The Spine Blog

Friday, September 14, 2018

It is well-established that greater surgical invasiveness and patient frailty are associated with increased complications. However, the interplay of those two factors has not been explored across different lumbar diagnoses. Additionally, the effect of frailty on patient reported outcomes is also not well-understood. In order to better understand these issues, Dr. Yagi and colleagues from Japan evaluated patient reported outcomes (PROs) and complications following surgery for adult spinal deformity (ASD), lumbar degenerative spondylolisthesis (DS), and lumbar spinal stenosis (SpS) in over 450 consecutive patients with at least 2 years of follow-up. They recorded baseline modified frailty index (mFI) and Charlson Comorbidity Index (CCI) and also recorded baseline and two year PROs (ODI and SRS scores for ASD patients, SF-36 PCS and MCS scores for DS and SpS patients). At baseline, they found the ASD patients had significantly higher mFI and CCI than DS and SPS patients. When stratifying by frailty and CCI, they found that patients with greater frailty and comorbidities tended to have worse baseline PROs. For the ASD patients, two year ODI scores were significantly worse for the most frail patients, while the 2 year PCS scores for DS and SpS patients were not associated with frailty. The authors did not calculate change scores across frailty strata, though the figures appear to show similar change scores for the different frailty groups or possibly even greater improvements for the frailest patients given their worse baseline scores. Major complication rates were highest in the ASD group and increased with increasing frailty, while the patterns were less consistent for DS and SpS. The authors concluded that frailty was associated with worse PROs and increased complications in ASD surgery, but that those relationships did not hold in DS and SpS.

The authors have done a nice job analyzing the interaction between frailty and surgical invasiveness and have confirmed what surgeons would expect—namely that frail patients have worse outcomes and higher complication rates than healthier, more robust patients. It is somewhat difficult to interpret the results for DS and SpS given that only 10% of DS patients and 2% of SpS patients were classified as frail (compared to 24% for ASD). Additionally, the authors recorded different outcome measures for ASD (namely the ODI) than for DS and SpS (namely the PCS), making comparisons of PROs across the diagnostic categories impossible. Most surgeons would believe that increasing frailty would lead to worse outcomes and higher complications following any surgery, not just major deformity surgery. While the differences may be more pronounced with major surgery, the relationship should hold for any type of surgery as long as the study is appropriately powered. Given the low number of frail patients in the DS and SpS cohorts, this analysis was likely underpowered. The authors also did not provide many details about the ASD surgery other than stating that patients had Cobb angles over 20 degrees and fusions spanning at least five levels. While the authors conclude that PROs are worse for the frailest ASD patients as compared to the healthier patients, the change scores are actually greatest for this group. This suggests that frail patients still gain significant benefit from ASD surgery, but they experience a high rate of complications. The paper does not provide any clear cut decision rule or algorithm to help with the decision about proceeding with ASD surgery in the elderly population, though surgeons can inform these patients that surgery will likely lead to better pain and function than their baseline but that they may have a bumpy road getting there.

Please read Dr. Yagi’s paper in the September issue. Does this change how you view the role of frailty in surgical decision making? Let us know by leaving a comment on The Spine Blog.

Adam Pearson, MD, MS

Associate Web Editor

Friday, September 7, 2018

As the population continues to age, the number of fragility fractures including hip fractures, distal radius fractures, and Type II dens fractures is on the rise. The best treatment for Type II dens fractures in the elderly remains controversial, with both observational cohort studies and large database analyses suggesting a possible survival advantage associated with surgery.1-3 These studies have generally included all patients over 65, and approximately 50% of Type II dens fracture patients are over 80.  One prior study has looked at the over 80 age group exclusively and reported in-hospital mortality rates of 15% for non-surgical patients and 12.5% in surgical patients.4 Large series of octogenarians with Type II dens fractures treated surgically are rare, but the group at Shock Trauma in Baltimore was able to assemble a consecutive series of 43 such patients treated over a 10 year period. All patients had significantly displaced (> 5 mm) and/or angulated (> 15 degrees) fractures and underwent posterior C1-C2 Harms fusion. They noted the majority of patients with less displaced fractures were treated in a collar. The average age was 84, and patients had a mean Charlson Comorbidity Index of 1.4. They reported a 30 day mortality rate of 2.3% (one patient), with a 1 year mortality rate of 19%. There was a high rate of complications, notably delirium (42%), dysphagia (28%), feeding tube placement (14%), and reintubation (9%). They did not report any cases of infection or hardware failure. The authors concluded that this represented an acceptably low mortality rate in-line with or better than previous reports.

The authors have done a nice job putting together a large case series of octogenarians undergoing surgery for displaced Type II dens fractures. Such a series can probably only be put together at a handful of trauma centers in the United States, and they only had about 4 of these cases per year at a very busy institution. While the results are reassuring and suggest that this surgery can be done with a low short-term mortality rate in this population, it does not provide us with much guidance on how to treat the average, elderly patient with a Type II dens fracture. Most of these fractures are less displaced than those included in the current study, and it is not clear if surgery is advantageous in this population. While multiple studies have shown a potential survival advantage associated with surgery, these have all been observational studies subject to a high risk of selection bias. While statistical efforts can be made to control for potential confounders, unmeasured confounders always exist. The current study did not offer a control group to which outcomes could be compared, and it is possible that the patients treated with surgery were substantially healthier than those treated non-operatively. It is not clear if the low mortality rate observed in the current series is due to the high quality care provided or due to the selection of healthier patients than those included in prior studies. The low mortality rate is reassuring and should give surgeons treating these challenging patients some peace of mind when they decide to operate on a displaced Type II dens fracture. Until a high quality RCT is done to address this question, the optimal treatment for less displaced Type II dens fractures in the elderly will remain unknown. Given the challenges involved in completing such a study, I would not anticipate seeing Level 1 data anytime soon.

Please read Dr. Clark’s article in the September 15 issue. Does this change how you view surgical treatment of displaced Type II dens fractures in the elderly? Let us know by leaving a comment on The Spine Blog.

Adam Pearson, MD, MS

Associate Web Editor


1.            Pearson AM, Martin BI, Lindsey M, Mirza SK. C2 Vertebral Fractures in the Medicare Population: Incidence, Outcomes, and Costs. J Bone Joint Surg Am 2016;98:449-56.

2.            Schoenfeld AJ, Bono CM, Reichmann WM, et al. Type II Odontoid Fractures of the Cervical Spine: Do Treatment Type and Medical Comorbidities Affect Mortality in Elderly Patients? Spine (Phila Pa 1976) 2011;36:879-85.

3.            Vaccaro AR, Kepler CK, Kopjar B, et al. Functional and quality-of-life outcomes in geriatric patients with type-II dens fracture. The Journal of bone and joint surgery American volume 2013;95:729-35.

4.            Smith HE, Kerr SM, Maltenfort M, et al. Early complications of surgical versus conservative treatment of isolated type II odontoid fractures in octogenarians: a retrospective cohort study. J Spinal Disord Tech 2008;21:535-9.


Friday, August 31, 2018

The spine literature is replete with basic and clinical science papers that frequently do not offer much practical assistance in the day to day practice of a spine care provider. One of the challenges facing spine centers is matching new patients to the appropriate provider. Many multidisciplinary spine centers combine non-interventional spine care providers, pain physicians who perform injections, and orthopaedic and neurosurgical spine surgeons into a large group practice. Given that only about 10% of spine patients will go onto spine surgery, some type of triage process is necessary to identify new patients who are appropriate to see a surgeon on their first visit. An alternative approach is to have non-operative providers screen all patients prior to surgeon referral, but this is less efficient than identifying likely surgical candidates in the scheduling process. Most spine centers have a non-evidence-based triage process that is designed using empiric principles rather than actual data. In order to improve this process, Dr. Boden and colleagues from Emory Spine Center in Atlanta reviewed over 8,000 patients who had seen their spine surgeons for lumbar problems over an 11 year period. They created a multivariate model to determine the strongest predictors of undergoing lumbar surgery within a year of presentation to the surgeon. All of the analyzed factors came from a patient questionnaire about their demographics characteristics, history, and current symptoms. No data from radiographic reports, validated spine specific outcome questionnaires, or patient treatment preferences were included. They found that the presence of leg symptoms was by far the strongest predictor of undergoing surgery, with an odds ratio of 45 compared to patients with only low back pain. As such, the authors limited the analysis to only the patients with leg symptoms. In this group, the strongest predictors were the presence of leg pain (OR = 4.1), leg pain worse than back pain (OR = 2.0), non-smoker (OR = 1.4), worsening leg pain (OR = 2.1), and age over 65 (OR = 1.2). Six other factors were also significant predictors of undergoing surgery. Based on these results, they created scoring systems based on 11 and 5 questions, which stratified patients into low, medium, and high likelihood of undergoing surgery. These groups had surgery rates of approximately 33%, 43%, and 55% (5 question)-58% (11 question), respectively. This compares to a baseline surgery rate of 40% after using the traditional triage process.

The authors have done nice work creating an evidence-based approach to triaging patients in a multidisciplinary spine clinic. They validated the model in their own population, suggesting that it is at least internally valid. As they point out, the rule may not work in different patient populations or with surgeons with different indications for surgery. The rule has face validity given that leg pain is present in patients with radiculopathy or neurogenic claudication, the two main clinical conditions for which lumbar surgery is performed. Questions about smoking, employment, and BMI suggest that this group of surgeons is less likely to operate on smokers, the obese, and patients on worker’s comp or disability, groups that are all known to have worse surgical outcomes. Other surgeons may operate on these patients more frequently. Two other important factors that could be evaluated in the scheduling process that were not included in this analysis are radiographic findings and patient enthusiasm for surgery. While only clinically trained providers can evaluate imaging definitively, certain buzzwords in MRI reports can be used by administrative staff to increase the likelihood of a patient being a surgical candidate (i.e. disk extrusion, severe stenosis, spondylolisthesis, etc.) Additionally, patients who report they have no interest in spine surgery are less likely to undergo surgery, and this could be used in the screening process. Whether or not this triage rule is helpful to other spine centers remains to be seen. However, it does provide some additional information to spine centers planning a triage questionnaire and may be motivation for other groups to perform a similar analysis of their practice. The Spine editors should be congratulated for publishing this type of practical health services research, which may directly impact how clinicians structure their practice.

Please read Dr. Boden’s article in the September 15 issue and the accompanying Point of View by Dr. Pugely. Would this type of triage process help your practice? Let us know by leaving a comment on The Spine Blog.

Adam Pearson, MD, MS

Associate Web Editor

Sunday, August 26, 2018

The opioid crisis has been in the news for years, with no indication that it will be successfully controlled anytime in the near future. Physicians have been a major contributor to the problem through their prescribing narcotics for chronic musculoskeletal and other non-cancer pain. Studies have shown that patients who are taking narcotics prior to spine surgery and those who remain on these medications for prolonged periods post-operatively have worse outcomes than patients not using narcotics. As such, it would be useful to determine risk factors for prolonged narcotic use following lumbar fusion so that modifiable factors could be addressed pre-operatively and patients could be counseled about their risk of long-term use. In order to better understand risk factors for long-term opioid use following lumbar fusion, Dr. Kalakoti and colleagues from Iowa used Humana claims data to determine risk factors for opioid use one year after lumbar fusion. They identified over 26,000 patients who underwent lumbar fusion (ALIF, P/TLIF, posterolateral fusion, or AP fusion) from 2007-2015 who had claims data available at least 3 months prior to surgery and for at least 12 months post-operatively. Patients were classified as opioid users (OU) who had received a prescription within the 3 months leading up to surgery (58%) or opioid naïve (ON). Opioid users were more likely to be under 50 (9.2% OU vs. 5.1% ON), male (42% OU vs. 40% ON), and live in the South (66% OU vs. 60% ON). At one year, 42% of OU and 9% of ON patients were still receiving prescriptions for narcotics. Multivariate regression demonstrated that pre-operative opioid use was the strongest predictor of opioid usage at 1 year following surgery (ORs between 4.6 and 7.8 for the different fusion techniques). Other well-known risk factors such as depression and fibromyalgia were also independent predictors of long-term opioid use, but pre-operative narcotic use was by far the strongest predictor.

The results of this study come as no surprise to spine surgeons and others that care for patients following lumbar fusion. The fact that the majority of patients had received a narcotic prescription before surgery was somewhat disconcerting given that back and radicular pain are not good indications for narcotic use. The use of a large claims database was a good method to gauge narcotic use across the United States over nearly a decade. The major limitations related to claims data are that details such as duration and dosage of narcotics, indications for surgery, patient-reported outcomes, smoking status, and worker's compensation status are lacking. Nonetheless, pre-operative narcotic use was such a strong risk factor for long-term use that inclusion of these other factors would be unlikely to change the primary conclusion. The main question raised by these findings is whether weaning off opioids pre-operatively—a difficult thing to do for patients in chronic pain—would decrease the risk of long-term use post-operatively. While the authors suggest this is true, this paper offers no data to indicate that is the case. It is possible that weaning opioids pre-operatively will not change patients' predisposition for long-term use post-operatively. Certain non-modifiable psychological and physiological factors may put some patients at risk for chronic pain and long-term narcotic use that would persist despite pre-operative weaning. Nonetheless, minimizing pre-operative opioid use makes good sense even without Level 1 evidence supporting it. The most important message from this and other papers on this topic is that narcotics are not appropriate for treating back pain, and providers caring for spine patients should generally not prescribe them other than for acute post-surgical pain.

Please read Dr. Kalakoti's article on this topic in the September 1 issue. Does this change how you view opioid prescribing for lumbar fusion patients? Let us know by leaving a comment on The Spine Blog.

Adam Pearson, MD, MS

Associate Web Editor​

Saturday, August 18, 2018

Spine surgeons are familiar with the pitfalls associated with thoracolumbar fusion in osteoporotic patients, namely screw loosening, cage subsidence, fracture, and pseudarthrosis. Given the increasing rate of spinal fusion in elderly patients, surgeons are encountering osteoporotic bone more frequently. Bisphosphonates and teriparatide are the most commonly used medications to treat osteoporosis. However, animal models have raised questions about bisphosphonates potentially interfering with bone healing and spinal fusion. On the other hand, teriparatide has been consistently shown to improve spinal fusion in animal models. In order to better clarify the literature on this topic, Dr. Buerba and colleagues performed a meta-analysis looking at the effect of bisphosphates and teriparatide on fusion rate, screw loosening, fracture, and patient reported outcomes. They identified 9 comparative studies, including 3 RCTs, 4 prospective cohort studies, and 2 retrospective cohort studies. Four compared bisphosphonates to controls and demonstrated trends towards increased fusion rate (OR = 2.2. p = 0.09) and lower screw loosening rate (OR = 0.45, p = 0.19) for the bisphosphonate group. Only one study compared teriparatide to a control group, and this showed higher fusion rates and lower screw loosening in the teriparatide group. In two studies comparing bisphosphonates to teriparatide, the teriparatide group had a significantly higher fusion rate (OR = 2.3, p < 0.0001) and trend towards lower rate of screw loosening (OR = 0.37, p = 0.09). Compared to controls, bisphosphonates were associated with a lower fracture rate at the fused or adjacent levels (OR = 0.18, p = 0.0007). Patient reported outcomes were generally not different between the groups.

The authors have done a nice job quantitatively summarizing a heterogeneous literature on this topic. This heterogeneity is also what makes interpreting the results difficult, as treatment duration, dose, type of bisphosphonate, and definition of osteoporosis varied across studies. Bisphosphates work on a complex pathway in which osteoclast inhibition can effect both bone resorption and formation, and we do not have a clear understanding of the effect of specific drug type, duration of use, or dosage on spinal fusion. The effect of teriparatide seems more straightforward, though duration of treatment remains variable across studies. There was also a heterogeneity of surgeries included in the studies, ranging from one level fusions for degenerative spondylolisthesis to long thoracolumbar fusions for deformity. Despite this being a meta-analysis, due to the different comparisons across studies (i.e. bisphosphonate or teriparatide vs. control, bisphosphonate vs. teriparatide), the actual number of patients in each comparison was relatively low and limited the study's power. Despite these limitations, the study does allow some big picture conclusions. The most important is likely that bisphosphonates due not seem to impair fusion, so it may be better to continue these medications in osteoporotic patients rather than stop them prior to fusion. Additionally, teriparatide seems to be favored over bisphosphonates if a new agent is going to be started peri-operatively. Based on the available literature, teriparatide appears to be indicated for osteoporotic patients undergoing spinal fusion. Future studies are needed to define when to start the medication pre-operatively, optimal dosage, and duration of treatment.

Please read Dr. Buerba's article on this topic in the September 1 issue. Does this change how you view the role of bisphosphonates and teriparatide in thoracolumbar spinal fusion? Let us know by leaving a comment on The Spine Blog.

Adam Pearson, MD, MS

Associate Web Editor