Dr. Shuvro Roy talks with Dr. Daniel Daneshvar about the prevalence of CTE at death among NFL players and the association between CTE severity and the risk of dementia. 

Read the related article in the BMJ. 

Disclosures can be found at Neurology.org. 

Podcast Transcript: 

Dr. Jose Merino: This is Jose Merino, Editor-in-Chief of the Neurology family of journals. The Neurology Podcast provides practical information to neurologists and other clinicians to help them provide better care for their patients. Thanks for listening, and have a great week.

Dr. Shuvro Roy: Hi, this is Shuvro Roy from the University of Washington, and welcome to today’s Neurology Podcast. We’ll be covering a study titled Prevalence of Chronic Traumatic Encephalopathy and Death in National Football League Players: A Retrospective Population-Based Cohort Study from 2008 to 2001, published in BMJ. We’re joined by lead author Daniel Daneshvar, Chair of Spaulding Rehabilitation at Harvard Medical School. Dan, thanks for joining us.

Dr. Daniel Daneshvar: Great to be here. Thanks for having me.

Dr. Shuvro Roy: So Dan, this study has already received widespread national attention, but please give us the headline. What did you find and what’s the number one thing that you want listeners to remember?

Dr. Daniel Daneshvar: So this was a study looking at how commonly this specific neurodegenerative disease called chronic traumatic encephalopathy or CTE, how commonly this occurs in a cohort with a lot of repeated traumatic brain injury exposure, namely former National Football League professional football players. And what we found was that between the years of 2016 and 2021, so for that five-year period, the absolute minimum prevalence of CTE at death in this group was 25%. The absolute maximum was 98%, and the true rate is somewhere in that range. Now, for context, in the general population in community cohort studies, the prevalence of this pathology is somewhere south of 1%.

Dr. Shuvro Roy: And what does this study add to our current understanding of CTE and neurologic disease associated with head trauma?

Dr. Daniel Daneshvar: So we’ve long known that individuals who have experienced repeated traumatic brain injury exposure, whether through sports like NFL, NHL players, professional soccer players, whether through occupational exposures like in the military or as in individuals who experience intimate partner violence, we’ve long known that these individuals have higher rates of neurodegenerative disease. But what we haven’t known is just how much higher that rate is. And what this study has shown us is first off that this is alarmingly high in former professional football players at death. And then the second thing that this study shows us is that this pathology is unsurprisingly very linked to clinical symptoms, including with more severe disease, cognitive impairment that’s so severe it results in dementia, meaning impairment in someone’s ability to perform activities of daily living, including feeding, bathing, dressing, grooming, toileting, those kinds of things.

Dr. Shuvro Roy: As you pointed out, your primary finding was that there was this maximum versus minimum prevalence, which ranged from anywhere from 98% to 25%, so obviously a wide range. But to help us understand this study, can you walk us through that methodology, the minimum and maximum prevalence approach? Why a range rather than a single estimate? And you mentioned that the number is likely somewhere in between. Where do you think that true number sits, if we even can estimate?

Dr. Daniel Daneshvar: So to understand why there’s a range, it’s helpful to understand how we got to these numbers. And so when we looked at the individuals who had CTE who played in the NFL across three brain banks, mostly the BU CTE Center, but also the Mount Sinai Brain Bank and the UCSF Alzheimer’s Disease Center Brain Bank, when we combined all NFL players across these brain banks, we found that NFL players had a 93% prevalence of CTE at death, but that 93% represents a selected cohort. These individuals who chose to donate their brains more often than not chose to donate because they or a loved one had some concerns about them, not always, but many times. And so we don’t believe that the true prevalence in NFL players is 93%, but what we wanted to get at is what could that true prevalence look like? And so we took advantage of the fact that amongst the NFL community, brain donation is something that has been less stigmatized over the past couple decades and is something that is very common. And so we said for the lowest, the conservative minimum prevalence of CTE, let’s say that everybody who was not studied by one of our brain banks, that every single professional footballer who died during that time period had no CTE at all in their brain. If we make that assumption, then the absolute minimum over those last six years of the study is that 25%. Now, it’s important to note what that assumption entails. That assumption assumes that, for example, the people who were diagnosed with CTE at other brain banks that were not included in this study, it assumes that they did not have CTE, even though they were diagnosed with CTE at other brain banks. That conservative minimum assumes that the 16% of people with dementia so severe, it was listed as the actual cause of death on their death certificates, the 16% who didn’t donate, that they did not have CTE either. Now, of the people who did donate, for the folks who had dementia, 96% of them had CTE. So that’s quite an assumption. We’re assuming that they were all clean too. Despite those assumptions, the conservative minimum of CTE prevalence amongst these NFL players is one in four. Now, the maximum on the other side assumes that the only individuals who played in the NFL who died during that time period that didn’t have CTE were the ones that we laid eyes on, and so that everybody else, they did have CTE. That’s how we get the upper bound. And so when I say that range is somewhere in the middle, based on those assumptions, it’s almost certainly higher than that 25% and likely substantially higher.

Dr. Shuvro Roy: So essentially trying to maybe correct for the possibility that maybe there’s a «sick bias» or selection bias of the brains that were donated, but even the 25% is almost certainly too conservative, considering it is likely that those folks with dementia, at least some of them had some degree of CTE. To that point, beyond prevalence, you showed that stage four CTE was strongly tied to dementia with 90 plus percent of those donors affected. What does that add to the argument about CTE pathology mattering clinically? Can you speak a little bit more to that?

Dr. Daniel Daneshvar: It’s pretty clear at this point that CTE pathology is associated with a substantial cognitive impairment, especially as it gets severe. In a different study, we took a look at individuals who had CTE in the absence of any other neurodegenerative disease, and we found in that study that individuals with more severe pathology were more likely to have dementia and specifically with stage three and stage four CTE. In this study, similarly, we also found that in individuals with CTE in the absence of other neurodegenerative pathology, which represented 53% of donors who had CTE that had no other neurodegenerative disease had dementia too. So CTE clearly can result in dementia when the pathology is severe. But that’s, again, not surprising for anybody in the neurodegenerative disease space because the pathologic marker that’s most closely linked to cognitive impairment in a disease like Alzheimer’s disease, it’s not the amyloid pathology, it’s the tauopathy, it’s the phosphorylated tau, the same neurofibrillary tangle pathology that we see in Alzheimer’s. When that disease occurs in a different pattern, but also in CTE, it’s not surprising that we see individuals with cognitive impairment.

Dr. Shuvro Roy: You also looked at the death certificates of the donors in these cases, and death certificates missed roughly 70% of the dementia that your clinicians identified. I think about CTE in the news and how many prior players who have since been found to have CTE posthumously were not necessarily identified to have dementia at the time. What does this finding mean for some of the earlier NFL mortality studies that relied on these numbers of the percentage of players who had dementia?

Dr. Daniel Daneshvar: Again, it’s not particularly surprising for anybody in the neurodegenerative disease space that death certificates under-report the prevalence of dementia, because many individuals die with dementia but may die of say an MI or something else. But we’ve both been the beleaguered intern in the ICU in the middle of the night that has unfortunately had to write a death certificate for a patient of ours. And oftentimes what we’re writing is the most proximal cause of death. So we might write that sepsis or pneumonia was the cause of death, but might not include on the death certificate that the sepsis was due to urosepsis related to dysphagia related to dementia. Those factors aren’t typically what we are including in the death certificates. And so we know that death certificates are under-counting the rate of neurodegenerative disease. But what this study allowed us to do was put a number to it to let us know that we’re missing the vast majority of individuals who are dying with dementia but of other processes if we only look at the death certificates. And so when we look at death certificates and we try to ascertain the causes of death for individuals, we might erroneously think that something like a heart attack is a bigger issue for NFL players than something like dementia, but that’s why it’s so important to take a look at what the population averages are. And so when we look at something like MI for NFL players, they die with MI, as we showed in a paper we published in Lani City Clinical Medicine last month. They die at a lower rate than the general population of MI, unlike dementia where neurodegenerative disease causes of death occur at a substantially over four times higher rate than the general population.

Dr. Shuvro Roy: Understood. Dan, I want to change tracks slightly. There’s some really interesting findings in these studies that build off of prior work, as you mentioned. In this study, for example, career length predicted how severe CTE was, but not whether it was actually present. Back a few years ago, we had the opportunity to interview Anne McKee, one of your co-authors on this study on the same topic. And one of the major takeaways, at least at that time, was that duration of career did predict CTE presence in those studies. How should we think about these newer data in that context of those prior studies?

Dr. Daniel Daneshvar: Yeah, so great question. In this study, 93% of the decedents who donated had CTE pathology. And so we’re looking at a profound ceiling effect rather than our previous work, which has looked at individuals across the entire range of exposure. To put it differently, what we effectively found in this study is that if you had 90 plus pack-years of smoking, that smoking 90 versus 95 pack-years didn’t really play as big a role in whether or not you got lung cancer. It played a bigger role in how severe that lung cancer was for the people who got it. But there are some people who for genetic or other reasons, they just didn’t get the disease. That’s what we found effectively in these analyses. When you restrict it to the most extreme outliers in exposure, the people who had the most exposure to repeated traumatic brain injuries, the relatively small differences between that extreme end, it didn’t predict CTE status, but it did predict the severity of the pathology. So if you had 16 years of American football versus 18 years of American football, that extra 10,000 Gs or so of head impact exposure, that didn’t predict CTE status. But for those who unfortunately did have CTE pathology, it very strongly linked to how severe that pathology was.

Dr. Shuvro Roy: Understood. That makes a lot of sense. So essentially, that signal may still exist. It’s more just with this particular population, given the sheer prevalence of CTE, we likely would not have the granularity to dig out the presence or lack thereof of CTE and how much that played into it, really just the severity. So that makes sense.

Dr. Daniel Daneshvar: It’s also just important to note that the majority of the people that were included in this cohort, they were part of the previous studies that included the full range of donors. For example, there were 338 people in this study that donated that played in the NFL. Now, in our previous work, we looked at not just the NFL players, but we looked at over 600, I think it was 636 or so people who played football across the entire range from one year of football all the way up to 30 years of football. And in that study, we found a very clear dose response relationship between both the duration of play and the cumulative force to the head that these individuals experienced, and both the likelihood of CTE and for those that got it, the severity of CTE. So the best predictor that we have right now for whether or not someone’s going to develop CTE, it appears to be the cumulative force to the head that someone experiences over their entire life. Of note, concussion did not predict CTE status. Of note, position played at the highest level did not predict CTE status. It really appears to be that cumulative force to the head.

Dr. Shuvro Roy: It’s those cumulative sub-concussive hits that seem to be the most predictive for the body of work that we’ve seen thus far. Now, in prior work in this space, as you just mentioned, there has seemingly not been a correlate between CTE and football position played. That was always surprising to me. You might think that positions, there is that repeated trauma, offensive line, defensive line, linebackers, running backs, et cetera, would seemingly be at higher risk. In this study, you had some newer findings. Could you take us through those findings related to position played and the implications there?

Dr. Daniel Daneshvar: In this study and in previous work, we found that there isn’t really a great link between position played at the highest level and CTE status. Now, of note, part of the issue is that we think of these repeated head impacts in terms of the number of hits, but what appears to matter more is the number times the force of those hits. And so folks like the linemen who are getting a lot of hits tend to be getting lower G-force hits. And in fact, actually, when you look at the cumulative force to the head, the relative difference between positions, they aren’t that high. They exist, but they aren’t that high. But it’s actually the running backs who get the highest combination of force and number of hits versus the lineman. Now, another thing that we note is we think of these other hits that don’t result in concussion. We often refer to them as sub-concuss, and that terminology is actually problematic because it masks the fact that so many of these head impacts are actually really substantial. For example, based on helmet sensor data, for a given offensive lineman in college who gets a concussion, going back and looking at all of the head impacts that that person experienced, for every head impact that resulted in a concussion, on average, that offensive lineman experienced 41 hits to the head of equal or greater force than the head impact that resulted in that concussion. And for that reason, we don’t think that the term sub-concussive is really the best name for those impacts that don’t result in concussions because sub implies that it’s less than, when in fact, very many of these head impacts are technically, if you had to give it a name, supra, they’re greater than the force of a concussion. And we published this in BJSM a couple years ago. We think the terminology should actually be non-concussive so that there are impacts that result in symptoms or signs, that’s a concussive injury, but that there are impacts that don’t result in signs or symptoms of concussion. Those are separate, those are non-concussive, and many of them can accumulate and cause subclinical traumatic brain injuries.

Dr. Shuvro Roy: I have certainly been guilty of using the term sub-concussive, so I will be correcting my terminology as well, and that’s quite helpful. One thing actually you mentioned with the running backs, while there was no difference in the rate of CT presence, was there any difference that we saw with the presence of stage four CTE by position?

 Dr. Daniel Daneshvar: One of the issues when we’re looking at NFL players is that most NFL players, they didn’t stick with one position. Most players, if you are a wide receiver at the NFL, many of them were the best player in their high school and were pretty much playing every position, but most often quarterback because you wanted your best player in the position that had the most touches. And so that’s one of the problems when we look at the highest levels of play. But when we did look at highest level of play, what we found is that individuals who had stage four CTE were actually most likely to be running backs and least likely to be defensive linemen. So these data lined up really well with our 2023 Nature Communications paper where we similarly found that it appeared that the running backs were the ones that got that highest cumulative G-force. These data support those.

Dr. Shuvro Roy: That is really helpful context. And actually to your point of players over the course of their careers, when we look at this cohort, obviously a cohort of professional athletes that develop CTE, the burning question for those listening is how should we counsel younger players, youth, high school, college football players from these data? Is there anything new to counsel them on based on these results?

Dr. Daniel Daneshvar: This was a study of NFL players, and so the data only explicitly applied to NFL players, but their cumulative head impact exposure didn’t start at the pro level. They had exposure in college, in high school, and many at the youth level. And so all of those head impacts matter. The accumulation of all of those, that cumulative force is what increases one’s risk of problems later in life. And so if the background prevalence of CTE is somewhere south of 1%, like 0.6, 0.8%, and the prevalence in NFL players was at an absolute minimum under the most conservative assumptions, 25%, that’s disconcerting then for those lower levels of play. And so that’s where we need to think about ways that we can be more intentional about that cumulative force to the head that athletes are experiencing. And what I mean by that is doing things that reduce that cumulative force.

Dr. Shuvro Roy: I just am curious, do you feel like we have made progress in that way? When I interviewed Dr. McKee three years ago, for example, we didn’t necessarily have these new Guardian caps that are placed over the helmet for practice, and now we even see them in games. Do you feel like we’ve made progress in that way?

Dr. Daniel Daneshvar: People are very excited about equipment and technological changes and their ability to influence cumulative force to the head, but I’m not convinced. I haven’t seen data that’s particularly convincing. What I mean by that is the Guardian caps that are used at the professional level, first off, those are ones that are not available below the professional level. So those are different than the ones that are being sold at the college and high school level. But those Guardian caps in lab testing have been shown to attenuate the force associated with the impacts by something like 7%. And so you’re taking on average a 27G impact and you’re making that 27G impact on average about a 25G impact. Pretty substantial head impact still. Now, the Guardian caps that are available at below the professional level, in lab testing, those have been shown to attenuate the forces associated by linear impact by about 3%. So you’re taking on average a 27G impact and you’re making it a 26 and change G impact. So I’m not particularly convinced that’s going to be where the bulk of the difference is. Now, on the other hand, we know that nearly 70% of head impacts occur in practice. We also know that at the professional level, there’s fewer than one contact practice per week. So if the college and high school and youth players practiced like the pros practice, we would eliminate more than half of the head impacts. We’d be making those 27G impacts, zero G impacts for up to 70% of them. That seems like it’s a lot more bang for our buck, so to speak. And when you’re talking about a few percent here or the majority there, a change that wouldn’t affect a single minute of game time can get us a lot more power.

Dr. Shuvro Roy: Those are eye-opening numbers. I was not aware, for example, that when we look at the guardian cap data, that it’s only a 7% reduction. And to your point, the vast majority of contact being in practice does seem like a lower hanging fruit. Finally, I would like to close with this. This study has placed a spotlight on the prevalence of CTE being far more common amongst NFL players than previously thought. I think this is the first time we’ve done a neurology podcast where we actually saw a study published in an ESPN article before we’ve even seen it in the journal itself. Where do you hope we can go from here? What do we still need to learn about CTE?

Dr. Daniel Daneshvar: Right now, CTE can only be diagnosed with known accuracy after someone passes away. And so we need to move from understanding CTE, which is critically important, to diagnosing and treating CTE, but we’re making incredible strides on that front. We are identifying factors that help us diagnose the disease both based on clinical markers and objective markers, and we’re also improving the ability to identify potential treatments for this disease. We should not get to the point where we can say an athlete has CTE in life and then not be able to provide that athlete with a treatment that meaningfully affects their risk of progression of disease. And so that’s where we really need to be focused.

Dr. Shuvro Roy: Dan, you’ve been generous with your time. Thank you so much. Again, the article is titled Prevalence of Chronic Traumatic Encephalopathy and Death in National Football League Players: A Retrospective Population-Based Cohort Study published in BMJ. That has been today’s Neurology Podcast. Thanks for tuning in.

Dr. Daniel Daneshvar: Real pleasure to be here. Thank you.

Dr. Stacey Clardy: This is Stacey Clardy, your podcast editor. If you’ve enjoyed the podcast, please take a few moments to subscribe, rate, and review the Neurology Podcast through Apple Podcasts, Google Podcasts, Spotify, or wherever you listen. And remember, you can always head to neurology.org/podcast for our full list of past episodes, or you can also search by keyword on your podcast app for any neurology-specific topics you want to learn about.