The results of two systematic reviews on the efficacy and safety of recombinant human bone morphogenetic protein-2 (rhBMP-2) in spinal fusion surgery have been published in the Annals of Internal Medicine. The reviews were conducted by two independent teams (the Centre for Reviews and Dissemination at the University of York and the Evidence-based Practice Center at Oregon Health and Science University) and were based on the same patient level data provided by the medical device company Medtronic. The company provided the two research teams with unprecedented access to all the data they had on their rhBMP-2 product (the Infuse Bone Graft), through the Yale University open data access (YODA) project.
As noted in a related BMJ news item, a previous review of industry-sponsored studies of rhBMP-2 and reports submitted to the US FDA found that data on important adverse effects was missing, and surgeons conducting the studies were found to have received large payments that were not declared. Following this publication, Medtronic announced that it would commission Yale to review all the data on its product; Yale then appointed the two teams to independently assess and reanalyse the provided data.
The authors of the York review embedded the individual patient data meta-analysis within a full systematic review. The analysis of effectiveness was limited to randomised controlled trials comparing rhBMP-2 with iliac crest bone graft (ICBG) in spinal surgery, whereas observational studies were also considered for safety. Although the results showed that rhBMP-2 increased the rates of successful fusion up to 24 months after surgery, there was substantial heterogeneity for this endpoint across the studies. In addition the difference between groups in terms of improvements in pain (3.5 percentage points on the Oswestry Disability Index [ODI; measures lower back pain on a scale of 0-100%]) was not considered to be clinically meaningful. There was no evidence of a consistent relationship between improvements in fusion due to rhBMP-2 and improvements in pain or function.
The review by the Oregon team included 13 RCTs – meta-analysis was conducted for the available individual patient data, and cohort studies were also described (total n=31). All but one RCT compared rhBMP-2 with ICBG. The results are presented according to fusion approach (cervical spine fusion) and the authors conclude overall that there was no difference between ICBG and rhBMP-2 in terms of fusion rates, pain or function, when used in anterior lumbar interbody fusion (ALIF) or posterior lumbar interbody fusion (PLIF). The data do not allow definitive conclusions to be made about effectiveness in other surgical approaches. The authors found that there was substantial evidence of reporting bias, with journal publications selecting analyses and results that favoured rhBMP-2 over ICBG.
The main sources of bias in the studies were lack of blinding of surgeons, patients, and outcome assessors (except for radiologic end points). The authors of the York assessment comment that the interpretation of pain could have been biased as participants were not blinded to the treatment received or their fusion status.
Safety findings from the studies show that a higher number of new cancer cases were reported in association with rhBMP-2 compared with ICBG. The overall absolute risk for cancer was however low in both groups, and it is uncertain whether this is a genuine risk (it is however consistent with the literature suggesting a possible link between BMP and cancer).
Data from observational studies suggest that rhBMP-2 may be associated with some specific adverse events (e.g. heterotopic bone formation, osteolysis, radiculitis), but these findings should be interpreted cautiously as they are based on heterogeneous non-randomised studies that provided little information about the comparability of groups.
The authors of the Oregon review conclude that it is difficult to identify a clear indication for rhBMP-2 in spinal fusion based on the currently available data, and they call for more research to provide more reliable estimates for risk for cancer and other adverse effects, and to identify patient populations for which it may be beneficial.
The authors of an accompanying editorial say that based on these findings, use of either autograft or rhBMP-2 to enhance fusion rates in patients having anterior lumbar interbody fusion or posterolateral fusion seems clinically reasonable. However rhBMP-2 should not be used in anterior cervical surgery without a compelling reason, given the higher complication rates noted. Patients should be counselled on the relative benefits and harms of each option and should be allowed to actively participate in decision making. As patients in the comparator groups in the RCTs received the gold standard ICBG, clinical decision making would need to take into account the probable lower fusion rates associated with the use of alternative graft materials (e.g. allograft; ceramic).
The author of another editorial states that this first YODA project “is a novel exercise that illustrates the value of evidence synthesis, data sharing, peer review and editing, and reproducible research in helping us get closer to the truth.”