I think I'm the last talk, so I want to thank Praveen and Pat for having me, and of course we should all make sure we thank Joni Shulman and the AANS staff for all the work they do to put this together. You've heard a bunch of great technique talks. I'm going to try and just touch a little bit about evidence-based medicine and guidelines. I have no disclosures. It's been over 20 years now since this article came out in JAMA that really sort of put the charge to training programs to incorporate evidence-based medicine into the routine part of how we train medical students and residents, but this is not a new idea. This actually dates back to Shogolay Maslow, who said that the practice of medicine is an art, but it's an art based on science. We'll start with a simple case, and these are all chip shots. We'll start with the first guy here in the row. So this is a 7-year-old lady that comes to your office with this MRI. What's the diagnosis here? She slipped and fell, or no, sorry, she bent over to pick something up and then felt this acute onset of back pain. So you've got the 3 scheduled MRs, and it looks like there's a compression deformity of L5. Okay, so most common, I mean I'll tell you she had no other significant past medical history. What's the most likely diagnosis, 7-year-old lady? After a fall, or she just fell? She was bending over to pick something up off the ground, and she felt this pop and then severe acute onset pain. Well, I'd be concerned about a metastasis. Okay, so she had no past medical history, otherwise previously healthy. Most common thing. Yeah, osteoporotic compression fracture. So what's the diagnosis, what are the options for management, and what do you think is the best option for management? We could observe her, try to get her pain under control, serve her with therapy first. She's neurologically intact. Neurologically intact, yeah, disabling back pain. These aren't meant to be tricky. Like you said, conservative management. Are there any other options? Let's say you try conservative therapy and it doesn't work. She has disabling pain, she can't get out of bed. We talked about how you can do a hypoplastic with a low-velocity injury. Okay, so cement augmentations. Is there any evidence that you're aware of to support one option over the other? I'm not particularly convinced. As many of you may know, there's these companion randomized controlled trials that came out in the New England Journal of Medicine in 2009. It stirred up a lot of controversy because they were a randomized controlled trial, multi-center studies. One of them is the Buckbinder trial, the other is the Kalametz trial. I won't go through both of them, but they were companion articles in the same journal. And they're very similar. They're randomized controlled trials. They looked at patients like this that had pain related to osteoporotic compression fractures. They were sort of innovative in the sense that the placebo arm of this was actually a sham procedure. Does anyone know what the sham procedure was? They opened up the cement so that the patients were awake, so they could smell the cement. And then they put a needle in and they injected some lidocaine or novocaine or something like that into the periosteum. And so these were patients that had one or two levels of pain from osteoporotic compression fractures, minimum of one year duration. For the Buckbinder trial, they actually used MRI to diagnose these. So this was a randomized controlled trial. About half of them had cement augmentation of the vertebroplasty, and the other half underwent the sham procedure. The results were surprising to many people who treat these patients, which is they were at virtually every single time point, virtually every single outcome measure, they failed to show any benefit. So the conclusion by the authors and the investigators were that there was no beneficial effect of vertebroplasty at any time point. The pain function and quality of life improved actually in everybody, regardless of what they did. So anybody here seen anybody post-op after vertebroplasty or kypoplasty? People do pretty well, right? So the patients that had disabling pain and the appropriate indications, appropriate diagnosis undergoing vertebroplasty, anecdotally at least, people seem to do very well. So why did these two randomized controlled trials fail to show any benefit when our clinical experience for those people who commonly treat these patients would suggest otherwise? How do you take this evidence? What does it mean for your practice? What are you going to do when you see a patient that has a painful osteoporotic compression fracture? These are the questions that a lot of the experts in this field brought up. There were a number of responses. This is just one of them. This was published in the Spine Journal by the authors who were part of the leadership of NASS. They asked these very specific questions. They looked at the methodology of these studies. They pointed out probably some fairly significant flaws in the design of those randomized controlled trials. These are things that we do intuitively when we see these patients. They looked at things like fracture acuity. In both these studies, they looked at patients up to one year out. So we know that if someone is a year out and still having back pain from an osteoporotic compression fracture, those are the people we typically treat with a vertebroplasty or cement augmentation. It's a good thing about acute fractures, people within the first four to six weeks or the first couple of months. In reality, in these studies, the people that they treated that were within the first couple of months were actually a fairly low percentage of the patients in the study. A high percentage of the patients were several months out. So that was one of the things they looked at. The other thing they looked at was, the Buckbinders say they used MRI, but in the CALMIS trial, there wasn't really any consistent diagnostic imaging used to determine who was the candidate for enrollment in this trial. They just used clinical history. Again, these were some of the flaws. They looked at other things. When you think about enrolling a patient into particularly surgical-type randomized controlled trials, there's this recruitment bias or this potential confounder-related recruitment. You guys take all the time, right? How often do you guys see a patient and you get a consult for a compression fracture? Pretty common, right? The Buckbinder trial took them four and a half years at several multi-center, high-volume centers to enroll 78 patients. It took them nearly five years to find less than 100 patients that would be willing to enroll in this trial. These were the kind of things that were sort of the issues with this. When you look at this, you consider randomized controlled trials to be the gold standard, but in reality, there were some methodological flaws with this study If you don't delve into that, then you can potentially come to some incorrect conclusions with regards to what the study said. This is another one. This is a 48-year-old guy who was bending over to lift a couch at home and felt acute onset back pain. A few days later, he felt pain radiating down his leg. What's the diagnosis? On the sagittal and axial MRI, it looks like there's an acute disc herniation at the L45 level. Good. These are chip shots. Acute disc herniation. He's got a left L45 disc herniation. What are the options for management? What do you think is the best option for management? I think for him, you could try conservative therapy, physical therapy, or you could consider surgery with an acute disc and a new onset of symptoms. That's what most of us do. Do you have any evidence to support your rationale or your thought process in terms of decision-making? Yeah, based on the SPORT trial. The SPORT study was a $30 million NIH-funded study to look at this. This is a very specific thing that we see all the time and I think that many of us who treat these patients feel like we have a very good understanding of how to manage this, which is exactly what I described. So that's what they did. This was published in JAMA 2006. It was a multi-center randomized controlled trial comparing standard open discectomy versus non-operative management. These were patients that had greater than six weeks of symptoms. Again, this would be who we would typically consider for possible surgical treatment. There was actually two parts to it. There was a randomized group and there was an observational cohort. That's how they felt was best to try and get people enrolled. The JAMA article, this primarily presented the randomized cohort. So there was 500 patients. About half of them underwent surgery. Half of them were assigned to surgery and half of them were assigned to non-operative treatment. But they allowed crossovers. So they did what was considered to be sort of the gold standard, most rigorous way of looking at a randomized study and they didn't attempt to treat analysis. I put this other stuff in white because that's not how it was actually presented in the JAMA article. They just looked at the randomized cohort and they only looked at the intent to treat analysis. Again, at virtually every time point, virtually every outcome measure they looked at, they showed that there was no statistically significant difference in terms of outcome when they did the intent to treat analysis. Again, this stirred up a lot of controversy because there were a lot of sound bites that went into headlines like Wall Street Journal, New York Times, and all that kind of stuff that said that there's no role for surgery in the treatment of acute lumbar radiculopathy from disc herniation. Yet we see these people all the time. We don't operate on everybody, but the people that we do feel would benefit from surgery with a good operation and appropriate indications seem to do well. So how do we reconcile these contradictory findings of what we consider to be the gold standard, a multi-center randomized controlled trial? How do we reconcile this? Can you back up a little bit to that graph? The rate of the spine is an important thing. You know what this graph, what these figures show? They show that these patients were being cared for by conscientious physicians who were helping the patients make the best decisions for themselves. Because ultimately, so many people did not approve that surgery and vice versa. So there were lots of people who randomized that surgery, and as their symptoms improved, appropriately did not get an operation. So when you see these things from an intent-to-treat analysis be perfectly lined up, what that really says is that the physicians and the patients were making the best decisions. Absolutely right. So David hit this nail on the head. So there was a great response. I mean, there were a bunch of responses, but there was a great response, I think, that Paul McCormick wrote that was published in JNS Spine. And he pointed out exactly, the conclusion was exactly what David said, which was what the sports study showed us, taught us, if you really boil it down to it, is that patients are very good at determining what is the best treatment for them. And if you have a conscientious physician that shares in the decision-making with these patients, that ultimately the right intervention will be. Again, Dr. McCormick went through all the methodological and design flaws in the study, and several of these were things like patients that got randomized, a high percentage of them actually at the time of randomization were having improving symptoms. Those aren't people we would operate on anyways. Those are people we continue with non-operative therapy. But really, again, the main issue with this was that there was an extremely high crossover rate, and yet they used an intent-to-treat analysis. And so we'll actually look later at the as-treated analysis in the observational cohort, which I think is more aligned with what we typically think of sort of the management and the expected outcomes for acute disc herniation. A 48-year-old male presents with several years of worsening chronic low back pain. It's worse with climbing stairs. He has no radicular pain, no leg pain. There's no stenosis at that level. So a partner in your practice shows you this case. He's planning an L4-5 ALIP. I'm going to ask this gentleman here in the other seat. He asks you about your thoughts about using recombinant human BMP2 as graft material. He tells you the patient has a history of melanoma treated in the past. Are there any risks associated with that? He'll say yes. He also tells you that the patient told him he was recently married and is planning on having kids. Are there any risks that you think you know of associated with that? For ALIP, absolutely. Good. Well, I won't go over the original randomized controlled trial about using BMP, but there was this controversial study paper that came out in the Spine Journal published by Eugene Kerrigy where he looked at this and he said, well, wait a minute. The industry-sponsored trials with regards to RhBMP2 showed zero adverse events or the conclusion that there were no adverse events, right? That raises some questions. How is it possible for these surgical studies that there were no adverse events that could be related to the use of BMP? That brought up some concerns about the integrity and safeguards within spinal research and all those kinds of things. He and the authors of this paper did a systematic review of the literature and some FDA data. What they concluded in this paper that was published was that there were significant risk of adverse events. This is what they published. They said that with a posterior lumbar interbody fusion, there could be up to 25% to 50% risk of adverse events with the use of BMP2. With ALIP, there could be up to 10% to 15% risk of adverse events. Postulateral fusion, 16% to 20% risk of adverse back and leg pain. What we all know, I think, through the use of BMP2, with anterior cervical fusion can be related to a fairly significant risk of adverse events. What they concluded was the true risk to patients receiving BMP2 is conservatively 10 to 50 times what was originally estimated or calculated. Just a show of hands, residents or trainees that have seen a case of BMP2 of any kind. Have you seen two cases where there is 50% risk with posterior lumbar interbody fusion? Would you expect that would be commensurate with your experience? Probably not. That seems a little questionable. There was a subsequent follow-up. These were the YODA trials, the Yale Open Data Access. There were two of them. They were done independently. They similarly did a systematic review of the literature. But they also had access to individual participant data from the original studies. So they got cooperation with the company that did the original industry-sponsored studies. They got individual participant data. They were able to look at that. They looked at not just adverse events. They did look at clinical outcomes as well. When they did this, again independently, adverse events, they found that actually this was the follow-up to this Carraghee study. They found that actually with ALIP there was no significant difference in terms of adverse events. That the retrograde ejaculation rate did seem to be higher in the cases that were using BMP2. But it was not statistically significant. When they looked at posterior lumbar fusion, those kinds of things, the risk of adverse events was not significantly higher with the use of BMP2. They also looked at cancer risk. There were these two studies. One was by Fu et al. and the other was Simmons et al. They did find that when they pulled all this data, looked at individual participant data, there probably was a higher relative risk of cancer in the cases where BMP2 was used. In the Fu study, they found it to be a relative risk of 3.45. But the data was insufficient to correlate dose with relationship to the risk. They concluded that this should be interpreted with caution because the type of cancer cases were fairly heterogeneous and could have been underreported. The Simmons study actually found that the relative risk was right around 2. Again, there might be an association with the risk for cancer, but the overall absolute risk is relatively low. It's uncertain whether or not there's a direct correlation. This is where we're at. Evidence-based medicine, are we in a movement in crisis? Twenty years ago this was sort of the movement towards evidence-based medicine, but how do we make sense of all this information? We have all this literature and so much of it is conflicting. How do we make sense of it? When we think about how we grade and evaluate evidence, we get beaten to our head. Randomized controlled trials are the gold standard, and yet particularly when we think about surgical randomized controlled trials or spine surgical randomized controlled trials, are those really the gold standard? Are those really going to tell us the information we want to know? When we think about it, there's actually a lot of issues. I think we just pointed them out in a few of these. A lot of these randomized controlled trials, the inclusion-exclusion criteria ends up being so limiting that the patients in those studies don't really have broad generalizability to the patients we actually treat. There's a lot of issues about design. In other words, if you want to do a randomized controlled trial, before you start the study, there has to be some presumed equipoise between the two options, the control or the standard treatment and the investigational treatment. There's ethical concerns about restricting or withholding treatment for certain types of trials. Technique and technology is rapidly evolving. These trials take years and years to basically, from the design stage to the enrollment, doing the study, and completion. Oftentimes, particularly for spine surgical trials, by the time the randomized controlled trial is completed, the data is analyzed and published, sometimes the techniques and the technology is obsolete. Does anyone know what the Everett Rogers diffusion of innovation theory is? What that is, is that basically, any kind of innovation, there's the people that pioneer it, there's the early adopters, and then it gets disseminated, you get the late adopters and the laggers. The idea here, as it relates to this, is oftentimes these randomized controlled trials are done by the experts. These are the people that know this, the innovators, the people that know the disease, the treatment, the best. Eventually, as those trials get published, and the conclusions get published, sometimes that trickles down to people that may not necessarily, the diffusion of that innovation may be then inappropriately used. The Hawthorne effect, do you know what the Hawthorne effect is? That's where basically people who know that they're being studied change, perhaps, the way they report their information because they know that they're being studied. So, randomized controlled trials do not provide a definitive answer, and many fail to change medical practice. When we think about the evidence, we also have to consider now, I think we're getting a good understanding that what we typically thought of as lower quality of evidence actually may have a significant amount of value to us. These are things like cohort studies, case control studies, prospective registries. It's all about context, though. It's contextualizing that data and knowing how to apply it. We always think about this, right? The six most dangerous orders, there's no evidence to suggest. There's never been a randomized controlled trial to show that wearing a parachute helps. So, that's what brings us to guidelines. That's why guidelines, in my opinion, are so important. There's so much evidence out there. How do you synthesize and take that information and apply it to your clinical practice? The guidelines effort is a massive effort done by many people, including a lot of the faculty here involved in that. The guidelines are best when you have an expert panel that really reviews, synthesizes, contextualizes the existing literature, and from that can provide graded recommendations. For guidelines it's important that they need to be a meaningful, relevant disease. They should formulate clinically relevant questions. There has to be a standardized, predetermined methodology. In other words, if you're going to do a guidelines, you have to have a methodology. You've got to stick to it. Ideally, they should minimize conflicts, full disclosure, transparency, and an extensive vetting process. As many of you know, I want to quickly go through this. There was a recent update to the lumbar fusion guidelines. This is, I think, one of the most relevant. Again, it's a very clinically relevant condition. The original guidelines were published in 2005. There were 16 topics, 50 recommendations, a review of literature over almost 50 years. The update was in 2009. It was just recently published. This was the expert panel involved. As in all things, old is good, new is better. I think impressively, despite the disco hair, Pat's weathered the time pretty well. Methodology. If you're going to read guidelines, a lot of people skip the introduction. You've got to read the introduction, know about the methodology, know about how they review the evidence. How do they classify it? It's not actually all randomized controlled trials at level one. It depends. If it's a therapeutic trial, then yeah, a randomized controlled trial might be level one. But if it's a diagnostic study, if it's a prognosis study, if it's a cost effectiveness study, then those aren't randomized controlled trials. There's different levels of grading depending on the type of study. What's really important to look at is how they potentially downgraded studies. When they looked at studies, were there limitations, were there flaws? They would downgrade it from a level one to a level two or a level two to a level three. I think that's an important part. From that, how do they grade their recommendations? We'll just do a couple of real quick cases. This is the case that we saw earlier. This is a 48-year-old male that had persistent back pain, chronic low back pain with no stenosis. Just show of hands, who would primarily treat this person with physical therapy and rehab? Show of hands, how many people would recommend fusion surgery? Most people, non-operative therapy. What were the guidelines for this? This is probably the one that I think is the most controversial out of all of this. This is the guidelines. This was a grade B recommendation, so it was based on multiple level two studies with a lumbar fusion or a comprehensive rehab program that involved cognitive therapy. Our recommended treatment alternatives, again, this was based on multiple level two studies. I'm not going to go through all of these, but I would just suggest for trainees, these are sort of these landmark studies that, despite their flaws and all that stuff, are good to know. The Fritzl study was a study that looked at randomization of surgical treatment versus some kind of non-operative treatment. It was a randomized controlled trial. They were able to show that there were significant benefits to surgery, but this was downgraded. This was downgraded to a level two because there was no consistent non-operative treatment. These are many patients that have failed non-operative treatment for years, and they were potentially randomized to basically more of the same. Then there are these two companion studies. These are the Brock studies. The first one was, again, a randomized controlled trial fusion versus an intensive rehab or comprehensive rehab program. This showed that there was no benefit, no significant difference between the outcomes, but this was also downgraded to level two because they only had 60 patients in the study. 30 of them underwent surgery and 30 of them underwent intensive rehab. This was probably underpowered. They did a companion study, same sort of approach, but these were patients that had chronic low back pain after prior discectomy. Again, no benefit, randomized controlled trial, but downgraded to level two because they only had about 60 patients in the whole study. This was the Fairbank study. Again, another randomized controlled trial, surgery versus a comprehensive rehab program. The conclusions of this is there wasn't really any significant difference between outcomes, but again, downgraded because there was a significant loss to follow-up rate. The surgical arm was very heterogeneous in terms of the approaches and techniques that were done. Again, these were the level two studies that suggest that perhaps for this particular condition, intractable low back pain without stenosis or spinal anesthesis, that lumbar fusion is a treatment alternative, but also intensive rehab with a cognitive therapy component might be an option as well. This is that case we saw earlier, acute lumbar disc herniation. Nonsurgical management, decompression, and fusion. Who would do a decompression and fusion for this? Okay, good, so nobody, right? I just want to show this. We talked about the support study earlier. They actually published the observational cohort. This is the level two evidence that confirms what we all know, which is for patients that have severe lumbar radiculopathy from acute disc herniation that failed conservative therapy, that surgery actually for virtually every single outcome measure for every time point shows improvement. So that's the level two evidence for that. The lumbar fusion guidelines, I didn't actually discuss this, but I just wanted to show as a follow-up. What did the lumbar fusion guidelines discuss? They said, well, based on grade C recommendations, this is based on multiple, I think level four evidence, that lumbar fusion is not recommended as a primary routine treatment for following lumbar discectomy, but there's level four evidence that suggests that for certain situations, there might be a role for lumbar fusion as an adjunct to the discectomy, and these are severe degenerative changes or instability, people with significant chronic axial back pain, or workers with manual labors. I'm not going to go through this, because this is all level four evidence, but there were some studies, again, to support those recommendations. So again, in summary, there's not purely good evidence to support the routine use of lumbar fusion in addition to discectomy. This is a lady that has severe leg greater than back pain. It's worse with standing and walking, or by sitting and leaning over the kitchen counter. So what is this? Lumbar stenosis, neurogenic claudication. Who would treat this with primarily decompression alone? Who would do a decompression infusion? So grade B recommendation, level two evidence to recommend surgical decompression is a primary treatment for symptomatic neurogenic claudication due to lumbar stenosis in the absence of spondylolisthesis. And then there was a grade C recommendation that in the absence of any deformity or documented deformity or instability, that lumbar fusion has not really been shown to improve outcomes when you add lumbar fusion in addition to decompression. So there was a sports study for stenosis without spondylolisthesis, and actually in that the intent to treat analysis still showed some benefit to surgery, but the as-treated analysis showed that there was a significant benefit with surgical treatment in these patients. With regards to the fusion component, there have been multiple studies that have shown that it's difficult to interpret the value of fusion because they've been very heterogeneous in terms of the approaches and the indications and whatnot. So based on the literature that we have now, decompression is recommended as a primary treatment for stenosis without spondylolisthesis. This is actually one of my patients I treated. She did well. I did all three, four laminectomies. She did well. And probably about a year and a half later she came back with worsening back greater than leg pain. So what's going on here? She's had a recurrence of spondylolisthesis. Anything else? So she's got a spondylolisthesis associated with that, right? So this is different than the initial scenario, right? So who would treat this now with a repeat decompression alone? And then who would do a decompression and fusion? I guess the majority, right? So this was lumbar fusion for stenosis with spondylolisthesis. This was grade B recommendation. Multiple level two studies that show that surgical decompression with fusion is recommended as the effective treatment alternative for symptomatic stenosis associated with spondylolisthesis. There was a support study, again, that looked at this randomized controlled trial. But if you looked at the observational cohort, level two evidence that showed that surgical treatment with decompression and fusion improved outcomes compared to non-surgical treatment. And so again, the evidence supports surgical treatment. Most of these studies were instrumented post-lateral arthrodesis, just so you know that. And so there's a variety of surgical alternatives that may be considered. And so the recommendation here was that each surgeon should choose the appropriate technique based on their experience, risk of complications, or whatnot. So knowing that, we talk about decompression. Who would do an un-instrumented post-lateral fusion? Nobody. How about post-lateral fusion with pedicle screws? Interbody fusion with pedicle screws? And then interbody fusion plus post-lateral fusion with pedicle screws? So pedicle screws, fixation as an adjunct, grade B recommendation based on a single level one study. This was the Fishgren study that did a randomized controlled trial of patients with degenerative spondylolisthesis randomized to instrumented fusion versus un-instrumented fusion. They showed that the fusion rate was higher with the addition of pedicle screws, but they were not able to show that the clinical outcomes were necessarily better. In terms of interbody techniques, interbody fusion, this is a great class B recommendation. Adding the interbody part enhances fusion rate, which lowers the re-operation rate in patients undergoing lumbar fusion. The addition of post-lateral fusion to the interbody is not recommended because the evidence seems to suggest that it indicates no benefit, but it adds an increased rate of complications. I think we're running out of time, so I'll just conclude that if you look at the recent cervical spinal cord injury guidelines, there's some forward from that. When we think about guidelines, Dan Resnick says evidence-based guidelines cannot exceed the strength of the literature. So medical evidence is really there not to provide rule following, but to help structure our decisions that will be also very much dependent on our clinical judgment and also our conversations and shared decision-making with the patient. Mark Hadley said guideline documents are not tools to be used by external agencies to measure or control the care provided by clinicians. Well, I will tell you that the reality is absolutely external agencies are trying to use the evidence to control the way that we provide care. So knowing the evidence is important. So as Mike pointed out, I think he started this by saying spine is awesome, I agree, but you have to know the evidence. So read critically, contextualize the data, apply your personal experience and expertise, because that's really how you provide better patient care and hopefully provide responsible health care. Just to prove how important evidence-based medicine is, all the societies and the committees that have invested in this, the AANS, the CNS, the Joint Guidelines Committee, N2QAD, NRAF supporting research, the Joint Spine Section, Washington Committee, Payor Policy Division, these are all the stakeholders that are very much invested in evidence-based medicine guidelines to again do these things to demonstrate that we can provide better patient care and provide responsible health care. Thanks.

evidence-based medicine

guidelines

medical practice

study design

patient population

lumbar fusion guidelines

shared decision-making