DHA changed a biological marker, but a two-year trial did not demonstrate cognitive benefit

High-dose DHA changed the biological marker under study without outperforming placebo on cognitive tests.

Changing a biomarker in cerebrospinal fluid is not, by itself, evidence of protecting memory. DHA supplementation increased the DHA-to-arachidonic-acid ratio in this fluid, but did not outperform placebo on cognitive tests over two years.

Story details

Researchers enrolled 365 adults aged 55–80 who did not have dementia and consumed little dietary DHA. Participants were randomly assigned to 2 g of DHA daily or placebo. The study first asked whether supplementation changed the DHA-to-arachidonic-acid ratio in cerebrospinal fluid after six months. It did, indicating biological target engagement. APOE ε4 carriers made up 47% of participants; no statistically significant difference by carrier status was detected in this ratio’s treatment response. That success did not translate into better cognitive performance over two years. The distinction matters when evaluating supplement claims: reaching a biological target is a step in testing a treatment, not proof of a meaningful health benefit. This trial also did not compare eating fish with taking supplements, so it cannot establish which dietary approach is better. Its findings should not be used as instructions to stop treatment prescribed for another condition. The original paper appeared in June; October coverage brings an earlier study back into the news.

What was found?

Brain-imaging measures also showed no treatment advantage over two years.

What does it mean?

These findings do not establish DHA supplements as an Alzheimer’s prevention treatment.

Limitations

Imaging measures were secondary outcomes and cognitive measures were exploratory; the primary endpoint concerned a biomarker change. Dropout was 38%, largely attributed by the researchers to the COVID-19 pandemic. The results cannot cover every age, formulation or duration.

Sources and review level

Original research · DOI: 10.1016/j.ebiom.2026.106316. First online publication: 2026-06-18; news coverage used for this package: 2026-10-22.

Original abstract and indexed original methods passages examined; complete uninterrupted full-text access was unavailable.

Keep this story

Feedback to the editor

Only the editor can read this; it is not a public comment. Do not include personal health information.

This story provides general information, not individual diagnosis or treatment advice. Findings should be considered within the conditions and limitations of the study.