Anterior Cervical Discectomy and Fusion Using Escherichia coli-Derived Recombinant Human Bone Morphogenetic Protein-2: A Pilot Study

The patients’ demographic data and baseline characteristics are presented in Table 1. The average patient age was 57.7 years, and the mean follow-up period was 15.0 months. All clinical outcomes significantly improved at the final follow-up compared with the preoperative indices (Table 2). Although several cases showed moderate to severe airway swelling in simple radiographs taken 1 to 3 days after surgery (Fig. 4), there was only 1 case of dysphagia at the time point of discharge (patient 11), 1 case of cerebrospinal fluid leakage (patient 11), and 1 case of pneumomediastinum (patient 4), which improved with conservative treatment without readmission or reoperation. However, no cases of airway compromise, wound infection, neurologic deficit, hoarseness, heterotopic ossification, seroma, or malignancy were observed during the follow-up period.

Fig. 4
A 54-year-old male patient diagnosed with cervical spondylotic myeloradiculopathy underwent anterior cervical discectomy and fusion with bilateral uncinectomy and received 0.3 mg of Escherichia coli-derived bone morphogenetic protein-2 per segment. Lateral radiographs of the cervical spine taken at postoperative day 1 (A), postoperative day 2 (B), postoperative day 3 (C), and 3-month follow-up (D). There were no symptoms such as dyspnea or dysphagia, and prevertebral soft-tissue swelling improved without dexamethasone administration.

• Click for larger image
• Download as PowerPoint slide

Table 1
Demographics and Baseline Characteristics of the Patients

• Click for larger image
• Click for full table
• Download as Excel file

Table 2
Clinical Outcomes

• Click for larger image
• Click for full table
• Download as Excel file

Of the 21 segments, 15 segments showed solid fusion at 3 months after surgery, 4 segments at 6 months, and 1 segment at 12 months. Only 1 segment showed pseudoarthrosis, resulting in a fusion rate of 95.2%. Although ADH and PDH significantly decreased postoperatively and at the final follow-up, the differences in ADH and PDH were 0.7 ± 0.7 and 0.4 ± 0.6, respectively; thus, no cases of cage subsidence were found. No significant differences were observed in postoperative and final cervical alignments and segmental angles (Table 3).

Table 3
Radiological Outcomes

• Click for larger image
• Click for full table
• Download as Excel file

BMP-2, which has a striking osteoinduction effect, is widely used in spine fusion surgery, including ACDF, to overcome the shortcomings of autologous IBG.8) Frenkel et al.5) reported that the overall fusion rate for the patients who underwent multilevel ACDF without CHO–BMP-2 was 82.6% compared with a 100% fusion rate in the group with 0.2–0.4 mg of CHO–BMP-2 per level. In a prospective study by Buttermann3), ACDF grafting of 0.9 mg of CHO–BMP-2 per segment with allograft was as effective as autologous IBG in terms of clinical outcomes and fusion rates.

However, there have been several reports regarding the life-threatening complications of BMP-2 in ACDF resulting from postoperative neck swelling. Shields et al.21) reported complications of 23.2% after the use of high-dose CHO–BMP-2 (2.1 mg/level) and 8.6% of prolonged hospital stay or readmission due to dysphagia or airway compromise. A meta-analysis showed that the use of BMP-2 is a risk factor for dysphagia after ACDF, with an odds ratio of 5.52.18) However, the incidence of dysphagia or hoarseness was not statistically different in 215,047 patients who used BMP or not in a U.S. database review. Despite reports from the U.S. FDA considering these complication rates, BMP-2 is still used off-label in ACDF.8)

Recently, E.BMP-2 has been used to overcome the limitations of the high cost and low yield rate of CHO–BMP-2. In a clinical study by Cho et al.,11) lumbar posterolateral fusion (PLF) using 6 mg of E.BMP-2 per level bilaterally showed a 100% fusion rate 24 weeks after surgery. Choi et al.10) reported a 100% PLF rate using 2.5 mg of E.BMP-2 per level unilaterally in addition to lumbar interbody fusion, and a solid fusion was observed at a mean of 4.5 months after surgery. Son et al.12) compared grafting autologous IBG and E.BMP-2 when performing lumbar interbody fusion and an additional one-sided PLF. They concluded that 1.0 mg of E.BMP-2 was effective and safe in short-level lumbar PLF, reporting a 100% fusion rate. Wang et al.22) reported a union rate of 82.9% at 6 months postoperatively and 100% at 12 months postoperatively when grafting E.BMP-2 with β-tricalcium phosphate granules in ACDF. In previous studies using BMP-2 in spine fusion surgery, a solid union was observed within a relatively short follow-up period after surgery.10, 11, 12, 22) Likewise, 15 of 21 segments (71.4%) showed solid fusion at 3 months after surgery and 19 of 21 segments (90.4%) at 6 months after surgery in the present study. This is considered to be a great advantage of BMP-2. And the final fusion rate of 95.2% using E.BMP-2 was not lower than that of 84.8% to 100% using autologous IBG in previous studies.23)

Although the exact cause of postoperative prevertebral soft-tissue swelling when using BMP-2 in ACDF is not known, it is related to a robust inflammatory reaction and an excessive dose of BMP-2.16) In a meta-analysis, complication rates including reoperation, readmissions, reintubations, or tracheostomies were higher in cohorts with higher BMP doses per level and were significantly positively correlated with the dose of BMP used per level. Moreover, 0.2–0.6 mg of BMP-2 per level was enough to gain good fusion in the multilevel fusion of the cervical spine.24)

In the present study, only 1 case of dysphagia and no case of airway compromise were observed. We attributed these favorable findings to the fact that we performed the following perioperative procedures. First of all, a low dose of 0.3 mg of E.BMP-2 per level was used. Secondly, E.BMP-2 was used with HA as a carrier. BMP-2 should be used with a carrier, such as collagen sponge, for the best outcomes. Because of the poor mechanical strength of the collagen sponge matrix, BMP-2 is easily released under pressure, which can lead to an ectopic bone formation and robust inflammation, followed by neck swelling.25) In contrast, the BMP-2 solution is not bound to the surface of HA but is physically entrapped in the highly porous structure of HA granules. After implantation, E.BMP-2 is expected to be slowly and continuously released from HA through diffusion, desorption, or degradation with lower risk of early leakage causing adjacent soft-tissue inflammation such as severe neck swelling.13, 26) Third, oxidized cellulose was used for hemostasis, and fibrin sealants were used to prevent the spread of E.BMP-2.15, 16) Finally, dexamethasone was administered intravenously, if needed.17, 18)

Of the 11 patients, 1 patient (9.1%) had dysphagia, which was operated at the C3/4 level (patient 11). Although the sample size was small, it was within the incidence range of previous studies that reported the incidence of dysphagia from 1% to 79% in ACDF with or without BMP-2. It is difficult to conclude that there is a definite correlation between E.BMP-2 use and dysphagia in this patient because the upper surgical level at C3/4 is one of the risk factors for dysphagia after ACDF.18) However, as several cases showed significant airway swelling, the following prospective large study should address the soft-tissue swelling and dysphagia in a quantitative way, using more systematic and objective tools. It was supposed that cerebrospinal fluid leakage (patient 11) and pneumomediastinum (patient 4) were not related to E.BMP-2 use.

To the best of our knowledge, this is the first study to report the clinical results of E.BMP-2 using HA as a carrier in ACDF. However, this study has several limitations, including the small sample size without a control group, the short-term follow-up period, and the inability to establish an appropriate dose of E.BMP-2 for a solid and safe union. In addition, the use of oxidized cellulose, fibrin sealants, and dexamethasone might have helped to lower the risk of E.BMP-2 causing severe neck swelling. However, omitting these perioperative procedures poses an ethical challenge because the safety of E.BMP-2 in ACDF has not yet been established. It will be necessary to establish the efficacy and safety of E.BMP-2 with HA in ACDF compared to IBG through longer follow-up studies with higher levels of evidence in larger patient groups.