- Email: [email protected]
Vertebral osteolysis after posterior interbody lumbar fusion with recombinant human bone morphogenetic protein 2: A report of five cases
The Spine Journal 7 (2007) 609–614
Case Reports
Vertebral osteolysis after posterior interbody lumbar fusion with recombinant human bone morphogenetic protein 2: A report of five cases Kai-Uwe Lewandrowski, MDa,*, Christopher Nanson, MDb, Robert Calderon, MDb a Center for Advanced Spinal Surgery, 3395 N Campbell Avenue, Tucson, AZ 85719, USA Department of Orthopaedic Surgery/AHSC, University of Arizona, PO Box 245064, Tucson, AZ 85724-5064, USA
b
Received 27 September 2006; accepted 23 February 2007
Abstract
BACKGROUND: Recombinant human bone morphogenetic protein 2 (rh-BMP-2) is frequently used in an off-label fashion. Its application for posterior interbody fusion appears intuitive because its use obviates the need for iliac crest bone graft and shows higher fusion rates than with the use of local autologous bone graft. To date, there is no report of adverse outcomes with such use of rh-BMP-2. PURPOSE: To draw attention to this unusual complication of posterior interbody lumbar fusion and to review the relevant literature. STUDY DESIGN: Clinical report of five cases of vertebral osteolysis that developed postoperatively from lumbar transforaminal interbody fusion of the L5/S1 motion segment using cages and rh-BMP-2. METHODS: Sixty-eight patients underwent transforaminal lumbar interbody fusion for spondylolisthesis or degenerative disc disease with discogenic back pain. Five of these 68 patients developed vertebral osteolysis within 4 months from their surgery. Their clinical presentation and radiographic findings are presented in this case series. RESULTS: Each one of these five patients had uneventful surgery and postoperative recovery. Their back and leg pain improved in the immediate postoperative period. However, each patient reported worsening back pain with variable radicular pain as early as 4 weeks and as late as 3 months after the index procedure. Diagnostic workup revealed evidence of vertebral osteolysis typically involving the L5 vertebral body. In all five patients, osteolytic defects filled in spontaneously, and symptoms typically resolved within an additional 3 months of nonoperative care. CONCLUSIONS: Vertebral osteolysis can occur with the use of rh-BMP-2 in posterior lumbar interbody fusions. Violation of the end plate during decortication may be a contributing factor. Symptoms often resolve spontaneously. Ó 2007 Elsevier Inc. All rights reserved.
Keywords:
Bone morphogenetic protein; Interbody fusion; Bone resorption
Introduction Recombinant human bone morphogenetic protein 2 (rhBMP-2) supersedes the use of iliac crest autograft bone because it induces the body to grow its own bone [1–4]. In 2000, the Food and Drug Administration approved the use of rh-BMP-2 for anterior interbody fusion when used FDA device/drug status: not applicable. Nothing of value received from a commercial entity related to this manuscript. * Corresponding author. Southwest Orthopaedic Surgery Specialists, P.L.C., Center for Advanced Spinal Surgery, 3395 N Campbell Avenue, Tucson, AZ 85719. Tel.: (520) 327-9677; fax:E-mail address: [email protected] (K.-U. Lewandrowski) 1529-9430/07/$ – see front matter Ó 2007 Elsevier Inc. All rights reserved. doi:10.1016/j.spinee.2007.01.011
in conjunction with the Lumbar Tapered Fusion Device; LT cage (Sofamor Danek) [5]. Since then, clinical studies have indicated equivalent interbody fusion rates between rh-BMP-2 and iliac crest bone graft [4,6]. As a result, rhBMP-2 has found a number of ‘‘off-label’’ applications, one of which includes its use for posterior interbody lumbar fusions [7]. The BMPs belong to the transforming growth factor super family [8]. They are involved in development and differentiation of skeletal tissues, as well as the brain, spinal cord, liver, kidneys, skeletal muscle, eyes, and epithelium [3,9]. Physiologic concentrations of BMP-2 are quite low. Although sufficient for normal fracture healing, only 0.002 mg of BMP-2 can be extracted from 1 kg of
610
K.-U. Lewandrowski et al. / The Spine Journal 7 (2007) 609–614
Table 1 #
Sex
Age
Preoperative diagnosis
Procedure
Previous surgery
1 2 3 4 5
F M M F M
50 50 50 50 51
L5/S1 L5/S1 L5/S1 L5/S1 L5/S1
L3–SI TLIF L5/S1 TLIF L5/S1 TLIF L5/S1 TLIF L5/S1 TLIF
L3–L5 PSF, L4/5 PLIF
adjacent level disease Spondylolisthesis Spondylosis Spondylosis Spondyolisthesis
TLIF 5 transforaminal lumbar interbody fusion; PLIF 5 posterior lumbar interbody fusion.
normal powdered bone [10]. To achieve demonstrable enhancement of fracture healing, much higher BMP-2 concentrations ranging from 0.01 mg/mL in rodents to 1.5 mg/ mL in nonhuman primate models are necessary. The commercially available form of rh-BMP-2, Infuse, contains 1.5 mg/mL of reconstituted collagen sponge [9,11]. More recently, even higher numbers have been suggested for posterolateral fusions. Boden et al. [12] found successful posterolateral lumbar spinal fusion with or without the use of internal fixation when rh-BMP-2 was delivered at a dose of 20 mg per side. It has been suggested that only these ‘‘.supraphysiologic concentrations.’’ are in fact capable of inducing the desired clinical effect by overcoming the tight regulation of BMP and its inhibitors [8]. Reports on complications related to the application of rh-BMP-2 in posterior lumbar interbody fusions are now emerging. Therefore, we describe five cases of vertebral osteolysis after transforaminal interbody fusion with rh-BMP2 (Infuse; Medtronic Sofamor Danek) used in conjunction with local bone graft and a single interbody fusion cage.
Materials and methods Study design and patients From November 2004 to March 2006, 68 patients consisting of 43 females and 25 males underwent minimally invasive lumbar fusions (TLIF) through 1-inch paraspinal incisions using the Wiltse approach. Surgical indications included spondylolisthesis (48 patients), degenerative disc disease resulting in discogenic low back pain (18 patients), recurrent herniated nucleus pulposus (1 patient), and adjacent level disease (1 patient). The procedures were performed by the senior author. Five of the 68 patients showed signs of vertebral osteolysis postoperatively. We retrospectively analyzed the operative reports, X-ray films, and medical records of these 5 consecutive patients. Instrumentation Sixty-three patients had bilateral posterior pedicle screw-rod instrumentation with Pathfinder titanium implants (Abbott Spine). An additional five patients had unilateral posterior pedicle screw-rod instrumentation with Pathfinder titanium implants (Abbott Spine) and with ipsilateral percutaneous placement of a translaminar facet
screws (AO Synthes 4.0-and 4.5-mm cannulated partially threaded screws). Two types of interbody fusion cages were used in conjunction with local bone graft and recombinant bone morphogenic protein 2 (Infuse): Traxis cage (Abbott Spine) made from polyetheretherketone (PEEK) and Saber cage (Depuy Spine) made from carbon fiber. Surgical technique Bilateral or unilateral Wiltse-type muscle-splitting approaches were used to perform the TLIF and to place the pedicle screws. In general, the TLIF was performed through the side on which the patients’ radicular symptoms were more severe. This involved removal of the entire facet joint, discectomy, and end plate preparation. Before inserting the interbody fusion cage, approximately 1.5 to 2 mL of morselized local bone graft and one Infuse collagen sponge were placed into the interspace. This was followed by insertion of one interbody fusion cage, which was filled with an additional Infuse collagen sponge and local bone graft. Cage position was checked under biplanar fluoroscopy to ensure anterior position across the midline. Then, Gel foam was injected into the disc interspace and the facetectomy site to minimize postoperative blood loss and leakage of rh-BMP-2 into the neuroforamen. It was removed by suction before wound closure. In this entire series of 68 patients, only small and medium kits with 2 and 4 absorbable collagen sponges (ACSs; 1’’ x 2’’) were used. These kits come with a vial 4.9 mg of rh-BMP-2. According to the manufacturer, only 4.2 mg of the 4.9 mg of rh-BMP-2 is applied to the ACS after reconstitution. Postoperative care Postoperatively, mobilization began as soon as tolerated by the patient either immediately postoperatively or on postoperative day 1 with formal gait training by the physical therapist. Most patients were discharged to their home on postoperative day 1. None of the patients received any lumbar supports or braces. Low-impact exercising, walking to tolerance, and swimming were encouraged immediately on the patients’ discharge to home. Heavy lifting and any other type of strenuous activity were discouraged for at least 6 weeks postoperatively. Patients were otherwise permitted to return to work as soon as they could tolerate it. All patients were counseled to avoid over-the-counter
K.-U. Lewandrowski et al. / The Spine Journal 7 (2007) 609–614
nonsteroidal anti-inflammatory medications during the first 6 weeks postoperatively.
Case presentation Five of the 68 patients who underwent posterior lumbar interbody fusion via the transforaminal approach developed osteolysis of the L5 vertebral body. The average age of these 5 patients was 50.2 years. They consisted of three men and two women. Four of the five patients underwent L5/S1 TLIF for single-level disease. A representative case is shown in Figure 1. These four patients had PEEK cages and two 1’’ x 2’’ absorbable collagen sponges from a small Infuse kit placed into the intervertebral disc space. These two ACSs were reconstituted with 4.2 mg rh-BMP-2 of the 4.9-mg vial. The remaining patient had revision surgery for adjacent level disease at L5/S1 and nonunion at L3/4 after a previous L3–L5 posterior spinal fusion with an interbody fusion cage at L4/5. She underwent removal of implants and reinstrumentation with placement of PEEK interbody fusion cages at L3/4 and L5/S1 via the TLIF approach (Fig. 2). This patient had a total of four 1’’ x
611
2’’ absorbable collagen sponges from a medium Infuse kit placed into the intervertebral disc space. These 4 ACSs were reconstituted with 4.2 mg rh-BMP-2 of the 4.9-mg vial. All five patients had degenerative disc disease with significant loss of height, osteophyte formation, and sclerosis of the end plates. To facilitate distraction for placement of a sizable interbody fusion cage, distracting paddle shavers were used to mobilize the L5/S1 interspace. The paddle shavers were placed as far as possible anteriorly to achieve distraction of the anterior vertebral body rim. This was done in an effort to minimize violation of the end plates. In each of the five patients, the height of the interbody fusion cage was no more than 9 mm. Each cage was trialed before insertion of the final implant. Each of these 5 patients did well initially until presenting with new onset severe low back pain. Their symptoms appeared anywhere between 4 weeks and 3 months postoperatively. None of the five patients showed any significant radicular symptoms. Typically, symptoms resolved with supportive nonoperative care within 3 months from their onset. A computed tomography (CT) scan was obtained. To our surprise, resorption of the inferior aspect of the L5
Figure 1. a-e Anterior-posterior and lateral radiographs of a 50-year-old women (Table 1, case 4) who underwent L5-S1 TLIF (a and b). Similar xrays taken 3 months after the index procedure were suggestive of osteolysis of in the L5 vertebral body (c and d). A CT scan confirmed osteolysis around the interbody fusion cage involving the inferior part of the L5 vertebral body (e and f). At 6 months postoperatively, her symptoms had resolved.
612
K.-U. Lewandrowski et al. / The Spine Journal 7 (2007) 609–614
Figure 2. a-d CT scans of the lumbar spine of a 50 year-old women following L3 - S1 TLIF with placement of rh-BMP2 and an interbody fusion cage at L5/ S1 (Table 1, case 1). The patient presented with severe low back pain 10 weeks after her surgery at which time saggital (a) and coronal (b) CT scans confirmed osteolysis. Similar scans obtained 6 months postoperatively showed at least partial reconstitution of the osteolytic defects. This patient’s symptoms eventually resolved without further treatment.
vertebral body had occurred in each of these five patients. There was one medially placed S1 pedicle screw in one patient. However, this patient did not have any evidence of S1 radiculopathy, and his acute low back pain resolved within an additional 2 months from its onset. Six months after his acute episode had resolved, the patient developed symptoms related to prominent spinal instrumentation with increasing activity. He improved with several trigger point injections and physical therapy. He underwent removal of the pedicle screws 9 months postoperatively. Histological microsection of a transpedicular biopsy of the osteolytic area showed a benign-appearing trabecular bone with small fragments of granulation tissue without evidence of osteomyelitis (Fig. 3).
Discussion A recombinant bone morphogenetic protein 2 is now widely used for posterior interbody fusions. Its use seems particularly advantageous when used in conjunction with minimally invasive access techniques because it obviates the need for a separate incision for iliac crest bone harvest
and its well-recognized morbidities including persistent pain, prolonged operating room time, and increased blood loss. Reports of high fusion rates reported with the use of rhBMP-2 in anterior lumbar interbody fusion have led to the expansion of its clinical use beyond the Food and Drug Administration–approved application for anterior lumbar interbody fusion with the LT cage [5]. Recent clinical evidence suggests that application of rhBMP-2 in posterior interbody fusion is as effective as TLIF procedures performed with the use of iliac crest autograft [7]. This case series of 5 patients with vertebral osteolysis after posterior interbody fusion with rhBMP-2 and PEEK cages is unique in several ways. First, osteolysis occurred only at the L5/S1 motion segment. Second, patients with vertebral osteolysis became symptomatic between 6 and 12 weeks postoperatively prompting workup of their acute low back pain. However, its clinical presentation was variable in duration and length of time from surgery, making assessment of its true prevalence difficult. In fact, other patients in our original series of 68 may have had vertebral osteolysis without clinical symptoms. Third, all 5 patients showed complete resolution of their symptoms within 12
K.-U. Lewandrowski et al. / The Spine Journal 7 (2007) 609–614
613
Figure 3. Histolopathologic photomicrograph (4 ) of a transpedicular biopsy of an osteolytic lesion of L5 obtained during routine hardware removal. The specimen was taken from a 50-year-old who previously underwent L5 - S1 TLIF with placement of rh-BMP2 and an interbody fusion cage (Table 1, case 1). The lesion had healed with formation of trabecular bone (B). As shown in the inset (10 ), granulation tissue (G) was found in close proximity to the trabecular bone.
weeks. As shown in Figure 2, follow-up computed tomography (CT) studies appeared to show at least partial reconstitution of the cavernous defects. Although there is no conclusive evidence to suggest any definitive cause for the development of vertebral osteolysis after a posterior interbody fusion through the TLIF approach, three possible explanations appear reasonable. First, violation of the end plates with paddle shavers and dilators during the interbody fusion may have occurred. This seems more probable at the L5/S1 level than at any other lumbar level because of the higher inclination of the interspace. Decortication through the subchondral bone of the end plate may have exposed bleeding cancellous bone from within the vertebral body to the rhBMP-2. Results of the histopathological evaluation of the transpedicular biopsy in one patient showed trabecular bone and granulation tissue without evidence of infection, suggesting the presence of an inflammatory process in the healing osteolytic defect. Second, ‘‘overstuffing’’ of the interspace may have contributed to vertebral osteolysis as well. Although it has been suggested that higher doses of rhBMP-2 are required for
posterolateral intertransverse fusions, it seems entirely possible that smaller doses than the 1.5 mg/mL of reconstituted collagen sponge contained in Infuse are sufficient to induce successful interbody fusions. This seems of particular importance if the collagen sponge containing rhBMP-2 is not only placed into an interbody fusion device, in which it is somewhat contained, but also into the intervertebral space. All of our five patients had local bone graft and one sponge of Infuse placed into the interspace before the insertion of the interbody fusion device. By using this ‘‘soft’’ interbody fusion technique, bony bridging between adjacent end plates may be observed sooner than through interbody fusion devices. Finally, dose-dependent cellular cascade activation of osteoclasts over osteoblasts may have occurred because of rh-BMP-2–induced inflammatory effects recently discussed in the pathogenesis of its adverse effects in anterior spinal fusion. This case series shows that vertebral osteolysis can occur after posterior lumbar interbody fusions with rh-BMP-2. Violation of the endplate during decortication
614
K.-U. Lewandrowski et al. / The Spine Journal 7 (2007) 609–614
may be a contributing factor. Patients often improve without operative intervention, and symptoms often resolve spontaneously. References [1] Khan SN, Lane JM. The use of recombinant human bone morphogenetic protein-2 (rhBMP-2) in orthopaedic applications. Expert Opin Biol Ther 2004;4:741–8. [2] Riley EH, Lane JM, Urist MR, Lyons KM, Lieberman JR. Bone morphogenetic protein-2: biology and applications. Clin Orthop Relat Res 1996;329:39–46. [3] Sandhu H. Spinal fusion using bone morphogenetic proteins. Orthopedics 2004;27:717–8. [4] Szpalski M, Gunzburg R. Recombinant human bone morphogenetic protein-2: a novel osteoinductive alternative to autogenous bone graft? Acta Orthop Belg 2005;71:133–48. [5] BMP 2–Genetics Institute/ Medtronic-Sofamor Danek/Integra. Bone morphogenetic protein 2–Genetics Institute/ Medtronic-Sofamor Danek/ Integra, INFUSE Bone Graft, recombinant human bone morphogenetic protein 2–Genetics Institute/Medtronic-Sofamor Danek/Integra, RhBMP 2–Genetics Institute/Medtronic-Sofamor Danek/Integra. BioDrugs 2002;16:376–7.
[6] Walker DH, Wright NM. Bone morphogenetic proteins and spinal fusion. Neurosurg Focus 2002;13:e3. [7] 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. J Neurosurg Spine 2004;1:19–23. [8] Chen D, Zhao M, Mundy GR. Bone morphogenetic proteins. Growth Factors 2004;22:233–41. [9] Zlotolow DA, Vaccaro AR, Salamon ML, Albert TJ. The role of human bone morphogenetic proteins in spinal fusion. J Am Acad Orthop Surg 2000;8:3–9. [10] Granjeiro JM, Oliveira RC, Bustos-Valenzuela JC, Sogayar MC, Taga R. Bone morphogenetic proteins: from structure to clinical use. Braz J Med Biol Res 2005;38:1463–73. [11] Sandhu HS, Khan SN, Suh DY, Boden SD. Demineralized bone matrix, bone morphogenetic proteins, and animal models of spine fusion: an overview. Eur Spine J 2001;10(suppl 2):S122–31. [12] 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.
Case Reports
Vertebral osteolysis after posterior interbody lumbar fusion with recombinant human bone morphogenetic protein 2: A report of five cases Kai-Uwe Lewandrowski, MDa,*, Christopher Nanson, MDb, Robert Calderon, MDb a Center for Advanced Spinal Surgery, 3395 N Campbell Avenue, Tucson, AZ 85719, USA Department of Orthopaedic Surgery/AHSC, University of Arizona, PO Box 245064, Tucson, AZ 85724-5064, USA
b
Received 27 September 2006; accepted 23 February 2007
Abstract
BACKGROUND: Recombinant human bone morphogenetic protein 2 (rh-BMP-2) is frequently used in an off-label fashion. Its application for posterior interbody fusion appears intuitive because its use obviates the need for iliac crest bone graft and shows higher fusion rates than with the use of local autologous bone graft. To date, there is no report of adverse outcomes with such use of rh-BMP-2. PURPOSE: To draw attention to this unusual complication of posterior interbody lumbar fusion and to review the relevant literature. STUDY DESIGN: Clinical report of five cases of vertebral osteolysis that developed postoperatively from lumbar transforaminal interbody fusion of the L5/S1 motion segment using cages and rh-BMP-2. METHODS: Sixty-eight patients underwent transforaminal lumbar interbody fusion for spondylolisthesis or degenerative disc disease with discogenic back pain. Five of these 68 patients developed vertebral osteolysis within 4 months from their surgery. Their clinical presentation and radiographic findings are presented in this case series. RESULTS: Each one of these five patients had uneventful surgery and postoperative recovery. Their back and leg pain improved in the immediate postoperative period. However, each patient reported worsening back pain with variable radicular pain as early as 4 weeks and as late as 3 months after the index procedure. Diagnostic workup revealed evidence of vertebral osteolysis typically involving the L5 vertebral body. In all five patients, osteolytic defects filled in spontaneously, and symptoms typically resolved within an additional 3 months of nonoperative care. CONCLUSIONS: Vertebral osteolysis can occur with the use of rh-BMP-2 in posterior lumbar interbody fusions. Violation of the end plate during decortication may be a contributing factor. Symptoms often resolve spontaneously. Ó 2007 Elsevier Inc. All rights reserved.
Keywords:
Bone morphogenetic protein; Interbody fusion; Bone resorption
Introduction Recombinant human bone morphogenetic protein 2 (rhBMP-2) supersedes the use of iliac crest autograft bone because it induces the body to grow its own bone [1–4]. In 2000, the Food and Drug Administration approved the use of rh-BMP-2 for anterior interbody fusion when used FDA device/drug status: not applicable. Nothing of value received from a commercial entity related to this manuscript. * Corresponding author. Southwest Orthopaedic Surgery Specialists, P.L.C., Center for Advanced Spinal Surgery, 3395 N Campbell Avenue, Tucson, AZ 85719. Tel.: (520) 327-9677; fax:E-mail address: [email protected] (K.-U. Lewandrowski) 1529-9430/07/$ – see front matter Ó 2007 Elsevier Inc. All rights reserved. doi:10.1016/j.spinee.2007.01.011
in conjunction with the Lumbar Tapered Fusion Device; LT cage (Sofamor Danek) [5]. Since then, clinical studies have indicated equivalent interbody fusion rates between rh-BMP-2 and iliac crest bone graft [4,6]. As a result, rhBMP-2 has found a number of ‘‘off-label’’ applications, one of which includes its use for posterior interbody lumbar fusions [7]. The BMPs belong to the transforming growth factor super family [8]. They are involved in development and differentiation of skeletal tissues, as well as the brain, spinal cord, liver, kidneys, skeletal muscle, eyes, and epithelium [3,9]. Physiologic concentrations of BMP-2 are quite low. Although sufficient for normal fracture healing, only 0.002 mg of BMP-2 can be extracted from 1 kg of
610
K.-U. Lewandrowski et al. / The Spine Journal 7 (2007) 609–614
Table 1 #
Sex
Age
Preoperative diagnosis
Procedure
Previous surgery
1 2 3 4 5
F M M F M
50 50 50 50 51
L5/S1 L5/S1 L5/S1 L5/S1 L5/S1
L3–SI TLIF L5/S1 TLIF L5/S1 TLIF L5/S1 TLIF L5/S1 TLIF
L3–L5 PSF, L4/5 PLIF
adjacent level disease Spondylolisthesis Spondylosis Spondylosis Spondyolisthesis
TLIF 5 transforaminal lumbar interbody fusion; PLIF 5 posterior lumbar interbody fusion.
normal powdered bone [10]. To achieve demonstrable enhancement of fracture healing, much higher BMP-2 concentrations ranging from 0.01 mg/mL in rodents to 1.5 mg/ mL in nonhuman primate models are necessary. The commercially available form of rh-BMP-2, Infuse, contains 1.5 mg/mL of reconstituted collagen sponge [9,11]. More recently, even higher numbers have been suggested for posterolateral fusions. Boden et al. [12] found successful posterolateral lumbar spinal fusion with or without the use of internal fixation when rh-BMP-2 was delivered at a dose of 20 mg per side. It has been suggested that only these ‘‘.supraphysiologic concentrations.’’ are in fact capable of inducing the desired clinical effect by overcoming the tight regulation of BMP and its inhibitors [8]. Reports on complications related to the application of rh-BMP-2 in posterior lumbar interbody fusions are now emerging. Therefore, we describe five cases of vertebral osteolysis after transforaminal interbody fusion with rh-BMP2 (Infuse; Medtronic Sofamor Danek) used in conjunction with local bone graft and a single interbody fusion cage.
Materials and methods Study design and patients From November 2004 to March 2006, 68 patients consisting of 43 females and 25 males underwent minimally invasive lumbar fusions (TLIF) through 1-inch paraspinal incisions using the Wiltse approach. Surgical indications included spondylolisthesis (48 patients), degenerative disc disease resulting in discogenic low back pain (18 patients), recurrent herniated nucleus pulposus (1 patient), and adjacent level disease (1 patient). The procedures were performed by the senior author. Five of the 68 patients showed signs of vertebral osteolysis postoperatively. We retrospectively analyzed the operative reports, X-ray films, and medical records of these 5 consecutive patients. Instrumentation Sixty-three patients had bilateral posterior pedicle screw-rod instrumentation with Pathfinder titanium implants (Abbott Spine). An additional five patients had unilateral posterior pedicle screw-rod instrumentation with Pathfinder titanium implants (Abbott Spine) and with ipsilateral percutaneous placement of a translaminar facet
screws (AO Synthes 4.0-and 4.5-mm cannulated partially threaded screws). Two types of interbody fusion cages were used in conjunction with local bone graft and recombinant bone morphogenic protein 2 (Infuse): Traxis cage (Abbott Spine) made from polyetheretherketone (PEEK) and Saber cage (Depuy Spine) made from carbon fiber. Surgical technique Bilateral or unilateral Wiltse-type muscle-splitting approaches were used to perform the TLIF and to place the pedicle screws. In general, the TLIF was performed through the side on which the patients’ radicular symptoms were more severe. This involved removal of the entire facet joint, discectomy, and end plate preparation. Before inserting the interbody fusion cage, approximately 1.5 to 2 mL of morselized local bone graft and one Infuse collagen sponge were placed into the interspace. This was followed by insertion of one interbody fusion cage, which was filled with an additional Infuse collagen sponge and local bone graft. Cage position was checked under biplanar fluoroscopy to ensure anterior position across the midline. Then, Gel foam was injected into the disc interspace and the facetectomy site to minimize postoperative blood loss and leakage of rh-BMP-2 into the neuroforamen. It was removed by suction before wound closure. In this entire series of 68 patients, only small and medium kits with 2 and 4 absorbable collagen sponges (ACSs; 1’’ x 2’’) were used. These kits come with a vial 4.9 mg of rh-BMP-2. According to the manufacturer, only 4.2 mg of the 4.9 mg of rh-BMP-2 is applied to the ACS after reconstitution. Postoperative care Postoperatively, mobilization began as soon as tolerated by the patient either immediately postoperatively or on postoperative day 1 with formal gait training by the physical therapist. Most patients were discharged to their home on postoperative day 1. None of the patients received any lumbar supports or braces. Low-impact exercising, walking to tolerance, and swimming were encouraged immediately on the patients’ discharge to home. Heavy lifting and any other type of strenuous activity were discouraged for at least 6 weeks postoperatively. Patients were otherwise permitted to return to work as soon as they could tolerate it. All patients were counseled to avoid over-the-counter
K.-U. Lewandrowski et al. / The Spine Journal 7 (2007) 609–614
nonsteroidal anti-inflammatory medications during the first 6 weeks postoperatively.
Case presentation Five of the 68 patients who underwent posterior lumbar interbody fusion via the transforaminal approach developed osteolysis of the L5 vertebral body. The average age of these 5 patients was 50.2 years. They consisted of three men and two women. Four of the five patients underwent L5/S1 TLIF for single-level disease. A representative case is shown in Figure 1. These four patients had PEEK cages and two 1’’ x 2’’ absorbable collagen sponges from a small Infuse kit placed into the intervertebral disc space. These two ACSs were reconstituted with 4.2 mg rh-BMP-2 of the 4.9-mg vial. The remaining patient had revision surgery for adjacent level disease at L5/S1 and nonunion at L3/4 after a previous L3–L5 posterior spinal fusion with an interbody fusion cage at L4/5. She underwent removal of implants and reinstrumentation with placement of PEEK interbody fusion cages at L3/4 and L5/S1 via the TLIF approach (Fig. 2). This patient had a total of four 1’’ x
611
2’’ absorbable collagen sponges from a medium Infuse kit placed into the intervertebral disc space. These 4 ACSs were reconstituted with 4.2 mg rh-BMP-2 of the 4.9-mg vial. All five patients had degenerative disc disease with significant loss of height, osteophyte formation, and sclerosis of the end plates. To facilitate distraction for placement of a sizable interbody fusion cage, distracting paddle shavers were used to mobilize the L5/S1 interspace. The paddle shavers were placed as far as possible anteriorly to achieve distraction of the anterior vertebral body rim. This was done in an effort to minimize violation of the end plates. In each of the five patients, the height of the interbody fusion cage was no more than 9 mm. Each cage was trialed before insertion of the final implant. Each of these 5 patients did well initially until presenting with new onset severe low back pain. Their symptoms appeared anywhere between 4 weeks and 3 months postoperatively. None of the five patients showed any significant radicular symptoms. Typically, symptoms resolved with supportive nonoperative care within 3 months from their onset. A computed tomography (CT) scan was obtained. To our surprise, resorption of the inferior aspect of the L5
Figure 1. a-e Anterior-posterior and lateral radiographs of a 50-year-old women (Table 1, case 4) who underwent L5-S1 TLIF (a and b). Similar xrays taken 3 months after the index procedure were suggestive of osteolysis of in the L5 vertebral body (c and d). A CT scan confirmed osteolysis around the interbody fusion cage involving the inferior part of the L5 vertebral body (e and f). At 6 months postoperatively, her symptoms had resolved.
612
K.-U. Lewandrowski et al. / The Spine Journal 7 (2007) 609–614
Figure 2. a-d CT scans of the lumbar spine of a 50 year-old women following L3 - S1 TLIF with placement of rh-BMP2 and an interbody fusion cage at L5/ S1 (Table 1, case 1). The patient presented with severe low back pain 10 weeks after her surgery at which time saggital (a) and coronal (b) CT scans confirmed osteolysis. Similar scans obtained 6 months postoperatively showed at least partial reconstitution of the osteolytic defects. This patient’s symptoms eventually resolved without further treatment.
vertebral body had occurred in each of these five patients. There was one medially placed S1 pedicle screw in one patient. However, this patient did not have any evidence of S1 radiculopathy, and his acute low back pain resolved within an additional 2 months from its onset. Six months after his acute episode had resolved, the patient developed symptoms related to prominent spinal instrumentation with increasing activity. He improved with several trigger point injections and physical therapy. He underwent removal of the pedicle screws 9 months postoperatively. Histological microsection of a transpedicular biopsy of the osteolytic area showed a benign-appearing trabecular bone with small fragments of granulation tissue without evidence of osteomyelitis (Fig. 3).
Discussion A recombinant bone morphogenetic protein 2 is now widely used for posterior interbody fusions. Its use seems particularly advantageous when used in conjunction with minimally invasive access techniques because it obviates the need for a separate incision for iliac crest bone harvest
and its well-recognized morbidities including persistent pain, prolonged operating room time, and increased blood loss. Reports of high fusion rates reported with the use of rhBMP-2 in anterior lumbar interbody fusion have led to the expansion of its clinical use beyond the Food and Drug Administration–approved application for anterior lumbar interbody fusion with the LT cage [5]. Recent clinical evidence suggests that application of rhBMP-2 in posterior interbody fusion is as effective as TLIF procedures performed with the use of iliac crest autograft [7]. This case series of 5 patients with vertebral osteolysis after posterior interbody fusion with rhBMP-2 and PEEK cages is unique in several ways. First, osteolysis occurred only at the L5/S1 motion segment. Second, patients with vertebral osteolysis became symptomatic between 6 and 12 weeks postoperatively prompting workup of their acute low back pain. However, its clinical presentation was variable in duration and length of time from surgery, making assessment of its true prevalence difficult. In fact, other patients in our original series of 68 may have had vertebral osteolysis without clinical symptoms. Third, all 5 patients showed complete resolution of their symptoms within 12
K.-U. Lewandrowski et al. / The Spine Journal 7 (2007) 609–614
613
Figure 3. Histolopathologic photomicrograph (4 ) of a transpedicular biopsy of an osteolytic lesion of L5 obtained during routine hardware removal. The specimen was taken from a 50-year-old who previously underwent L5 - S1 TLIF with placement of rh-BMP2 and an interbody fusion cage (Table 1, case 1). The lesion had healed with formation of trabecular bone (B). As shown in the inset (10 ), granulation tissue (G) was found in close proximity to the trabecular bone.
weeks. As shown in Figure 2, follow-up computed tomography (CT) studies appeared to show at least partial reconstitution of the cavernous defects. Although there is no conclusive evidence to suggest any definitive cause for the development of vertebral osteolysis after a posterior interbody fusion through the TLIF approach, three possible explanations appear reasonable. First, violation of the end plates with paddle shavers and dilators during the interbody fusion may have occurred. This seems more probable at the L5/S1 level than at any other lumbar level because of the higher inclination of the interspace. Decortication through the subchondral bone of the end plate may have exposed bleeding cancellous bone from within the vertebral body to the rhBMP-2. Results of the histopathological evaluation of the transpedicular biopsy in one patient showed trabecular bone and granulation tissue without evidence of infection, suggesting the presence of an inflammatory process in the healing osteolytic defect. Second, ‘‘overstuffing’’ of the interspace may have contributed to vertebral osteolysis as well. Although it has been suggested that higher doses of rhBMP-2 are required for
posterolateral intertransverse fusions, it seems entirely possible that smaller doses than the 1.5 mg/mL of reconstituted collagen sponge contained in Infuse are sufficient to induce successful interbody fusions. This seems of particular importance if the collagen sponge containing rhBMP-2 is not only placed into an interbody fusion device, in which it is somewhat contained, but also into the intervertebral space. All of our five patients had local bone graft and one sponge of Infuse placed into the interspace before the insertion of the interbody fusion device. By using this ‘‘soft’’ interbody fusion technique, bony bridging between adjacent end plates may be observed sooner than through interbody fusion devices. Finally, dose-dependent cellular cascade activation of osteoclasts over osteoblasts may have occurred because of rh-BMP-2–induced inflammatory effects recently discussed in the pathogenesis of its adverse effects in anterior spinal fusion. This case series shows that vertebral osteolysis can occur after posterior lumbar interbody fusions with rh-BMP-2. Violation of the endplate during decortication
614
K.-U. Lewandrowski et al. / The Spine Journal 7 (2007) 609–614
may be a contributing factor. Patients often improve without operative intervention, and symptoms often resolve spontaneously. References [1] Khan SN, Lane JM. The use of recombinant human bone morphogenetic protein-2 (rhBMP-2) in orthopaedic applications. Expert Opin Biol Ther 2004;4:741–8. [2] Riley EH, Lane JM, Urist MR, Lyons KM, Lieberman JR. Bone morphogenetic protein-2: biology and applications. Clin Orthop Relat Res 1996;329:39–46. [3] Sandhu H. Spinal fusion using bone morphogenetic proteins. Orthopedics 2004;27:717–8. [4] Szpalski M, Gunzburg R. Recombinant human bone morphogenetic protein-2: a novel osteoinductive alternative to autogenous bone graft? Acta Orthop Belg 2005;71:133–48. [5] BMP 2–Genetics Institute/ Medtronic-Sofamor Danek/Integra. Bone morphogenetic protein 2–Genetics Institute/ Medtronic-Sofamor Danek/ Integra, INFUSE Bone Graft, recombinant human bone morphogenetic protein 2–Genetics Institute/Medtronic-Sofamor Danek/Integra, RhBMP 2–Genetics Institute/Medtronic-Sofamor Danek/Integra. BioDrugs 2002;16:376–7.
[6] Walker DH, Wright NM. Bone morphogenetic proteins and spinal fusion. Neurosurg Focus 2002;13:e3. [7] 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. J Neurosurg Spine 2004;1:19–23. [8] Chen D, Zhao M, Mundy GR. Bone morphogenetic proteins. Growth Factors 2004;22:233–41. [9] Zlotolow DA, Vaccaro AR, Salamon ML, Albert TJ. The role of human bone morphogenetic proteins in spinal fusion. J Am Acad Orthop Surg 2000;8:3–9. [10] Granjeiro JM, Oliveira RC, Bustos-Valenzuela JC, Sogayar MC, Taga R. Bone morphogenetic proteins: from structure to clinical use. Braz J Med Biol Res 2005;38:1463–73. [11] Sandhu HS, Khan SN, Suh DY, Boden SD. Demineralized bone matrix, bone morphogenetic proteins, and animal models of spine fusion: an overview. Eur Spine J 2001;10(suppl 2):S122–31. [12] 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.