Chronic traumatic encephalopathy was present in an estimated 74.5% of former National Football League players who died between January 2008 and December 2023, according to a study published on October 28, 2026 in JAMA Network Open. The figure matters less for its size — researchers have been reporting high CTE rates in football players for two decades — than for surviving the single strongest objection raised against all of that earlier work.
That objection is selection bias. CTE can only be confirmed by examining brain tissue after death, so the disease has been measured almost entirely in brains that families chose to donate. A family that watched a former player lose his memory, his temper or his independence is far more likely to make that donation than a family that did not. Every raw percentage drawn from a brain bank is therefore inflated by an unknown amount. The new paper's contribution is an attempt to estimate that amount and subtract it.
2,362 deaths, 396 of them donors
The Boston University-led team — Julia Gallini is the lead author and Yorghos Tripodis the corresponding author, with co-authors including Ann C. McKee, Jesse Mez, Michael L. Alosco, Robert A. Stern and Christopher Nowinski — assembled a population-based cross-sectional sample rather than a donor sample. Using National Death Index records, the researchers identified all 2,362 former NFL players recorded as having died in that 16-year window. Of those, 396 had donated their brains to the university's UNITE Brain Bank; 1,966 had not.
Among the 387 donors whose CTE status was known, 362 had the disease. That is 93.5% — the kind of number that has produced a decade of headlines, and the number critics have rightly described as unrepresentative of anything beyond the brain bank itself.
Pricing in the distortion
To get from 93.5% to a population estimate, the researchers had to quantify how much more likely a symptomatic player was to reach the brain bank at all. Their analysis put that selection ratio at roughly five: a former player with cognitive symptoms was about five times likelier to become a donor than one without. Applying that correction yields the headline estimate of 74.5%, with a 95% confidence interval running from 69.6% to 78.3%.
The paper also reports a deliberately pessimistic version of the same calculation — a conservative lower bound of 63.7% (95% CI, 57.2% to 68.9%). And it runs the arithmetic again on an assumption the authors consider implausibly strong: even if symptomatic players were ten times likelier to donate, rather than five, the estimated prevalence would still land above 60%.
A third figure gives the clinical picture some shape. The same bias-adjusted method put the share of these men who had a history of all-cause dementia during life at 27.2% (95% CI, 22.6% to 31.6%) — well below the share with CTE pathology in their brain tissue, a reminder that the two are not the same measurement.
For scale, Gallini noted that CTE turns up in fewer than 1% of donors to community brain banks, which draw from the general population rather than from former contact-sport athletes.
What changed since August
An earlier analysis released in August had estimated that somewhere between 25% and 98% of former NFL players who died from 2016 to 2021 had CTE. That range was so wide as to be almost unusable, and Boston University said at the time that it reflected genuine uncertainty about how representative brain donors were. The new study's narrower interval is the product of modeling that representativeness directly instead of leaving it open.
The limits the authors set themselves
The paper is explicit about what it does not show. It describes men who have died, not men who are alive: it makes no prediction about how many current or retired players carry the disease now. It does not extend to college, high school or youth football, where exposure is lower. The proxy measures used to flag cognitive symptoms in the nondonor group are imperfect, the authors write, and the adjustment is only as good as the estimated selection ratio behind it. The conclusion they do draw is narrower than the headline: selection bias inflates donor-based estimates, but it is unlikely to explain away the burden of CTE pathology found in deceased NFL players.
Reaction
The NFL said it "continuously strives to make the game of football safer," pointing to concussion-reduction protocols, equipment such as Guardian Caps and rule changes, and encouraged former players to use the health resources available to them. The NFL Players Association called the study one of many that are critical to understanding football's long-term health effects.
Former quarterback Matt Hasselbeck, an ambassador for the research, said the sport has to "learn from this data to make the game safer," according to CBS News. Patriots head coach Mike Vrabel, who played linebacker for 14 seasons, told reporters that players of his era "knew what the risks were" and that he "knew what I was signing up for," while agreeing the league should keep working on safety.
Gallini framed the practical consequence in terms of what the number implies for the living: more support for families, better tools to diagnose CTE before death, and treatment for the symptoms that come with it.
Reporting from JAMA Network Open, Boston University and CBS News.
