Imagine you're trying to figure out how many houses on a street have termites, but you can only check the ones whose owners called the exterminator. You'd massively overcount — worried homeowners self-select into inspection. That's the core problem in CTE research: brains donated to science come disproportionately from families who noticed something was wrong. This study's contribution is the statistical exterminator-bias correction. The committed claim: after adjusting for donation bias, approximately 75% of NFL players who died between 2008 and 2023 had signs of CTE. The previous estimate from the same Boston University group, published in August, put the figure at 25% over a shorter five-year window. The raw donated-brain figure — 93% CTE prevalence among 396 brains — has long been understood as inflated by selection bias. The new number sits between the raw rate and the general population's ~1% baseline, but much closer to the former. The method combines two datasets: 396 autopsied brains from the UNITE Brain Bank plus national death records for 1,966 other deceased former NFL players who were never autopsied. The researchers estimated that families of symptomatic players were about five times more likely to donate brains, then performed a statistical correction to back out what the population-level rate would be. From 2,362 total deaths, they estimate roughly 1,760 had CTE. A separate estimate puts the dementia rate at 27%, acknowledging that CTE presence doesn't automatically mean cognitive symptoms. The integrity question is where this gets interesting. The validation is entirely model-based — no one independently checked those 1,966 non-donated brains. The 5× donation-bias multiplier is itself an estimate derived from comparing cognitive-symptom rates across donated vs. non-donated cohorts. If that multiplier is off by even 30%, the headline number swings substantially. The study acknowledges this: lead author Yorghos Tripodis calls it a 'best guess.' Independent experts like Dr. Jim Kelly note that the team 'had no autopsies on five times as many players who died as they did have autopsies for.' The biggest structural limitation remains the same one that's plagued the field for decades: CTE cannot be diagnosed in living patients. Every prevalence study is working backward from autopsies and forward from statistical models, with no ground-truth check available in between. Until a reliable biomarker or imaging technique can detect tau accumulation patterns in vivo, the entire subfield is stuck estimating through the fog. The milestone to watch isn't a bigger sample — it's a diagnostic breakthrough that makes autopsy-only detection obsolete. The NFL's response is boilerplate about helmet redesigns and rule changes like the Guardian Cap and dynamic kickoff rule. What the league notably does not engage with is the core implication: if 75% is even roughly right, then CTE is not a rare complication of professional football but a near-default outcome. That reframes the policy question from 'how do we protect vulnerable players' to 'is the activity itself fundamentally brain-damaging at the professional level.' The 27% dementia estimate — meaning roughly one in four dead former players — makes the downstream human cost concrete. The study was published in JAMA Network Open and led by Yorghos Tripodis, director of statistics and data management at the Boston University Alzheimer's Disease and CTE Center. It extends work from the UNITE Brain Bank, the same institution whose earlier 99%-of-donated-brains finding became a landmark (and frequently misinterpreted) statistic. The methodological advance here is real: attempting to correct for the bias that made that 99% figure misleading. Whether 75% holds up to independent replication is the next question that matters.
