Eating Ultra-Processed Foods Raises Dementia Risk by Up to 58% in Large Study
A 9-year longitudinal study links high ultraprocessed food intake to sharply higher dementia and cognitive impairment risk in older adults.
Summary
A large longitudinal study following over 5,000 older U.S. adults found that those eating the most ultraprocessed foods — including sugary drinks, packaged snacks, and processed meats — were 58% more likely to develop dementia and 46% more likely to develop mild cognitive impairment over nearly nine years. Published in the American Journal of Public Health and led by Harvard T.H. Chan School of Public Health researchers, the study also found a protective effect: people eating the most minimally processed whole foods had a 41% lower dementia risk. While the study is observational and diet was self-reported, findings align with a growing body of evidence linking ultraprocessed food to chronic disease and now brain health.
Detailed Summary
A major longitudinal study has added brain health to the growing list of risks associated with ultraprocessed food consumption, finding that older adults who ate the most heavily processed foods faced dramatically higher odds of dementia and cognitive decline over nearly a decade of follow-up.
Researchers from Harvard T.H. Chan School of Public Health tracked 5,370 older U.S. adults for nearly nine years. Those in the highest quintile of ultraprocessed food consumption were 58% more likely to develop dementia, 46% more likely to develop mild cognitive impairment, and 47% more likely to experience either outcome compared to those who ate the least ultraprocessed food. These foods included sugar-sweetened beverages, packaged snacks, and processed meats — staples of the modern American diet.
Equally striking was the protective signal from whole foods. Participants eating the most minimally processed foods — fruits, vegetables, whole grains, and unprocessed meats — had a 41% lower risk of dementia, a 24% lower risk of mild cognitive impairment, and a 26% lower risk of either outcome. This bidirectional finding strengthens the overall case that diet quality is a meaningful lever for cognitive aging.
The study was published as part of a special issue of the American Journal of Public Health featuring more than a dozen ultraprocessed food papers. Researchers also cited biologically plausible mechanisms, noting that ultraprocessed foods are already associated with cardiovascular disease, diabetes, obesity, and several cancers — all conditions with known links to brain health and neurodegeneration.
Important caveats apply. The study is observational, and diet was self-reported, meaning causation cannot be established. The linear trend for dementia specifically did not reach statistical significance, though the trend for mild cognitive impairment and composite outcomes did. Independent replication and mechanistic research are still needed, but the consistency with prior research makes these findings hard to dismiss.
Key Findings
- Highest ultraprocessed food consumers were 58% more likely to develop dementia over ~9 years
- High ultraprocessed intake raised mild cognitive impairment risk by 46% in older adults
- Eating the most whole, minimally processed foods lowered dementia risk by 41%
- Findings published in American Journal of Public Health alongside 12+ related ultraprocessed food studies
- Observational design limits causation claims, but results align with broader chronic disease evidence
Methodology
This is a news report summarizing a peer-reviewed longitudinal cohort study published in the American Journal of Public Health, led by Harvard T.H. Chan School of Public Health researchers. The evidence basis is observational with a large sample (~5,370 older U.S. adults) and nearly 9 years of follow-up; diet was self-reported, which is a standard but imperfect measurement method.
Study Limitations
The study is observational and cannot establish causation between ultraprocessed food and dementia. Self-reported dietary data introduces recall and reporting bias. The linear trend for dementia risk specifically did not reach statistical significance, warranting cautious interpretation pending further replication.
Enjoyed this summary?
Get the latest longevity research delivered to your inbox every week.