retrospective cohort study; a review of complications. METHODS: Patients who underwent a single-level TLIF between January 2004 to May 2007 with either autograft iliac crest or rhBMP-2 were identified. A retrospective review of these patients included operative reports, pre- and postoperative medical records, most recent postoperative dynamic and static lumbar radiographs, and computed tomography scans (when available). RESULTS: A total of 130 patients met the study criteria; 119 patients were available for follow-up, with an average radiographic follow-up of 19.1 months and an average clinical follow-up of 27.6 months. Thirty-three patients received iliac crest autograft and 86 patients received rhBMP-2. Complications occurred in 40 of the 119 study patients (33.6%). The autograft group had a higher complication rate (45.5% vs. 29.1%), but the difference was not statistically significant (p5.09). Complications in the autograft group included persistent donor-site pain (30.3%), donor-site infection (3.1%), lumbar wound infection (6.1%), and postoperative radiculitis (3.0%). Complications in the rhBMP-2 group included postoperative radiculitis (14.0%), vertebral osteolysis (5.8%), ectopic bone formation (2.3%), and lumbar wound infection (3.5%). A hydrogel sealant (Duraseal; Confluent Surgical Inc., Waltham, MA, USA) was used in 37 out of 86 patients in the rhBMP-2 group. The use of this sealant decreased the rate of postoperative radiculitis in the rhBMP-2 group from 20.4% to 5.4% (p5.047). The radiographic nonunion rate at most recent follow-up was 3.0% in the autograft group and 3.5% (p5.90) in the rhBMP-2 group. CONCLUSIONS: The most common complications in the autograft group were related to the donor site. The most common complication in the rhBMP-2 group was postoperative radiculitis, the incidence of which is reduced by the use of a hydrogel sealant. Ó 2009 Elsevier Inc. All rights reserved.

Keywords:

Complications; Transforaminal lumbar interbody fusion; TLIF; BMP; Bone morphogenetic protein

FDA drugs: not approved for this indication (rhBMP-2 and Duraseal [off-label use]). Author disclosures: TNA (royalties, consulting fees, speaking arrangements from DePuy Spine; stock ownership, K2M; scientific advisory board of Inuve Gertis); ARV (consulting fees from Depuy and Medtronic); DGA 1529-9430/09/$ – see front matter Ó 2009 Elsevier Inc. All rights reserved. doi:10.1016/j.spinee.2009.04.004

(royalties, consulting fees from Depuy from Medtronic); JAR (speaking arrangements, Synthes Speakers; trips/travel, Stryker Travel Grant). * Corresponding author. Tel.: (267) 339-3500; fax: (215) 503-0054. E-mail address: [email protected] (J.A. Rihn)

624

J.A. Rihn et al. / The Spine Journal 9 (2009) 623–629

Context Surgical manipulation, scarring, and bone formation near the dorsal root ganglion have been concerns when considering the TLIF procedure. The off-label use of rhBMP-2 in TLIF has become increasingly popular, though the effects of this technique, compared with autograft use, are not well studied.

Contribution In this retrospective case series from one institution, the authors have found a relatively high complication rate using the TLIF technique, with specific complications associated with both rhBMP and autograft methods. Osteolysis and radiculitis rates in the rhBMP group (6% and 14%, respectively) are reported.

Implications While TLIF technique is in some ways attractive, this study indicates adverse events are common. Operating through the foramen adjacent to sensory ganglion is associated with the risk of both direct intra-operative injury and the delayed radiculitis. The use of rhBMP in this area has additional unique problems, as noted in this study. The sobering complications reported in this study suggest the popularity of the TLIF, with or without rhBMP, may need reassessment and comparative studies evaluating competing strategies. dThe Editors

Introduction The transforaminal lumbar interbody fusion (TLIF) procedure is traditionally performed with the use of iliac crest autograft. Although clinical success has been documented with iliac crest autograft, its use is not without consequence [1–3]. Approximately 2% to 3% of patients require reoperation at the donor site because of wound complications and/ or infection and up to 58% of patients report persistent donor-site pain [3–6]. Because of the morbidity associated with iliac crest harvest, the ‘‘off-label’’ use of recombinant human bone morphogenetic protein type 2 (rhBMP-2) (InFuse; Medtronic Sofamor Danek, Memphis, TN, USA) in the TLIF procedure has become increasingly popular. Although several studies have demonstrated the successful use of rhBMP-2 for this indication [3,7–9] concerns remain regarding the safety of its use. Prior studies have reported overall complication rates of up to 35% after the use of rhBMP-2 in the TLIF procedure

[2,3,7,10,11]. These complications include pedicle screw malposition, interbody cage migration, hematoma, infection, CSF leak, transient and persistent nerve-related symptoms, heterotopic bone formation, and vertebral osteolysis. Villavicencio et al. [7] reported a 12.7% (9/71 patients) incidence of postoperative neural injury after TLIF performed with rhBMP-2. Although most had resolution of their symptoms within 3 months, two of these patients had radiculopathy that persisted at most recent follow-up [7]. Potter et al. [2] in a retrospective study of 100 TLIF procedures performed with bone morphogenetic protein (BMP), reported a 7% incidence of postoperative transient radiculopathy. The most common nerve root affected was L5. Most cases resolved completely between 1 and 6 weeks postoperatively [2]. The use of rhBMP-2 for interbody fusion has raised concern for heterotopic bone formation in the epidural space. In a study by Paramore et al. [12], a canine model was used to investigate the effect of BMP that was intentionally placed into the subarachnoid space and fusion bed. At 16 weeks, new bone had developed in the subarachnoid space of all animals. Spinal cord compression and mild spinal stenosis were detected in all treated animals, but no clinical or pathological features of neurotoxicity were noted [12]. A recent study by Joseph and Rampersaud [13] assessed the incidence and clinical sequelae of heterotopic bone formation with the use of rhBMP-2 for posterior lumbar interbody fusion (PLIF) and TLIF. Heterotopic bone formation occurred in 20.8% with the use of BMP compared with 8.3% without the use of BMP. Despite the higher incidence, there does not seem to be any associated clinical sequelae [13]. Several past studies have demonstrated similar findings [3,12–15]. The reported rate of vertebral end plate resorption and osteolysis after TLIF using rhBMP-2 varies widely. This variability is likely because of differences in surgical technique, diagnostic imaging, and the definition of vertebral osteolysis used to make the diagnosis. A 2006 study by McClellan et al. [16] demonstrated vertebral bone resorption at 22 lumbar levels of the 32 reviewed by computed tomography (CT) scan (68.8%). The defects were characterized as severe in 31% (7 of 22), which were often associated with graft subsidence and loss of end plate integrity (5/7, 74%) [16]. Lewandrowski et al. [10] reported a 7.4% incidence (5 of 68) after single-level TLIF using cages and rhBMP-2. Each patient reported worsening back pain with variable radicular pain as early as 4 weeks and as late as 3 months from his or her date of surgery. The defects filled in spontaneously, and symptoms typically resolved within an additional 3 months of nonoperative care [10]. A study by Vaidya et al. [17] evaluating the complications with the use of rhBMP-2 in polyetheretherketone (PEEK) cages for interbody fusions noted an 82% (41/50) rate of lumbar end plate resorption. In this study, the diagnosis and extent of resorption was based on plain radiographic findings and even minor indistinctness of end

J.A. Rihn et al. / The Spine Journal 9 (2009) 623–629

plates was considered a positive finding [17]. Most studies report that vertebral osteolysis is a self-liming phenomenon that does not affect fusion rate or clinical outcome. Most existing studies on the TLIF procedure are limited in that they contain small and/or heterogeneous groups of patients. It is important to study the complications of such a procedure that is increasingly using rhBMP-2 in an ‘‘off-label’’ fashion. The purpose of this study was to document the complications that occur after a single-level TLIF procedure performed at a single institution. A comparison of the rate and type of complications that occur with the use of iliac crest autograft versus rhBMP-2 will be made. We hypothesize that the two study groups will have similar complication rates.

Material and methods Study design After approval was obtained from the Institutional Review Board, a computerized surgical case log database search was performed to identify all patients who underwent a TLIF procedure for the treatment of a degenerative condition between January 2004 and May 2007. Patients between the ages of 18 and 80 years who underwent single-level TLIF procedure using either rhBMP-2 or iliac crest autograft were considered for inclusion in the study. Those patients who had degenerative or isthmic spondylolisthesis and/or who had prior lumbar surgery were considered eligible for the study. Those patients who underwent a multilevel TLIF procedure, who received a bone graft substitute or extender other than rhBMP-2, or who had operative treatment for nondegenerative conditions (ie, tumor, infection, or trauma) were excluded. In the autograft group, bone graft was harvested by removing both cortical bone (ie, posterior iliac crest and the superior portion of the outer table of the ileum) and the underlying cancelous bone. This was performed through a separate incision. Beginning in August 2006, Duraseal (Confluent Surgical Inc., Waltham, MA, USA) was used in patients who underwent TLIF using rhBMP-2. Duraseal is a polyethylene glycol hydrogel sealant that is FDA approved for use in the treatment of intracranial durotomy and dural tear sites. A thin layer of Duraseal was applied posterior to the interbody cage at the annulotomy site and over the exposed dura and nerve root, theoretically providing a watertight seal at the annulotomy site. The rational behind the use of Duraseal in this fashion was to prevent rhBMP-2 from leaking through the annulotomy into the spinal canal and around the nerve root. For patients that met the study inclusion criteria, a retrospective review of operative reports, pre- and postoperative clinic records, dynamic and static lumbar radiographs, and CT scans was performed. Intraoperative and postoperative complications were documented. Complication was defined as any adverse event that occurred in the intraoperative or postoperative period and that had a negative impact on

625

Table 1 Patient characteristics (n5119) Average age (y) Sex (% male) Diagnosis (%) DDD DDD/HNP RHNP IS DS Failed laminectomy and fusion Average radiographic follow-up (mo) Average clinical follow-up (mo) Previous lumbar surgery (%) TLIF level (%) L2–L3 L3–L4 L4–L5 L5–S1

47.4 52.9 10.9 12.6 27.7 32.8 15.1 0.8 19.1 27.6 37.0 1.7 6.7 47.9 43.7

DDD, degenerative disc disease; HNP, herniated nucleus pulposus; RHNP, recurrent herniated nucleus pulposus; IS, isthmic spondylolisthesis; DS, degenerative spondylolisthesis; TLIF, transforaminal lumbar interbody fusion.

patient recovery and/or clinical or radiographic outcomes. Demographic information was recorded. Patients were administered a questionnaire via a telephone interview that inquired about iliac crest donor-site pain. Fusion was evaluated using the most recent static and dynamic lumbar radiographs and, when available, lumbar CT scans. The presence of a successful fusion was assessed based on previously described radiographic criteria [3,7,18,19]. Data were analyzed using SPSS v12.0.1 (SPSS Inc., Chicago, IL, USA). Descriptive statistics were calculated, including frequencies for categorical and ordinal variables and means, standard deviations, and ranges for continuous variables. Univariate analyses were performed using Table 2 Complications Complication Persistent donor-site pain Donor-site infection Lumbar infection Lumbar hematoma Lumbar seroma Radiculitis Ectopic bone formation Vertebral osteolysis Dural tear Nonunion Malpositioned instrumentation Urinary tract infection Ileus Retained drain Number (%) of patients with complicationsa Total number (%) of complications a

Autograft, % (n533)

rhBMP-2, % (n586)

10 (30.3) 1 (3.0) 2 (6.1) 1 (3.0) 0 1 (3.0) 0 0 0 1 (3.0) 0 1 (3.0) 1 (3.0) 0

N/A N/A 3 (3.5) 1 (1.2) 1 (1.2) 12 (14.0) 2 (2.3) 5 (5.8) 4 (4.7) 3 (3.5) 2 (2.3) 2 (2.3) 1 (1.2) 1 (1.2)

15 (45.5)

25 (29.1)

18 (54.5)

37 (43.0)

Ten patients had more than one complication.

626

J.A. Rihn et al. / The Spine Journal 9 (2009) 623–629

independent t tests for the continuous outcome measures and chi-squared tests for categorical outcome measures. Statistical significance was set at a p value of !.05.

Results Patient data A total of 130 consecutive patients underwent a singlelevel TLIF procedure using either rhBMP-2 or iliac crest autograft between January 2004 and May 2007 and met the study inclusion criteria. All surgeries where performed by one of the senior authors at a single institution. Of the 130 patients, 119 patients were available for follow-up, with an average radiographic follow-up of 19.1 months (range, 10–35 months) and an average clinical follow-up of 27.6 months (range, 15–49 months). Follow-up was not obtained in 11 patients due either to inability to contact/change in contact information (n58) or refusal to participate in the study (n53). Thirty-three patients received iliac crest autograft and 86 patients received rhBMP-2. A summary of patient characteristics can be seen in Table 1. There were no significant differences between the autograft and rhBMP-2 groups regarding patient characteristics with the exception of duration of clinical follow-up (autograft 35.8 months vs. rhBMP-2 24.4 months, p!.001). Complications Overall, 40 out of the 119 study patients (33.6%) had a total of 55 complications. Ten patients had more than one complication. Although the patients in the autograft group had a higher complication rate (45.5% for the autograft group vs. 29.1% for the rhBMP-2 group), this difference was not statistically significant (p5.09). A summary of the complications can be seen in Table 2. The reoperation rate was 10.1%, with no statistically significant difference between the two groups (12.1% for the autograft group vs. 9.3% for the rhBMP-2 group, p5.65). Reasons for reoperation included lumbar wound infection (n55), iliac crest donor-site infection (n51), lumbar hematoma (n52), retained drain (n51), lumbar seroma (n51), malpositioned screw with radiculitis (n51), and ectopic bone

formation within the neuroforamen with postoperative radiculitis (n51). The most common complication in the autograft group was donor-site morbidity; 30.3% of patients reported that they still had pain in the location of the donor site at most recent follow-up. One of the 33 autograft patients (3.1%) required reoperation (ie, irrigation and debridement and primary wound closure) on the donor site 16 days after the initial procedure because of the development of an infection. The cultures were positive for Coryneform bacteria and the patient was treated successfully with 6 weeks of intravenous vancomycin. Postoperative radiculitis was defined as worsening leg pain after surgery in a dermatomal distribution. Overall, this was diagnosed in 10.9% (13/119) of the study patients, ipsilateral to the side on which the TLIF was performed. No patients developed contralateral radiculitis. The rate of postoperative radiculitis was higher in the rhBMP-2 group, with 14.0% of the rhBMP-2 patients (12/86) and 3.0% of the autograft patients (1/33) affected, but this difference did not reach significance (p5.08). Within the rhBMP-2 group, 37 (43.0%) of the patients were treated with Duraseal after the TLIF procedure. The incidence of radiculitis in patients in whom Duraseal was not used was 20.4% (10/49), compared with 5.4% (2/37) in those in whom Duraseal was used (p5.047). All patients who had postoperative radiculitis had postoperative CT scans available for review. Four patients, all in the rhBMP-2 group, had identifiable reasons for the postoperative radiculitis. Two patients were noted on CT scan to have ectopic bone formation in the neuroforamen ipsilateral to side of interbody cage placement, consistent with the dermatomal distribution of the leg pain. One of these patients underwent reoperation and decompression of the affected neuroforamen (L5–S1) leading to improvement of the radiculitis. The other patient elected not to undergo reoperation. One patient was noted to have a malpositioned pedicle screw on postoperative CT scan. This patient’s radiculitis did not improve, however, after screw removal. The fourth patient had radiculitis that was felt to be related to the development of a postoperative infection that required reoperation and long-term intravenous antibiotics. The radiculitis persisted in this patient despite successful treatment

Table 3 Lumbar infection data in transforaminal lumbar interbody fusion patients Patient

Onset

Organism

Intervention

Antibiotics course

1

4 wk

2

10 d

Irrigation and debridement 1 and anterior-posterior reconstruction Irrigation and debridement 2

6 wk intravenous ceftazidime and ciprofloxacin 6 wk intravenous vancomycin

3

15 d 5d

5

7d

Staphylococcus aureus

Irrigation and debridement 2 and anterior-posterior reconstruction Irrigation and debridement 2; vacuum-assisted closure device placement Irrigation and debridement 4

6 wk intravenous vancomycin

4

Pseudomonas (developed osteomyelitis) Escherichia coli; coagulase-negative Staphylococcus aureus Methicillin-resistant Staphylococcus aureus (developed osteomyelitis) Staphylococcus aureus

6 wk intravenous ancef 6 wk intravenous ancef

J.A. Rihn et al. / The Spine Journal 9 (2009) 623–629

627

Fig. 1. (Left) Lateral lumbar radiograph, (Center) sagittal, and (Right) coronal lumbar CT reconstructions 8 months after an L4–L5 TLIF procedure performed with rhBMP-2, which demonstrate a successful interbody fusion.

of the infection. A reason for the radiculitis could not be identified in the other nine patients. The majority of cases of radiculitis (10/13) occurred within the first 2 weeks after surgery, with one patient developing radiculitis 5 weeks after surgery (ie, patient with infection) and two patients developing radiculitis more 2 months after surgery (both with ectopic bone in neuroforamen). Five out of the 86 patients in the rhBMP-2 group (5.8%) and none of the patients in the autograft group developed vertebral osteolysis diagnosed by CT scan. The diagnosis was made between 1 month and 5 months after surgery. All five patients presented with increased low back pain. Two patients went on to fuse without complication and with resolution of back pain at 1 year after surgery. Two patients were eventually diagnosed with osteomyelitis and required revision anterior/posterior debridement and reconstruction as well as long-term intravenous antibiotics. The fifth patient with vertebral osteolysis was diagnosed with a nonunion but at the time of this study was refusing reoperation. Two autograft patients and three rhBMP-2 patients developed lumbar infections, which required reoperation(s). Details of the infections and treatments can be seen in Table 3.

Radiographic outcome A comparison of most recent follow-up radiographs between the two groups revealed a nonunion rate of 3.0% (1/33) in the autograft group and 3.5% (3/86) (p5.90) in the rhBMP-2 group. CT scans were available for review in 16 of the 119 patients, and were obtained for patient complaints of persistent back and/or leg pain. Four of these patients who underwent CT scan were diagnosed with a nonunion. The remaining 12 patients with CT scans available were determined to have a solid fusion (Fig. 1).

Discussion TLIF is a procedure that provides a circumferential fusion while avoiding the morbidity of an anterior lumbar approach. Several clinical studies have demonstrated high fusion rates and successful clinical outcomes after the TLIF procedure, both with the use of iliac crest autograft and with the use of rhBMP-2 [1–3,7–9]. It is important, however, to carefully review the complications associated with the TLIF procedure. A thorough review of the complications provides information that can be used when deciding on a treatment and when counseling patients in the preoperative and postoperative period. To date, there are few studies that provide a thorough evaluation of the complications of the TLIF procedure. Existing studies have significant variability in patient population, including the TLIF technique (eg, minimally invasive vs. open), number of levels (eg, single- and multilevel TLIF procedures), and source of bone graft or bone graft alternative/extender (eg, rhBMP-2, iliac crest autograft, local autograft). In our study, we investigated the difference in complications observed with single-level TLIF using either iliac crest autograft or rhBMP-2 at an average follow-up of 27.6 months. The most notable complications to occur included persistent donor-site pain, postoperative radiculitis, heterotopic bone formation, vertebral osteolysis, and infection. Iliac crest bone grafting is a common surgical procedure with recognized complications that may have lasting adverse effects [3–6]. Minor and major complication rates have been reported to range from 4% to 21.8% and 0% to 8.6%, respectively [5,20–22]. In our study, greater than 30% of patients in the iliac crest autograft group reported persistent donor-site pain at their most recent clinical follow-up. It is important to note, however, that the method that we used to evaluate donor-site pain (ie, a numeric rating scale from 0 to 10) may not be entirely accurate. Because of the proximity of the posterior iliac crest harvest

628

J.A. Rihn et al. / The Spine Journal 9 (2009) 623–629

site to the lumbosacral spine, it is questionable as to whether patients can reliably differentiate between donorsite and surgical-site pain. A recent study compared iliac crest donor-site pain in patients who undergo thoracic fusion and those who undergo lumbar fusion. These authors found that the thoracic fusion patients had significantly less donor-site pain, suggesting that it may be difficult for lumbar fusion patients to differentiate surgical-site pain from posterior iliac crest donor pain. For this reason, the incidence and severity of donor-site pain reported in the current study may be an overestimate [23]. Postoperative radiculitis is a well-described complication associated with the use of rhBMP-2. The cause remains unclear but is thought to be because of a BMP-related inflammatory reaction and/or heterotopic bone formation near the nerve root and intervertebral foramen. The rate of postoperative radiculitis reported in this study is greater than previous reports [2,10,11]. We found a higher rate of postoperative radiculitis in the rhBMP-2 than the autograft group. The rate of postoperative radiculitis, however, was significantly reduced when a hydrogel sealant was used (20.4% compared with 5.4%). Theoretically, Duraseal serves to seal the annulotomy through which the TLIF is performed and coat/protect the exposed nerve root and dura from the rhBMP-2. This is the first study to report that the use of Duraseal in this fashion can reduce the incidence of postoperative radiculitis. Because of the absorption of water, Duraseal does expand after its application. For this reason, there is some concern it can lead to increased neural compression with subsequent neurological deficit and/or radiculopathy. We did not see this clinically; however, as the Duraseal was actually associated with a reduction in the rate of postoperative radiculitis. Heterotopic bone formation remains a significant concern with the use of rhBMP-2 in lumbar spine surgery. Several studies evaluating the use of BMP in PLIF and TLIF have demonstrated high rates of new bone formation in the spinal canal but have not reported clinical symptoms related to the ectopic bone [3,12–15]. In our study, two patients in the rhBMP-2 group with postoperative radiculitis were noted on CT scan to have heterotopic bone formation in the neuroforamen ipsilateral to side of interbody cage placement. The mechanism of ectopic bone formation when using BMP remains unknown. It is thought to be the result, at least in part, of end plate bleeding that passes through the annulotomy into the spinal canal. This serves as a conduit for osteoprogenitor cells and rhBMP-2 [12,14,15]. Although no formal studies have been performed to support this notion, placement of the rhBMP-2 sponge and interbody cage in the anterior 1/3 of the intervertebral space, away from the annulotomy, is thought to minimize the risk of ectopic bone formation in the spinal canal and/or intervertebral foramen. Vertebral osteolysis has been documented after the use of rhBMP-2 in interbody fusion [10,16,17,24,25]. At the current time, the pathophysiology of vertebral osteolysis remains unknown, although it is suspected that violation of the end plate

during decortication may be a contributing factor [24]. Although vertebral osteolysis is well described in the literature as a self-limiting event, it must be a diagnosis of exclusion. It is important to rule out infection and follow the patient until he/she develops a solid fusion. Further basic science and clinical studies are necessary to further define the cause and clinical effects (if any) of vertebral osteolysis. This study was limited by its retrospective design and the lack of clinical follow-up on 11 patients who were determined to be eligible for this study but either did not have follow-up available or refused to participate. Additionally, we did not have postoperative CT scan data on all patients. Typically, CT scan data were only obtained if the patient was having persistent postoperative back and/or leg pain. CT was not, however, performed routinely to assess fusion. Furthermore, this study was limited to patients undergoing single-level TLIF procedures for degenerative conditions. The results of this study cannot be generalized to patients who undergo multilevel TLIF surgery or those who undergo TLIF for other pathological conditions (eg, infection). In general, this study reports a similar complication rate after the TLIF procedure whether autograft or rhBMP-2 is used. The use of rhBMP-2 eliminates the potential for donor-site morbidity, but introduces additional concerns, including postoperative radiculitis and ectopic bone formation. Further prospective randomized clinical outcome studies with long-term follow-up are needed, however, to help determine if rhBMP-2 offers significant advantages over iliac crest or even local autograft.

Conclusion The procedure, whether performed using iliac crest autograft or rhBMP-2, provides over a 95% fusion. Although the use of rhBMP-2 eliminates autograft donor-site morbidity, additional complications such as postoperative radiculitis and ectopic bone are associated with its use. Further studies are needed to determine the mechanism of these complications and methods to minimize their occurrence. References [1] Lowe TG, Tahernia AD, O’Brien MF, Smith DA. Unilateral transforaminal posterior lumbar interbody fusion (TLIF): indications, technique, and 2-year results. J Spinal Disord Tech 2002;15:31–8. [2] Potter BK, Freedman BA, Verwiebe EG, et al. Transforaminal lumbar interbody fusion: clinical and radiographic results and complications in 100 consecutive patients. J Spinal Disord Tech 2005;18:337–46. [3] Mummaneni PV, Pan J, Haid RW, Rodts GE. Contribution of recombinant human bone morphogenetic protein-2 to the rapid creation of interbody fusion when used in transforaminal lumbar interbody fusion: a preliminary report. Invited submission from the Joint Section Meeting on Disorders of the Spine and Peripheral Nerves, March 2004. [4] Fernyhough JC, Schimandle JJ, Weigel MC, et al. Chronic donor site pain complicating bone graft harvesting from the posterior iliac crest for spinal fusion. Spine 1992;17:1474–80.

J.A. Rihn et al. / The Spine Journal 9 (2009) 623–629 [5] Goulet JA, Senunas LE, DeSilva GL, Greenfield ML. Autogenous iliac crest bone graft. Complications and functional assessment. Clin Orthop Relat Res 1997;339:76–81. [6] Sawin PD, Traynelis VC, Menezes AH. A comparative analysis of fusion rates and donor-site morbidity for autogenetic rib and iliac crest bone grafts in posterior cervical fusions. J Neurosurg 1998;88:255–65. [7] Villavicencio AT, Burneikiene S, Nelson EL, et al. Safety of transforaminal lumbar interbody fusion and intervertebral recombinant human bone morphogenetic protein-2. J Neurosurg Spine 2005;3: 436–43. [8] Boden SD, Zdeblick TA, Sandu HS, et al. The use of rhBMP-2 in interbody fusion cages: definitive evidence of osteoinduction in humans: a preliminary report. Spine 2000;25:376–81. [9] Kleeman TJ, Ahn UM, Talbot-Kleeman A. Laproscopic anterior lumbar interbody fusion with rhBMP-2: a prospective study of clinical and radiographic outcomes. Spine 2001;26:2751–6. [10] Lewandrowski KU, Nanson C, Calderon R. Vertebral osteolysis after posterior interbody lumbar fusion with recombinant human bone morphogenetic protein 2: a report of five cases. Spine J 2007;7:609–14. [11] Villavicencio AT, Burneikiene S, Nelson EL, et al. Perioperative complications in transforaminal lumbar interbody fusion versus anterior-posterior reconstruction for lumbar disc degeneration and instability. J Spinal Disord Tech 2006;19:92–7. [12] Paramore CG, Lauryssen C, Rauzzino MJ, et al. The safety of OP-1 for lumbar fusion with decompressionda canine study. Neurosurgery 1999;44:1151–6. [13] Joseph V, Rampersaud YR. Heterotopic bone formation with the use of rhBMP2 in posterior minimal access interbody fusion: a CT analysis. Spine 2007;32:2885–90. [14] Canady JW, Zeitler DP, Thompson SA, Nicholas CD. Suitability of the iliac crest as a site for harvest of autogenous bone grafts. Cleft Palate J 1993;30:579–81.

629

[15] Schwender JD, Holly LT, Rouben DP, Foley KT. Minimally invasive transforaminal lumbar interbody fusion (TLIF): technical feasibility and initial results. J Spinal Disord Tech 2005;18(Suppl):S1–6. [16] McClellan JW, Mulconrey DS, Forbes RJ, Fullmer N. Vertebral bone resorption after transforaminal lumbar interbody fusion with bone morphogenetic protein (rhBMP-2). J Spinal Disord Tech 2006;19: 483–6. [17] Vaidya R, Sethi A, Bartol S, et al. Complications in the use of rhBMP-2 in PEEK cages for interbody spinal fusions. J Spinal Disord Tech 2008;21:557–62. [18] Boden SD, Kang J, Sandhu H, Heller JG. Use of recombinant human bone morphogenetic protein-2 to achieve posterolateral lumbar spine fusion in humans: a prospective, randomized clinical pilot trial: 2002 Volvo Award in clinical studies. Spine 2002;27:2662–73. [19] Burkus JK, Dorchak JD, Sanders DL. Radiographic assessment of interbody fusion using recombinant human bone morphogenetic protein type 2. Spine 2003;28:372–7. [20] Cockin J. Autologous bone graftingdcomplications at the donor site. J Bone Joint Surg 1971;53B:153. [21] Younger EM, Chapman MW. Morbidity at bone graft donor sites. J Orthop Trauma 1989;3:192–5. [22] Carreon L, Glassman SD, Brock D, Dimar J, et al. Adverse events in patients re-exposed to bone morphogenetic protein for spine surgery. Spine 2008;33:391–3. [23] Delawi D, Dhert WJ, Castelein RM, et al. 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 2007;32: 1865–8. [24] Burkus JK, Sandhu HS, Gornet MF. Influence of rhBMP-2 on the healing patterns associated with allograft interbody constructs in comparison with autograft. Spine 2006;31:775–81. [25] Davne SH, Myers DL. Complications of lumbar spinal fusion with transpedicular instrumentation. Spine 1992;17:S184–9.