The Sao Paulo Autopsy Service is the only morgue serving the 12 million people of the metropolitan area, and in Brazil the law requires a general autopsy for anyone who dies without a death certificate. That administrative fact has quietly produced something rare in dementia science: a brain collection that nobody volunteered for out of worry. Nobody's family called because of a boxing career or a decade of headers. The University of Sao Paulo Biobank for Aging Studies draws from that service, and its shelves therefore look less like a hypothesis and more like a city.

A team led by Roberta Diehl Rodriguez, with Lea Grinberg as senior author, went back through 1,151 of those brains, collected between 2004 and 2022, and reviewed 14,963 histological slides stained for phosphorylated tau, looking for chronic traumatic encephalopathy neuropathological change (CTE-NC): the perivascular tau at the depth of a cortical sulcus that consensus criteria treat as the signature lesion. Seven people had it. That is 0.6 percent. Their mean age at death was 73.4 years, and only one was a woman. Two were staged as CTE stage I, three as stage II and two as stage III; under the severity criteria from the Second NINDS/NIBIB consensus meeting, six of the seven were rated high severity. The work appeared in the Journal of Neuropathology and Experimental Neurology, online on 2 February 2026 and in the May 2026 issue, and the authors describe it as the largest single cohort assessed for CTE-NC to date.

Put that next to the numbers most families have actually seen. In 2017, in JAMA, Jesse Mez, Ann McKee and colleagues reported CTE in 177 of 202 donated brains of former American football players, including 110 of 111 former NFL players. In 2022, in the New England Journal of Medicine, David Priemer and colleagues examined 225 consecutive brains from a brain bank dedicated to deceased service members and found CTE in 10 of them, 4.4 percent. In December 2025, again in JNEN, the Late Effects of TBI project reported CTE-NC in 7 of 47 consecutive donor brains from a community cohort assembled around head trauma. None of these are in conflict. They are answers to different questions, and only the Sao Paulo series is asking what a general population looks like.

What the seven cases show is how seldom the pathology travelled alone. Alzheimer disease-type changes, aging-related tau astrogliopathy and argyrophilic grain disease were the most common co-pathologies; all seven had subpial and perivascular ARTAG, and TDP-43 pathology turned up in the stage III cases. Dementia was present in 43 percent of the CTE-NC cases and 57 percent had neuropsychiatric symptoms, figures drawn from seven people and therefore closer to case description than to epidemiology. For a family, the practical translation is unglamorous but important: a mixed brain is the normal brain of a person in their seventies, and when CTE-NC does appear in a community sample, it appears amid other things that also damage memory and mood.

The honest weakness sits in the paper's own abstract. Clinical, cognitive and neuropsychiatric information came from interviews with informants, and information on sports exposure and traumatic brain injury history was limited; comprehensive exposure histories were unavailable for most cases, with only two having documented soccer involvement. So this study cannot tell you what repetitive head impacts do, because it never had a reliable count of who took them. It tells you the base rate in an unselected city, which earlier series had estimated at 0 percent to under 1 percent in roughly 1,800 people, mostly from dementia clinics or cohorts skewed toward highly educated, non-Hispanic White donors. Sao Paulo's contribution is that the same low number holds in an admixed, socioeconomically mixed population where both dementia and head injury are common. The authors' own reading is that the finding supports primary prevention: reducing exposure to repetitive head impacts in sport and other settings.