Key takeaways:

• A diet with lower inflammatory potential was associated with lower dementia risks in adults with elevated biomarkers.
• For Mediterranean and generally healthy diets, associations were less significant.

Healthier dietary patterns with lower inflammatory potential were associated with lower dementia risk for people with elevated Alzheimer’s disease and neurobiological risk biomarkers, according to a study published in JAMA Network Open.

The study explored how diet quality relates to dementia risk, specifically examining phosphorylated tau at threonine 217 (p-tau217), which the authors described as one of the most accurate blood-based biomarkers for the early detection of Alzheimer’s disease.

Data derived from Mrhar A. et al. JAMA Network Open. 2026:9(6):e2620254:doi:10.1001/jamanetworkopen.2026.20254

Taking those biomarkers into account allowed researchers to examine whether the association between diet quality and dementia risk differed according to underlying biological risk profiles, said Anja Mrhar, MSc, a researcher with the Aging Research Center at Karolinska Institutet, told Healio.

“The key new aspect is that diet quality appeared to remain relevant even among older adults with elevated biomarker levels,” Mrhar said. “In particular, the most consistent associations among people at elevated biological risk were observed for a dietary pattern with lower inflammatory potential. This suggests that the inflammation-related dimension of diet may be especially relevant to study in future dementia prevention research in this high-risk group.”

Although previous research has found healthier dietary patterns are associated with better brain health and lower dementia risk, there has been little focus specifically on people who already show biological signs of dementia risk, Mrhar said.

The methods

The study analyzed 1,865 adults (mean age at baseline, 70.5 years; 60.3% female) who did not have dementia from the population-based Swedish National Study on Aging and Care in Kungsholmen, which started enrolling participants in March 2001 and examined them up to six times through November 2019.

Mrhar and colleagues tracked baseline serum p-tau217, neurofilament light chain (NFL) and glial fibrillary acidic protein (GFAP) concentrations to see how they changed in relation to three dietary patterns: Alternate Mediterranean Diet (AMED), Alternative Healthy Eating Index (AHEI) and reversed Empirical Dietary Inflammatory Index (rEDII).

“These three patterns were useful because they allowed us to compare different aspects of diet quality: Mediterranean-style eating, general healthy eating and inflammatory potential,” Mrhar said. “Importantly, the patterns overlap in some foods, but they are not identical and were developed for different purposes. This may help explain why their associations with dementia risk differed across biomarker groups.”

The big picture

Mrhar and colleagues analyzed data between September 2024 and August 2025 and reanalyzed it in March and April 2026.

Over the 15-year follow-up period, with a mean follow-up of 8.4 years, 240 of the participants developed dementia (incidence rate of 15.4 per 1,000 person years; 95% CI, 13.5-17.4).

The authors found that each 1-z score increase in adherence to the rEDII diet was associated with lower dementia risk among participants with elevated p-tau217, NFL and GFAP levels. When analyzing cumulative diet quality and 15-year dementia risk, hazard ratios for participants following the rEDII diet were 0.71 for the higher p-tau217 group (95% CI, 0.58-0.88), 0.79 for the higher NFL group (95% CI, 0.66-0.9) and 0.73 for the higher GFAP group (95% CI, 0.60-0.89).

For participants following AMED and AHEI diets, associations between lower dementia risk and high biomarker levels generally were not significant, but associations with lower dementia risk were seen among participants with lower biomarker levels, according to the study.

Participants following the AHEI diet with lower p-tau217 levels had a hazard ratio of 0.66 (95% CI, 0.58-0.88), compared with a hazard ratio of 0.8 (95% CI, 0.69-1) among participants with higher p-tau217 levels (P = .04 for the interaction).

Participants following the AHEI diet with lower NFL levels had a hazard ratio of 0.57 (95% CI, 0.44-0.73), compared with a hazard ratio of 0.89 (95% CI, 0.72-1.04) among participants with higher NFL levels (P = .008 for the interaction).

Mrhar emphasized that the study was observational and does not prove that diet prevents dementia or that changing diets would reduce risk. Intervention studies are needed to test that directly, she said.

Mrhar and colleagues said the next steps should involve studying more diverse populations to see if the findings apply more broadly. They should also identify which specific food combinations and nutrients may be driving the associations, which she said could help future research move from broad dietary associations toward more targeted prevention strategies.

“It may also help clarify whether the observed associations are mainly related to inflammation, cardiovascular and metabolic health, or other mechanisms involved in neurodegeneration,” Mrhar told Healio.

For more information:

Anja Mrhar, MD, can be reached at anja.mrhar@ki.se.