What Are the Types of Complications That Can Result From the Use of BMP in Spinal Fusion Surgery?
Twenty-three cohort studies reported on BMP-related complications after lumbar spinal fusion surgery. These studies are relatively heterogeneous in their methods: the surgical approach [i.e., anterior lumbar interbody fusion (ALIF), posterior lumbar interbody fusion (PLIF), transforaminal lumbar interbody fusion (TLIF), or posterolateral fusion (PLF)], number of surgical levels, operative indications, dose and regional (i.e., where within the interbody space) placement of BMP, and type of cage used varied by study. Seventeen studies used rhBMP-2 (InFUSE), whereas 5 studies assessed rhBMP-7 (OP-1 Putty); 1 large registry study did not differentiate between use of rhBMP-2 and rhBMP-7. The details (i.e., dose, type of implant) for each study are summarized in the Supplemental Digital Content, Table 4, available at: http://links.lww.com/BRS/A416.
Extradiscal, Ectopic, or Heterotopic Bone Formation.
Thirteen studies reported rates of heteroptic bone formation after lumbar spine surgery with BMP (Figure 3). Of these studies, 3 identified patients who developed extradiscal, ectopic, or heterotopic bone formation after lumbar spine surgery with rhBMP-2, and one identified patients with extradiscal bone formation after fusion with rhBMP-7. Extradiscal, ectopic, and heterotopic bone formation presumably occurs when BMP leaks from the carrier into the epidural space and could result in canal or foraminal compression or in fusion of unanticipated levels.13 In a small prospective cohort study, Haid et al14 identified bone formation outside of the disc space in 24 of 32 patients (75%) who underwent single-level PLIF with rhBMP-2 using thin-cut 1-mm CT scans and plain radiographs. All radiographs and CT scans were interpreted by 2 independent blinded radiologists. Patients received ACS soaked with 4.0 to 8.0 mg rhBMP-2 placed exclusively within the cage; the dosage of rhBMP-2 varied by cage size. There was statistically significantly more extradiscal bone formation in the BMP group (75%; 24 of 32 patients) compared with the control group (13%; 4 of 31 patients) (P < 0.0001). Interestingly, cage placement at or within 2 mm of the margin of the posterior vertebral cortex was associated with bone formation in the spinal canal in 23 of 30 rhBMP-2-treated patients (77%), compared with 4 of 31 control patients (12%) with similar cage placement. Bone formation outside the disc space was not associated with increased leg pain or any other patient outcomes. Although 3 of 34 patients required reoperation at the same spinal level, the reasons for reoperation were not stated, and no association was stated for the presence of extradiscal bone formation and reoperation.
Using thin-cut helical CT scans, Joseph and Rampersaud15 identified heterotopic bone formation in 5 of 23 patients (21%; n = 1 central canal and n = 4 neural foramens) who underwent 1- to 2-level PLIF or TLIF with rhBMP-2 and local autogenous bone graft as part of a small prospective study. Patients received an ACS soaked with 4.2 mg rhBMP-2/level, which was placed in the anterior disc space and within the cage (TLIF) or within the cage only (PLIF). In addition, local autogenous bone graft was placed in the disc space. It is not clear whether CT scans were independently reviewed. Foraminal bone formation occurred in 4 of 5 of these patients, all of whom underwent TLIF. The remaining patient developed intracanalar bone formation. No adverse clinical outcomes were associated with this bone growth; although 3 patients required additional surgery, it was not associated with the presence of heterotopic bone. The reasons for reoperation were low back pain associated with instrument prominence in 2 patients and radiculopathy resulting from a herniated disc in 1 patient. There was no statistically significant difference in rates of heterotopic bone formation between the investigational and the control group. Ectopic or paraspinal bone formation was not reported.
Meisel et al16 detected intracanalar bone formation by plain radiographs and thin-cut 1-mm CT scans (both studies reviewed by 2 independent radiologists) in 1 of 17 patients (6%) who underwent one- to two-level PLIF with rhBMP-2 as part of a small prospective cohort study. The patient remained asymptomatic. RhBMP-2 was used at a dose of 12 mg/level (6 mg in each cage) in most patients and was placed exclusively within the cage.
Kanayama et al17 conducted a prospective study that included 10 patients who underwent single-level posterolateral fusion with pedicle screws with 7 mg/level rhBMP-7 mixed with a collagen carrier placed in equal amounts on both sides of the posterolateral fusion. Biopsies were taken after fusion in 7 patients, and hematoxylin/eosin-stained sections showed evidence of bone (hypothesized to be of endochondral origin) surrounding hyaline cartilage in 2 patients (29%). No associated outcomes were discussed nor was it reported who examined the biopsies.
Nine studies reported an absence of extradiscal, ectopic, or heterotopic bone formation (placement of BMP and other experimental details are noted in Supplemental Digital Content, Table 3, available at: http://links.lww.com/BRS/A416): Burkus et al,18,19 Mummaneni et al,20 Singh et al,21 Slosar et al,22 and Villavicencio et al23 did not identify any cases of extradiscal, ectopic, or heterotopic bone formation in rhBMP-2-treated patients, whereas Furlan et al24 and Vaccaro et al25,26 did not identify any such cases in patients treated with rhBMP-7.
Seven studies reported on resorption of the vertebrae, graft, or endplates, 6 of which identified patients with this. Although osteolysis or resorption are a normal part of the remodeling process that often leads to fusion, it may also result in mechanical failure (i.e., cage migration, excessive endplate subsidence, fracture of allograft).13 Higher rates of resorption have been noted with BMP and this is presumed to result from enhanced osteoclastic activity by the BMP.27 In a prospective study in which patients underwent single-level ALIF using rhBMP-2, Burkus et al18 identified localized bone remodeling zones that extended <4 mm into the trabecular host bone of the adjacent vertebrae in 14 of 79 rhBMP-2-treated patients [18%; vs. none of the 52 patients (0%) in the control group]. RhBMP-2 of an unspecified dose was placed within the dowel. These zones were observed 3 months after surgery by thin-cut CT scans and were not detectable by plain radiographs. CT scans and radiographs were interpreted by 2 independent blinded radiologists, and a third was available to resolve any disagreement. The remodeling zones remained within the periphery of the diameter of the central opening of the allograft dowel in all cases and occurred in the superior vertebral body of 12 of 14 patients (86%). All remodeling zones were completely healed and filled with new trabecular bone by 24 months. The presence of the bone remodeling zones was not associated with fusion level (P = 0.21) or patient demographics and had no effect on fusion or clinical outcomes. Furthermore, no correlation was made between the presence of bone remodeling zones and the need for reoperation. Using thin-cut CT scans, Meisel et al16 identified transient bone resorption in all patients who underwent 1- to 2-level PLIF with rhBMP-2 [100% (15/15)]. A dose of 12 mg/level was used in most patients and placed exclusively within the cage. Patients remained asymptomatic, and bone formation was detected in the areas of resorption at 6 months' follow-up. “Clinical success” was achieved in all patients, and the authors concluded that endplate resorption had no negative effect on clinical outcomes. Other details of this study are described above.
In a retrospective study, McClellan et al28 reported vertebral resorption in 22 of 32 (69%) of vertebral levels. Patients were treated with 1- to 2-level TLIF with rhBMP-2 placed both within and adjacent to the interbody implant; the dose varied between patients and was not controlled. Resorption was detected at a mean of 4.4 (range, 3–7) months after surgery using thin-cut 1-mm CT scans, which were interpreted by a nonindependent blinded radiologist. The osteolytic defects were characterized as mild (50%), moderate (18%), or severe (32%). Severe defects were defined as vertebral osteoclasis that was >75% of the allograft area or >1 × 1 cm on CT reconstruction and were associated with graft subsidence and loss of endplate integrity. Mild defects were defined as vertebral osteoclasis that was <25% of the allograft area and <3 × 3 mm on CT reconstruction, and moderate defects as osteoclasis between 25% and 75% of the allograft area and <5 × 5 mm on CT reconstruction. A lack of bridging bone between facet articulations or vertebral endplates was associated with both mild and moderate osteoclastic defects. No associations with clinical outcomes were made. In addition, because patients were not followed long enough to be able to detect fusion, it was not possible to correlate possible osseous defects with fusion rate. The authors noted that selection bias may have occurred, because postoperative CT scans were limited to those patients who were experiencing postoperative symptoms.
In a prospective study of 9 patients who underwent stand-alone ALIF with femoral ring allograft and BMP-2 (dose not reported) or autograft, Pradhan et al29 noted that graft and endplate resorption occurred earlier and was more severe in patients who received rhBMP-2 compared with control patients. The rhBMP-2/ACS was placed both within and around the allograft. All radiographs and CT scans were assessed by an independent, blinded radiologist. The authors did not report the total number of patients who developed such resorption; however, extensive erosion of and surrounding the allograft was detected in radiographs and thin-slice 1-mm CT scans in the 5 patients (56%) who had failed fusion. Three of these 5 patients underwent revision fusion surgery.
Vaidya et al30 reported resorption in 41 of 50 (82%) treated levels in a prospective study that evaluated rhBMP-2 use for single- or multilevel lumbar (ALIF, PLIF, and TLIF) interbody fusion surgery. Vertebral resorption was detected using plain radiographs and, in 10 patients, CT scans images were evaluated by 3 independent observers. RhBMP-2 was used at 2 mg/level and placed both within and posterior to the cage. The transition from resorption to bone formation occurred between 6 and 9 months after surgery. The degree of resorption varied between patients as well as between vertebral levels in patients who underwent fusion of more than 1 vertebral level. The authors speculated that erosion of the endplates led to an increase in disc space size and consequently cage migration and subsidence (discussed below). In case examples disc removal was limited and endplate violation was evident highlighting the question of association to surgical technique. Clinical outcomes were not reported.
In a prospective study of TLIF and ALIF with pedicle screws, and cervical interbody fusion, Vaidya et al31 reported at 12-month follow-up significant lucency and subsidence in 53% (9 of 17) of TLIF levels treated with BMP-2 (24% had >10% subsidence) and in 70% (14 of 20) of ALIF levels with BMP-2 (27% had >10% subsidence). The mean subsidence for TLIF and ALIF without BMP-2 was 12% and 6%, respectively. The authors hypothesized that erosion of the endplates directly contributed to graft subsidence. Fusion was assessed by CT scans that were evaluated by 2 independent reviewers. Because the fusion rate was 100%, the authors concluded that endplate erosion did not affect fusion success. In the lumbar spine, RhBMP-2 was used at a dose of 2 mg/level and placed within the allograft spacer and disc space (ALIF) or anteriorly within the disc space (TLIF). Results for the cervical spine are presented below. Clinical outcomes were not reported.
Finally, Slosar et al22 did not detect allograft osteolysis or graft fragmentation in any patients. In this prospective study, patients were treated with 1- to 3-level ALIF with pedicle screws and 3.0 mg/level rhBMP-2 placed in the center of the femoral ring. Three independent observers analyzed CT scans for the presence of osteolytic activity. Clinical outcomes were not reported. Solid fusion was achieved in 81 of 103 patients (79%) at 6 months, 96 of 103 patients (93%) at 12 months, and 99 of 103 patients (98%) at 24 months.
Graft Subsidence and Cage Migration.
There is a paucity of studies that have reported the precise incidence of graft subsidence and/or cage migration with lumbar interbody fusion without BMP with pedicle screw or transfacet or translaminar screws. There are 2 studies32,33 that report a range of 1% to 4.8% of cage migration after TLIF without BMP. One study reported a 0% incidence of cage migration after PLIF without BMP.34 Several studies have reported on cage migration with PLIF when posterior screw fixation is not used; however, this does not provide a valid comparison with the studies presented herein with BMP as the majority included posterior fixation. Subsidence is a normal part of the biology of fusion. In cases without BMP, it is typically not pronounced and does not contribute to cage migration. However, the lack of data on subsidence and cage migration for interbody fusion without BMP means it is underreported or its occurrence is insignificant. The rate of graft subsidence with BMP was reported in 5 studies, 4 of which identified patients with this complication. Using plain radiographs and thin-cut CT scans, Haid et al14 identified graft subsidence at least 3 mm from the posterior margin of the vertebral body in 2 of 33 PLIF patients (6%) treated with 4.0 to 8.0 mg rhBMP-2. The authors did not offer any direct correlations between graft subsidence and clinical outcomes or reoperation. Further details of this study are described above. McClellan et al28 noted graft subsidence that accompanied a loss of endplate integrity and severe vertebral resorption in 5 of 32 (16%) vertebral levels in patients treated with TLIF and a varied and unspecified dose of rhBMP-2. The degree of subsidence is provided in Table 4. The authors attributed the graft subsidence to vertebral resorption. Although patients were not followed long enough to determine whether graft subsidence affected fusion success, the authors noted that graft subsidence could theoretically result in pseudarthrosis. The types of cages used varied and included (but were not limited to) Tangent (Medtronic), TPLIF (Synthes Spine), and PEEK (Medtronic). No information was provided regarding cage placement; other details of this study are described above. Using plain radiographs and CT scans in some patients, Vaidya et al30 identified disc subsidence in 11 of 50 (22%) ALIF, TLIF, and PLIF levels (2 mg rhBMP-2/level); the mean disc subsidence was 17.8%. Interbody PEEK cages containing rhBMP-2 were used in conjunction with posterior pedicle screws (TLIF and PLIF) and anterior buttress plates or a screw with a washer and posterior pedicle screw construct were used in the ALIF cohort. Further details regarding associated cage migration and clinical outcomes are discussed below. Vaidya et al31 reported significant subsidence of >10% in 23 of 37 (62%) treated lumbar levels 12 months after fusion. The mean subsidence in ALIF and TLIF was 27% and 24%, respectively. The authors noted that the degree of subsidence varied between patients as well as between vertebral levels in patients who underwent fusion in more than 1 level. Furthermore, disc subsidence did not seem to affect fusion success or clinical outcomes. Although 4 patients underwent a second surgery, no connections were made to reoperation and graft subsidence. Details of the cage type and placement were not disclosed. Vaidya et al30,31 attributed graft subsidence to erosion of the endplates. Boden et al4 did not detect any clinically significant subsidence (i.e., ≥1 mm) on CT scans (0/11); these scans were interpreted by 5 presumably nonindependent blinded physicians. In this RCT, 11 patients had received single-level ALIF with 3.9 mg or 7.8 mg rhBMP-2 on an ACS.
The incidence of cage migration was reported in 3 studies in which patients were treated with rhBMP-2; 2 of these studies identified patients with this outcome. Haid et al14 reported shifting of the rhBMP-2-filled cage marginally into the spinal canal or neuroforamina in 12 of 34 patients (35%); similar rates of cage migration were reported in 10 of 33 patients (30%) control group. Patient outcomes associated with cage migration were not reported. Cage placement was determined primarily by thin-cut CT scans, all of which were interpreted by 2 independent blinded radiologists. Vaidya et al30 identified cage migration by 6 weeks after surgery in 10 of 36 patients (28%), 8 of whom had a TLIF. Case examples in this publication demonstrated very limited disc space preparation and evidence of endplate violation on postop imaging. Eight patients needed revision surgery because of neurologic symptoms caused by the cage migration. Although all cages had fit snugly within the intervertebral space at the close of the primary surgery, the authors found during re-exploration that the cages had become loose. These authors attributed the cage migration to an increase in the size of the intervertebral space caused by the endplate resorption that frequently accompanies rhBMP-2 use. They further speculated that placing the rhBMP-2 in the middle of the cage may have caused resorption on either end of the cage, resulting in cage loosening, migration, and subsidence. These authors did not cite the surgical technique issues evident in their case examples. Conversely, Boden et al4 did not identify any cage migration on CT scans (0/11); each patient received 2 NOVUS LT (Sofamor Danek) tapered interbody cages.
Hematomas, Seromas, and Wound Complications.
Three studies reported hematoma formation after fusion of lumbar vertebrae with rhBMP-2. Boden et al35 identified hematomas in 2 of 22 (9%) patients after single-level posterolateral fusion with 40 mg/level rhBMP-2; 20 mg rhBMP-2 was applied to 10 cm3 biphasic calciumphospate (60% hydroxyapatite:40% tricalcium phosphate) on both sides of the spine. Both hematomas required evacuation 4 to 5 days after operation; 1 patient had lingering numbness in both legs and the other had no other symptoms. None of the 22 patients developed seromas (Figure 4). In a study that evaluated BMP-2 use in deformity surgery, Luhmann et al36 noted 1 of 70 patients (1%) developed a deep wound hematoma that required a drainage procedure. It was not associated with any long-term negative clinical outcomes. In this study, patients were treated with ALIF, posterior fusion, or circumferentially with varying doses of BMP-2 and carriers. Villavicencio et al23 conducted a retrospective study in which up to 3-level TLIF was performed and reported hematomas in 2 of 74 patients (3%). RhBMP-2 was used at either 4.2 or 12.0 mg/level and was placed anteriorly against the anterior anulus, and if posterolateral fusion was performed, the rhBMP-2/ACS sponge was wrapped around cancellous bone granules and placed posterolaterally in the decorticated intertransverse process area.
Cahill et al37 reviewed the discharge records of the 13,972 patients in the Nationwide Inpatient Database that underwent lumbar fusion with either rhBMP-2 or rhBMP-7 in 2006. Wound complications, including hematomas, seromas, and wound infections or dehiscence, occurred in 281 of 13,972 patients (2%). This rate was not statistically different than the 507 of 22,835 patients (2.1%) who did not receive BMP for lumbar fusion surgery (OR, 0.90; 95% CI, 0.78, 1.04). Because these rates are from a large registry study, no details regarding patient outcomes or BMP type, dose, placement were reported.
Wound infections or dehiscence were additionally reported in 8 studies. Boden et al4 reported 1 of 11 patients (9%) with wound dehiscence (9% (1/11) after ALIF with rhBMP-2. Luhmann et al36 reported wound infections in 3% of patients after ALIF or posterior fusion with rhBMP-2. One patient developed superficial wound dehiscence after ALIF, whereas another patient had a deep wound infection after posterior fusion that required irrigation, debridement, and intravenous antibiotics. No long-term negative outcomes occurred in these patients. Slosar et al22 reported 1 of 45 patients (2%) with deep wound infection after ALIF with 3.0 mg rhBMP-2/level. The patient was treated with irrigation, debridement, and an intravenous course of antibiotics. Villavicencio et al23 reported 2 of 74 patients (3%) who developed postoperative infections after TLIF with or without posterolateral fusion with rhBMP-2; no other details were disclosed. Furlan et al24 noted that 2 patients developed superficial infections after posterolateral fusion with rhBMP-7; however, the authors did not provide a separate rate of infection for patients who underwent lumbar versus cervical fusion. The infections were successfully treated with debridement and antibiotics. Vaccaro et al25 also reported wound infections in 4 of 24 patients (17%) after posterolateral fusion with rhBMP-7, whereas none of the 12 patients who received an allograft developed this complication. The infections resolved without surgical treatment. Two studies noted no postoperative wound infections: Boden et al35 (0/22 patients) and Burkus et al18 (0/79 patients).
Two studies reported rates of dysphagia after lumbar fusion with BMP. Cahill et al37 reported that 36 of 13,972 patients (0.25%) reported in the U.S. Nationwide Inpatient Database to have received lumbar fusion surgery with rhBMP-2 in 2006 or rhBMP-7 developed dysphagia or hoarseness. No significant difference was identified in 1639 of 22,835 patients (0.21%) who did not receive BMP [OR, 0.9 (2009); 95% CI, 0.89, 1.05]. None of the 36 patients reported in the small prospective study conducted by Vaidya et al31 developed dysphagia after ALIF or TLIF with rhBMP-2.
Mindea et al38 evaluated the incidence of postoperative radiculitis after minimally invasive TLIF in a small retrospective cohort study. RhBMP-2 was used at 4.2 mg/level, half of which was combined with morselized local autograft and placed in the interbody device; the other half was placed anterior to the cage in the disc space. Four of 35 patients (11%) developed new symptoms of radiculitis 2 to 4 days after surgery on the fused (ipsilateral) side that included pain along the dermatomal distribution and subjective parethesias without lower extremity paresis. The radiculitis was treated nonoperatively in all patients and resolved within 6 weeks after surgery. No structural basis for these symptoms was identified on CT scans; the authors did not report whether the scans were interpreted by an independent or blinded radiologist. Radiculitis did not develop in any of the 8 patients who did not receive rhBMP-2.
Five studies assessed antibody responses to BMPs or bovine collagen after surgery. Boden et al35 reported a transient antibody response to rhBMP-2 in 1 of 22 patients (5%) 3 months after surgery. No clinical sequelae were noted. Boden et al4 did not detect an elevated antibody response to rhBMP-2 in any of the 11 patients, although 3 patients (27%) developed antibodies to bovine type I collagen. This response was persistent in 1 patient who also had a positive antibody titer before surgery. No complications were associated with this antibody responses. Burkus et al18 did not identify an elevated antibody response to rhBMP-2 in any patients, although 14 of 79 patients (18%) developed a postoperative elevated antibody response to bovine collagen. Similarly, 3 months after PLIF with rhBMP-2, Haid et al14 found that no patients had an elevated antibody response against rhBMP-2, and 3 of 34 patients (9%) had developed antibodies against bovine type I collagen. There were no signs of any negative clinical sequelae in patients who tested positive for antibodies against bovine collagen. Vaccaro et al39 tested for the presence of neutralizing antirhBMP-7 antibodies and found that although 26% of patients developed such antibodies at any time during follow-up, although there was no association with this neutralizing activity with any clinical parameters. Further, no neutralizing antibodies were detected in the serum of patients at the 24 or 36 month follow-up appointments.
Luhmann et al36 did not detect any evidence of local or systemic toxicity after lumbar spine fusion with rhBMP-2, nor did Furlan et al24 or Vaccaro et al25,26 identify any such responses after lumbar spine fusion with rhBMP-7.
Inflammation Associated With the Collagen Carrier.
Kanayama et al17 described evidence of an inflammatory response to the collagen carrier in 2 of 7 patients (29%) assessed by histology. Residual pieces of collagen carrier were identified and associated with chronic moderate inflammation in 1 patient and mild inflammation in another. No other details or outcomes were mentioned.
Twelve cohort studies reported on BMP-related complications after anterior cervical discectomy fusion (ACDF) surgery. The methods in these studies were relatively heterogeneous and varied in terms of the number of spinal levels being operated on, surgical indications, dose and placement of BMP, and type of cage used. RhBMP-2 (InFUSE) was used in ADCF surgeries in 8 studies, whereas 1 study used rhBMP-7 (OP-1 Putty); another study did not differentiate which type of BMP was used (Supplement Digital Content, Table 5, details for each study, available at: http://links.lww.com/BRS/A416).
A review of dysphagia rates as reported in prospective studies that evaluated ACDF patients treated without BMP is included in this issue.40 Overall rates of dysphagia declined over time from 33% at 1 week after surgery and plateaued at 1 year to a rate of 13 to 21%; however, it is difficult to directly compare these rates to those presented here. Dysphagia rates are confounded by the number of surgical levels treated, surgical approach, patient gender and age, timing postsurgery, the method by which dysphagia was determined, and how dysphagia is defined. As a result, it may be more appropriate to compare rates of dysphagia in patients treated with or without BMP from individual studies in order to avoid bias; in the 5 studies reviewed below that included control (no BMP) groups, rates of dysphagia were consistently higher in BMP-treated patients compared with patients who did not receive BMP (see below for details). For this review, dysphagia or neck swelling was identified in patients in 8 studies. Dysphagia has been associated with cervical spinal fusion with BMPs and is believed to occur as a result of a BMP-induced local inflammatory response.
Buttermann41 conducted a prospective cohort study in which 30 patients received up to 3-level ACDF with 0.9 mg/level rhBMP-2; the rhBMP-2/ACS was placed both within and posterior to the allograft in the interbody space. Although all patients in the BMP group were prescribed a low-dose oral steroid (methylprednisolone) after hospital discharge, neck swelling that resulted in dysphagia was identified in 50% of rhBMP-2-treated patients (Figure 5). In contrast, only 3 of 36 (14%) control ICBG patients had the same complication. The authors noted that the severity of symptoms was considerably higher in rhBMP-2-treated patients. Of patients treated with rhBMP-2, 20% required readmission to the intensive care unit (ICU) for intravenous steroid treatment, although reoperation was not required. Onset of dysphagia occurred at 4 (±3) days after ACDF; complete resolution of symptoms occurred 21 (±16) days later. The incidence of dysphagia was highest in patients who had undergone 2-level ACDF (63%); 50% of patients treated with one-level and 30% of patients treated with 3-level ACDF developed dysphagia.
In a large registry study of 2886 patients who underwent anterior or posterior cervical fusion surgery with rhBMP-2 or rhBMP-7 in 2006, Cahill et al37 found that dysphagia or hoarseness of voice was reported in 110 of 2886 patients (3.81%). This outcome occurred in 100 of 2299 patients (4.35%) who underwent ACDF and in 10 of 478 patients (2.1%) who received posterior cervical fusion. The rate of dysphagia or voice complications was significantly higher in BMP-treated patients who underwent ACDF than the 608 of 24,768 patients (2.45%) who did not receive BMP (OR, 1.80; 95% CI, 1.45, 2.24). In contrast, there was no significant difference in dysphagia rates in posterior cervical fusion patients treated with versus without BMP [no BMP, 1.63% (39/2391); OR, 1.28; 95% CI, 0.63, 2.59]. Again, because of the large nature of this registry study, no other details were disclosed.
Hiremath et al42 reviewed the postoperative morbidity rates of 16 posterior cervical fusion patients who received rhBMP-2. Although 1 patient developed neck swelling (6%), it was not significant enough to cause dysphagia and resolved with a short course of steroids. The patient had no evidence of a hematoma, and no long-term negative outcomes occurred. Patients received an average of 1.95 mg rhBMP-2/level (range, 0.75–4.05 mg/level); the placement of BMP was not reported. There were no associations reported between neurologic demise and the use of BMP in this study.
In a retrospective cohort study of 151 patients, Shields et al43 reported that 13 of 151 patients (9%) developed dysphagia, respiratory difficulties, or incisional swelling. Patients had been treated with 1- to 3-level ACDF (N = 138) or corpectomy (N = 13) with up to 2.1 mg/level rhBMP-2 (more was used in corpectomies). The rhBMP-2/ACS was placed within as well as often lateral and anterior to the graft. Of these 13 patients, 5 required a prolonged hospital stay (>48 hours) and 8 were readmitted to the hospital (6 of these were readmitted 3 to 4 days after surgery). Single-level ADCF had been performed in 3 patients and 2-level ACDF in 8 patients; single-level vertebrectomy had been used in 2 patients and 2-level vertebrectomy in 1 patient. Only 3 patients underwent drain placement during the primary procedure. No mention was made as to whether patients were treated with steroids at any point.
In a retrospective cohort study, Smucker et al44 reported that 19 of 69 patients (28%) treated with rhBMP-2 developed prevertebral swelling complications, which were defined as complications that developed within 6 weeks of surgery and included visible swelling at the surgical site, difficulty in swallowing (dysphagia) or breathing, and resulted in medical evaluation, additional time in the hospital, or surgical drainage due to compromised airways. Patients had been treated with single- or multilevel ACDF using varied and undetermined doses of rhBMP-2; rhBMP-2/ACS was placed within the allograft or interbody spacer and/or surrounding the graft in the disc space. In contrast, only 6 of 165 control patients (3.6%) had the same response (P < 0.0001). In the rhBMP-2-group with 69 patients, 9 patients (13%) had a delay in discharge, 5 patients (7%) had severe dysphagia, 2 patients (3%) had visible neck swelling, 2 patients (3%) were reintubated, 1 patient (1%) required PEG (percutaneous endoscopic gastrostomy) placement, and 1 patient had a tracheostomy. Exploration and drainage was performed in 3 of 69 patients (4%); diffuse swelling was found in the neck, and all 3 patients recovered without further swelling complications. Readmission was necessary for 2 of 69 patients (3%), whereas 3 patients (4%) were prematurely seen at a clinic or the emergency room and 2 patients (3%) required an outpatient otolaryngology consultation. BMP-treated patients developed swelling complications at a mean of 4.2 (range, 2–7) days after surgery. Besides rhBMP-2 use, no other variables (e.g., smoking status, advanced age, revision surgeries, etc.) were significantly associated with cervical swelling after controlling for variables using a logistic regression analysis (OR = 10.1; 95% CI, 3.8–26.6; P = 0.0001). The authors noted that at least some patients developed swelling complications even though the rhBMP-2 had been placed exclusively within the central portion of the graft. Administration of steroids was not reported.
In a retrospective study, Tumialan et al45 reported that 14 of 200 patients (7%) developed postoperative dysphagia following 1- to 4-level ACDF with up to 2.1 mg rhBMP-2/level; rhBMP-2/ACS was placed only within a PEEK cage. The potential effect of rhBMP-2 dose in this study is discussed in the next section. Dysphagia was characterized as severe in 5 of 14 patients (35%) with dysphagia; these patients were unable to swallow and required supplemental nutrition and placement of a PEG tube (in 4 patients). One patient with amyotrophic lateral sclerosis required a permanent PEG tube; the others were able to swallow solid food within 2 to 3 months. Of those who required a PEG tube, 4-level ACDF had been performed in 1 patient, 3-level ACDF in 2 patients, and 2-level ACDF in 1 patient. Dysphagia was characterized as moderate in 3 of the 14 of patients with dysphagia; these patients were examined by a speech therapist and/or required alterations in their diet and had a delay in discharge up to 72 hours. Three-level ACDF had been performed in 2 of these patients and 2-level ACDF in the other. Finally, dysphagia was considered mild in 6 of the 14 patients with dysphagia; patients remained in the hospital up to 48 hours longer than anticipated due to swallowing difficulties. Three- and 4-level ACDFs had been performed in 3 and 2 patients with mild dysphagia, respectively. Symptoms of dysphagia resolved within 6 weeks (mild cases), and 6 to 12 months (moderate cases). Timing of resolution was not noted for severe cases, although 1 patient required a permanent feeding tube.
Vaidya et al46 conducted a retrospective cohort study in which 22 patients were treated with up to 3-level ACDF using 1 mg/level rhBMP-2; rhBMP-2/ACS was placed exclusively within the cage. Dysphagia was experienced by 85% of patients treated with rhBMP-2. At 2 weeks after surgery, dysphagia affected 85% of patients. In this and the immediate postoperative period, 10% of BMP-2 patients developed severe dysphagia (prolonged hospitalization required, patients were unable to swallow fluids), 30% moderate dysphagia (patients experienced difficulty swallowing), and 45% mild dysphagia (patients experienced discomfort while swallowing). At 6 weeks after surgery, 65% of patients had dysphagia, which was classified as severe in 5%, moderate in 20%, and mild in 40%. By 3 months after operation, no patients had severe dysphagia. Two years after ACDF, 20% of patients still experienced mild dysphagia. Three patients had a delayed hospital charge and 1 patient required a PEG tube for 6 weeks after the index procedure. The incidence of dysphagia was significantly higher in the BMP versus control group at both 2 weeks and 6 weeks after ACDF (P = 0.009 and P = 0.019, respectively) (control group: 39% (7/18 patients) and 22% (4/18 patients), at 2 weeks and 6 weeks, respectively). Furthermore, patients treated with 2- or 3-level ACDF experienced higher rates of dysphagia than those who underwent 1-level ACDF. No patients treated with single-level ACDF developed severe dysphagia. Steroid or other treatment was not discussed.
Vaidya et al31 identified “prolonged” dysphagia in 6 of 11 patients (55%) treated with 1 mg/level rhBMP-2 as part of a small prospective study, all of whom had postoperative prevertebral swelling. The authors noted that the cases of dysphagia were more severe than those they had previously seen in their practice; however, they did not discuss whether patients required steroids or other treatment such as ICU admission, reintubation, etc. None of 12 patients in the control group developed dysphagia. Although some patients received multilevel fusion, no correlation was made between development of dysphagia and the number of levels treated. The placement of rhBMP-2 was not disclosed.
Extradiscal, Ectopic, or Heterotopic Bone Formation.
Extradiscal, ectopic, or heterotopic bone formation was reported in 2 studies in which patients received rhBMP-2 and 1 study in which patients received rhBMP-7. Baskin et al47 conducted a RCT in which 18 patients were treated with up to 2-level ACDF with 0.6 mg rhBMP-2/level; rhBMP-2/ACS was placed within the allograft. Two independent, blinded radiologists interpreted all radiographs. Bone formation anterior to segments adjacent to the fused vertebrae was detected in 2 of 18 patients (11%) on radiographs obtained 12 months after surgery; this outcome was also detected in 1 of 15 patients (6%) control patients. No patients required reoperation at the same spinal level; no other outcomes were discussed for patients who developed extradiscal bone. Tumialan et al45 noted extradiscal interbody bone formation in radiographs of 3 of 200 patients (2%), all of whom remained asymptomatic. Radiographs were interpreted by the authors. Furlan et al24 noted asymptomatic heterotopic ossification in 1 of 14 patients (7%), detected by an independent radiologist on a radiograph. Patients were treated with ACDF for this prospective cohort study; patients received 7 mg/level rhBMP-7 placed in equal amounts on each side of the spine in the posterolateral gutters.
Four studies identified vertebral, graft, or endplate resorption after ACDF with rhBMP-2. Vaidya et al30 reported resorption of both the superior and inferior endplates in radiographs of all vertebral levels (32/32). Patients had been treated with ACDF and 1 mg/level rhBMP-2, which was placed exclusively within the cage. Radiographs were evaluated by 3 independent observers. Erosion of the endplates developed between 2 weeks and 6 weeks, which was notably earlier than it occurred in patients who underwent fusion of the lumbar vertebrae. The transition to bone formation occurred primarily between 3 and 6 months, and no complications were associated with endplate resorption. Vaidya et al46 detected endplate resorption in all of 55 vertebral levels treated by 6 weeks after fusion; the resorption did not affect fusion success, however, and had resolved by 6 months. Radiographs were used to identify resorption and were evaluated by 3 independent observers as well as a musculoskeletal radiologist. In a prospective study in which 11 patients were treated with ACDF using 1 mg/level rhBMP-2, Vaidya et al31 found significant endplate erosion on CT scans of all 25 patients tested. However, the authors did not disclose how many of these patients received cervical versus lumbar fusion. Measurements were made on all scans by 2 independent observers. As discussed above, all patients had successful fusion. Finally, Shields et al43 noted graft resorption in 1 of 151 patients (1%) after single-level ACDF. No details were reported regarding the clinical outcome of this patient or the type of radiology used to identify resorption.
Graft Subsidence and Cage Migration.
Two studies detected graft subsidence after ACDF with rhBMP-2. Vaidya et al30 identified subsidence of the disc space in 13 of 32 (41%) treated levels, with an average subsidence of 12.8%. The clinical outcome of these patients was not reported. This study used PEEK cages, which were filled with rhBMP-2 and used in conjunction with anterior locking plates. No other details regarding cage placement were given. Vaidya et al31 reported early lucency and graft subsidence in 6 of 18 (33%) treated levels at 12 months after surgery, with a mean subsidence of 53%. Subsidence was not associated with an increase in pain that occurred up to 6 weeks after surgery. The authors did not provide details on cage type or placement.
Graft migration was noted in 2 studies. Shields et al43 reported implant dislodgement in 2 patients (1%); the outcomes of these patients was not reported. The authors used a variety of cages, including a resorbable poly(d,l-lactic acid) cage (HYDROSORB 228, Medtronic), homologous bone graft (CORNERSTONE, Medtronic), or a titanium mesh cage (PYRAMESH 228, Medtronic). Vaidya et al30 identified minimal cage migration in 1 of 59 patients (4%) that was not associated with any clinical sequelae. Again, Vaidya et al30,31 hypothesized that resorption was responsible for graft subsidence and migration.
Hematomas, Seromas, and Wound Complications.
Hematomas or seromas were identified in 3 studies. Shields et al43 reported that 15 of 151 patients (10%) developed hematomas within 5 days of ACDF or vertebrectomy with rhBMP-2. Eight patients required surgical drainage of the hematoma. Four patients underwent 1-level ACDF, 3 patients had noncontiguous 2-level ACDF, whereas another 2 patients received 2-level adjacent ACDF. In addition, 2 patients underwent a 1-level vertebrectomy and 2 patients underwent a 2-level vertebrectomy. Previous ACDFs had been performed in 6 of these patients. Seven of 15 patients (47%) who developed hematomas had received a drain during the primary procedure. Vertebrectomies and previous ACDFs seemed to predispose patients to hematomas, although the statistical significance of this observation was not determined. Tumialan et al45 noted hematoma formation that required surgical evacuation in 2 patients (1%). Two patients (1%) were also reported who developed seromas; both required surgical drainage.45 Hiremath et al42 did not identify hematomas in any of the 16 patients treated with ACDF and rhBMP-2.
Cahill et al37 reported that wound complications, including hematomas, seromas, and wound infections or dehiscence, occurred in 42 of 2886 patients (1.46%) reported in the Nationwide Inpatient Database who received anterior or posterior cervical fusion surgery with rhBMP-2 or rhBMP-7 in 2006. This outcome occurred at a significantly higher rate in the 28 of 2229 ACDF patients (1.22%) who received BMP versus 160 of 24,786 patients (0.65%) who did not receive BMP (OR, 1.89; 95% CI, 1.26, 2.83). There was no significant difference between the BMP and no BMP patients who underwent posterior cervical fusion [2.93% (14 of 478 patients) vs. 2.51% (60 of 2391 patients); OR, 1.17; 95% CI, 0.64, 2.11].
In a retrospective review, Crawford et al48 compared the perioperative complications of 77 consecutive patients who underwent instrumented posterior cervical fusion; 41 patients received an average of 3.6 (range, 1.05–6.0) mg rhBMP-2/level, whereas 36 patients received ICBG without BMP. Twenty-eight patients had a laminectomy as part of the procedure, 16 in the BMP group. The authors reported more postoperative wound complications (i.e., prolonged drainage and/or deep infection) requiring treatment in the BMP group (14.6%) compared with the ICBG group (2.8%) (6 patients vs. 1 patient; P = 0.113). There were no statistically significant differences reported regarding the dose of rhBMP-2 used and postoperative complication. Postoperative neurologic demise was not reported in any of the patients. Hiremath et al42 noted that none of the 16 patients who received ACDF with rhBMP-2 developed wound infections. Finally, Furlan et al24 noted superficial infections after vertebral fusion with rhBMP-7 in 2 of 30 patients (7%) but did not report a separate rate of infection for patients who underwent cervical versus lumbar fusion. The infections were successfully treated with debridement and antibiotics.
Baskin et al47 tested patients treated with 1- to 2-level ACDF and rhBMP-2 for elevated antibody responses to rhBMP-2 and bovine type I collagen. Although none of 15 patients had a higher antirhBMP-2 antibody level compared with preoperative levels, 1 patient did develop an antibody response to bovine collagen.
One study reported no evidence of systemic or local toxicity in any patients.24
One retrospective cohort study using registry data reported on BMP-related complications after thoracic fusion surgery. This study did not differentiate which type of BMP was used (Supplemental Digital Content, Table 6, available at: http://links.lww.com/BRS/A416).
Cahill et al37 reported that dysphagia or voice hoarseness occurred in 6 of 747 patients (0.8%) in the Nationwide Inpatient Database who underwent thoracic fusion surgery with BMP in 2006. This rate was not significantly different than the 34 of 2511 patients (1.3%) who did not receive BMP (OR, 0.59; 95% CI, 0.24, 1.41).
Hematomas, Seromas, and Wound Complications.
Cahill et al37 did not find a significant difference in the rate of wound complications (hematomas, seromas, infections, and dehiscence) in the 35 of 746 patients (4.7%) who received BMP versus the 146 of 2511 patients (5.8%) who did not as part of thoracic fusion surgery (OR, 0.79; 95% CI, 0.54, 1.16).
The most commonly reported complications associated with BMP use in lumbar, cervical, or thoracic spine surgery include the following: resorption or osteolysis; extradiscal, ectopic, or heterotopic bone formation; graft subsidence; graft or cage migration; an elevated antibody response to BMP or collagen; wound infection; and radiculitis.
What Are the Rates of BMP-Related Complications in Patients Who Undergo Lumbar, Cervical, and Thoracic Spinal Fusion Surgery?
Table 1 summarizes the mean rates and ranges of BMP-related complications after lumbar spinal fusion. Of note, the incidences of resorption [0%–100% reported by 7 studies, mean of 44% (114 of 257)]; extradiscal, ectopic, or heterotopic bone formation [0%–75% reported by 13 studies, mean of 8% (32/417)]; graft subsidence [0%–62% reported by 5 studies, mean of 25% (41/163)]; and cage migration [0%–35% reported by 3 studies, mean of 27% (22/81)] ranged widely among studies. Eight studies reported wound infections or dehiscence, which occurred in 12 of 355 patients (3%; range, 0%–17%). Another study defined hematomas, seromas, and wound infections as wound complications, which occurred in 381 of 13,972 patients (2.0%). Rates of other BMP-related complications had narrower ranges, including hematomas [1%–9% reported by 3 studies; mean of 4% (5 of 136 patients)], elevated antibody responses to rhBMP-2 [0%–5% reported by 4 studies; mean of 1% (1 of 146 patients)], or bovine type I collagen [9%–27% reported by 3 studies; mean of 16% (20 of 124 patients)]. Two studies, including 1 large registry study, reported the rates of dysphagia or neck swelling; the mean rate was 0.3% (36 of 14,008 patients). One study reported a 26% (53 of 208 patients) incidence of neutralizing antibodies against rhBMP-7, and another study noted an inflammatory response to the collagen carrier after rhBMP-7 use in 2 of 7 patients (29%). Four studies reported no cases of local or systemic toxicity after BMP use (0 of 122 patients).
The mean rates and ranges of BMP-related complications after cervical spinal fusion are found in Table 2. As was found for studies that reported on complications after lumbar fusion with BMP, ranges of BMP-related complication rates were greater when more studies reported on a given complication. Eight studies, including a large registry study, reported dysphagia, neck swelling, or respiratory difficulties in 3%.8 to 85% of patients after cervical fusion with BMP [mean of 5.8% (195 of 3383 patients]. Resorption was identified in 1% to 100% of patients as reported by 4 studies [mean of 43% (113 of 263 patients)]. Extradiscal, ectopic, or heterotopic bone formation was reported by 3 studies and occurred in 2% to 13% of patients [mean 3% (6 of 229 patients)]. Other types of complications had narrower ranges of rates: graft subsidence [41%–44% reported by 2 studies; mean of 43% (37 of 87 patients)], graft migration [1%–4% reported by 2 studies; mean of 2% (3 of 174 patients)], hematomas [0%–10% reported by 3 studies; mean of 5% (17 of 367 patients)], and seromas [1% reported by 1 study (2 of 200 patients)]. Wound infections occurred in a mean of 11% (6 of 57 patients; 0%–15%) as reported by 3 studies. One study categorized hematomas, seromas, and wound infections together as wound complications, which occurred in 1.4% (42 of 2886) of patients. One study reported an elevated antibody response to bovine type I collagen in 7% (1 of 15 patients) of patients, whereas none experienced similar activity against rhBMP-2 (0 of 15 patients). One study noted no instances of local or systemic toxicity (0 of 14 patients).
Only 1 study reported the rates of BMP-related complications after thoracic spinal fusion (Table 3). Dysphagia, neck swelling, or respiratory difficulties occurred in 0.8% (6 of 746) of patients, whereas wound complications (hematomas, seromas, infections, or dehiscence) occurred in 4.7% (35 of 746) of patients.
BMP-related complication rates range from 0% to 100%. Some events occur in more narrow ranges such as extradiscal, ectopic, or heterotopic ossification (cervical spine); graft subsidence (cervical surgery); graft migration (cervical surgery); cage migration (lumbar surgery); elevated antibody response to BMP or collagen (lumbar surgery); and hematoma (cervical surgery); otherwise, rates ranged widely among studies.
Is There a Dose-Response Relationship Associated With Complications After the Use of BMP in Spinal Fusion Surgery?
We identified 1 human study in which varying doses of rhBMP-2 were used for spinal fusion surgeries (Supplemental Digital Content, Table 7, http://links.lww.com/BRS/A416). Tumialan et al45 conducted a retrospective cohort study on a series of 200 patients who underwent single- or multilevel ACDF. The first 24 patients in this series were treated with 2.1 mg rhBMP-2/level (group A); these patients were also reported by Boakye et al49 Because of the high level of excess interbody bone formation in these patients, the authors reduced the dosage of BMP to 1.05 mg/level for the next 93 patients (group B) and then again to 0.7 mg/level for the final 83 patients (group C). The excess interbody bone formation they had observed previously (13%, 3/24) at the highest dosage did not occur in any patients treated with the lowest dosage of rhBMP-2.45,49 Patients in group A also experienced transient dysphagia (2/24), CSF leak (1/24), transient C-5 paresis (1/24), and transient right vocal cord paresis with hoarseness (1/24), as reported by Boakye et al.49 Five patients from group C had complications, including 3 (of 4 total) PEG placements for severe dysphagia and 2 (of 4 total) re-explorations for seromas and hematomas. The remaining complications all must have occurred in patients from group B, although this was not explicitly stated. Excess interbody bone formation was not observed at the lower dose level, and perioperative complication rates were highest in the higher dose groups; however, the authors noted that reducing BMP to the lowest dose had no apparent effect on overall complication rates, although there were many confounding variables (e.g., repeated operations, single- and multilevel cases) that prevented stratification of the relationship between dose and postoperative rates of complications. It is therefore difficult to draw any conclusions as to a possible relationship between BMP dose and the rate complications. However, the authors concluded that when placed in a PEEK spacer, the lowest dose of rhBMP-2 (0.7 mg/level) provided “good-to-excellent” outcomes with 100% radiographic fusion and a low rate of complications (7%).
The overall strength of the evidence for studies evaluating types of specific complications associated with BMP use in lumbar surgery is “high,” i.e., future research is very unlikely to change our confidence in the estimate of effect. For the cervical spine surgery, it is “low,” i.e., future research is very likely to have an important impact on our confidence in the estimate of effect and is likely to change the estimate. For thoracic spine surgery, it is “very low,” i.e., any estimate of effect is very uncertain, Table 4. The overall strength of the evidence for summarizing rates of complications associated with BMP use in lumbar spine surgery is “moderate,” i.e., future research is likely to have an important impact on our confidence in the estimate of effect and may change the estimate. For cervical spine surgery, it is “low.” For thoracic spine surgery, it is “very low.” The overall strength of the evidence for studies evaluating the dose-response relationship between BMP use rates of specific complications is also “very low.”
The purposes of this systematic review were to identify the character and rates of complications in patients after the use of BMP in spinal surgery and to determine whether there is a dose-response relationship of BMP with complications. The rates of BMP-related complications in the 31 studies reviewed here are summarized in Tables 1 (lumbar spine), 2 (cervical spine), and 3 (thoracic spine).
The normal degrees of resorption or osteolysis for lumbar interbody fusion with pedicle screws without BMP have not been defined in the literature to date. Resorption of the vertebrae, graft, and/or endplate was reported in 8 of the 23 studies identified that evaluated the use of BMP in the lumbar spine16,18,22,28–31; on average, 44% of the 257 total patients from these studies developed this complication (range, 0%–100%). Similarly, resorption was reported in 4 of 12 studies identified in which patients underwent fusion of the cervical spine30,31,43,46; a mean of 43% of the 263 total patients had detectable resorption (range, 1%–100%). Although BMP use seems to be related to higher rates of bone resorption, it is important to note that allograft resorption occurs as a normal part of early graft consolidation. Successful graft consolidation results from an initial inflammatory response with bone resorption that is followed by new bone formation.50 The use of BMP-2, particularly within the interbody space, may result in an accelerated and exaggerated resorptive phase, presumably from a relative increase in osteoclastic activity. It has been suggested that this response is dose dependent; however, this has not been adequately tested and optimal dosing for various fusion applications remains to be defined. Most of the literature suggests that it is best prevented or minimized with careful endplate preparation avoiding endplate violation, and possibly, by using smaller doses. Studies have not defined the natural history of this problem after fusion, although 2 of the included studies suggested that graft subsidence and migration occurred as a result of resorption the issues of disc preparation were not addressed.30,31 Hence, conclusions cannot be made at this time regarding the clinical corollary of radiographic resorption or its management, and time to union and ultimate union rates.13 Further, there is a lack of substantive data or consensus among the reviewed studies to determine the ideal accompanying grafting material (e.g., allograft) or interbody cage (e.g., synthetic, metallic, allograft), carrier, optimal dosing, and placement location of BMP-2 in the interbody space. Appropriately designed studies are necessary clarify these issues.
There are no published series that have reported on heterotopic bone formation in lumbar interbody fusion without the use of BMP-2. Extradiscal, ectopic, or heterotopic bone formation was reported in 13 of the 23 studies identified that evaluated the use of BMP in the lumbar spine14–26; however, 9 of these 13 studies reported no such cases. This unintended bone formation developed in an average of 8% of the 417 total patients reported in these 13 studies (range, 0%–75%). Similarly, extradiscal, ectopic, or heterotopic bone formation was reported in 3 of 12 cervical studies24,31,45,47; of the 229 total patients in these 3 studies, a mean of 3% developed extradiscal bone formation (range, 2%–13%). In the lumbar spine, patients who developed this heterotopic bone formation had received ALIF, PLIF, or TLIF with 4.0 to 12.0 mg BMP/level; in the cervical spine, patients underwent ACDF with 0.6 to 7.0 mg BMP/level. Heterotopic bone formation was identified either by plain radiographs or thin-cut CT scans. The formation of extradiscal, ectopic, or heterotopic bone has been attributed to elution of BMP from the ACS outside of the disc space. The resulting bone growth could potentially lead to fusion of other vertebral levels, as well as neurologic impairment due to canal or foraminal stenosis.13 In all reports presented here, however, patients with extradiscal, ectopic, or heterotopic bone formation remained asymptomatic, and no reoperations were attributed to such bone formation. It should be noted, although, that the studies were not adequately descriptive regarding the precise reasons for reoperations at the index level. Larger studies are needed to define the incidence of extradiscal, ectopic, or heterotopic ossification per procedure and its relationship to dose, carrier, regional placement, cage and instrumentation, and radiographic and clinical outcome.
Dysphagia was reported in 8 of 12 studies in which patients underwent cervical fusion with rhBMP-231,41,43–46; this complication developed in a mean of 5.9% of the 3383 total patients from these 8 studies (range, 3.8–85%). Two of the 23 studies in which patients were treated with lumbar fusion and BMP reported this complication; a mean of 0.26% (36/14,008) of patients developed neck swelling, dysphagia, or respiratory difficulties. The onset of dysphagia usually occurred within 2 to 7 days after surgery. The symptoms of dysphagia ranged from mild, in which swallowing difficulties prolonged a patient's hospital stay, to severe, in which a patient's inability to swallow necessitated the placement of a temporary or permanent feeding tube.45 Neck swelling and respiratory difficulties were also reported. These complications are due to an inflammatory response incited by BMP.13 The BMP dosage, carriers, optimal cages, regional placement, and indications for ventral surgery to be further defined. Larger controlled studies are needed to determine the specific chemical mediators of dysphagia, and ways to minimize or avoid it. In response to the reported complications associated with off-label BMP use in the ventral cervical spine, Food and Drug Administration issued a warning in 2008,51 and recommended “that practitioners either use approved alternative treatments or consider enrolling as investigators in approved clinical studies.”
The incidence of BMP-associated radiculopathy after lumber fusion is poorly characterized from the literature published to date. Only 1 small study evaluated the incidence of new-onset radiculitis after lumbar fusion with rhBMP-2; 11% (4/35) of patients developed this complication. CT scans indicated that none of the patients had any structural basis for these symptoms. A number of authors have noted the evolution of radiculopathy after lumbar interbody fusion procedures with rhBMP-2.14,15,23,35 However, the studies' design and execution do not permit the conclusion of a true association between the use of BMP-2 or −7 and postoperative radiculopathy. In other words, it is plausible that the radiculopathies without demonstrable mechanical compression (i.e., due to cage migration, cyst formation) in the reported literature represent the natural history of this problem that is known to occur in a minority of patients following either TLIF or PLIF. Appropriately designed and executed studies will hopefully further characterize radiculopathy after BMP-enhanced fusions.
To date there are only 3 studies that report specifically on the complications associated with the use of the rhBMP-2 in posterior cervical fusion surgery.37,42,48 Two of these represent Level III evidence and both report discordant associations between wound complications (i.e., drainage and/or infection); Crawford et al48 reported a higher wound complication rate with rhBMP-2, and Hiremath et al42 reported an absence of complications with the product. However, Cahill et al reported that wound complications, including hematomas, seromas, infections, and dehiscence, occurred at a rate of 1.4% (42/2886) in all patients entered in the Nationwide Inpatient Database who underwent cervical fusion with BMP in 2006; 2.9% (14/478) of those who received posterior cervical fusion developed at least one of these complications. In addition, dysphagia or voice complications were reported for 3.8% (110/2886) of these patients; in patients who underwent posterior cervical fusion, 2.1% (10/478) of patients developed dysphagia. This last report represents a Level II evidence study. Given the limited number of reports and the lack of higher level evidence on the topic, meaningful conclusions cannot be made at this time about the safety of rhBMP-2 for use in posterior cervical fusion surgery.
An objective of this study was to determine whether there is a dose-response relationship between BMP dose and the rate of complications. One study was identified that used varying dosages of rhBMP-2 for ACDF.45 However, its design and execution prevents drawing conclusions regarding a potential dose-response relationship. The authors originally lowered the dose from 2.1 mg/level to 1.05 mg/level after observing a high rate of asymptomatic extraneous interbody bone formation [13% (3/24)].45,49 After lowering the dose, no additional cases of unwanted bone formation occurred in the remaining 176 patients. However, even at the lowest dose of BMP used (0.7 mg/level), 3 of the 5 cases of severe dysphagia occurred. The remaining studies included in this review for both the cervical and lumbar spine are heterogenous in design, execution, and reported results. As such, it is not possible to establish a dose-response relationship for complications for neither the cervical nor the lumbar spine. More research is necessary to determine if such an association exists.
Five studies were excluded from this review because they included 5 or less patients. Lewandrowski et al52 reported case series on 5 patients who developed resorption of the inferior aspect of the L5 vertebrae. Osteolysis was detected at 1 to 3 months after TLIF with rhBMP-2, and patients presented with severe low back pain. Symptoms resolved within 3 months and new bone formation spontaneously filled in the areas of osteolysis. Laursen et al53 noted 1 case of severe resorption in a case series of 5 patients who presented with an acute unstable burst fracture and received rhBMP-7 during PLIF. Osteolysis at the graft site was detected at 3 and 6 months after procedure and resolved by 12 months. The patient remained asymptomatic. Wong et al54 published a chart review of 5 patients who developed ectopic bone in the lumbar canal after PLIF/TLIF with rhBMP-2. These patients all developed neurologic complications, including chronic pain. This is stark contrast to the study by Joseph and Raampersaud15 who reported no clinical morbidity due to heterotopic bone following the same procedure performed via a tubular retractor. The authors cautioned that neurologic impairment is a rare but serious event that can follow spinal fusion using BMP. Brower and Vickroy55 reviewed a case in which the patient developed psoas ossification associated with severe and persistent pain after posterolateral fusion with rhBMP-2. Finally, Perri et al56 presented a patient who presented to the ER with dysphagia and severe swelling 5 days after ACDF using rhBMP-2. The patient received steroids and was extubated and recovered following hospital release 4 days later.
It has been hypothesized that large doses of BMP may lead to higher rates of complications, such as resorption and extradiscal, ectopic, or heterotopic bone formation. However, there is no level I evidence that addresses directly whether increased doses of rhBMP-2 are associated with higher rates of complications in instances when BMP is used off-label. FDA-approved use of rhBMP-2 in the lumbar spine is limited to anterior lumbar interbody fusion with LT-cage; a dosage of 4.2 to 12 mg rhBMP-2/level is recommended and varies with cage size selected. However, BMP use for other types of lumbar fusion and for cervical fusions remains off-label, and the collective results from the available level 2 to 5 evidence vary substantially. Therefore, well-designed and executed controlled studies are needed to define the various aspects of BMP use (i.e., time to fusion, fusion rates, optimal dose and carrier for the different types of fusion, ideal type of interbody graft, role and type of stabilization) before practice guidelines regarding its use can be put forth.
The half-life of BMP in different surgical environments may vary by the vascularity and the rate at which the body may “clear” the BMP from the surgical site. This rapid or slow clearance of the protein results in differing concentrations that may be different from various surgical sites and thus may affect the dosage at any given time from the point of implantation. Certain areas, such as the interbody space where the protein/carrier combination may be within a cage implant and perhaps more sheltered from the surrounding environment, may allow for a higher concentration and thus dosage, and may affect the chance of a dose-dependent complication. This may differ from a posterior lateral fusion bed where the blood and fluids from the surgical site may flow into contact with the posterolateral gutter and dilute the concentration of the protein from the fusion area. This may result in a different complication rate or a vastly different complication occurring due to surgical site, concentration/dosage, and the surrounding environment.
The complications that are associated with BMP can be substantial and require further investigation and better characterization. They also mandate a heightened informed awareness on the part of the surgeon using these products. Exchange of information is limited by FDA regulations that prohibit commercial vendors from providing information or promoting use outside of the FDA approved indications. In addition to the peer-reviewed medical literature, physician directed use of BMP and the associated morbidities have been at the forefront of national media. A public health notification warning from the FDA regarding its use in the cervical spine was forwarded in 2008. This was done to heighten awareness of potential problems for surgeons and patients. Further studies are needed to determine the optimal usage and the best way to minimize complications.
- Given the potential complications related to the use of BMP-2 in ventral cervical spine surgery, its use is not recommended until its clinical efficacy and safety is adequately defined by well-designed and executed studies.
- Given the potential complications related to the use of BMP-2 in posterior lumbar interbody fusion, there are concerns with regard to its routine use in this fusion application.
- There is insufficient data to validate the use of BMP-2 for posterior cervical or for thoracic fusion surgery.
The off-label use of rhBMP-2 has been associated with a wide range of complications. There are many variables that contribute to the physiologic effect of osteobiologics including regional use, concentration, and the carrier as well as host responses. Because of the paucity of higher level studies, the incidence of complications relative to concentration, carrier, type of implants, and natural history are not clearly defined. Therefore, it is important to understand the osteobiologic being used and its known effect on the region of the spine it is being applied to and to follow the patients closely clinically, and radiographically as indicated.
- Multiple complications are associated after the use of rhBMP-2 in both cervical and lumbar spine fusion surgery:
- Complications associated with BMP-2 use in the lumbar spine include excessive vertebral body resorption, interbody space subsidence, graft migration and extradiscal bone formation though reoperation is rare.
- Reported complications after the use of BMP-2 for ventral subaxial cervical fusion surgery include a heightened incidence and degree of dysphagia and soft tissue swelling.
- Based on the level of evidence of the peer-reviewed literature that report on complications after lumbar fusion surgery with BMP, it is unlikely that future research will impact the types of complications; however, the rates vary enough that future research will likely impact these findings. For the cervical and thoracic spine, future research will likely impact both types and rates associated with BMP use.
- Further research is necessary to define the precise incidence, natural history, radiographic and clinical outcomes of complications associated with the use of BMP for the various types of cervical and lumbar fusion.
- Further study is indicated to define the optimal dose and carrier, and regional placement of product, in addition to the optimal interbody devices, surgical techniques and fixation methods for lumbar and cervical fusion.
Supplemental digital content is available for this article. Direct URL citations appear in the printed text and are provided in the HTML and PDF versions of this article on the journal's Web site (www.spinejournal.com).
The authors are indebted to Ms. Nancy Holmes, RN, for her administrative assistance and to Mr. Jeff Hermsmeyer, BS, for his assistance in searching the literature, abstracting data, and proofing.
1. Hsu WK, Wang JC. The use of bone morphogenetic protein in spine fusion
. Spine J
2. Boden SD. The ABCs of BMPs. Orthop Nurs
3. Urist MR. Bone: formation by autoinduction. Science
4. Boden SD, Zdeblick TA, Sandhu HS, et al. The use of rhBMP-2 in interbody fusion
cages. Definitive evidence of osteoinduction in humans: a preliminary report. Spine (Phila Pa 1976)
5. Burkus JK, Gornet MF, Dickman CA, et al. Anterior lumbar interbody fusion
using rhBMP-2 with tapered interbody cages. J Spinal Disord Tech
6. U.S. Food and Drug Administration. InFUSE bone graft/LT-CAGE lumbar tapered fusion
devices. Approval Letter 2002.
7. U.S. Food and Drug Administration. OP-1 Putty. Approval Letter 2003.
8. Ebara S, Nakayama K. Mechanism for the action of bone morphogenetic proteins and regulation of their activity. Spine (Phila Pa 1976)
9. U.S. Food and Drug Administration. InFUSE bone graft/LT-CAGE lumbar tapered fusion
devices: summary of safety and effectiveness. 2002.
9a. Dettori JR, Norvell DC, Dekutoski M, et al. Methods for systematic reviews on patient safety during spine surgery. Spine
10. Wright JG, Swiontkowski MF, Heckman JD. Introducing levels of evidence to the journal. J Bone Joint Surg Am
11. van Tulder M, Furlan A, Bombardier C, et al. Updated method guidelines for systematic reviews in the Cochrane collaboration back review group. Spine (Phila Pa 1976)
12. West S, King V, Carey TS, et al. Systems to rate the strength of scientific evidence. Evidence Report/Technology Assessment No. 47 (Prepared by the Research Triangle Institute-University of North Carolina Evidence-based Practice Center, Contract No. 290–97–0011). Rockville MD: Agency for Healthcare Research and Quality, 2002.
13. Benglis D, Wang MY, Levi AD. A comprehensive review of the safety profile of bone morphogenetic protein in spine surgery. Neurosurgery
14. Haid RW Jr, Branch CL Jr, Alexander JT, et al. Posterior lumbar interbody fusion
using recombinant human bone morphogenetic protein type 2 with cylindrical interbody cages. Spine J
15. Joseph V, Rampersaud YR. Heterotopic bone formation with the use of rhBMP2 in posterior minimal access interbody fusion
: a CT analysis. Spine (Phila Pa 1976)
16. Meisel HJ, Schnoring M, Hohaus C, et al. Posterior lumbar interbody fusion
using rhBMP-2. Eur Spine J
17. Kanayama M, Hashimoto T, Shigenobu K, et al. A prospective randomized study of posterolateral lumbar fusion
using osteogenic protein-1 (OP-1) versus local autograft with ceramic bone substitute: emphasis of surgical exploration and histologic assessment. Spine (Phila Pa 1976)
18. Burkus JK, Sandhu HS, Gornet MF. Influence of rhBMP-2 on the healing patterns associated with allograft interbody constructs in comparison with autograft. Spine (Phila Pa 1976)
19. Burkus JK, Dorchak JD, Sanders DL. Radiographic assessment of interbody fusion
using recombinant human bone morphogenetic protein type 2. Spine (Phila Pa 1976)
20. Mummaneni PV, Pan J, Haid RW, et al. 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
21. Singh K, Smucker JD, Gill S, et al. 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
22. Slosar PJ, Josey R, Reynolds J. Accelerating lumbar fusions by combining rhBMP-2 with allograft bone: a prospective analysis of interbody fusion
rates and clinical outcomes. Spine J
23. 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
24. Furlan JC, Perrin RG, Govender PV, et al. Use of osteogenic protein-1 in patients at high risk for spinal pseudarthrosis: a prospective cohort study assessing safety, health-related quality of life, and radiographic fusion
. Invited submission from the Joint Section on Disorders of the Spine and Peripheral Nerves, March 2007. J Neurosurg Spine
25. 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
26. Vaccaro AR, Patel T, Fischgrund J, et al. A pilot safety and efficacy study of OP-1 putty (rhBMP-7) as an adjunct to iliac crest autograft in posterolateral lumbar fusions. Eur Spine J
27. McKay B, Sandhu HS. Use of recombinant human bone morphogenetic protein-2 in spinal fusion
applications. Spine (Phila Pa 1976)
28. McClellan JW, Mulconrey DS, Forbes RJ, et al. Vertebral bone resorption after transforaminal lumbar interbody fusion
with bone morphogenetic protein (rhBMP-2). J Spinal Disord Tech
29. Pradhan BB, Bae HW, Dawson EG, et al. Graft resorption with the use of bone morphogenetic protein: lessons from anterior lumbar interbody fusion
using femoral ring allografts and recombinant human bone morphogenetic protein-2. Spine (Phila Pa 1976).
30. 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
31. Vaidya R, Weir R, Sethi A, et al. Interbody fusion
with allograft and rhBMP-2 leads to consistent fusion
but early subsidence. J Bone Joint Surg Br
32. Dhall SS, Wang MY, Mummaneni PV. Clinical and radiographic comparison of mini-open transforaminal lumbar interbody fusion
with open transforaminal lumbar interbody fusion
in 42 patients with long-term follow-up. J Neurosurg Spine
33. Taneichi H, Suda K, Kajino T, et al. Unilateral transforaminal lumbar interbody fusion
and bilateral anterior-column fixation with two Brantigan I/F cages per level: clinical outcomes during a minimum 2-year follow-up period. J Neurosurg Spine
34. Brantigan JW, Steffee AD, Lewis ML, et al. Lumbar interbody fusion
using the Brantigan I/F cage for posterior lumbar interbody fusion
and the variable pedicle screw placement system: two-year results from a Food and Drug Administration investigational device exemption clinical trial. Spine (Phila Pa 1976)
35. Boden SD, Kang J, Sandhu H, et al. 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 (Phila Pa 1976)
36. Luhmann SJ, Bridwell KH, Cheng I, et al. Use of bone morphogenetic protein-2 for adult spinal deformity. Spine (Phila Pa 1976)
37. Cahill KS, Chi JH, Day A, et al. Prevalence, complications
, and hospital charges associated with use of bone-morphogenetic proteins in spinal fusion
38. Mindea SA, Shih P, Song JK. Recombinant human bone morphogenetic protein-2-induced radiculitis in elective minimally invasive transforaminal lumbar interbody fusions: a series review. Spine (Phila Pa 1976)
39. Vaccaro AR, Lawrence JP, Patel T, et al. The safety and efficacy of OP-1 (rhBMP-7) as a replacement for iliac crest autograft in posterolateral lumbar arthrodesis: a long-term (>4 years) pivotal study. Spine (Phila Pa 1976)
40. Riley LH, Vaccaro AR, Dettori JR, et al. Postoperative dysphagia in anterior cervical spine surgery. Spine
41. Buttermann GR. Prospective nonrandomized comparison of an allograft with bone morphogenic protein versus an iliac-crest autograft in anterior cervical discectomy and fusion
. Spine J
42. Hiremath GK, Steinmetz MP, Krishnaney AA. Is it safe to use recombinant human bone morphogenetic protein in posterior cervical fusion
? Spine (Phila Pa 1976)
43. Shields LB, Raque GH, Glassman SD, et al. Adverse effects associated with high-dose recombinant human bone morphogenetic protein-2 use in anterior cervical spine fusion
. Spine (Phila Pa 1976)
44. Smucker JD, Rhee JM, Singh K, et al. Increased swelling complications
associated with off-label usage of rhBMP-2 in the anterior cervical spine. Spine (Phila Pa 1976)
45. Tumialan LM, Pan J, Rodts GE, et al. The safety and efficacy of anterior cervical discectomy and fusion
with polyetheretherketone spacer and recombinant human bone morphogenetic protein-2: a review of 200 patients. J Neurosurg Spine
46. Vaidya R, Carp J, Sethi A, et al. Complications
of anterior cervical discectomy and fusion
using recombinant human bone morphogenetic protein-2. Eur Spine J
47. Baskin DS, Ryan P, Sonntag V, et al. A prospective, randomized, controlled cervical fusion
study using recombinant human bone morphogenetic protein-2 with the CORNERSTONE-SR allograft ring and the ATLANTIS anterior cervical plate. Spine (Phila Pa 1976)
48. Crawford CH III, Carreon LY, McGinnis MD, et al. Perioperative complications
of recombinant human bone morphogenetic protein-2 on an absorbable collagen sponge versus iliac crest bone graft for posterior cervical arthrodesis. Spine (Phila Pa 1976)
49. Boakye M, Mummaneni PV, Garrett M, et al. Anterior cervical discectomy and fusion
involving a polyetheretherketone spacer and bone morphogenetic protein. J Neurosurg Spine
50. Rasmussen PA, Trost GR, Tribus C. Use of bone as an interbody fusion
device. Tech Neurosurg
51. U.S. Food and Drug Administration. FDA Public Health Notification: Life-threatening complications
associated with recombinant human bone morphogenetic protein in cervical spine fusion
52. 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
53. Laursen M, Hoy K, Hansen ES, et al. Recombinant bone morphogenetic protein-7 as an intracorporal bone growth stimulator in unstable thoracolumbar burst fractures in humans: preliminary results. Eur Spine J
54. Wong DA, Kumar A, Jatana S, et al. Neurologic impairment from ectopic bone in the lumbar canal: a potential complication of off-label PLIF/TLIF use of bone morphogenetic protein-2 (BMP
-2). Spine J
55. Brower RS, Vickroy NM. A case of psoas ossification from the use of BMP
-2 for posterolateral fusion
at L4–L5. Spine (Phila Pa 1976)
56. Perri B, Cooper M, Lauryssen C, et al. Adverse swelling associated with use of rh-BMP
-2 in anterior cervical discectomy and fusion
: a case study. Spine J
BMP; fusion; complications; adverse events
Supplemental Digital Content
© 2010 Lippincott Williams & Wilkins, Inc.