Secondary Logo

Journal Logo

The Spine Blog

Friday, May 29, 2020

Very few studies have followed lumbar spinal stenosis (SpS) patients for 8-10 years after surgery, and these have generally showed a persistent long-term benefit of surgery but somewhat worsening outcomes over time.1,2 Studies evaluating outcome predictors have generally looked at medium-term outcomes out to 4 years.3 In order to address the gap in the literature evaluating long-term outcome predictors in SpS, Dr. Tuomainen and colleagues from Finland evaluated the associations between smoking status, BMI, history of prior lumbar surgery, self-rated health status, and duration of painkiller use prior to surgery and post-operative Oswestry Disability Index (ODI) and visual analog scale (VAS) pain scores out to ten years after decompressive surgery. Of the original cohort of 102 patients, 72 were available for follow-up at 10 years after surgery (17 had died and 13 were unavailable). Overall, patients improved significantly from a mean baseline ODI score of 43 to 29 at 10-year follow-up. Similarly, VAS scores improved significantly from 56 to 33. In looking at the subgroup analyses, patients with worse self-rated health and use of painkillers for over 12 months had worse baseline ODI and VAS scores. Univariate analyses demonstrated that smokers' ODI scores at 10 years were only 6 points better than baseline compared to a 15-point improvement in non-smokers. Those with poor self-rated health had significantly worse 10-year ODI scores compared to those with good self-rated health (38 vs. 26), though a similar change score from baseline. Patients who had been taking painkillers for less than 3 months at baseline had significantly better ODI scores at 10 years compared to those taking painkillers for over 12 months (19 vs. 35) and improved about 5 points more compared to the chronic use group. Multivariate analysis indicated that not smoking, no prior lumbar surgery, better self-rated health, and use of painkillers for less than 12 months were associated with better ODI and VAS scores over the 10 year follow-up period.

The authors have made a nice contribution to the literature by looking at outcome predictors over 10 years following lumbar decompression surgery for SpS. Their findings support similar findings in other studies that have identified smoking, revision surgery, poor general health status, and pre-operative opioid use as predictors of worse surgical outcomes. Like all long-term follow-up studies, this one was limited by loss to follow-up. However, 70% follow-up at 10 years is relatively good, especially considering that more than half of those lost to follow-up died. Additionally, those who remained in the study were relatively similar to those who were lost to follow-up. Most subgroup analyses have limited statistical power due to the low numbers in many subgroups, and this study is no exception. There were only 18 smokers and 12 patients who had prior lumbar surgery. The authors did not make it clear if the multivariate analysis controlled for baseline differences, so it is hard to determine if the observed outcome differences were due to the baseline differences or differential response to surgery. Finally, there was no non-operative control group, so it could not be determined if the subgroups with worse surgical outcomes would actually have a lower treatment effect (i.e. change score in the surgery group – change score in the non-operative group). The Spine Patient Outcomes Research Trial (SPORT) reported extensive subgroup analyses evaluating outcome predictors following surgical and non-operative treatment for SpS based on 4-year longitudinal follow-up of ODI scores.3 Patients with revision surgery were excluded from SPORT, but all of the subgroups identified as predicting worse outcomes in the current study also predicted worse surgical outcomes in SPORT (i.e. smoking, multiple medical comorbidities, and taking opioids at baseline). However, only smoking was associated with a worse treatment effect of surgery. This was because those with multiple medical comorbidities and opioid users had extremely poor non-operative outcomes, so they still improved more with surgery. Smokers had particularly poor outcomes in SpS and were the only subgroup that did not improve significantly from baseline following surgery. This is interesting as fusion was not performed in the SPORT SpS cohort, so increased pseudarthosis rate was not the driver of worse outcomes. Future studies should evaluate the cause of worse surgical outcomes following SpS surgery in smokers and determine if pre-operative smoking cessation improves outcomes.

Please read Dr. Tuomainen's article on this topic in the June 1 issue. Does this change how you consider long-term outcome predictors in SpS?

Adam Pearson, MD, MS

Associate Web Editor



1.            Atlas SJ, Keller RB, Wu YA, Deyo RA, Singer DE. Long-term outcomes of surgical and nonsurgical management of lumbar spinal stenosis: 8 to 10 year results from the maine lumbar spine study. Spine (Phila Pa 1976) 2005;30:936-43.

2.            Lurie JD, Tosteson TD, Tosteson AN, et al. Surgical versus nonoperative treatment for lumbar disc herniation: eight-year results for the spine patient outcomes research trial. Spine (Phila Pa 1976) 2014;39:3-16.

3.            Pearson A, Lurie J, Tosteson T, Zhao W, Abdu W, Weinstein JN. Who should have surgery for spinal stenosis? Treatment effect predictors in SPORT. Spine 2012;37:1791-802.


Friday, May 22, 2020

Adjacent segment disease (ASD) is a common complication following ACDF, occurring at a rate of approximately 3% per year. Traditionally, surgeons treat this with ACDF at the adjacent level using a plate, screws, and bone graft. They frequently have to remove the original plate in order to accomplish this, which requires greater dissection and can be technically difficult with some plate designs. Zero-profile stand-alone cages (SAC) with integrated screws were initially promoted for having lower dysphagia rates, and surgeons subsequently realized they were advantageous in treating ASD as the index plate did not have to be removed to place the device. As such, SACs were rapidly adopted to treat ASD, without much data supporting their use in that application. Many SAC designs include only 2 screws compared to the typical plate with 4 screws, so there is some concern that the SAC could be biomechanically inferior to a plate and screws in the ASD environment where stresses are higher. In order to compare SACs to the traditional plate and screws for ASD, Dr. Gandhi and colleagues from William Beaumont Hospital retrospectively reviewed 46 patients undergoing ACDF for ASD, 17 of whom were treated with plate and screws and 29 of whom were treated with a SAC. They also included a control group of 40 patients undergoing primary single level ACDF with plate and screws. Their primary outcome was fusion rate at a minimum follow-up of 6 months, and this was determined based on bridging bone and/or a lack of motion (<2 mm between spinous processes) on flexion-extension x-rays. Patients suspected of pseudarthrosis were also evaluated with a CT scan. They found an 82% fusion rate in the plate and screw cohort compared to a 69% fusion rate in the SAC cohort, a non-significant difference. The primary ACDF cohort had a fusion rate of 95%, which was significantly higher than the SAC ASD cohort. There were zero revisions in the patients treated with a plate and screws in both the primary and ASD groups, and 14% of the SAC ASD patients underwent revision (2 for pseudarthrosis and 2 for hardware migration).

The authors did a nice job creating a small retrospective cohort study comparing fusion rates between SAC and plate and screw constructs when treating ASD. The study has all of the limitations typical for a small, retrospective study, including lack of power and selection bias. While the plate and screw cohort had a fusion rate 13% higher rate of fusion, which translates to a 72% increase in relative risk of pseudarthrosis, this difference was not significant due to the low numbers involved (3 non-unions in the plate group compared to 9 in the SAC group). Additionally, the groups were likely treated by different surgeons and were different at baseline in both measured and unmeasured ways, so selection bias could be confounding the results. The average duration of follow-up was not reported, and a minimum follow-up of 6 months was likely not sufficient to diagnose pseudarthrosis. It has been shown that radiographic fusion can take up to 2 years to occur, so some patients classified as having non-unions in this study may have gone onto solid fusion over time. While 14 of 86 patients were classified as having non-union in this study, only 2 went onto revision surgery for this indication, so the majority of the non-union patients were likely relatively asymptomatic. The limited and variable follow-up duration also makes it difficult to interpret the revision rates as they undoubtedly would increase with longer follow-up. Despite its significant limitations, this paper does raise the question of whether or not SACs are the best option for ASD given that they may be biomechanically inferior to a plate and screws in this application. Clearly more study is needed to answer the questions, and the authors point out that SACs that incorporate more screws may have better fusion rates. It seems as though a plate and screws may be the best option in situations where the original plate is easily removed. However, in cases where the plate cannot be easily removed (i.e. index 3 level plate, stripped screws, etc.), the SAC may have an important role.

Please read Dr. Gandhi's article on this topic in the June 1 issue. Does this change how you view the use of SACs for ASD?

Adam Pearson, MD, MS

Associate Web Editor

Friday, May 15, 2020

When treating adult spinal deformity, there is a substantial minority of patients in whom stopping the fusion at L1 or T12 would be appealing other than concerns related to failure at the thoracolumbar junction (TLJ). For patients who require fusion above the L2 level, surgeons tend to reflexively extend the fusion to T10 in order to prevent proximal junctional failure (PJF). In order to define a subgroup of patients for whom stopping at the TLJ is acceptable, Dr. Park and colleagues from Seoul retrospectively reviewed a case series of 63 adult deformity patients over age 50 who underwent fusion from T11, T12 or L1 to the sacrum or pelvis. About 75% had pelvic fixation, and 71% had L1 as the upper instrumented vertebra (UIV). They defined PJF as a proximal junctional kyphosis greater then 20 degrees, UIV or UIV+1 fracture, or screw pullout. The average follow-up was over four years. Ninety percent of patients were female, and the average age was 67. They found that 37% of patients developed PJF at a mean of 9 months after surgery, and one third of these patients had revision surgery with fusion to T3 or T4. Univariate analysis revealed that increasing age, osteoporosis, increased pre-operative pelvic tilt, and baseline kyphosis at the junctional level were significantly associated with PJF. There were trends suggesting that a greater correction of pelvic tilt and lumbar lordosis  were also associated with PJF. In multivariate analysis, increasing age, osteoporosis, and baseline kyphosis at the junctional level were all independent predictors of PJF. The PJF patients had significantly worse ODI scores (49 vs. 31) and SRS-22 scores at final follow-up. The authors suggested that stopping at the TLJ should be avoided in patients over 70, osteoporotic patients, and those with baseline kyphosis at the junctional level as 56% of patients with just one of these risk factors experienced PJF. Conversely, none of the patients without these risk factors had PJF.

The authors have done a nice analysis of their case series of adult spinal deformity patients with the UIV at T11, T12, and L1. Over one third of these patients developed PJF, indicating that avoiding stopping at the TLJ is generally a good idea. Their analysis of risk factors for PJF helps to define the group of patients for whom stopping  at the TLJ is reasonable, namely younger patients with good bone quality and no baseline kyphosis at the junctional level. Similar to prior studies, they identified more aggressive correction as a risk factor for PJF, though this did not reach statistical significance. This study was limited by relatively small numbers, and the PJF group included only 23 patients. Such a small group limits the power of risk factor analysis, so some PJF risk factors may have been missed due to Type II error. Additionally, the study did not include a comparison group in whom fusion was performed to T10 or above. It is possible that extending the fusion to T10 in the PJF patients may not have prevented the complication. While performing an RCT to answer this question would be difficult, the authors likely had a large cohort of patients fused to T10, and this could have been used as a comparison group to help answer the question. This paper does suggest that stopping at the TLJ is a reasonable option for a subgroup of adult deformity patients who are younger, have good bone, no baseline kyphosis at the junctional level, and who do not need a major correction of sagittal deformity.

Please read Dr. Park's article on this topic. Does this change your opinion of stopping a fusion at the TLJ?

Adam Pearson, MD, MS

Associate Web Editor


Friday, May 8, 2020

Spine fellowship directors have a major role in the professional development of the surgeons they train. As a result, they have a significant influence on the overall field as they mold the next generation of spine surgeons. In order to understand the field of spine surgery, one must study the process through which spine surgeons are educated, as this process passes along the knowledge, technical skills, and culture that define the subspecialty. Dr. Donnally and colleagues from Thomas Jefferson University and the University of Miami sought to determine the educational and professional characteristics of spine fellowship directors. They used the NASS Fellowship Directory to identify 103 fellowship leaders and then contacted them in order to obtain their CVs. When this was unsuccessful, they used on-line resources to obtain data. They found that 96% of fellowship leaders were male, with an average age of 53. The mean duration of time in the position was about 10 years, indicating that fellowship leaders were promoted to the position at an average age of 43. Interestingly, about 25% of fellowship leaders completed fellowships at one of three institutions:  Case Western Reserve (10), Washington University (9), and Thomas Jefferson University (7). The mean Scopus H index was 23.8, substantially higher than the mean reported for spine surgery fellowship faculty overall (13.6) by a prior study.1 About 20% of fellowship directors had been on the Board of Directors of either NASS, CSRS, or SRS.

This is a novel study looking at the relatively unstudied topic of leadership in spine surgery education. While scholarly interest in orthopedic residency education has increased over the past decade, very little has been written about subspecialty fellowship education. Spine fellowship training lasts only one year, however, this year likely has a disproportionately large influence on the development and practice patterns of spine surgeons. This paper was a simple cross sectional survey of fellowship leaders, and it had some methodological limitations. The authors did not report how many fellowship leaders actually supplied them with their CVs, so we do not get a sense of how much missing data were present. In some ways, the paper just scratches the surface of the issue and does not address the deeper questions such as what personal and leadership qualities do fellowship directors possess. The results do highlight the homogeneity and lack of diversity in the field. Very few fields remain over 95% male, though this will likely change slowly over time as more female spine surgeons are trained. The authors did not ask the fellowship directors to report their race, though there is likely a lack of racial diversity in this cohort as well. Just as important, and possibly more striking, is the lack of educational diversity among the fellowship leaders. The outsize influence of a few institutions in training future fellowship directors may limit the diversity of thought in the field. There is a large body of spine surgery dogma passed along from one generation to the next, oftentimes with very little evidence supporting it. This study offers some interesting insights about the characteristics of spine fellowship directors. To truly understand who these people are and how they got there would require a deep sociological dive.

Please read Dr. Donnally's article in the May 15 issue. Does this change how you view the role of spine fellowship directors?

Adam Pearson, MD, MS

Associate Web Editor


1.            Schoenfeld AJ, Bhalla A, George J, Harris MB, Bono CM. Academic productivity and contributions to the literature among spine surgery fellowship faculty. Spine J 2015;15:2126-31.


Friday, May 1, 2020

Lumbar spine and hip pathology frequently co-exist, posing a challenge to joint replacement and spine surgeons. Sometimes the concomitant pathology leads to misdiagnosis and results in a failed operation. In other cases, patients clearly have symptoms attributable to both regions, and it can be difficult to determine which pathology to address first. The joint replacement literature has identified lumbar spine pathology, and specifically prior lumbar fusion, as a risk factor for dislocation. While hip replacement surgeons would prefer never to encounter patients with lumbar spine pathology, they do not have this luxury and need to know how to proceed when they encounter patients with both diagnoses. Hip pathology does not necessarily predispose spine surgeons to complications, and they generally worry little about hip pathology other than to warn patients that they might need a hip replacement after their lumbar fusion. In order to shed some light on this issue, Daniel Yang and colleagues from Brown University used the PearlDiver database to identify patients with concomitant hip and lumbar pathology. They created a cohort of about 86,000 patients who underwent total hip arthroplasty (THA) from 2007-2017. Ninety-four percent of these THA patients never underwent a lumbar fusion in the time interval under study. Less than 1% (about 600) of THA patients had undergone a remote lumbar fusion more than 2 years prior to THA, 2.4% (about 2,000) underwent THA within 2 years following a lumbar fusion, and 1.6% (about 1,400) underwent THA followed by lumbar fusion. Thirty-day complication rates including dislocation, infection, and DVT occurred at similar rates in the THA alone and THA after a remote lumbar fusion groups, though the latter group had a higher THA revision rate during the study period (8% vs. 4%, OR=1.9). In looking at the dual diagnosis patients who underwent THA and lumbar fusion relatively close together, the patients who underwent THA first tended to have worse outcomes. Compared to the THA only group, the THA first group had a higher revision rate (9.0% vs. 4.4%, OR=1.9), higher dislocation rate (5.4% vs. 2.0%, OR=2.5), and higher infection rate (6.7% vs. 2.3%, OR=2.7). The fusion first group had a higher revision rate than the THA only group (7.6% vs. 4.4%, OR=1.6), though 30 day complications were not significantly higher. All three groups undergoing both THA and lumbar fusion had higher pre- and post-operative opioid use compared to the THA alone group.

The authors have done a nice job using an administrative database to create a large enough cohort of patients undergoing both THA and lumbar fusion to attempt to answer the question of which operation should be done first. Given that only 6% of THA patients in this cohort had undergone lumbar fusion, it would be nearly impossible to have a large enough sample using anything other than an administrative database to address the question. However, this study design has significant limitations. For one, patients with both back and hip pathology were not randomized to the order of operations, so this choice could have been confounded by many unmeasured variables. The authors did attempt to control for some potential confounders, but the database did not include details such as severity of disease, length of fusion, or BMI. Additionally, patient reported outcomes were not included, so it is unclear if patients had better subjective outcomes (rather than just fewer complications) with one approach or the other. The limited time horizon likely affected the observed revision rates, as patients who had lumbar fusion first likely had THA later in the period under study and had less time in which to undergo a revision. A survival analysis may have been more appropriate for this analysis. The study design also missed THA patients who had underwent lumbar fusion prior to 2007. Additionally, the statistical comparisons were between the THA only group and the dual diagnosis groups, not between the THA first and fusion first groups. Despite these limitations, the data do suggest that dual diagnosis patients may have a lower rate of dislocations in the 30 days after surgery if lumbar fusion is performed prior to THA (3.3% vs. 5.4%). While patients with concomitant lumbar and hip pathology have higher complication rates than those with isolated hip disease, performing lumbar fusion prior to THA may be the preferred strategy for many patients in this challenging cohort.

Please read Mr. Yang's article on this topic in the May 15 issue. Does this change how you consider the order in which to perform lumbar fusion and THA?

Adam Pearson, MD, MS
Associate Web Editor