Skip Navigation LinksHome > Blogs > The Spine Blog
The Spine Blog
Friday, February 5, 2016

The Spine evidence-based medicine group published their most recent biannual evidence-based recommendations in the February 1 issue. These reviews evaluate six potentially high impact articles and make practice recommendations based on the strength of the evidence. Dr. Vaccaro and his colleagues reviewed articles on wide-ranging topics from bracing to sagittal balance parameters. The bracing article was a Level 1 RCT comparing clinical and radiographic outcomes between elderly patients with osteoporotic compression fractures randomized to a rigid brace, soft brace or no brace. They found no differences in clinical or radiographic outcomes and made a strong recommendation that bracing was not necessary for these injuries. Another RCT compared outcomes between cervical radiculopathy patients randomized to cervical epidural steroid injection (CESI), PT plus medication, or a combination of CESI, PT, and medication. They found no significant differences on the patient reported outcome measures, though the reviewers raised the question of the study being underpowered and noted the lack of a non-treatment control. Due to these limitations, they recommended no change to practice. Two articles focusing on adult deformity from the International Spine Study Group (ISSG) were included. Both of these studies used the ISSG database that includes prospectively collected patient reported outcomes, complications, and radiographic images. One study showed that a pedicle subtraction osteotomy was the only independent risk factor for rod breakage, and the reviewers made a weak recommendation to consider the use of satellite rods to reduce the risk. They noted that an RCT comparing the use of satellite rods to two rod constructs would be necessary to make a stronger recommendation. The other deformity paper reported that the T1 pelvic angle (T1PA) was strongly correlated to baseline patient reported outcomes and might be more reliable than other measures of sagittal balance that are more dependent on patient posture. Given that the paper did not report the effect of changing T1PA surgically on outcomes, the reviewers did not make any recommendation to change practice. The debate over the role of fusion for degenerative spondylolisthesis (DS) continues, and a retrospective cohort study comparing outcomes between DS patients treated with minimally invasive decompression alone and those undergoing traditional decompression and fusion was reviewed. While the article showed no outcome differences between the two groups, the study likely suffered from significant selection bias that precluded any firm conclusions. The last article reviewed was a prospective cohort study comparing outcomes between lumbar stenosis patients treated with decompression alone versus decompression plus the Coflex interspinous device. While the paper had significant methodological limitations, there were no differences between the groups, and the authors made a strong recommendation concluding that there was no evidence supporting the use of the device.


These biannual articles by the Spine evidence-based medicine board offer high quality, critical reviews of papers focusing on clinically relevant topics. While the recommendations for clinical practice are extremely valuable, the reviews also offer insight to the quality of the latest spine literature. Of all six articles, only the bracing article represented a true Level 1 trial without significant methodological flaws. The CESI article was likely underpowered and did not have a true control group and the other studies were all observational studies with the typical limitations inherent in such designs. The DS and Coflex studies that compared two surgical techniques could have been performed as RCTs but instead were designed as cohort studies. As a result, both were subject to potential selection bias that markedly limited the strength of their conclusions. Performing high quality RCTs is challenging, but the spine surgery community should be motivated to design and undertake Level 1 trials when possible given the limitations in observational studies. Until such studies are performed, most clinical questions will remain without definitive answers.

Please read the evidence-based medicine board's review published in the February 1 issue. Does this change how you view any of the topics included in the article? Let us know by leaving a comment on The Spine Blog.

Adam Pearson, MD, MS

Associate Web Editor

Friday, January 29, 2016

There is a large volume of literature showing better surgical outcomes for high volume hospitals and surgeons, and spine surgery is no exception. These data should encourage patients and payers to seek out these high volume "centers of excellence". However, patients frequently prefer to have their care done locally, even if it is at a medical center with worse outcomes. As these findings have been published over the past 15 years, there has also been consolidation in the hospital industry, with large hospitals aligning themselves with smaller community hospitals into large health systems. In some cases, this has allowed these integrated health systems to concentrate certain procedures at high volume hospitals with high volume surgeons. While one would assume these trends would result in more spine procedures being done at high volume institutions, this topic has not been extensively studied. Given the gap in the literature, Dr. Bosco and his colleagues analyzed a New York state administrative database including over 200,000 spine surgeries performed from 2005-2014 to determine if spine surgery was being concentrated in high volume institutions. While the overall rate of surgery increased only 2.5% from 2005 to 2014, the number of procedures performed at high volume hospitals (defined as being in the top tertile of hospitals in terms of spine cases in 2005) increased 50%. During the same period of time, the spine procedure volumes at the low and medium volume hospitals decreased by 13% and 30%, respectively. The number of high volume hospitals increased from 10 to 16, while the number of low volume hospitals decreased from 114 to 104. By 2014, 48% of all spine surgery was being performed in high volume centers.


Given the knowledge that high volume centers have better outcomes and the trends in hospital consolidation, it is not surprising that spine surgery started to be concentrated in high volume centers across New York State, generally in Buffalo, Rochester, Syracuse, Albany, and New York City. The fact that it happened without any regulation or formal planning is fascinating. Some combination of market forces, patient consumerism, and surgeon subspecialization seems to have driven this phenomenon. The observational nature of this study precludes any firm conclusions regarding the explanation of the concentration of spine care in high volume institutions, and, given the multitude of factors that in play, it seems unlikely that any definitive conclusions will be possible. This paper does a very nice job demonstrating the regionalization of spine care in New York State over the past decade. It seems likely that similar trends are occurring nationwide, and it would be interesting to see these analyzed. The paper did not evaluate the trends at the surgeon-level, though it seems likely that care is getting more concentrated with high volume surgeons. All of this is generally good news for patients, payers, and hospitals given that high volumes are generally associated with better outcomes, fewer complications, and lower cost. The downside may be felt by rural patients who live distant from high volume institutions as they may have to travel long distances to obtain the best care or receive care at a local center which could have worse results. It seems likely that this trend will continue into the future, though eventually concentration will likely reach a limit in terms of how far patients are willing to travel.

Please read Dr. Bosco's article on this topic in the January 15 issue. Does this change how you view the regionalization of spine surgery? Let us know by leaving a comment on The Spine Blog.

Adam Pearson, MD, MS

Associate Web Editor


Friday, January 15, 2016

The spine surgery community has yet to define the clinical situations in which bone morphogenetic protein (BMP) is beneficial and appropriate. It is FDA-approved for use in anterior lumbar interbody fusion, which seems a reasonable indication given that it obviates the need for iliac crest bone graft (ICBG), and local bone graft options are limited in this approach. It is not approved for posterolateral or posterior lumbar interbody fusion, and, given the high fusion rates with the use of local bone graft and instrumentation for one and two level fusion, it is probably not indicated for this use.1 Long fusion constructs from the thoracic spine to the sacrum in adult spinal deformity (ASD) represent a challenging fusion environment, with a high rate of pseudarthrosis at the L5-S1 level, even with the use of ICBG. Recent studies from Washington University have shown a significantly lower rate of pseudarthrosis with the use of BMP compared to ICBG.2 Given the potential to reduce reoperation rates for pseudarthrosis following adult deformity surgery, Dr. Paul and his colleagues analyzed the New York state inpatient database to determine the rate of reoperation for pseudarthrosis in ASD patients treated with and without BMP. They identified over 3,700 adult patients with a diagnosis of scoliosis undergoing at least 3-level surgery in 2008, with 4 years of follow-up. They stratified the patients into a 3-8 level fusion group and a greater than 8-level fusion group and found that the reoperation rate for pseudarthrosis in the greater than 8-level fusion group was 34% in the no-BMP group and 5% in the BMP group (relative risk=7.5, p<0.001). In the 3-8 level fusion group, the differences were less pronounced but still notable, with a reoperation rate for pseudarthrosis of 6% in the no-BMP group and 1% in the BMP group (relative risk=4.1, not-significant). A survival analysis controlling for potential confounders gave similar results. There were no differences in reoperation rates for wound problems or infection.


This study represents a high quality database analysis which adds further support to the use of BMP in long fusions from the thoracic spine to the sacrum in ASD patients. All retrospective database studies have inherent limitations that must be considered prior to drawing strong conclusions, and this study is no exception. The results are always dependent on the accuracy of the coding, though the rates of revision for pseudarthrosis are in line with prior literature.2 The algorithm to define pseudarthrosis was not explicitly defined, and it is possible that the cause of reoperation was misclassified. However, this classification error would likely affect the two groups evenly, so it is unlikely to have substantially biased the results. While the authors did attempt to control for potential confounders included in the database (i.e. age, gender, race, insurance status), other unmeasured confounders (i.e. BMI, smoking status, revision surgery, surgeon experience, etc.) could affect the results. Nonetheless, the data reported here as well as from Washington University strongly suggests that BMP reduces the pseudarthrosis rate and subsequent reoperation rate for ASD patients undergoing long fusions from the thoracic spine to the sacrum. An RCT or FDA IDE study evaluating BMP for this application would provide stronger evidence supporting its use, but the currently available observational evidence is fairly convincing. Questions still remain on how to optimize fusion rates and patient outcomes such as BMP dosage and the role of pelvic fixation and interbody support at L5-S1. The currently available literature—consisting of retrospective, observational data--suggests that the best results might be obtained using BMP and pelvic fixation without the need for an interbody fusion.2,3 We will await higher level evidence before making any definitive conclusions.


Please read Dr. Paul’s article on this topic in the January 1 issue. Does this change how you view the role of BMP in adult deformity surgery? Let us know by leaving a comment on The Spine Blog.

Adam Pearson, MD, MS

Associate Web Editor




1.            Inage K, Ohtori S, Koshi T, et al. One, two-, and three-level instrumented posterolateral fusion of the lumbar spine with a local bone graft: a prospective study with a 2-year follow-up. Spine 2011;36:1392-6.

2.            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.

3.            Annis P, Brodke DS, Spiker WR, Daubs MD, Lawrence BD. The Fate of L5-S1 With Low-Dose BMP-2 and Pelvic Fixation, With or Without Interbody Fusion, in Adult Deformity Surgery. Spine (Phila Pa 1976) 2015;40:E634-9.


Friday, January 8, 2016

The effect of nonunion on patient reported outcomes (PROs) following lumbar fusion has long been debated. Fischgrund et al. wrote a classic paper on degenerative spondylolisthesis (DS) that demonstrated patients undergoing instrumented fusions had a higher fusion rate but similar clinical outcomes as patients undergoing uninstrumented fusions at 2 year follow-up.1 Kornblum et al. subsequently analyzed the uninstrumented cohort at over 7 years of follow-up and found that those who had gone onto nonunion had worse outcomes than those with a solid fusion.2 A Japanese paper reported similar findings, with nonunion in uninstrumented fusions for DS unrelated to early outcomes but associated with worse long-term outcomes.3 Less has been published about the relationship between fusion status and outcomes following fusion for degenerative disk disease (DDD). Given this void in the literature, Dr. Noschenko and colleagues from the University of Colorado analyzed the Yale Open Data Access Project (YODA) database that included outcomes for 496 DDD patients enrolled in FDA trials comparing BMP-2 and iliac crest bone graft in lumbar interbody fusion. Three of the trials were for anterior lumbar interbody fusion (ALIF) and one was for posterior lumbar interbody fusion (PLIF). After excluding patients who underwent reoperation for nonunion, they divided patients into those with radiographic fusion and nonunion at 1 and 2 years. This yielded 29 (5.5%) patients who had a radiographic nonunion at 1 or 2 years who did not undergo reoperation for nonunion. They performed a meta-analysis of the 4 trials and reported that the ODI score improved 13 points more and the numerical rating score (NRS) for low back pain improved 5 points more, and the NRS leg pain score improved 3.1 points more in the fused group compared to the nonunion group. The fusion patients were significantly more likely to meet the minimum clinically important difference (MCID) on all three outcome measures than the nonunion patients (75% vs. 44%). However, fusion status was clearly not a perfect predictor of PROs given that 25% of patients with a solid fusion failed to meet the MCID on all three outcome measures while 44% of patients with nonunion did.


This study represents a good use of a database to try to answer a clinically important question. If fusion rate was not related to PROs, one could argue that all of the efforts to obtain a solid fusion (i.e. pedicle screws, interbody fusion devices, BMP-2, etc.) are irrelevant and less time, energy, and money should be wasted to avoid a complication that does not adversely affect the patient. The data presented suggest that fusion does matter—to some degree. Before jumping to any strong conclusions, the limitations of this study need to be considered. A major issue is the exclusion of patients undergoing reoperation for nonunion. These excluded patients were likely doing worse than those who did not go onto reoperation and were included in the current study, likely biasing the results by suggesting better outcomes for the nonunion patients. While this problem could have been mitigated by providing some data on the excluded patients or a separate analysis of their outcomes up to the point they underwent reoperation, the authors failed to tell us anything about this group, not even how many they numbered. Additionally, there was no traditional “Table 1” comparing the patient characteristics of the fusion and nonunion patients. There were likely many potential confounders that could explain some of the differences in outcomes, though these were not considered. The relatively low number of nonunion patients also limits the analysis to some degree, though 29 nonunion patients is a larger sample than has been analyzed in the past. Finally, the authors combined all the data from the 1 and 2 year points, probably to get over some of their issues related to the study being underpowered. They acknowledged that the nonunion rate increased from 1 to 2 years, yet outcomes tended to improve over this time period, especially among the nonunion patients. As such, they should have reported outcomes for 1 and 2 years separately. Despite these limitations, this paper adds to the literature suggesting that fusion status matters. However, we do not know how much it matters, and this likely varies depending on the indication for surgery, the fusion technique, the definition of nonunion, and patient characteristics. Given what we know, surgeons should make the effort to obtain a solid fusion, though how strong of an effort remains to be determined.

Please read Dr. Noschenko’s article on this topic in the January 1 issue. Does this change how you view the role of fusion status in determining outcomes? Let us know by leaving a comment on The Spine Blog.

Adam Pearson, MD, MS

Associate Web Editor





1.            Fischgrund JS, Mackay M, Herkowitz HN, Brower R, Montgomery DM, Kurz LT. 1997 Volvo Award winner in clinical studies. Degenerative lumbar spondylolisthesis with spinal stenosis: a prospective, randomized study comparing decompressive laminectomy and arthrodesis with and without spinal instrumentation. Spine 1997;22:2807-12.

2.            Kornblum MB, Fischgrund JS, Herkowitz HN, Abraham DA, Berkower DL, Ditkoff JS. Degenerative lumbar spondylolisthesis with spinal stenosis: a prospective long-term study comparing fusion and pseudarthrosis. Spine 2004;29:726-33; discussion 33-4.

3.            Tsutsumimoto T, Shimogata M, Yoshimura Y, Misawa H. Union versus nonunion after posterolateral lumbar fusion: a comparison of long-term surgical outcomes in patients with degenerative lumbar spondylolisthesis. Eur Spine J 2008;17:1107-12.


Sunday, January 3, 2016

The New Year always provides an opportunity for reflection on what happened in the preceding year and as well as on what might attract our attention in the upcoming year. 2015 was an interesting year in the spine world, and many of the controversial topics debated in 2014 continued to generate discussion over the past year. Hot topics in the cervical spine included cervical disk replacement and cervical myelopathy. The use of interbody fusion in the lumbar spine continues to gain popularity, and the literature focused on the state of the evidence behind this practice. Four years after the original articles raised questions about bone morphogenetic protein (BMP), it remains a controversial albeit less popular molecule.1 These issues were all discussed on The Spine Blog, and responses to Quick Polls on the topics provide us some insight to how spine specialists are thinking about them. One of the potential advantages of cervical disk replacement over ACDF is the elimination of pseudarthrosis as a complication that can lead to reoperation. Most of the industry sponsored trials comparing ACDF to disk replacement have reported significantly lower rates of reoperation following disk replacement, with reoperation for pseudarthrosis being one of the main drivers of this difference. Readers were queried about the likelihood of reoperation for pseudarthrosis following one level ACDF, and their responses showed the variation in beliefs on this topic. A plurality (38%) answered less than one percent, however, nearly one third believed the rate was over 5%. The remaining third answered between 1% and 5%. Dr. Guppy’s article from Kaiser Permanente on this topic reported a 0.2% rate of reoperation for pseudarthrosis, a rate markedly lower than that reported in the FDA studies comparing ACDF to disk replacement.2 These discrepancies underscore how defining pseudarthrosis and the need for reoperation remain subjective, and the reoperation rate is likely driven as much by surgeon and patient preference as it is by actual biology. The best surgical technique for cervical myelopathy with multilevel spondylosis remains debatable, and this is reflected by answers to Quick Polls on the topic. In asking about the preferred treatment for 4-level disease with maintained lordosis, the majority (57%) favored laminectomy and fusion, with 29% favoring laminoplasty. No respondent favored 4-level ACDF. The responses were much different for the cervical myelopathy scenario with 3-level disease, with ACDF being the most popular response (55%), followed by laminectomy and fusion (36%), and laminoplasty (9%). Though still debated, surgeons seem to prefer laminectomy and fusion for 4-level disease, and an anterior approach 3-level disease. Laminoplasty remains popular in Asia but has not been widely adopted in North America or Europe.


The use of interbody fusion in the lumbar spine was the focus of much literature on the topic. In 2014, Kepler et al. published a review of the American Board of Orthopaedic Surgeons database showing that the use of interbody fusion for degenerative spondylolisthesis increased from 14% in 2000 to 37% in 2011 despite a lack of evidence supporting this approach.3 In 2015, Gottschalk et al. reviewed the data from Emory  that showed no advantage for the addition of interbody fusion to posterolateral fusion for L4-L5 degenerative spondylolisthesis.4 The Quick Poll on this topic demonstrated that the majority of surgeons preferred some form of interbody fusion when treating L4-L5 degenerative spondylolisthesis, with 38% favoring a TLIF/PLIF operation and 23% preferring a lateral interbody fusion. Only 31% preferred a posterolateral instrumented fusion. The only Level 1 study on this topic demonstrated no differences between patients treated with TLIF versus posterolateral instrumented fusion.5 Interbody fusion at L5-S1 has become a popular way to support fusion at the lumbosacral junction at the caudal end of a long fusion construct for adult deformity patients. With the increased adoption of pelvic fixation, the need for an interbody device has been questioned. Annis et al. found no differences in L5-S1 fusion rates in adult deformity patients undergoing long fusions to the pelvis treated with or without L5-S1 interbody fusion, along with low dose BMP-2.6 Quick Poll respondents felt differently, with 75% favoring the use of L5-S1 interbody fusion in these patients. The enthusiasm for BMP-2 use in spinal fusion has certainly waned, but it remains a tool in the surgeon’s armamentarium to deal with challenging fusion environments (i.e. long fusions to the sacrum/pelvis, nonunions). It remains to be seen if BMP-2 use is associated with cancer, though no evidence has emerged suggesting it does beyond the borderline findings from some of the FDA IDE trial data1. Quick Poll respondents believe that BMP-2 does not cause cancer, with 100% responding that it does not.


It is likely that more questions face spine specialists at the end of 2015 compared to at the end of 2014. Scientific investigation almost always yields more questions than answers. Hopefully in 2016 spine care providers will use the best available evidence to guide their decision-making and also be motivated to perform studies to answer some of our many remaining questions.


Adam Pearson, MD, MS

Associate Web Editor




1.            Carragee EJ, Hurwitz EL, Weiner BK. A critical review of recombinant human bone morphogenetic protein-2 trials in spinal surgery: emerging safety concerns and lessons learned. The spine journal : official journal of the North American Spine Society 2011;11:471-91.

2.            Guppy KH, Harris J, Paxton LW, Alvarez JL, Bernbeck JA. Reoperation Rates for Symptomatic Nonunions in Anterior Cervical Fusions from a National Spine Registry. Spine (Phila Pa 1976) 2015.

3.            Kepler CK, Vaccaro AR, Hilibrand AS, et al. National trends in the use of fusion techniques to treat degenerative spondylolisthesis. Spine (Phila Pa 1976) 2014;39:1584-9.

4.            Gottschalk MB, Premkumar A, Sweeney K, et al. Posterolateral Lumbar Arthrodesis With and Without Interbody Arthrodesis for L4-L5 Degenerative Spondylolisthesis: A Comparative Value Analysis. Spine (Phila Pa 1976) 2015;40:917-25.

5.            Hoy K, Bunger C, Niederman B, et al. Transforaminal lumbar interbody fusion (TLIF) versus posterolateral instrumented fusion (PLF) in degenerative lumbar disorders: a randomized clinical trial with 2-year follow-up. Eur Spine J 2013;22:2022-9.

6.            Annis P, Brodke DS, Spiker WR, Daubs MD, Lawrence BD. The Fate of L5-S1 With Low-Dose BMP-2 and Pelvic Fixation, With or Without Interbody Fusion, in Adult Deformity Surgery. Spine (Phila Pa 1976) 2015;40:E634-9.


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.