The use of recombinant human bone morphogenetic

Transcription

1 J Neurosurg Pediatrics 14: , 2014 AANS, 2014 Bone morphogenetic protein associated complications in pediatric spinal fusion in the early postoperative period: an analysis of 4658 patients and review of the literature Clinical article Brandon G. Rocque, M.D., 1 Mick P. Kelly, B.S., 2 Joseph H. Miller, M.D., 1 Yiping Li, M.D., 3 and Paul A. Anderson, M.D. 2 1 Pediatric Neurosurgery, Children s Hospital of Alabama, Birmingham, Alabama; and Departments of 2 Orthopedics and Rehabilitation and 3 Neurological Surgery, University of Wisconsin, Madison, Wisconsin Object. Use of recombinant human bone morphogenetic protein 2 has risen steadily since its approval by the FDA for use in anterior lumbar interbody fusion in The FDA has not approved the use of bone morphogenetic protein (BMP) in children. Age less than 18 years or lack of evidence of epiphyseal closure are considered by the manufacturer to be contraindications to BMP use. In light of this, the authors performed a query of the database of one of the nation s largest health insurance companies to determine the rate of BMP use and complications in pediatric patients undergoing spinal fusion. Methods. The authors used the PearlDiver Technologies private payer database containing all records from United Health- Care from 2005 to 2011 to query all cases of pediatric spinal fusion with or without BMP use. A review of the literature was also performed to examine the complications associated with BMP use in pediatric spinal fusion. Results. A total of 4658 patients underwent spinal fusion. The majority was female (65.4%), and the vast majority was age years (94.98%) and underwent thoracolumbar fusion (93.13%). Bone morphogenetic protein was used in 1752 spinal fusions (37.61%). There was no difference in the rate of BMP use when comparing male and female patients or age 10 years or older versus less than 10 years. Anterior cervical fusions were significantly less likely to use BMP (7.3%). Complications occurred in 9.82% of patients treated with versus 9.88% of patients treated without BMP. The complication rate was nearly identical in male versus female patients and in patients older versus younger than 10 years. Comparison of systemic, wound-related, CNS, and other complications showed no difference between groups treated with and without BMP. The reoperation rate was also nearly identical. Conclusions. Bone morphogenetic protein is used in a higher than expected percentage of pediatric spinal fusions. The rate of acute complications in these operations does not appear to be different in patients treated with versus those treated without BMP. Caution must be exercised in interpreting these data due to the many limitations of the administrative database as a data source, including the short length of follow-up. ( Key Words spine fusion bone morphogenetic protein 2 complications bone morphogenetic protein 7 pediatric allograft The use of recombinant human bone morphogenetic protein 2 ([rhbmp-2] or INFUSE; Medtronic, Inc.) has risen steadily since its approval by the FDA for use in anterior lumbar interbody fusion in Off-label use (anterior and posterior cervical, thoracolumbar, and posterolateral lumbar fusions) has accounted for the majority of the clinical use of rhbmp. 14 Recently there has been controversy regarding the increased risk of complications with bone morphogenetic protein (BMP) use compared with autograft bone, although 2 large population-based studies found no increase in the risk of complication with BMP use over other fusion material. 2,19 In the pediatric population, BMP is not approved for Abbreviations used in this paper: BMP = bone morphogenetic protein; ICD-9-CM = International Classification of Diseases, Ninth Revision, Clinical Modification; KID = Kids Inpatient Database; rhbmp-2 = recombinant human BMP 2; SRS M&M = Scoliosis Research Society Morbidity and Mortality. use in any indication. The package insert states that its use is contraindicated in patients who are skeletally immature (age less than 18 years or with no radiographic evidence of epiphyseal closure). In addition, in 2008 when the FDA issued a black box warning of the risk of rhbmp-2 causing airway complications in anterior neck surgery, they also reaffirmed that rhbmp-2 was contraindicated in skeletally immature patients. 16 However, there has been limited evaluation of the use of BMP in the pediatric population. Existing reviews of this topic are limited either by small sample size or by lack of longitudinal follow-up. Data from the Kids Inpatient Database (KID) for the year 2009 show that BMP was used in 9.2% of the 8289 pediatric spinal fusions performed that year. 6 Thus, in 2009 alone, more than 700 pediatric spinal fusions were performed using contraindicated BMP. Missing from the literature is any comprehensive evaluation of adverse effects associated with the use of BMP. The goal of this study was to better characterize 635

2 B. G. Rocque et al. the use of BMP in the pediatric population by using a large administrative database. Furthermore, we aimed to perform an analysis of acute complications and adverse events associated with the use of BMP in this population. In addition, we have performed a review of the literature regarding BMP use for pediatric spinal fusion. We hypothesize that the use of rhbmp-2 in children is not associated with a higher risk of acute complications in pediatric spinal fusion. Methods Review of the Literature Search Strategy. A search of PubMed ( ), Cochrane Database of Systematic Reviews and Cochrane Central Register of Controlled Trials ( ), and Google Scholar databases was conducted to identify articles of interest. These databases were searched using the key words spine, spine fusion, cervical spine, cervical spine fusion, thoracic, or lumbar combined with bmp, bmp 2, bmp 7, bone morphogenic protein, bone morphogenetic protein, or allograft and pediatric, pediatric use, child, childhood, or children. Searches were limited to the English language and to clinical studies rather than basic science data. The reference sections of included studies were also searched to identify relevant articles. The search was performed independently by 2 authors (B.G.R., Y.L.) and the results were combined. Study Selection. Search results were first screened by title and abstract. The abstracts of articles of interest were reviewed in detail. Case series, cohort studies, clinical trials, meta-analyses, and systematic reviews were all deemed eligible for inclusion. Individual case reports were excluded from this analysis (fewer than 5 individual case reports excluded). In addition, studies focusing on adult populations were excluded. Institutional Database We retrospectively analyzed the PearlDiver Technologies private payer database containing all orthopedicrelated diagnosis and procedure codes from all United HealthCare patient records from 2005 to United HealthCare is the largest private insurance carrier in the US, covering more than 70 million people in all 50 states. Detailed information about the geographic distribution of covered individuals is not available. The ICD-9-CM (International Classification of Diseases, Ninth Revision, Clinical Modification) procedure codes that identified patients who underwent spinal fusion are found in Table 1. The ICD-9-CM code and CPT code were used to specifically identify patients who underwent spinal fusion with BMP. The absence of this code implied that the fusion was done without the use of BMP (control cohort). Available covariates reported in the database were sex and age (grouped as age less than 10 or age years). Each patient was only enrolled once. Data capture includes both the initial hospital encounter and 90-day follow-up of each patient. Follow-up data are captured if the encounter includes at least one of the more than 2000 diagnosis or procedure codes considered orthopedic-related by the PearlDiver database. This includes all codes for neoplasm, hematoma, diabetes, atherosclerosis, embolism, arthritis, arthropathy, trauma, bone and joint disease, rheumatism, and complications. (See Appendix for a comprehensive list of included codes.) The privacy policy of PearlDiver Technologies, Inc., does not allow population of a given field in a database query if there are fewer than 11 patients who would meet those specific criteria. Therefore, some fields of inquiry cannot be evaluated. The University of Wisconsin Institutional Review Board provided exemption for this project. Postoperative Complications The ICD-9-CM diagnosis codes identified complications in the patient cohorts that occurred during the initial hospital stay and in the subsequent 90-day follow-up in which an orthopedic diagnosis or procedure code was entered. Patient data are available longitudinally as long as the patient remains insured by United HealthCare. Longitudinal follow-up on patients who change insurance providers is not available. Complications were categorized into 13 groups: cardiac; respiratory; peripheral vascular; central nervous system; hematoma; accidental cut, puncture, or hemorrhage during procedure; complications of operative wound; postoperative infection; other; CSF leak; deep vein thrombosis; mechanical complication of implant or graft; and death (Table 2). Repeat Operations The ICD-9 procedural codes were used to identify reoperations within 90 days of the index procedure. A list of ICD-9 codes commonly associated with reoperations was developed from a database search at the senior author s institution (Table 3). Statistical Analysis Chi-square analysis was used to compare the overall complication rates between the 2 treatment groups. Relative risk values and 95% confidence intervals were determined for the association of any complication with BMP usage compared with control. Statistical significance was defined as a Type I error (a) < Results Literature Review Our review of the literature identified 8 studies addressing the use of BMP with pediatric spinal fusion (see Table 4). Dodwell et al. used the KID, which is a sample of 4121 US hospitals, and identified 8289 spinal fusions performed in children in 2009; 6 BMP was used in 9.2% of these cases. In-hospital complications occurred in 3.0% of cases in which BMP was used and in 3.6% of cases in which no BMP was used. This difference was not statistically significant. Use of BMP was associated with older age, lumbosacral level of arthrodesis, fewer vertebrae 636

3 Complications associated with BMP in pediatric spinal fusion TABLE 1: The ICD-9-CM procedure codes identifying patients who underwent spinal fusion Procedure Group & Code Procedure Definition primary code atlas-axis spinal fusion other cervical fusion of the posterior column, posterior technique refusion of atlas-axis spine refusion of other cervical spine, posterior column, posterior technique other cervical fusion of the anterior column, anterior technique refusion of other cervical spine, anterior column, anterior technique dorsal and dorsolumbar fusion of the anterior column, anterior technique dorsal and dorsolumbar fusion of the posterior column, posterior technique lumbar and lumbosacral fusion of the anterior column, anterior technique lumbar and lumbosacral fusion of the posterior column, posterior technique lumbar and lumbosacral fusion of the anterior column, posterior technique refusion of lumbar and lumbosacral spine, anterior column, anterior technique refusion of lumbar and lumbosacral spine, posterior column, posterior technique refusion of lumbar and lumbosacral spine, anterior column, posterior technique use of BMP insertion of bone morphogenetic protein 20930* allograft, morselized, or placement of osteopromotive material, for spine surgery only * CPT (Current Procedural Terminology) code. fused, spondylolisthesis, neurofibromatosis, and surgery performed in the Midwest. Use of BMP was less likely in patients with Medicare or Medicaid health coverage and those treated in specialized pediatric hospitals. Because this sample was drawn from an inpatient database, only those complications that occurred during the initial hospital stay were captured. Similar findings are reported in another study of the KID records for Like Dodwell and colleagues, Jain et al. 10 report 9.3% overall BMP use in 2009, but also note a steady increase in BMP use, from 2.7% in Use of BMP in this study was more common with thoracolumbar fractures, patients with spondylolisthesis, private insurance, nonteaching hospitals, and no use of autograft. No analysis of complications was performed in this study. A report from the Scoliosis Research Society Morbidity and Mortality (SRS M&M) Committee examined the incidence of perioperative complications in 55,862 cases of spinal fusion with and without BMP. 19 Of these cases, 17,513 were pediatric scoliosis operations, and 9% (1576) used BMP. This rate of BMP use is consistent with the rate reported by Dodwell et al. The complication rate for these patients was 7.0% without BMP and 8.8% with BMP (difference not statistically significant). Unlike the KID database, the SRS M&M database contains data entered by the primary surgeon. Therefore, it may be affected by a selection bias, but it also contains follow-up data, rather than only information about the initial hospitalization and in-hospital complications. A series of 81 pediatric patients underwent 91 orthopedic procedures including the use of BMP. 13 This included indications other than spine surgery, but the majority of procedures (52) were spinal fusion. In the total cohort, there were 9 operative site complications and 7 other complications (17.5% total complication rate). The authors note no systemic toxicity and no excessive bone growth. A patient with optic glioma had enlargement of the tumor, and another patient developed dural fibrosis requiring reoperation after BMP exposure. A single-center case series of 48 patients, ages 3 to 18 years, undergoing occipital cervical decompression and fusion with rhbmp (4.2 mg, INFUSE; Medtronic, Inc.) found 6 complications that were presumed to be related to the use of BMP. 12 Five of these were epidural fluid collections, with 2 requiring operative intervention for neurological compromise. Although there is no comparison group in this publication, the authors note that in their previous experience with this procedure in this population of patients, no epidural fluid collections had been encountered. They conclude that the use of rhbmp was the most likely factor leading to this complication. In contrast, another single center reviewed 17 cases of pediatric spinal arthrodesis, 7 of which included the occiput. 1 These authors concluded that no BMP-related complications occurred. A third single-institution study included 19 patients undergoing occipitocervical, cervical, thoracic, lumbar, or lumbosacral fusion with BMP. 7 There was 1 case of bony overgrowth causing neural compression and requiring reoperation. There were also 2 superficial wound infections. Finally, in a series of 11 patients with neuromuscular scoliosis, deemed to be high risk for pseudarthrosis at L5 S1, BMP was used at that level only in the setting of a long fusion to the pelvis. 9 One patient had a pseudarthrosis and one had a wound infection. There were no other identified BMP-related complications. 637

4 B. G. Rocque et al. TABLE 2: The ICD-9-CM codes identifying complications* Category & Diagnosis Code Diagnosis Definition cardiac complications cardiac complications acute myocardial infarction postoperative shock respiratory complications respiratory complications from a procedure other respiratory complications, Mendelson s syndrome respiratory complications, pulmonary embolism peripheral vascular complications peripheral vascular complications (includes phlebitis, thrombophlebitis) CNS complications (CNS) nervous system complications nervous system complications, unspecified central nervous system complication: anoxic brain damage or cerebral hypoxia (not stroke or infarction [997.02]) other nervous system complications complications of hematomas hemorrhage or hematoma complicating a procedure, NEC hemorrhage complicating a procedure hematoma complicating a procedure seroma complicating a procedure complication of accidental cut, puncture, or hemorrhage during a procedure accidental puncture or laceration during a procedure E870.0 accidental puncture or laceration during a surgical operation complication of operative wound disruption of operative wound disruption of internal operation wound disruption of external operation wound non-healing surgical wound postoperative infection 998.5, postoperative infections infected postoperative seroma other postoperative infection (abscess) other complications other specified complications of procedures, NEC other specified complications unspecified complication of procedure, NEC unspecified or unclassified complications of medical care, NEC; unspecified misadventure of medical care E878.8 surgical operation and other surgical procedures as the cause of abnormal reaction of patient, or of later complication, without mention of misadventure at the time of operations; other specified surgical operations and procedures E878.9 surgical operation and other surgical procedures as the cause of abnormal reaction of patient, or of later complication, without mention of misadventure at the time of operations; unspecified surgical operations and procedures CSF leak reaction to spinal or lumbar puncture persistent postoperative fistula, NEC (continued) 638

5 Complications associated with BMP in pediatric spinal fusion TABLE 2: The ICD-9-CM codes identifying complications* (continued) Category & Diagnosis Code Diagnosis Definition deep vein thrombosis (DVT) acute venous embolism and thrombosis of unspecified deep vessels of lower extremity acute venous embolism and thrombosis of deep vessels of proximal lower extremity acute venous embolism and thrombosis of deep vessels of distal lower extremity embolism and thrombosis of unspecified site mechanical complication of implant or graft mechanical complication of internal orthopedic device, implant, and graft unspecified mechanical complication of internal orthopedic device, implant, and graft mechanical loosening of prosthetic joint dislocation of prosthetic joint broken prosthetic joint implant peri-prosthetic fracture around prosthetic joint peri-prosthetic osteolysis articular bearing surface wear of prosthetic joint other mechanical complication of prosthetic joint implant other mechanical complication of other internal orthopedic device, implant, and graft death instantaneous death * NEC = not elsewhere classified. Database Query Utilization of BMP. A total of 4658 pediatric patients underwent spinal fusion between 2005 and The majority of the patients were female (3048, 65.44%), and the vast majority of the patients were between 10 and 19 years of age (4424, 94.98%). There were 234 patients (5.02%) younger than 10 years. Further subdivision of patient ages was not possible using the PearlDiver database. Thoracolumbar fusion surgery accounted for 4338 (93.13%) of the operations, with anterior and posterior cervical surgery accounting for the remainder of the database entries. There were 193 anterior cervical fusion operations. The exact number of posterior cervical operations was not available. Table 5 shows the sex, age, and location of surgery. Of the 4658 pediatric patients who underwent spinal fusion, 1752 (37.6%) had operations in which BMP was used. A higher proportion of female patients compared with male patients undergoing spinal fusion had operations in which BMP was used (39.4% vs 34.2%, p = 0.001). The rate of BMP use was exactly the same in patients less than 10 years compared with those years (37.6%, p = 0.998). Thoracolumbar spinal fusions comprised the majority of operations in the database. Of the 4338 thoracolumbar fusion operations, 1714 (39.5%) used BMP. There was no difference in BMP use comparing male with female patients undergoing thoracolumbar fusion (male 39.4%, female 40.5% BMP use; p = 0.499). Likewise, there was no difference between patients less than 10 years compared with those years who underwent thoracolumbar fusion (39.8% vs 39.5% BMP use, p = 0.996). Regarding anterior cervical fusions, 14 of 193 patients (7.3%) had documented BMP use. This is a significantly lower rate than for thoracolumbar fusions (39.5%, p < ). Insufficient details are available to compare male and female patients in this group. In patients less than 10 years of age, 2 of 10 had fusions in which BMP was used, compared with 12 of 183 patients years old (20% vs 6.6%, p = 0.111). Thirty-five patients underwent posterior cervical fusion in which BMP was used. Data regarding posterior cervical fusion without BMP are not available due to the PearlDiver privacy policy (fewer than 11 cases). Complications. Of the 1752 cases in which spinal fusion involved BMP, 172 patients (9.8%) had a complication. Of the 2906 patients without BMP, 287 (9.9%) had a complication (Fig. 1). This difference was not statistically significant (p = 0.948). The complication rate for male patients was 11.4% in those with BMP versus 10.7% in those without BMP. For female patients, complications with BMP occurred in 9.1% versus 9.4% in those without BMP. Neither of these differences was statistically significant (p = for male patients, p = for female patients). The difference between the percentage of male and female patients treated with BMP was also not significant. Insufficient data are available to analyze the complication rate for anterior or posterior cervical spinal fusions. Considering only thoracolumbar fusions, the complication rate is almost identical to the complication rate for the entire sample. This is not surprising, given the large proportion of the sample that is made up of thoracolumbar fusions. We have attempted to better characterize complications by grouping them into systemic, wound, CNS, and other complications. Among the 1752 patients in whose fusions BMP was used, there were 38 systemic complica- 639

6 B. G. Rocque et al. TABLE 3: The ICD-9-CM codes identifying reoperation within 90 days of index procedure ICD-9-CM Code Description excision of bone for graft, unspecified site excision of other bone for graft, except facial bones internal fixation of other bone, except facial bones, without fracture reduction spinal fusion, not otherwise specified atlas-axis spinal fusion other cervical fusion of the anterior column, anterior technique other cervical fusion of the posterior column, posterior technique dorsal and dorsolumbar fusion of the anterior column, anterior technique dorsal and dorsolumbar fusion of the posterior column, posterior technique lumbar and lumbosacral fusion of the anterior column, anterior technique lumbar and lumbosacral fusion of the posterior column, posterior technique lumbar and lumbosacral fusion of the anterior column, posterior technique refusion of atlas-axis spine refusion of other cervical spine, anterior column, anterior technique refusion of other cervical spine, posterior column, posterior technique refusion of dorsal and dorsolumbar spine, anterior column, anterior technique refusion of dorsal and dorsolumbar spine, posterior column, posterior technique refusion of lumbar and lumbosacral spine, anterior column, anterior technique refusion of lumbar and lumbosacral spine, posterior column, posterior technique refusion of lumbar and lumbosacral spine, anterior column, posterior technique fusion or refusion of 2 3 vertebrae fusion or refusion of 4 8 vertebrae fusion or refusion of 9 or more vertebrae insertion of interbody spinal fusion device insertion of other spinal device 3.09 other exploration and decompression of spinal canal 3.59 other repair and plastic operations on spinal cord structures excision of intervertebral disc other incision of bone without division, unspecified site other incision of other bone, except facial bones, without division local excision of lesion or tissue of bone, unspecified site local excision of lesion or tissue of other bone, except facial bones other incision of soft tissue excision of lesion of other soft tissue other fasciectomy other myectomy other incision with drainage of skin and subcutaneous tissue excisional debridement of wound, infection, or burn nonexcisional debridement of wound, infection, or burn closure of skin and subcutaneous tissue of other sites removal of implanted device from other bone tions (2.2%), 72 wound complications (4.1%), 12 CNS complications (0.68%), and 108 other complications (6.2%). In the group without BMP (2906 total patients), there were 60 systemic (2.1%), 107 wound (3.7%), and 201 (6.9%) other complications. None of these differences between groups was statistically significant. There were insufficient CNS complications in patients treated without BMP for this number to be reported in our database query. Twenty-four of 1752 patients (1.4%) treated with BMP required reoperation compared with 40 of 2906 (1.4%) treated without BMP. Due to the length of follow-up (90 days), no assessment of cancer or tumor incidence was possible. Discussion Bone morphogenetic protein, specifically rhbmp-2, 640

7 Complications associated with BMP in pediatric spinal fusion TABLE 4: Systematic review of the literature on BMP use in pediatric spinal fusion Authors & Year No. of Patients Complication Notes Dodwell et al., % with BMP; 3.6% without BMP KID study 2009 Jain et al., no analysis KID records Williams et al., ,513; scoliosis surgery 7.0% without BMP; 8.8% with BMP SRS M&M database Oetgen & Richards, ; 52 spinal fusions 17.5% single institution; no control Lindley et al., ; occipital cervical only 6 (12.5%) single institution; no control Abd-El-Barr et al., ; 7 included occiput none single institution; no control Fahim et al., (15.7%) single institution; no control Gressot et al., ; neuromuscular scoliosis 2 (18%) single institution; no control has been proven to increase the fusion rate in selected spinal fusion operations. However, complications such as seroma formation, tissue edema, and heterotopic bone formation have been attributed to BMP use. In the adult population these and other complications have been extensively studied, and controversy persists regarding the reported rate of adverse effects of BMP in adults. 4,8 In children, there is no large, controlled study assessing potential adverse effects of BMP use. The FDA has not approved BMP use in children, and product labeling lists among the contraindications age less than 18 or no radiographic evidence of epiphyseal closure. Bone morphogenetic protein plays an important role in skeletogenesis as well as in the development of various other tissues. 15 How surgical administration of BMP may influence the developing skeletal system (or any other immature organ system in children) is unknown. In addition, BMP is a member of the transforming growth factor beta family, dysregulation of which has been implicated in oncogenesis. Data from multiple studies pooled by the FDA showed a statistically insignificant increased cancer rate in patients with BMP exposure compared with those treated without BMP ( downloads/advisorycommittees/committees%20meet ingmaterials/medicaldevicesadvisorycommittee/ortho paedicand%20rehabilitationdevicespanel/ucm pdf). Likewise, a re-evaluation of the publicly available data on high-dose rhbmp-2 (AMPLIFY; Medtronic, Inc.) shows a significant increase in the incidence of new cancer events at 2 years for patients exposed to BMP compared with controls (3.37 new cancer events per 100 person-years vs 0.50 events per 100 person-years). 3 Another administrative database study of more than 4000 patients found an increased risk of benign neoplasms with BMP exposure (OR 1.31, 95% CI ), but not of malignancies. 11 None of these studies included pediatric patients, and the cancer risk for children has not been assessed. Importantly, the methodology of the present study did not allow for longterm follow-up, and thus cannot address cancer risk. An extensive literature review confirmed the dearth of studies evaluating the safety of BMP in children. Three reports used large administrative databases to evaluate use of BMP in spinal fusion in children, and 2 of these studies assessed complications. Neither study showed a difference in complication rate with BMP use. Six single-institution studies reported a complication rate with BMP use from 0% to 18%, but all were retrospective and none included a control arm. From available data, conclusions about BMPrelated complications in children are hard to draw. We have queried a large, private payer database and report the use of BMP in 37.61% of pediatric spinal fusions performed between 2005 and This rate of BMP utilization is markedly higher than the published rate of approximately 9% in previous studies based on administrative databases. 6,10,19 Previous work has suggested a higher use of BMP in patients covered by private insurance. 10 The current study is based on a database composed solely of patients from a single private insurance corporation. This is one factor that may contribute to the 4-fold increase in utilization. Other factors previously noted to be associated with BMP use, such as teaching hospital, region of the country, and presence of spondylolisthesis cannot be analyzed using this dataset. Dodwell, Snyder, and Wright found older age to be associated with increased BMP use. 6 TABLE 5: Summary of patients with and without BMP use Factor BMP Used No BMP Used % w/ BMP Total total patients male female age 10 yrs age yrs thoracolumbar fusion posterior cervical 35 * * 35 anterior cervical * Data not available due to PearlDiver privacy policy. 641

8 B. G. Rocque et al. Fig. 1. Bar graph showing complications of spinal surgeries performed with and without BMP use. Age is given in years. Although we are unable to report the ages of included patients, the majority of the patients in the present dataset are age years. It is possible that age is another factor contributing to the high rate of BMP use. The purpose of this study has been to compare the rate of complication with BMP use to that of spinal fusion without BMP use. We have found no difference. Furthermore, there is no subgroup of patients for whom there is a noticeable or significant difference in the complication rate with versus without BMP use. Importantly, however, due to the small numbers, an analysis of the complications occurring after cervical spine fusions, either anterior or posterior, cannot be performed using this database. Previous work has suggested that a higher rate of complications is likely with BMP use in the cervical spine. 16 The overall complication rate of 9.8% is higher than what has previously been reported from administrative database research, but close to the rate reported by singleinstitution retrospective case series. We have been as inclusive as possible in our definition of complications and reoperation (see Tables 2 and 3) in an attempt to capture as many potential complications as possible. Because of this, it is not surprising that the complication rate in this study is higher than in previous database studies. Furthermore, unlike studies based on inpatient samples, our methodology allowed for the collection of complications occurring later than the initial hospitalization. This is important for the assessment of an agent like BMP, in which wound complication or neurological compromise from seroma or inflammation might become apparent after the initial hospital discharge. Limitations of the Study The PearlDiver database contains data from a single health care provider. Thus, the patients included probably share demographic and socioeconomic factors that are not quantified. The vast majority of patients are years old, and most are girls. This severely limits the generalizability of this study to the pediatric population at large. United HealthCare insures individuals throughout the US, but we have no information about the geographic distribution or socioeconomic status of patients included in this database query. We have no information about the number of spinal levels fused, the indications for fusion, the exact surgical approach, the dose of BMP used, or the exact BMP product used (although it is presumed that the BMP product used is nearly exclusively rhbmp-2, or INFUSE). In addition, our methodology does not provide any way to determine if autologous bone is harvested for autograft. Presumably, the use of BMP would minimize the need for autologous bone graft harvest. Complications that result from the harvest of autograft would probably be higher in the group without BMP, but there is no way to evaluate this effect. Ninety-day follow-up allows the identification of complications outside the initial hospitalization, but may miss long-term complications such as cancer. There is no method available to determine if a patient is lost to follow-up, nor is the rate of cohort maintenance available. The database is constructed from claims made to insurance, so a patient who has lost or changed coverage and then had a complication would not be captured. This limitation would apply to patients treated both with and without BMP, and would probably not represent a significant proportion of patients, given that this entire cohort is pediatric patients covered under parental insurance. Complication identification is performed using ICD-9 codes. It is possible that this leads to underreporting due to missed codes or to overreporting on a hospital s part due to common efforts to maximize reporting of comorbid conditions. Although prospective data would be preferred, ICD-9 codes have been shown to be a reliable way to identify adverse events. 17,18 A central assumption of this work is that the ICD-9 codes associated with the use of BMP are accurately reported. We believe that this assumption is valid given the high cost of BMP. This assumption has been made in similar large database studies of adult populations. 5 Since this database is drawn from an insurance provider who is ultimately paying for the use of BMP, we believe that the coding for BMP is likely to be correct. However, due to PearlDiver s privacy policy, no individual field can be populated if the total of patients is fewer than 11. This will result in failure to report low-frequency events that may bias the results. The primary objective of the use of BMP is to promote bony fusion. Since the length of follow-up in the present study is limited to 90 days, there is no way to assess the success of fusion; we are only able to assess short-term complications. Finally, some complications such as ectopic bone formation and radiculitis are difficult to capture using ICD-9 codes. The only covariates available for analysis are age 10 or more years or less than 10 years, location of surgery, and sex. Comparison of complications within and between these subgroups yields no difference in complication rate. We are unable to perform more rigorous statistical analysis of risk due to small numbers of patients in the less than 10- year age group and in the cervical surgery groups. The results of this study can be viewed as reassuring; increased rates of complications do not appear to occur in pediatric patients, although case reports of the use of BMP at the craniocervical junction show potential for significant neurological complications. The FDA considers this product to be contraindicated, and therefore the surgeon should only use it when absolutely needed. If used, the lowest dose 642

9 Complications associated with BMP in pediatric spinal fusion possible to achieve the desired effect should be administered. Unfortunately, dosing has not been critically evaluated in pediatric patients and would be expected to vary by age and location of the fusion. Additionally, a proper informed consent should include a discussion of the offlabel risks of BMP use in pediatric patients. Conclusions We have queried a large administrative database to examine complications after spinal fusion in children with and without the use of BMP. Whereas we found utilization of BMP in this population to be surprisingly high, we found no difference in the rate of complication. It is apparent from this work and that of others that BMP is used frequently in pediatric spinal fusion. The present work suffers many limitations and cannot adequately address the safety of this agent in children. Although adverse effects of BMP use in children have been reported, they have not been adequately studied: prospective or registry-based assessments of outcomes are essential. Given the large numbers of spinal fusions performed each year in which BMP is used, accrual to such a study should be feasible and would contribute valuable information regarding both the beneficial and adverse effects of this biological agent on the young spine and the young patient. Disclosure The authors report no conflict of interest concerning the materials or methods used in this study or the findings specified in this paper. Author contributions to the study and manuscript preparation include the following. Conception and design: Anderson, Rocque. Acquisition of data: Kelly, Miller, Li. Analysis and interpretation of data: Anderson, Rocque, Kelly, Miller. Drafting the article: Anderson, Rocque, Miller. Critically revising the article: Anderson, Rocque. Reviewed submitted version of manuscript: all authors. Approved the final version of the manuscript on behalf of all authors: Anderson. Statistical analysis: Rocque. Administrative/technical/ material support: Anderson, Li. Study supervision: Anderson. Appendix This article contains an appendix that is available only in the online version of the article. References 1. Abd-El-Barr MM, Cox JB, Antonucci MU, Bennett J, Murad GJA, Pincus DW: Recombinant human bone morphogenetic protein-2 as an adjunct for spine fusion in a pediatric population. Pediatr Neurosurg 47: , Cahill KS, Chi JH, Day A, Claus EB: Prevalence, complications, and hospital charges associated with use of bone-morphogenetic proteins in spinal fusion procedures. JAMA 302: 58 66, Carragee EJ, Chu G, Rohatgi R, Hurwitz EL, Weiner BK, Yoon ST, et al: Cancer risk after use of recombinant bone morphogenetic protein-2 for spinal arthrodesis. J Bone Joint Surg Am 95: , Carragee EJ, Hurwitz EL, Weiner BK: A critical review of recombinant human bone morphogenetic protein-2 trials in spinal surgery: emerging safety concerns and lessons learned. Spine J 11: , Cooper GS, Kou TD: Risk of cancer after lumbar fusion surgery with recombinant human bone morphogenic protein-2 (rh-bmp-2). Spine (Phila Pa 1976) 38: , Dodwell E, Snyder B, Wright J: Off-label use of bone morphogenetic proteins in pediatric spinal arthrodesis. JAMA 308: , Fahim DK, Whitehead WE, Curry DJ, Dauser RC, Luerssen TG, Jea A: Routine use of recombinant human bone morphogenetic protein-2 in posterior fusions of the pediatric spine: safety profile and efficacy in the early postoperative period. Neurosurgery 67: , Glassman SD, Howard J, Dimar J, Sweet A, Wilson G, Carreon L: Complications with recombinant human bone morphogenic protein-2 in posterolateral spine fusion: a consecutive series of 1037 cases. Spine (Phila Pa 1976) 36: , Gressot LV, Patel AJ, Hwang SW, Fulkerson DH, Jea A: Rh- BMP-2 for L5-S1 arthrodesis in long fusions to the pelvis for neuromuscular spinal deformity in the pediatric age group: analysis of 11 patients. Childs Nerv Syst 30: , Jain A, Kebaish KM, Sponseller PD: Factors associated with use of bone morphogenetic protein during pediatric spinal fusion surgery: an analysis of 4817 patients. J Bone Joint Surg Am 95: , Lad SP, Bagley JH, Karikari IO, Babu R, Ugiliweneza B, Kong M, et al: Cancer after spinal fusion: the role of bone morphogenetic protein. Neurosurgery 73: , Lindley TE, Dahdaleh NS, Menezes AH, Abode-Iyamah KO: Complications associated with recombinant human bone morphogenetic protein use in pediatric craniocervical arthrodesis. Clinical article. J Neurosurg Pediatr 7: , Oetgen ME, Richards BS: Complications associated with the use of bone morphogenetic protein in pediatric patients. J Pediatr Orthop 30: , Ong KL, Villarraga ML, Lau E, Carreon LY, Kurtz SM, Glassman SD: Off-label use of bone morphogenetic proteins in the United States using administrative data. Spine (Phila Pa 1976) 35: , Thawani JP, Wang AC, Than KD, Lin CY, La Marca F, Park P: Bone morphogenetic proteins and cancer: review of the literature. Neurosurgery 66: , US Food and Drug Administration: FDA Public Health Notification: Life-Threatening Complications Associated With Recombinant Human Bone Morphogenetic Protein in Cervical Spine Fusion. Silver Spring, MD: FDA, 2008 ( publichealthnotifications/ucm htm) [Accessed August 19, 2014] 17. Wang MC, Chan L, Maiman DJ, Kreuter W, Deyo RA: Complications and mortality associated with cervical spine surgery for degenerative disease in the United States. Spine (Phila Pa 1976) 32: , Weingart SN, Iezzoni LI, Davis RB, Palmer RH, Cahalane M, Hamel MB, et al: Use of administrative data to find substandard care: validation of the complications screening program. Med Care 38: , Williams BJ, Smith JS, Fu KMG, Hamilton DK, Polly DW Jr, Ames CP, et al: Does bone morphogenetic protein increase the incidence of perioperative complications in spinal fusion? A comparison of 55,862 cases of spinal fusion with and without bone morphogenetic protein. Spine (Phila Pa 1976) 36: , 2011 Manuscript submitted December 28, Accepted August 18, Please include this information when citing this paper: published online October 10, 2014; DOI: / PEDS Address correspondence to: Paul A. Anderson, M.D., Department of Orthopedics & Rehabilitation, University of Wisconsin, UWMF Centennial Bldg., 1685 Highland Ave., 6th Fl., Madison, WI anderson@ortho.wisc.edu. 643