The Spine Blog

Friday, February 17, 2017

As bundled payments begin to include more aspects of post-operative care, healthcare systems are under increasing pressure to reduce inpatient length of stay and the rate of discharges to care facilities, while also preventing readmissions and maintaining quality measures and patient satisfaction. These goals can seem to be opposed to each other, and new payment models have shifted the risk for balancing these outcomes from payors to hospitals. Lumbar fusion is a big-ticket healthcare intervention that is increasingly utilized, which puts it in the crosshairs of agencies trying to reduce healthcare spending. As such, Ms. Bradywood and her colleagues from Virginia Mason Medical Center created a lumbar fusion care pathway in an effort to standardize post-fusion care, decrease length of stay, and maintain or improve quality measures. The group that developed the pathway included multiple stakeholders, including nursing, physical therapy, physician assistants, and spine surgeons. They recognized the substantial variation in post-fusion care and sought to come to consensus about evidence-based post-operative care. They used Lean methodology and developed an order set embedded in the electronic medical record, a patient handout describing expectations for their recovery, and instituted multidisciplinary rounds focused on discharge planning. This paper is a before and after study comparing outcomes for over 450 lumbar fusion patients who underwent surgery before or after the adoption of the care pathway. They found that length of stay decreased from 3.9 to 3.4 days, the rate of discharge to a care facility decreased from 36% to 25%, and the rate of foley catheter removal by the morning of post-operative day 2 increased from 77% to 87%. They reported no change in the rate of readmission, inpatient pain scores, or patient satisfaction.


This paper is a nice demonstration of how adoption of a care pathway can result in standardized, more efficient care without compromising quality or patient satisfaction. While a half day decrease in length of stay may seem modest, this translates to thousands of dollars of savings. The significant decrease in the rate of discharge to a care facility probably translates to even more savings. While the authors do not discuss this, the largest barrier to adoption of a care pathway is likely surgeon buy-in. Differences in post-operative care between individual surgeons is substantial, and there is very little evidence to guide decisions on topics such as bracing, drain use, and activity restrictions. The fact that this group was successful in having eight spine surgeons come to consensus about best practices on these issues—and abandon some of their practice patterns in the process—is testament to a group committed to improving the post-operative care process. Another significant barrier is changing patient, nursing and physical therapy expectations. If patients and caregivers expect slow mobilization and the use of high dose narcotic PCAs for days after surgery, it will be difficult to move the needle on length of stay. By setting expectations early in the process—this is best started at the time of the pre-operative office visit—patients will not be surprised when they start mobilizing soon after surgery. The authors should be congratulated on the creation of their care pathway and for publishing their results. The field of healthcare delivery science is in its infancy, and groups doing this type of work should be encouraged to share their results. It is difficult to predict how this approach to creating a lumbar fusion care pathway would translate in other healthcare systems’ cultures. Hopefully other groups can share their experiences in creating care pathways, allowing others to learn from their processes.


Please read Ms. Bradywood’s article in the February 1 issue. Does this motivate you to create a lumbar fusion pathway at your institution, or have you already done so? Let us know by leaving a comment on The Spine Blog.

Adam Pearson, MD, MS

Associate Web Editor

Monday, February 13, 2017

Minimally invasive surgery (MIS) techniques have been developed and widely adopted to treat lumbar degenerative conditions. However, there are no high quality RCTs comparing outcomes and complications between MIS and open surgery (OS) for lumbar spondylolisthesis, the most common indication for lumbar fusion. In an effort to fill this gap, Dr. Lu and colleagues from the Mayo Clinic performed a meta-analysis comparing MIS to OS for lumbar spondylolisthesis (both degenerative and isthmic spondylolisthesis were included). They identified 10 observational cohort studies, 5 of which were prospective and 5 of which were retrospective. The studies included 602 patients treated with MIS and 274 with OS. According to GRADE classification system, the 5 prospective studies offered low quality evidence and the 5 retrospective studies were of very low quality. Overall, the ages, gender distribution, and levels fused were similar for the MIS and OS groups, though some individual studies did have more pronounced baseline differences. Meta-analysis revealed the mean blood loss was 330 mL less and LOS was 1.7 days less for the MIS patients. No significant differences were observed for operative time, complications, or patient reported outcomes on the Oswestry Disability Index or visual analog scale.

This meta-analysis likely represents the highest quality evidence available comparing MIS to OS for lumbar spondylolisthesis. However, given the low and very low quality of the included papers, the overall strength of the evidence provided by this meta-analysis is also relatively low. It comes as no surprise that MIS offers the advantage of less blood loss, though given that transfusion following one level open decompression and fusion is uncommon, it is not clear that this provides a clinically meaningful benefit. The LOS advantage also needs to be scrutinized as half of the included studies had mean LOS over 6 days for both the MIS and OS groups, and one had a mean LOS of 10 days for MIS and 14 days for OS. Length of stay tends to be much lower in the United States, with 1-3 day stays being more typical for MIS or OS. Additionally, LOS is frequently affected by provider and patient expectations rather than more objective characteristics like mobility or analgesic requirements. An unbiased comparison of LOS differences would only be possible with a study design in which the patient and inpatient care team were blinded. The major limitation of the included studies, and thus the meta-analysis, is the risk of selection bias. Treatment type was determined by the patient and surgeon rather than by randomization, so the MIS and OS patients may have been very different. The authors provided very little baseline data (i.e. age, gender, number of fused levels), so it is not possible to compare factors known to affect outcomes such as comorbidities, psychosocial characteristics, body mass index, smoking status, work status, and baseline patient reported outcomes. The authors have done a nice job working with the low quality evidence that exists, though meta-analysis is not statistical alchemy, and the strength of its conclusions are limited by the quality of the included studies. This and other studies make it clear that MIS techniques decrease blood loss. Whether or not MIS leads to faster mobilization, decreased post-operative pain, and long-term outcomes equal to OS remains to be seen. The spine community awaits an RCT to answer these questions.

Please read Dr. Lu’s paper in the February 1 issue. Does this change how you view the benefits of MIS? Let us know by leaving a comment on The Spine Blog.

Adam Pearson MD, MS

Associate Web Editor

Friday, February 3, 2017

S2-Alar-Iliac (S2AI) screws have gained popularity, primarily due to some of the complications associated with traditional iliac screws (IS) such as loosening in cancellous bone, screw prominence, technical difficulties with cross-connectors, and the morbidity associated with a wide dissection to the iliac wing. The S2AI technique addresses most of these problems, though there is a learning curve to placing S2AI screws, and some surgeons may find it technically more demanding than IS placement. The group at Johns Hopkins, which popularized the technique, decided to review their experience to compare outcomes between patients treated with IS and S2AI screws. This retrospective cohort study identified 25 IS patients and 65 S2AI patients treated between 2010 and 2014. The average fusion length was 7 levels, and, overall, the two cohorts were relatively similar. The IS group included a greater proportion of tumor patients (20% vs. 6%), including 5 patients who were not S2AI candidates due to partial sacrectomy for tumor. There were also trends towards increased BMP-2 use and L5-S1 interbody fusion in the S2AI group. Additionally, most of the IS patients were from the earlier years in the series. They found a markedly higher reoperation rate (48% vs. 8.8%) and infection rate (44% vs. 1.5%) in the IS group. Additionally, there was also a trend towards a higher rate of distal device failure and screw loosening in the IS cohort. There were no differences in the rate of L5-S1 pseudarthrosis, visual analog scale improvement, or ambulatory status between the two groups.

This paper most likely represents the largest comparative study on this topic. However, the authors probably overreach when they conclude that the lower reoperation rate and lower infection rate is due to the use of S2AI screws. While S2AI screws were associated with significantly lower reoperation and infection rates, this retrospective cohort design precludes any conclusions about causation. This study design is at high risk for confounding by factors like underlying diagnosis, treating surgeon, and unmeasured confounders. While they did perform a multivariate analysis, the number of patients is probably too small to successfully control for multiple covariates. It would have been interesting to see an analysis of reoperations related specifically to the distal fixation (i.e. screw loosening or breakage, screw prominence or pseudarthrosis). It seems like the differences in infection rate may have been more related to patient and disease characteristics than pelvic fixation technique. These data serve as a good starting point to compare outcomes between the two fixation techniques. In order to make any strong conclusions about the relative effectiveness of the two techniques, a randomized trial or at least a prospective study would be necessary. Any such study would require multiple sites in order to generate a sufficient number of patients, and it would be a logistically challenging study to perform. Until such data are available, surgeons will have to rely on retrospective data like the current paper and their judgment in order to select the best pelvic fixation for each of their patients.

Please read Dr. Elder’s article on this topic in the February 1 issue. Does this change your perspective on pelvic fixation? Let us know by leaving a comment on The Spine Blog.

Adam Pearson, MD, MS

Associate Web Editor

Thursday, January 26, 2017

Patient reported outcomes (PROs) have become the standard metric in clinical research and are frequently used in routine clinical practice to get a more quantitative picture of symptom severity before and after treatment. Sometimes baseline PROs are not available, and retrospective studies have used patient recall to obtain baseline values. Such designs are clearly subject to recall bias. To better understand the risk of recall bias, Dr. Aleem and colleagues from the Mayo Clinic performed a prospective study in which 59 spinal stenosis patients undergoing decompression or decompression and fusion (95% had a fusion) reported baseline and final follow-up numeric back and leg pain scores and Oswestry Disability Index (ODI) scores. All patients had at least one year follow-up, with the mean follow-up at 43 months post-surgery. In addition to recording their final follow-up scores, the investigators also asked the patients to recall their pre-operative status and answer the questionnaires the way they did at baseline. On average, patients tended to overestimate the severity of their pain and disability at baseline:  recalled back pain scores were 2.3 points higher than actual baseline scores, leg pain scores 1.8 points higher, and ODI scores 9.6 points higher. There was no correlation between actual and recalled baseline leg and back pain scores, and only a moderate correlation (r=0.44) between actual and recalled baseline ODI scores. There were no patient characteristics associated with more accurate recall. Patients who failed to improve on back or leg pain scores tended to overestimate the severity of their baseline symptoms to a greater degree than patients who did improve.


This is an interesting paper that demonstrates clear recall bias towards a recollection of worse baseline symptoms and disability in a population of patients who underwent spinal stenosis surgery. The clear message is that studies cannot rely on recalled baseline PROs, and these must be recorded prospectively. A more complex issue that arises from these data is patient perception of improvement following surgery. The findings suggest that patients bias their recollection of pre-operative symptoms towards them being more severe than they actually were, which gives them the impression that they had a greater degree of improvement following surgery. In fact, the patients who failed to improve with surgery were the ones with the greatest recall bias towards increased severity of baseline symptoms, suggesting that they shifted their baseline symptoms in order to feel that surgery was effective. Such tendencies might partially explain the weak correlation between improvements in PROs and patient satisfaction. A more detailed subgroup analysis in which the relationship between patient characteristics like education, worker’s compensation status, depression, and narcotic use and severity of recall bias would have been interesting, but such an analysis would have required many more patients. An analysis of the relationship between patient satisfaction, change in PROs, and degree of recall bias would have also been interesting. This paper did have some important limitations, like the use of a written baseline survey and a telephone follow-up survey, though the effect of this was likely small. More concerning was the highly variable duration between baseline survey and final follow-up survey, which likely varied between 1 and 5 years. While the authors noted no relationship between this duration and degree of recall bias, it seems likely that recall bias may be different at one year and five years. Additionally, one third of patients did not complete either the baseline or final PRO surveys, which increases the risk of bias and potentially limits the generalizability of the study. The good news for surgeons out of this study is that patients tend to bias their recollections to believe they were worse off at baseline, which should lead to a higher level of satisfaction with surgery. While such rosy misperceptions can lead to happier patients and surgeons, it is a reminder of how we should measure baseline and post-operative PROs in order to get a more objective sense of how we are doing.

Please read Dr. Aleem’s article in the January 15 issue. Does this change how you consider patient recollection of their baseline symptoms? Let us know by leaving a comment on The Spine Blog.

Adam Pearson, MD, MS

Associate Web Editor

Friday, January 13, 2017

Fortunately for society, gunshot wounds (GSWs) involving the spine and spinal cord are relatively rare. Unfortunately for spine surgeons and researchers, this makes it a particularly difficult topic to study. Unlike blunt force trauma that tends to cause spinal fractures with discoligamentous disruption and instability, penetrating GSWs do not typically cause instability as the soft-tissue stabilizers tend to remain intact. The major clinical problem resulting from GSWs involving the spine is spinal cord or nerve root injury, and these injuries tend to occur at the time of the shooting rather than in a delayed fashion. Given this background, the literature has suggested the surgical treatment of low velocity, civilian GSWs involving the spine is generally unindicated. Given the low number of GSWs involving the spine, there are only a few series in the literature involving large number of patients. In order to fill this void, Dr. Nwosu and colleagues reviewed 489 patients with civilian GSWs resulting in spinal cord or cauda equina injury who presented to Rancho Los Amigos National Rehabilitation Center following acute treatment elsewhere. Of these 489 patients, 91 initially underwent spine surgery. The authors classified the surgery as indicated or not-indicated according to literature-based standards, with indicated surgery including removal of bullets from the lumbosacral spinal canal, removal of subcutaneous symptomatic bullets, and persistent CSF leak beyond 5 days. Surgery for removing a bullet from the cervical canal for root escape, for decompression of neural elements from non-bullet lesions (i.e. retropulsed bone, hematoma), and for progressive neurological deficit was classified as indicated to give the benefit of the doubt to the decision to operate for these debated indications. Of the 91 index surgeries, 69 (75%) were classified as unindicated. Of these unindicated surgeries, 45% of bullet removals, 95% of decompression/fusions, and 66% of CSF leak repairs were deemed unindicated. Telephone follow-up was performed in a subset of patients, and this revealed a trend towards less neurological improvement and greater pain in the patients treated surgically.

This paper represents the largest series of civilian GSWs involving the spine that has ever been assembled, and the authors should be congratulated on their efforts. The high rate of unindicated surgery revealed by this study reflects the poor understanding of these injuries by the spine surgery community. Given that most spine surgeons do not encounter GSWs frequently, they may apply treatment principles for blunt trauma, with which they are more familiar. As this and other studies have shown, instability due to GSWs is incredibly rare, and even fractures involving all three columns are generally stable. Additionally, the spinal cord injury is generally caused at the time of the shooting, and decompression rarely results in neurological improvement. While this study represents the best study on this topic to date, the limitations inherent in a retrospective cohort study need to be considered. Treatment was not randomized, so the patients taken to surgery were likely substantially different from those treated non-operatively. As such, selection bias may have affected the results. Additionally, only a small proportion of patients provided long-term telephone follow-up, so attrition bias could have affected these data. Nonetheless, this and other studies have suggested that surgical treatment for most civilian GSWs involving the spine is rarely indicated. Hopefully this message will get out to the spine surgery community so that unnecessary surgery and the associated costs and complications can be avoided in this unfortunate group of patients.

Please read Dr. Nwosu's article on this topic in the January 15 issue. Does this change your view on the role of surgery for civilian GSWs to the spine? Let us know by leaving a comment on The Spine Blog.

Adam Pearson, MD, MS

Associate Web Editor