Trends in the Use of Bone Morphogenetic Protein–2 in Adult Spinal Deformity Surgery: A 10-Year Analysis of 54 054 Patients

INTRODUCTION

First discovered in 1965, bone morphogenetic protein–2 (BMP-2) is a protein that functions as a differentiation factor and acts on mesenchymal stem cells to produce bone formation.¹ Although only approved by the US Food and Drug Administration (FDA) for anterior lumbar interbody fusion with a cage,² the off-label use in spinal fusion surgery has tremendously increased in the past decade.³ However, BMP-2 use has been associated, among others, with complications such as seroma, hematoma, and ectopic bone formation, which led the FDA to announce a public health notification in 2008 regarding potential hazards of its use.⁴

Adult spinal deformity is known to affect up to 70% of the elderly patients, though only a small percentage ultimately require surgical intervention. Long-segment fusions are commonly used for scoliosis correction, but nonunion may affect approximately 17% of patients.⁵ BMP-2 has been promoted as a potential factor in decreasing the rate of pseudoarthrosis.⁶ Ruofeng et al² found in a 2015 study that the rate of BMP-2 use in patients over 65 years of age who underwent scoliosis surgery was 35.5%.

The purpose of our study is to analyze the trends in BMP-2 use for scoliosis surgery in patients over 21 years of age over a 10-year period, using a large nationwide inpatient database. Trends in BMP-2 use based on patient and operative characteristics are also reported, as well as factors associated with its use.

METHODS

RESULTS

A total of 54 054 patients who underwent adult spinal deformity (ASD) surgery between 2002 and 2011 were identified (Table 1). Average age at surgery was 59 ± 15 years; 27% of patients were male and 73% were female. Almost half of patients had Medicare as insurance (45.1%) and the other half had private insurance (46.9%). Almost half of patients had surgery in a Southern hospital (41.8%), followed by the Midwest region (27.6%). Over two-thirds of procedures were performed in teaching hospitals (63.1%), 11.6% were spinal refusion procedures, osteotomy was performed in 6.6% of cases, and fusion of 8 or more spinal segments was performed in 21.7% of all cases.

The overall rate of BMP-2 use was 39.7% (95% confidence interval 35.0%-44.3%). Based on age groups, the overall rate of BMP-2 use was 34.5% for patients aged 21 to 54, 42.6% for patients aged 55 to 64, 42.0% for patients aged 65 to 74, and 41.7% for patients aged 75 or older (P < .001). The rate of BMP-2 use was 40.5% for females and 37.6% for males (P = .009), 41.4% for white patients, 29.5% for black patients, and 36.4% for Hispanic/others (P = .006). When stratified by insurance status, BMP-2 use was 42.7% for Medicare, 31.8% for Medicaid, 37.9% for private insurance, and 34.4% for others (P = .006); its use was 28.6% in the Northeast, 45.2% in the Midwest, 38.1% in the South, and 48.2% in the West (P = .035). Teaching hospitals and nonteaching hospitals had similar rates of BMP-2 use (38.1% versus 42.6%, P = .388). Its use in primary fusion procedures was 39.3% versus 42.3% in revision procedures (P = .129); 39.7% in nonosteotomy procedures and 38.9% in osteotomy procedures (P = .785); 39.0% in posterior-only procedures, 33.6% in anterior-only procedures, and 44.3% in circumferential procedures (P = .045); and 42.4% in 1- to 7-segment fusions and 47.3% in 8+ level fusion procedures (P = .190). Average total hospital charges were $152,403 ± 117,454 for patients who did not receive BMP-2 and $205,426 ± 137,561 for patients who did (P < .001); this corresponded to a $53,023 difference.

Overall, there was a significant change in the rate of BMP-2 use over time (Figure 1). Figure 2, A to F, depicts rates of BMP-2 use stratified by age group (A), sex (B), race (C), insurance (D), hospital region (E), and hospital teaching status (F). On the other hand, Figure 3, A to C, depicts rates of BMP-2 use stratified by primary versus revision status (A), use of osteotomy (B), and number of levels fused (C). All changes over time were found to be statistically significant (all P < .001).

• Download figure
• Open in new tab
• Download powerpoint

Figure 1

Rate of BMP-2 use over time (P < .001).

• Download figure
• Open in new tab
• Download powerpoint

Figure 2

Rates of BMP-2 use stratified by (A) age group, (B) sex, (C) race, (D) insurance, (E) hospital region, and (F) hospital teaching status. All changes over time were found to be statistically significant (all P < .001).

• Download figure
• Open in new tab
• Download powerpoint

Figure 3

Rates of BMP-2 use stratified by (A) primary versus revision status, (B) use of osteotomy, and (C) number of levels fused. All changes over time were found to be statistically significant (all P < .001).

DISCUSSION

Although initially approved only for anterior lumbar interbody fusion, BMP-2 has been used for posterior lumbar fusion, posterior thoracic fusion, anterior cervical fusion, and posterior cervical fusion; for ASD patients, rates of BMP-2 use are approximately 35.5% to 61.6%.^2,8 In this study, we found a similar rate of 39.7%, with a steady increase in its use up to 2009 and then a decrease thereafter up to 2011.

Ruofeng et al² reported similar findings in their study of 29 787 patients over the age of 65 who underwent posterior long-segment fusion for scoliosis; the rate of BMP-2 use decreased also after 2009. However, Ruofeng et al² reported an increase after 2010, which is different from our findings. One of the possible explanations for this is that Ruofeng et al² only looked at posterior instrumentations, whereas our present study also included anterior and circumferential fusions. In fact, the highest rate of BMP-2 use was in patients who underwent anterior-posterior fusion (44.3%). Similarly, Martin et al³ reported a rapid increase in BMP-2 use for cervical and lumbar fusion operations (degenerative conditions) up to 2008 (involving up to 45.2% of lumbar and 13.5% of cervical operations); there was an average annual decrease of 11.7% thereafter. However, in our current study there was an increase between 2008 and 2009, which likely reflects different practices when using BMP-2 in degenerative conditions (as in the investigation by Martin et al³) versus deformity. The decrease in the use of BMP-2 after 2008/2009 is most likely the result of the FDA warning in 2008, which stated that the use of the osteoinductive agent for cervical spine operations was associated with serious and potentially life-threatening complications such as breathing/swallowing difficulty.² Likewise, McKie et al⁹ noted a decrease, albeit small, in the use of BMP-2 for anterior cervical fusion after the 2008 warning (5% decrease compared to the preadvisory era). Although the warning was issued primarily for cervical spine surgery, it is interesting that the use of BMP-2 use also decreased in lumbar degenerative disease and adult deformity surgery.

Nonetheless, the use of BMP-2 in adult scoliosis surgery has not been associated with a significant risk of complications in other investigations. In a report of 279 ASD patients from the International Spine Study Group, authors noted 172 patients (61.6%) in whom BMP-2 use was used.⁸ Although patients who received BMP had higher rates of total complications (1.4 versus 0.6 complications per patient) and minor complications (0.9 versus 0.2), BMP did not increase the risk for acute major, neurological, or wound complications.⁸ Furthermore, the use of BMP-2 has been associated with a significantly decreased risk of reoperation due to pseudoarthrosis after ASD surgery (5.0% versus 33.9% for patients in whom BMP-2 was not used).⁶

When looking at factors associated with BMP-2 use, we found that the use of BMP was higher in patients above 55 years of age, higher in females, white patients, and Medicare patients, and in the Midwest and West regions. Interestingly, the rate of use was not different for primary versus revision procedures, osteotomy versus no osteotomy, or fusion procedures involving less than 8 versus 8 or more segments. These findings suggest that there are certain patient/regional discrepancies in the use of BMP, but the exact cause is unknown and warrants further investigation.

Interestingly, the impact of BMP-2 use in hospital charges was significant, increasing the total charges on average by $53,023. While it has been shown that its use significantly decreases the need for further revision surgery due to pseudoarthrosis,⁶ a 2007 systematic review on BMP-2 use concluded that there is limited evidence to support greater improvement in clinical outcome for patients who use this agent, and that “the use of BMP for spinal fusion is unlikely to be cost-effective” given the high cost of this product.¹⁰ Although revision fusion procedures have been found to significantly increase the cost of surgery compared to primary procedures ($176,809 versus $161,791 in fusion procedures with 4+ levels),¹¹ no formal cost-effectiveness analysis in the ASD population has been performed yet, and this may be the next step in evaluating the long-term efficacy of BMP-2 use in scoliosis surgery. Specifically, it would be interesting to evaluate whether utilizing BMP and potentially preventing need for further revision surgery is more cost effective than not utilizing BMP and potentially increasing the risk for pseudoarthrosis/instrumentation failure.

CONCLUSION

After analysis of a large nationwide database, it was found that the rate of BMP-2 use in ASD surgery is approximately 40%. There was a significant increase in use from 2002 to 2009, and a decrease thereafter. The highest rates of use were found in older patients, female patients, white patients, Medicare patients, circumferential procedures, and patients undergoing surgery in the Midwest and West regions.