Skip Navigation LinksHome > Blogs > The Spine Blog
The Spine Blog
Friday, October 17, 2014

The indications for the use of bone morphogenetic protein (BMP) in spinal fusion remain unclear. While prior studies have shown an increased radiographic fusion rate, there is little data to suggest lower re-operation rates or better patient reported outcomes when used in good fusion environments such as single level posterolateral instrumented lumbar fusion.1,2 Additionally, BMP may be associated with complications such as increased post-operative pain, edema and airway compromise in the anterior cervical spine, retrograde ejaculation in the anterior lumbar spine, graft subsidence, and possibly cancer. In order to add further data to the BMP discussion, Dr. Guppy and his colleagues at Kaiser Permanente used their spine surgery registry to determine reoperation rates for nonunion among over 9,000 spinal fusion patients, including every region of the spine. For patients with at least one year of follow-up, the overall reoperation rate for nonunion was 1.9% for BMP patients and 2.2% for non-BMP patients. These rates increased to 2.3% and 2.6%, respectively, for those with at least 2 years of follow-up. The BMP and non-BMP groups were markedly different, with the BMP patients being older, including more females, more deformity, spondylolisthesis, and post-laminectomy syndrome patients, more likely to have undergone long thoracolumbar fusion, and less likely to have undergone cervical fusion. The authors created a Cox regression model to control for these differences, which demonstrated a 33% reduction in reoperation for nonunion for the BMP patients, but this difference was not significant.


This paper supports prior studies that have suggested that BMP use does not reduce reoperation rates in most fusion environments. The conclusions that can be made from this study are limited by the typical shortcomings of administrative databases, which are even more pronounced when the two groups being compared are so different at baseline. Given these differences in measured characteristics, one must assume that they are also different in unmeasured characteristics (i.e. smoking, prior surgery, psychosocial characteristics, etc.) for which the analysis cannot be controlled. Additionally, the indications for reoperation for nonunion are very variable among surgeons and also depend on patients’ willingness to undergo reoperation. The use of BMP is also markedly variable among surgeons, so the relationship between BMP use and reoperation is likely confounded by the individual surgeons making these decisions. Reoperation for nonunion is also a relatively rare event (under 3% over 2 years in this study), which results in very limited power to perform subgroup analyses even when starting with nearly 10,000 patients. While this study does have limitations, it adds data to the growing pile of evidence suggesting that BMP is not indicated for straightforward fusion operations in good fusion environments where the likelihood of fusion is high without it. The spine community has yet to determine where it might be of benefit, though some data as well as general principles suggest that it may make a positive difference in challenging fusion environments such as long deformity constructs or established nonunion.3


Please read Dr. Guppy’s article in the October 15 issue. Does this paper change how you see the role of BMP in spinal fusion? Let us know by adding a comment on The Spine Blog.

Adam Pearson, MD, MS

Associate Web Editor




1.            Fu R, Selph S, McDonagh M, et al. Effectiveness and harms of recombinant human bone morphogenetic protein-2 in spine fusion: a systematic review and meta-analysis. Annals of internal medicine 2013;158:890-902.

2.            Simmonds MC, Brown JV, Heirs MK, et al. Safety and effectiveness of recombinant human bone morphogenetic protein-2 for spinal fusion: a meta-analysis of individual-participant data. Annals of internal medicine 2013;158:877-89.

3.            Kim HJ, Buchowski JM, Zebala LP, Dickson DD, Koester L, Bridwell KH. RhBMP-2 is superior to iliac crest bone graft for long fusions to the sacrum in adult spinal deformity: 4- to 14-year follow-up. Spine 2013;38:1209-15.


Friday, October 10, 2014

Surgical treatment of metastatic and primary spinal tumors is complex and associated with high complication rates. While surgery of this magnitude would be expected to result in a high complication rate in healthy patients, spinal tumor patients typically have multiple comorbidities and in many cases compromised immune systems and soft-tissues due to chemotherapy and radiation. This creates a perfect storm for generating complications by combining complex surgery with sick, deconditioned patients. Recently, the Center for Medicare and Medicaid Services (CMS) has begun to levy financial penalties on hospitals with high readmission rates for certain diagnoses, which has created an even heightened awareness about readmissions. On this background, Dr. Ames and his colleagues from UCSF reviewed 197 cases of metastatic (83%) and primary (17%) spinal tumors from 2005-2011 in order to evaluate 30 day and 1 year readmission rates as well as risk factors for readmission in this population. The overall 30 day readmission rate was 14% (17% for metastatic patients), with a one year overall readmission rate of 31% (38% for metastatic patients).  Patients with metastatic disease had a one year mortality of 30%, and 40% of Tokuhashi Group 0 patients (i.e. osteosarcoma, lung, stomach, bladder, esophagus, and pancreas cancer) died within a year of surgery. Medical complications were the most common cause of readmission (43%), while 33% of readmissions were due to recurrent disease and 23% due to surgical complications. Diabetes, obesity, pulmonary circulatory disease, depression, and operative time over ten hours were all independent predictors of readmission, and normal ambulatory function at baseline decreased the odds of readmission by 50%.


The high rate of readmission and mortality following surgical treatment of spinal tumors comes as no surprise. No other result could be expected when deconditioned patients with metastatic cancer—many of whom also have neurological deficits—are subjected to surgeries that involve one to three liters of blood loss and 8-14 hours in the operating room. These results are likely generalizable to other tertiary care centers performing spine tumor surgery. While UCSF represents a high volume center, they only performed about 30 cases of spinal tumors per year, indicating that these cases are not that common. The results of this paper, which might be used as a benchmark against which other centers could be compared, could have policy implications. If CMS begins to penalize hospitals for all readmissions without appropriate case mix adjustment, tertiary care centers will be disincentivized to provide this type of complex care. With relatively few centers performing spine tumor surgery, patients could find even less access to care for this challenging problem. Hopefully these data will illustrate the high rate of complications, readmission, and mortality inherent in spinal tumor surgery and prevent the creation of financial penalties for institutions and providers that care for cancer patients. Society has difficult decisions to face in cancer care. Chemotherapeutic drugs and surgical treatment of cancer are oftentimes not cost effective in the traditional sense (i.e. cost per quality adjusted life year), though life can sometimes be prolonged and quality of life improved. The amount we should spend on terminal cancer care will continue to be debated as healthcare resources become scarcer.

Please read Dr. Ames article on this topic in the October 1 issue. Does this change how you view the surgical treatment of spinal tumors? Let us know by leaving a comment on The Spine Blog.

Adam Pearson, MD, MS

Associate Web Editor

Friday, October 03, 2014

Most previous literature has indicated that smoking increases pseudarthrosis rates in both the lumbar and cervical spine.1,2 The effect of smoking on complications is less clear, in part because literature on this topic has typically controlled for the effects of medical comorbidities that result from smoking, an analytical approach that yields results that are difficult to interpret (i.e. controlling for COPD and heart disease results in no association between smoking and complications) .3 In an effort to better understand the effect of smoking on one and two level lumbar fusion, Dr. Bydon and colleagues from Johns Hopkins retrospectively reviewed 281 cases from 1990-2011 to determine the association between smoking and post-operative complications, including pseudarthrosis. Approximately half the patients had one level fusions, about half received BMP, and 18% were smokers (n=50). The authors reported a non-significant increased risk of pseudarthrosis in smokers overall (18% smokers vs. 10% non-smokers), with the strongest association between smoking and pseduarthrosis observed in patients undergoing two level fusion (29% vs. 10%, p = 0.02). The overall risk of complications was relatively low—around 5%--and was not significantly different between smokers and non-smokers. There appeared to be a non-significant trend towards an increased risk of complications for smokers in one level fusion (OR=1.9), however, only 2 smokers in this group had complications, so the comparison was clearly underpowered. Based on these data, the authors concluded that smokers were not at higher risk for complications other than pseudarthrosis after two level fusion.


This paper does add support to the already large body of evidence indicating that smoking is associated with pseudarthrosis. However, this paper has marked limitations making it difficult to interpret the comparisons that showed no significant differences (i.e. pseudarthrosis rates for single level fusions and overall complication rates). The study is substantially underpowered, especially for the subgroup analyses looking at one and two level fusions separately—there were only about 25 smokers in each subgroup. Given the relatively low rate of complications (about 5-10%), there were nowhere close to enough patients to avoid Type II error. Additionally, BMP was used in about 50% of cases, and smokers were about 25% more likely to receive it. This confounds the results and could underestimate the effect of smoking on pseudarthrosis. The paper is limited by its retrospective nature, and data such as smoking status and complications had to be identified by chart review, a notoriously unreliable method of capturing data that may or may not have been accurately recorded. Even the definition of pseudarthrosis was somewhat vague and likely depended on what imaging studies were done and at what time point after surgery. This study does support the idea that the effect of smoking is likely modified by the specific fusion environment, with smoking being more strongly related to pseudarthrosis in challenging fusion environments such as multilevel fusions or uninstrumented fusions. While this and other studies may indicate that fusion rate may be less affected by smoking in one level instrumented fusions, it should not be used as a justification for patients to continue to smoke after fusion.


Please read Dr. Bydon’s paper on this in the October 1 issue. Does this change your view on how smoking affects fusion outcomes? Let us know by leaving a comment on The Spine Blog. 

Adam Pearson, MD, MS
Associate Web Editor




1.            Glassman SD, Anagnost SC, Parker A, Burke D, Johnson JR, Dimar JR. The effect of cigarette smoking and smoking cessation on spinal fusion. Spine 2000;25:2608-15.

2.            Hilibrand AS, Fye MA, Emery SE, Palumbo MA, Bohlman HH. Impact of smoking on the outcome of anterior cervical arthrodesis with interbody or strut-grafting. J Bone Joint Surg Am 2001;83-A:668-73.

3.            Seicean A, Seicean S, Alan N, et al. Effect of smoking on the perioperative outcomes of patients who undergo elective spine surgery. Spine (Phila Pa 1976) 2013;38:1294-302.


Friday, September 26, 2014

The best bone graft to achieve fusion in different fusion environments remains unknown, and the use of iliac crest bone graft (ICBG), once the standard graft choice, has markedly decreased with the development of bone graft substitutes and the increased use of instrumentation and local bone graft. There is good evidence that the use of allograft in the cervical spine and local bone in the lumbar spine yields fusion rates close to ICBG for one and two level fusions when instrumentation is used. The degree of morbidity associated with ICBG harvest is also unclear, with traditional teaching suggesting a relatively high rate of long-term donor site pain, but more recent studies suggesting this may not be the case. If the morbidity associated with ICBG harvest is actually low, it may be a more appealing source of graft given higher fusion rates and potentially lower costs. On this background, Mr. Gruskay and his colleagues from Yale analyzed all fusion patients captured by the National Surgery Quality Improvement Project (NSQIP) database from 2010-2012. This yielded nearly 14,000 patients undergoing anterior or posterior procedures in the cervical, thoracic, and lumbar spine. Only 6% of patients underwent ICBG harvest, with ICBG use most common in posterior fusion procedures (approximately 10% of posterior fusion patients had ICBG). Iliac crest was harvested more often in multilevel than single level cases, though it was used in a small minority of patients even in fusions involving three or more levels. Unadjusted analyses demonstrated ICBG was associated with a 130% increase in transfusion rate, a 0.6 day increased length of stay (LOS), and 35 additional minutes in the OR. Given that these data are confounded by surgical approach and number of levels fused, multivariate analysis controlling for these and other factors showed a 50% increase in transfusion rate, a 0.2 day increased LOS, and 22 additional minutes of operating time if ICBG was harvested.


The results of this paper confirm what most spine surgeons would expect—ICBG harvest takes 20 or 30 minutes, increases blood loss, and might prolong hospital stay in a small number of patients who get a transfusion or have increased post-operative pain. The real questions about ICBG—does it cause long-term donor site pain, decrease re-operation rate due to higher fusion rates or save money—cannot be answered with this study design. Most surgeons would expect that there is some downside to ICBG harvest, namely that it takes some time, increases blood loss, and might increase pain, and this study confirms some of this. In one and two level instrumented fusions, it seems as though the potential benefits of ICBG probably do not outweigh the negatives when compared with allograft or local bone graft. The most burning question is whether ICBG, local bone graft plus extenders, or BMP is the best graft for fusions including at least 3 levels. That question remains open, though there is some recent data from the deformity world suggesting BMP may have some advantages over ICBG. Given the current data that suggests fusion rates for one and two level instrumented fusions are reasonably high with allograft and/or local bone, it is not surprising that ICBG has fallen out of favor.


Please read Mr. Gruskay’s article on this topic in the September 15 issue. Does this change your views on ICBG? Let us know by leaving a comment on The Spine Blog.

Adam Pearson, MD, MS

Associate Web Editor


Friday, September 19, 2014

The biannual publication of “Evidence-Based Recommendations for Spine Surgery” is designed to review six recent potentially high impact papers. This month’s installment discusses articles that address questions that arise on a regular basis in a typical spine surgery practice. The first article reviewed was an RCT comparing TLIF to posterolateral instrumented fusion. The study reports no differences in outcomes or complications between the two procedures, though TLIF was associated with greater blood loss and longer operating room time. The major limitation of the article was the inclusion of essentially all fusion patients, regardless of diagnosis, number of levels fused, and revision status. Given the heterogeneity of the patient population, it is not surprising that few significant differences were noted. Most spine surgeons would agree that interbody fusion is not indicated in most cases of fusion, though it might improve outcomes in certain conditions such as a hypermobile isthmic spondylolisthesis, a focal scoliosis or a recurrent disk herniation requiring fusion. Future studies need to evaluate interbody fusion in specific diagnostic groups. The authors also reviewed a paper reporting a higher rate of cancer in patients undergoing fusion using a high-dose BMP-2  formulation compared to those undergoing fusion with autograft, but the numbers are small, cancer is poorly defined, and many of the included cancers were skin cancers. The recent AOSpine North America study on myelopathy was reviewed, and this large cohort study did demonstrate improvement with surgery for patients across the myelopathy severity spectrum. Unfortunately, we still do not know if patients with mild myelopathy benefit from surgery compared to non-operative treatment or if surgery is worth the risk in elderly, debilitated patients with severe symptoms. Another interesting study compared outcomes between patients undergoing an L4-5 posterior interbody fusion for disk pathology in whom the L3-4 interspinous complex was removed or left intact. The authors reported higher rates of adjacent segment degeneration and re-operation for adjacent segment disease in those who had the L3-4 interspinous complex taken down.  This study supports preserving as much normal anatomy as possible while sufficiently decompressing the neural elements. A Swedish study looked at predictors of success following ACDF and found that non-smokers, men, and those with the highest baseline neck pain scores improved the most. Unfortunately, these groups are also the ones most likely to improve with non-operative treatment, so studies looking at surgical outcome predictors are not that helpful if a non-operative arm is not included. Finally, a definitive RCT looking at bracing for adolescent idiopathic scoliosis did show a clinically and statistically significant benefit, a finding that supports a widespread practice.


Reviews such as this are instructive, both in the sense they inform practice when good evidence is available as well as making it clear when evidence does not exist and the best clinical course is ambiguous. Of these six articles that were reviewed, the scoliosis trial was the only one that led to a definitive conclusion about the best treatment, and bracing for scoliosis is already widely accepted. This study’s impact is similar to the Spine Patient Outcomes Research Trial (SPORT) in that both provide strong evidence to support currently accepted practice. The other studies provide some insight to questions that remain unanswered such as the “best” lumbar fusion technique or how to treat patients with mild myelopathy. Critical review of these studies makes it clear that the literature is giving us a sense of the best way to treat the average degenerative spondylolisthesis patient or the average myelopathy patient, but the “average” patient never walks into clinic. In order for research to have more of an impact on clinical decision making, future studies will need to evaluate predictors of outcomes based on both patient- and disease-specific characteristics. Given that treatment decisions always involve at least two options (i.e. surgical and non-operative), future work will need to look at predictors of outcomes for all of the treatments under consideration.


Please read this installment of Evidence-Based Recommendations for Spine Surgery. Let us know how these recommendations change your view of these topics by leaving a comment on The Spine Blog.

Adam Pearson, MD, MS

Associate Web Editor

About the Blog

Spine Journal
This Blog provides a forum for discussion about high impact articles published in Spine, including the bi-annual publication of "Evidenced-Based Recommendations for Spine Surgery." Website users can use this forum to discuss how the articles have affected their practice and query the authors about their findings and recommendations.