Abstract
Bone morphogenetic protein (BMP) is commonly used as an ICBG substitute for transforaminal lumbar interbody spine fusion (TLIF). However, multiple recent reports have raised concerns regarding a substantial incidence of perioperative radiculopathy. Also, given the serious complications reported with anterior cervical BMP use, risks related to swelling and edema with TLIF need to be clarified. As TLIF related complications with rhBMP-2 have generally been reported in small series or isolated cases, without a clear denominator, actual complication rates are largely unknown. The purpose this study is to characterize perioperative complications and complication rates in a large consecutive series of TLIF procedures with rhBMP-2. We reviewed inpatient and outpatient medical records for a consecutive series of 204 patients [113 females, 91 males, mean age 49.3 (22-79) years] who underwent TLIF using rhBMP-2 between 2003 and 2007. Complications observed within a 3-month perioperative interval were categorized as to etiology and severity. Wound problems were delineated as wound infection, hematoma/seroma or persistent drainage/superficial dehiscence. Neurologic deficits and radiculopathies were analyzed to determine the presence of a clear etiology (screw misplacement) and identify any potential relationship to rhBMP-2 usage. Complications were observed in 47 of 204 patients (21.6%) during the 3-month perioperative period. Major complications occurred in 13 patients (6.4%) and minor complications in 34 patients (16.7%). New or more severe postoperative neurologic complaints were noted in 13 patients (6.4%), 6 of whom required additional surgery. These cases included one malpositioned pedicle screw and one epidural hematoma. In four patients (2.0%), localized seroma/hematoma in the area of the foramen caused neural compression, and required revision. In one additional patient, vertebral osteolysis caused foraminal narrowing and radiculopathy, but resolved without further surgery. Persistent radiculopathy without clear etiology on imaging studies was seen in six patients. Wound related problems were seen in six patients (2.9%), distributed as wound infection (3), hematoma/seroma (1) and persistent drainage/dehiscence (2). Overall, this study demonstrates a modest complication rate for TLIF using rhBMP-2. While perioperative complications which appeared specific to BMP usage were noted, they occurred infrequently. It will be necessary to weigh this incidence of complications against the complication rate associated with ICBG harvest and any differential benefit in obtaining a solid arthrodesis.