A study published Tuesday, Aug. 25, in The BMJ found that at least one in four former NFL players who died between 2016 and 2021 had chronic traumatic encephalopathy, a degenerative brain disease linked to repeated head impacts. The work was led by Daniel H. Daneshvar of Mass General Brigham and Harvard Medical School, with co-authors from Boston University, the Concussion & CTE Foundation, Mount Sinai and the University of California, San Francisco.

Brains came from three banks: Boston University’s UNITE Brain Bank supplied 333 of the 338 donors, Mount Sinai one, and UCSF’s Alzheimer’s Disease Research Center four.

The work was funded by the National Institute of Neurological Disorders and Stroke, the National Institute on Aging, and the Maloney/Carpenter Trauma-Related Neurodegenerative Disease Research Fund. The authors state the funders had no role in the study’s design, conduct, analysis or the decision to publish. The paper was externally peer reviewed; The BMJ published a linked editorial alongside it, which was not.

Several authors disclose ties to the parties the findings bear on. Daneshvar reports research funding from the Football Players Health Study at Harvard, which is funded by the NFL Players Association, clinical funding from an NFLPA-funded program, and volunteer service on the union’s Mackey-White Committee. Co-author Christopher Nowinski reports travel support from the NFL, the NFLPA and World Rugby, and work as an expert witness in concussion and CTE cases. Co-author Robert Cantu reports serving as a paid consultant to the NFL’s Head, Neck and Spine Committee and as vice president of the National Operating Committee on Standards for Athletic Equipment, the body that certifies football helmets. Ann McKee also serves on the Mackey-White Committee.

The NFL Players Association, which funds CTE research at Boston University, called the results a sobering reminder that football carries real risks and said the findings “should serve as a call to action across the entire football ecosystem.”

An NFL spokesperson said the league continues to work on making the game safer through strategies aimed at reducing concussions and head impacts, and pointed former players toward league health resources.

That call to action raises a question the sport has circled for a decade without resolving: if the helmet is what allows a player to lead with his head, would taking equipment away  or reducing it  produce better tackling and fewer head impacts? And does rugby, played without helmets or shoulder pads, offer a natural experiment?

The evidence answers the first question partially and the second question in a way that complicates the premise.

Background: What the Study Measured, and What It Could Not

Researchers identified every former NFL player who debuted after 1949  the year plastic helmets became mandatory in the league  and died between 2008 and 2021. That is 1,712 men. Of those, 338 brains, or 19.7%, were examined, and 315 of them, 93.2%, showed CTE. Across that full window the possible prevalence ranges from 18.5% to 98.7%.

The headline figure comes from a narrower window. Between 2016 and 2021, when donation was most frequent, 878 former players died. Of those, 235 donated their brains  26.8%  and 215, or 91.4%, were found to have CTE.

In simple terms: the researchers did not test 878 brains. They tested 235 and then asked what the 878 figure could support. The most conservative assumption  that the confirmed donors were the only decedents with CTE  produces a prevalence of 24.5%. The opposite extreme produces 97.7%. Neither bound is an estimate of the truth; they are the outer walls of what the data can rule out.

This is why the number is best read as a floor. Brain donation is not random, and the paper documents exactly how: compared with non-donors, donors had longer NFL careers, more Pro Bowl selections and more Hall of Fame inductions. The authors suggest longer careers may have produced more symptoms, prompting families to donate, and that public visibility may also have played a role.

Three details push the floor upward. At least 11 CTE cases in NFL players who died during the study period were reported by other brain banks and excluded here to keep the methodology consistent. At least 12 more were suspected of having CTE but never examined after death. And 215 non-donors had a neurodegenerative cause of death listed on their death certificate. The authors write that these omissions would raise the lower bound.

For context, community brain bank studies find CTE in a fraction of a percent. The Framingham Heart Study found one case among 164 donors; a Kaiser Permanente study found three among 532; an Austrian cohort found none among 310. Across community cohorts without documented repetitive head-impact exposure, prevalence ran from 0% to 1.3%.

Dementia was common among donors  202 of 338, or 59.8%. Among donors with CTE, 193 of 315 had it. Among those with the most severe form, stage IV, 94 of 104, or 90.4%, had dementia. After weighting to account for donation bias and adjusting for age, education, age of first exposure to tackle football and other factors, stage IV CTE carried a 44% higher risk of dementia than other donors (risk ratio 1.44, 95% CI 1.16 to 1.78).

The study’s most quietly consequential finding may be about paperwork. Of the 202 donors clinicians judged to have had dementia, death certificates recorded a dementia or neurodegenerative primary cause in only 62  30.7%. Counting secondary causes brings it to 82, or 40.6%. Nearly seven in ten dementia cases never appeared on the death record at all, which means prior research built on death certificates has understated the problem.

Two caveats deserve equal billing with the headline. Within this cohort, longer play was not associated with whether a donor had CTE at all, only with how severe it was. The authors suggest that above a certain exposure threshold, other factors  genetic or environmental  may govern who develops the disease. And they state plainly that because every man in the study played professionally, the findings cannot necessarily be generalized to the broader football community.

The Case for Less Equipment

The argument that helmets enable bad tackling is not fringe. It has a name in the injury-prevention literature  risk compensation  and a documented history in this sport.

When hard-shell helmets became standard, spear tackling emerged: players driving the crown of the helmet into an opponent, using the head as the point of contact because the head suddenly felt indestructible. Catastrophic head and neck injuries spiked. The NCAA and the National Federation of State High School Associations banned the technique in 1976, and head and spine injuries fell by more than half.

The National Athletic Trainers’ Association addressed this directly in its 2021 position statement on head-first contact, attributing the persistence of spearing behavior to the false sense of security created by hard outer shells, and recommending helmetless training as one counter.

That recommendation rests on actual trials. Erik Swartz, then at the University of New Hampshire, ran a randomized controlled trial in which 50 Division I players were split into two groups. The intervention group performed five-minute tackling drills without helmets or shoulder pads twice weekly in preseason and once weekly in season. The control group did non-contact skill work for the same duration. By season’s end, the helmetless-training group recorded roughly 30% fewer head impacts per exposure.

The program, called Helmetless Tackling Training, was later tested in Hawaiian high school football. Players who adhered to the 12-week program experienced 34% to 37% fewer head impacts across a season. That study won the Journal of Athletic Training’s 2025 outstanding research award.

The mechanism is not complicated. Remove the helmet and the athlete’s own reflexes take the head out of the collision. The technique is explicitly modeled on rugby.

There is a critical distinction here that the popular version of this argument tends to lose. Every one of these studies removed equipment in drills, not in games. No peer-reviewed trial has tested playing tackle football without helmets, and none is likely to. Helmets exist because they prevent skull fractures, subdural hematomas and deaths  acute injuries that killed players in the leather era. Stripping them from live play would trade a diffuse, long-horizon risk for an immediate catastrophic one.

Shoulder pads sit in a similar position. They are removed in HuTT drills alongside helmets, but there is no body of research on removing them from competition.

The Case Against: More Padding Also Shows Results

Football’s most-watched recent equipment experiment ran in the opposite direction  adding padding, not subtracting it.

The NFL mandated Guardian Cap soft-shell helmet covers for linemen, tight ends and linebackers in 2022 preseason practices. An analysis published in The American Journal of Sports Medicine in 2025, using NFL Injury Surveillance System data from 2018 through 2023, found concussion rates in the affected position groups fell 54% in univariate analysis and 62% in a regression model.

The same paper found no significant association for concussions involving the helmet shell itself, which suggests the mechanism is less straightforward than the headline number implies.

Independent results have been less favorable. A prospective cohort study of 2,610 Wisconsin high school players across 41 teams, published in the British Journal of Sports Medicine in January 2025, found no reduction in concussion risk from wearing the caps in practice  33 concussions among 1,422 cap wearers versus 31 among 1,188 who went without, and no reduction in game concussions either. The study also measured severity, and found none: no difference in initial SCAT5 symptom scores and no difference in days to return to sport, in practice or in games. The authors cautioned that the devices may offer false reassurance, that players may compensate by playing more aggressively, and that money spent on unproven equipment can crowd out measures with real evidence behind them, such as athletic trainer staffing and rule changes. They also noted their results cannot be generalized upward: high school players use a thinner, lighter cap than the NFL and college model. A separate 2025 study in Medicine & Science in Sports & Exercise instrumented 54 college players across more than 7,500 head impacts and found the cap had no effect on impact magnitude or frequency.

In simple terms: the honest reading of this literature is that equipment  more of it or less of it  is not where the large effects live. Behavior is.

Does Rugby Prove the Point?

This is where the comparison becomes uncomfortable for the argument that helmets are the root cause. Rugby is played without helmets or shoulder pads. Rugby also has CTE.

A 2023 study in Acta Neuropathologica, drawing on brain banks at the University of Glasgow, Boston University and the University of Sydney, examined 31 former rugby union players. CTE was found in 21 of them  68%  among both amateur and elite players. Risk tracked with career length, with each additional year of play associated with a 14% increase in CTE risk.

Lead author Willie Stewart’s group had previously reported that former Scottish international rugby union players faced neurodegenerative disease risk roughly two and a half times higher than expected.

Rugby union also carries high concussion incidence, with reported rates ranging from 4.1 concussions per 1,000 player hours in the community game to 22.2 per 1,000 hours in the professional game.

The legal picture mirrors the NFL’s. On Aug. 21, 2026  four days before the BMJ study appeared  London’s High Court ruled that more than 500 former rugby union and rugby league players could take brain-injury claims against the sport’s governing bodies to trial in Hughes & Ors v World Rugby Ltd & Ors, part of a group action involving roughly 1,000 claimants. Named claimants include England’s 2003 World Cup-winning hooker Steve Thompson, who has early-onset dementia, and former Wales captain Ryan Jones. World Rugby has denied liability, arguing that brain injury is a “foreseeable and inherent risk” of the sport.

So the absence of a helmet did not prevent CTE. What rugby does offer American football is not a lesson about equipment. It is a lesson about rules and technique  and that evidence is strong.

What Rugby Actually Demonstrates

Ireland lowered the legal tackle target from armpit level to the base of the sternum in 2023. A seven-season prospective study published in the Journal of Science and Medicine in Sport tracked 239 teams across 8,805 player-seasons. Concussion rates fell 38% in amateur men’s rugby and 63% in women’s rugby. Overall injuries dropped 30% and 18% respectively.

The schoolboy results were weaker. Overall injuries fell 19%, but concussion reductions were not statistically significant  a gap the authors attribute to differences in technique, physical development and stability among younger players.

In Scotland, University of Edinburgh researchers analyzed 18,702 tackles across 60 men’s community matches before and after the 2023/24 law change. Players were 22% more likely to bend at the waist, and head-on-head contact fell 45%.

World Rugby’s Executive Board reviewed roughly 150,000 community tackles and found 8% to 10% fewer high-risk upright tackles. The lower tackle height became full law of the community game on July 1, 2026.

Football has already imported the technique, if not the rule. Pete Carroll’s Seattle staff developed shoulder-leverage “Hawk tackling,” explicitly rugby-inspired and built around taking the head out of the play, and released it publicly in 2014. The NFL distributed it. Ohio State, Nebraska and others adopted it. In 2017, USA Football partnered with the NFL, USA Rugby and the Seahawks on a shoulder-tackling curriculum, later relaunched as the Shoulder Tackling System.

What football has not done is what rugby did: change the rules of the game so that the safer technique becomes the only legal option, then measure the result.

Analysis

The argument that reducing equipment would teach better tackling is defensible in its narrow form and indefensible in its broad one.

The narrow form  that helmetless and pad-less drills produce measurably safer tackling behavior  is supported by randomized controlled trials at the college level and adherence-based results at the high school level, with effect sizes in the 28% to 37% range for head-impact reduction. That is a real finding with a real mechanism.

The broad form  that removing helmets and shoulder pads from live play would solve CTE  collides with rugby’s own numbers. Two-thirds of a rugby brain series showed CTE. Career length drove risk. Helmetless play did not protect those athletes.

The synthesis is more tentative than either camp would like, and this study is part of why. Within these 338 NFL donors, longer play predicted how severe the disease was but not whether a man had it. The authors read that as a possible threshold effect: past some level of exposure, other factors may determine who develops CTE at all. If that holds, the lever is not shaving impacts off the career of a man already headed to the NFL. It is keeping total exposure under the threshold in the first place  which points at youth, high school and college play, where the impacts accumulate for a decade before anyone signs a professional contract.

That reasoning is an inference, not a finding. The authors state plainly that because every man studied played professionally, the results cannot necessarily be generalized to the wider football community, and the dose-response evidence from cohorts with a broader range of exposure comes from other papers, not this one. But if it holds, the interventions with the most leverage are the ones that reduce the number of impacts a fourteen-year-old absorbs: contact limits in practice, legal-target rules that make head-level contact structurally difficult, enforcement of existing head-first contact penalties, and technique instruction that starts before a player ever puts a helmet on.

That is a coaching, officiating and youth-policy agenda. It is considerably cheaper than a helmet redesign and considerably harder to implement, because it requires changing what adults teach children rather than what manufacturers sell them.

Conclusion

The one-in-four figure will be argued over, and it should be. It is a lower bound built on a non-random sample, and the researchers were explicit about that. But a lower bound of 24.5% against community cohorts running between 0% and 1.3% is not a number that improves with better methodology  and the paper’s own accounting of excluded and unexamined cases points the floor upward, not down.

The NFLPA’s framing  a call to action across the entire football ecosystem  is the right scope, if it is taken literally. The ecosystem that produced these brains was not primarily the NFL. It was Pop Warner, Friday nights and Saturday afternoons, accumulating impacts for a decade before anyone signed a professional contract.

Rugby is not proof that helmets cause CTE. Rugby is proof that a governing body can change a rule, measure the result across 150,000 tackles, and publish what it finds. That is the part worth copying.

Key Takeaways

• At least 24.5% of former NFL players who died between 2016 and 2021 had CTE, per a study published Aug. 25, 2026 in The BMJ. The figure is a conservative floor, not a point estimate; the study’s upper bound is 97.7%.
• Of 235 brains donated in that window, 215 (91.4%) had CTE  but brain donation is not random, which is why the 24.5% floor is the defensible number.
• Randomized trials of helmetless tackling drills show 28% to 37% reductions in head impacts. No trial supports removing equipment from live play.
• Guardian Cap evidence is mixed: a 54% to 62% reduction in NFL preseason practice concussions, but no effect on rate, symptom severity or recovery time in a 2,610-player Wisconsin high school study.
• Rugby, played without helmets, showed CTE in 21 of 31 donated brains (68%), with risk rising 14% per year played  undercutting the claim that helmets are the root cause.
• Rugby’s transferable lesson is rules, not gear: lowering the legal tackle height cut concussions 38% (men) and 63% (women) in Irish amateur rugby and reduced head-on-head contact 45% in Scotland. It became World Rugby law on July 1, 2026.
• Within this cohort, longer play predicted CTE severity but not whether a player had CTE at all  the authors suggest a possible exposure threshold. They also caution the findings cannot necessarily be generalized to non-professional football.
• Death certificates recorded a dementia or neurodegenerative cause for only 40.6% of donors clinicians judged to have had dementia, meaning earlier research built on death records understated the burden.

Sources

Daneshvar DH, Nowinski CJ, Abdolmohammadi B, et al. “Prevalence of chronic traumatic encephalopathy at death in National Football League players: retrospective population based cohort study, 2008-21.” BMJ 2026;394:e100418. doi:10.1136/bmj-2026-100418

Kurth T, et al. “Chronic traumatic encephalopathy in former NFL players.” BMJ 2026 (linked editorial). doi:10.1136/bmj-2026-100526

Hammer E, Mosiman S, Joachim MR, et al. “The association between Guardian Cap use during practices and sport-related concussion risk in high school American football players.” Br J Sports Med 2025;59:257–262. doi:10.1136/bjsports-2024-108945

Funk JR, et al. “An Analysis of Guardian Cap Use and Changes in the Concussion Rate in NFL Preseason Practices From 2018 to 2023.” Am J Sports Med, 2025. PubMed

Sinnott A, et al. “Guardian Cap and head impact magnitude in college football.” Med Sci Sports Exerc, November 2025.

Swartz EE, et al. “Early Results of a Helmetless-Tackling Intervention to Decrease Head Impacts in Football Players.” J Athl Train, 2015. PMID 26651278

Swartz EE, et al. J Sci Med Sport, 2019. PMID 31204104

National Athletic Trainers’ Association Position Statement: Reducing Intentional Head-First Contact Behavior in American Football Players. PMC8876879

Stewart W, et al. Chronic traumatic encephalopathy in former rugby union players. Acta Neuropathol, 2023. doi:10.1007/s00401-023-02644-3

Comyns T, Kenny I, et al. Seven-season tackle-height study. J Sci Med Sport, 2026.

Scottish Rugby / University of Edinburgh  18,702-tackle video analysis, November 2025.

World Rugby Executive Board advisory, December 2025  community tackle height law.

Hughes & Ors v World Rugby Ltd & Ors  High Court ruling, Aug. 21, 2026.

USA Football  Shoulder Tackling System; ESPN  NFL distribution of the Carroll tackling video.