Complications of rhBMP-2 utilization for posterolateral lumbar fusions requiring reoperation: a single practice, retrospective case series report

Background context: Recombinant human bone morphogenetic protein-2 (rhBMP-2) (INFUSE, Medtronic, Memphis, TN, USA) has been used off-label for posterolateral lumbar fusions for many years.

Purpose: The goal of this study was to evaluate the complications requiring reoperation associated with rhBMP-2 application for posterolateral lumbar fusions.

Study design/setting: During a 7-year period of time (2002-2009), all patients undergoing lumbar posterolateral fusion using rhBMP-2 (INFUSE) were retrospectively evaluated within a large orthopedic surgery private practice.

Patient sample: A total of 1,158 consecutive patients were evaluated with 468 (40.4%) males and 690 (59.6%) females.

Outcome measures: Complications related to rhBMP were defined as reoperation secondary to symptomatic failed fusion (nonunion), symptomatic seroma formation, symptomatic reformation of foraminal bone, and infection.

Methods: Inclusion criteria were posterolateral fusion with rhBMP-2 implant and age equal to or older than 18 years. Surgical indications and treatment were performed in accordance with the surgeon's best knowledge, discretion, and experience. Patients consented to lumbar decompression and arthrodesis using rhBMP-2. All patients were educated and informed of the off-label utilization of rhBMP-2. Patient follow-up was performed at regular intervals of 2 weeks, 6 weeks, 12 weeks, 6 months, 1 year, and later if required or indicated.

Results: Average age was 59.2 years, and body mass index was 30.7 kg/m². Numbers of levels fused were 1 (414, 35.8%), 2 (469, 40.5%), 3 (162, 14.0%), 4 (70, 6.0%), 5 (19, 1.6%), 6 (11, 0.9%), 7 (7, 0.6%), 8 (4, 0.3%), and 9 (2, 0.2%). Patients having complications requiring reoperation were 117 of 1,158 (10.1%): symptomatic nonunion requiring redo fusion and instrumentation 41 (3.5%), seroma with acute neural compression 32 (2.8%), excess bone formation with delayed neural compression 4 (0.3%), and infection requiring debridement 26 (2.2%). Nonunion was related to male sex and previous BMP exposure. Seroma formation was significantly higher in patients with higher doses of rhBMP-2 (p=.050) and with more than 12 mg of rhBMP-2 (χ(2)=0.025). Bone reformation and neural compression at the laminectomy and foraminotomy sites occurred in a delayed fashion. Infection was associated with obesity and respiratory disease. Infections were noted with a greater BMP dose (p<.001), more than 12 mg (χ(2)<0.001), fusion more than three levels (χ(2)<0.001), and reexposed to BMP (χ(2)=0.023).

Conclusions: rhBMP-2 utilization for posterolateral lumbar fusions has a low symptomatic nonunion rate. Prior rhBMP-2 exposure and male sex were related to symptomatic nonunion formation. rhBMP-2-associated neural compression acutely with seroma formation and delayed with foraminal bone formation is concerning and associated with higher rhBMP-2 concentrations.