Antioxidants don't slow Alzheimer's cognitive decline — but they may help with behavior
Li et al. 2026 (Frontiers in Nutrition, 23 RCTs, n=10,537) is the largest RCT meta-analysis on antioxidant supplements in Alzheimer's disease. Core finding: no improvement in cognitive scores (ADAS-Cog SMD 0.01, MMSE SMD 0.19 — both non-significant), no change in amyloid or tau biomarkers, and a paradoxical increase in a urinary oxidative-damage marker. Neuropsychiatric symptoms did improve (NPI SMD −0.85). For prevention, dietary vitamin C ≥75 mg/day from food — not supplements — is the only antioxidant with significant observational support. Article discloses abstract-only status throughout; full text pending.
The largest randomized-trial meta-analysis ever conducted on antioxidant supplements in Alzheimer's disease has delivered a split verdict: no improvement in cognition, no change in the amyloid and tau biomarkers that define the disease — and a paradoxical signal suggesting oxidative damage may actually increase. The behavioral picture is more complicated in one specific direction: neuropsychiatric symptoms improved meaningfully.
Li et al. 2026, accepted June 1 in Frontiers in Nutrition, pooled 23 randomized controlled trials and 10,537 patients with diagnosed Alzheimer's disease. 1 The finding matters because antioxidant supplements — vitamins C and E, polyphenols, and other antioxidant formulations — are among the most commonly taken supplements by older adults and AD caregivers hoping to slow neurodegeneration. This is the clearest evidence yet that supplemental antioxidants do not function as disease-modifying agents in established Alzheimer's disease.
Important disclosure: the paper was accepted on June 1, 2026, and only the abstract is currently available on the Frontiers in Nutrition platform. The formatted full text, including individual forest plots, subgroup analyses, GRADE certainty ratings, risk-of-bias assessment, and funding and conflict-of-interest disclosures, had not yet been published as of this writing. All findings below are drawn from the abstract. Where the full text would normally clarify a result — particularly the paradoxical oxidative-stress finding — this is noted explicitly.
Study at a glance
Study type: Systematic review and meta-analysis of randomized controlled trials · Venue: Frontiers in Nutrition (Q1, IF 5.1) · DOI: 10.3389/fnut.2026.1878597 · Accepted: June 1, 2026 · Peer-review status: accepted (formatted full text pending)
Li, Wang, and 12 co-authors at Xiangtan Central Hospital and the Hengyang Medical School of University of South China searched PubMed, Embase, and Web of Science from inception through October 2, 2025. 1 The 23 included trials enrolled 10,537 adults with diagnosed Alzheimer's disease. Interventions covered a wide range of antioxidant types — vitamin-class antioxidants (vitamins C and E), polyphenols, and other antioxidant preparations — administered at various doses and durations. Where heterogeneity across studies was high (I² > 50%), the analysis used a random-effects model. All effect sizes are reported as standardized mean differences (SMDs), a unit-free scale where values above 0.8 are conventionally considered large.
Loading content card…
What antioxidants did not do
On every cognitive and pathological outcome, antioxidant supplementation produced no statistically significant effect.
| Outcome | SMD | 95% CI | Result |
|---|---|---|---|
| ADAS-Cog (cognitive function) | 0.01 | −0.21 to 0.23 | No significant change |
| MMSE (cognitive function) | 0.19 | −0.17 to 0.56 | No significant change |
| Aβ42 (amyloid biomarker) | — | — | No significant change |
| Phosphorylated tau (p-tau) | — | — | No significant change |
| Total tau (t-tau) | — | — | No significant change |
The Alzheimer's Disease Assessment Scale — Cognitive Subscale (ADAS-Cog; range 0–70, higher = worse) and the Mini-Mental State Examination (MMSE; range 0–30, higher = better) are the two most widely used trial endpoints for AD cognitive function. 1 An ADAS-Cog SMD of 0.01 — a confidence interval spanning nearly zero on both sides — is as close to a null result as meta-analyses produce. The MMSE point estimate is directionally positive but the confidence interval includes zero, so it cannot be distinguished from chance.
The biomarker results are equally flat. Aβ42 (amyloid-beta 42, the aggregation-prone amyloid fragment that accumulates in AD plaques), p-tau (phosphorylated tau, the pathological form associated with neurofibrillary tangles), and t-tau (total tau) all showed no significant change. These three cerebrospinal-fluid or blood-based markers are the closest available proxies for the underlying pathological process of Alzheimer's disease. The authors' conclusion is direct: current evidence does not support antioxidants as disease-modifying therapies. 1
What they did do — and one troubling signal
Against the backdrop of cognitive and pathological null results, one symptom-level outcome stands out. Antioxidant supplementation produced a significant, medium-to-large reduction in neuropsychiatric symptoms, measured by the Neuropsychiatric Inventory (NPI): SMD −0.85 (95% CI −1.13 to −0.57). 1 An NPI score decrease means improvement in the behavioral and psychological symptoms of dementia — agitation, anxiety, depression, delusions, hallucinations, and sleep disturbances that affect both patients and caregivers. An SMD of 0.85, with a confidence interval that clears zero by a wide margin, is a clinically meaningful signal.
Li et al. interpret this cautiously: antioxidants "may serve as adjunctive interventions to improve quality of life and behavioral symptoms," not as treatments that alter the disease's course. 1 That distinction matters for how caregivers and clinicians should frame the potential role of antioxidants.
The troubling signal is a second oxidative-stress result. Urinary 8-iso-prostaglandin F2α (8-iso-PGF2α), a validated marker of lipid peroxidation and oxidative damage, was significantly elevated following antioxidant supplementation: SMD +0.75 (95% CI 0.12 to 1.38). 1 This is the opposite of the expected direction. Antioxidants are used precisely because they are expected to reduce oxidative stress markers. An increase in a lipid peroxidation marker raises the possibility that supplemental antioxidants, in this population, may produce net oxidative harm rather than benefit.
Interpreting this finding requires the full text, which is not yet available. The abstract does not specify which antioxidant types drove the result, whether one subgroup dominated the effect, or whether publication bias may be a factor. The high heterogeneity reported across several outcome measures (I² > 50%) further complicates any single interpretation. The result is a caution signal — not a definitive finding — but it cannot be dismissed.
Context from prior evidence
The null cognitive result for antioxidant supplements in established Alzheimer's disease is consistent with what the existing evidence base already suggested.
Farina et al. 2017, a Cochrane systematic review, examined the specific case of high-dose vitamin E (α-tocopherol 2,000 IU/day) across four double-blind RCTs. 2 In 304 AD patients from one eligible study, vitamin E produced no significant improvement in ADAS-Cog (MD −1.81, 95% CI −3.75 to 0.13; p = 0.07). The minimum clinically important difference for ADAS-Cog is conventionally 4 points — the upper bound of that interval falls well short. In 516 participants with mild cognitive impairment (MCI), vitamin E showed no effect on progression to AD dementia (RR 1.03, 95% CI 0.79–1.35). The Cochrane team concluded they found "no evidence that the alpha-tocopherol form of vitamin E given to people with MCI prevents progression to dementia, or that it improves cognitive function in people with MCI or dementia due to AD." 2
Loading content card…
The prevention picture is different, and that distinction matters. Hu et al. 2025, a dose-response meta-analysis of 11 prospective cohort studies, found that dietary vitamin C intake ≥ 75 mg/day was associated with a 16% lower risk of Alzheimer's disease (HR 0.84, 95% CI 0.76–0.93). 3 Vitamin E dietary intake (HR 0.90, 95% CI 0.60–1.34) and β-carotene (HR 1.02, 95% CI 0.85–1.22) showed no significant association with AD risk. Notably, vitamin C supplements showed only a borderline association (HR 0.85, 95% CI 0.72–1.00). This is an observational finding — cohort studies cannot establish causality — but the signal for dietary vitamin C specifically is one of the more consistent associations in the recent antioxidant-prevention literature.
Loading content card…
Limitations
Several limitations apply to the Li et al. 2026 findings and should inform how much weight to place on any individual result:
- Abstract-only evidence: The formatted full text was not available at acceptance. Subgroup analyses by antioxidant type, dose, trial duration, and baseline severity — each of which could shift the clinical interpretation — have not been examined here. The specific antioxidants driving the NPI benefit and the paradoxical oxidative-stress finding are unknown until the full text is published.
- High heterogeneity: The abstract reports I² > 50% across multiple outcomes. High heterogeneity means the included trials differed substantially in ways that matter (antioxidant type, dose, patient population, outcome measurement protocols). A pooled SMD from a high-heterogeneity meta-analysis describes the central tendency of a heterogeneous set — it cannot be reliably applied to any specific antioxidant or clinical scenario. 1
- No GRADE assessment available: Evidence certainty ratings, risk-of-bias summaries, and sensitivity analyses await the full text.
- Established AD only: All 23 trials enrolled patients with diagnosed Alzheimer's disease. The results say nothing about antioxidant supplementation in MCI or in cognitively healthy adults at elevated AD risk.
- Intervention heterogeneity: "Antioxidants" is not a single substance. The pooled estimates reflect a mix of vitamin C, vitamin E, polyphenols, and combination preparations at varying doses. A null result for the pooled estimate does not rule out specific antioxidants at specific doses producing meaningful effects — it means the literature as a whole does not support that conclusion.
Dietary recommendation
The evidence from Li et al. 2026, read alongside Farina 2017 and Hu 2025, points in three practical directions depending on who is asking.
For individuals with Alzheimer's disease and their caregivers: antioxidant supplements should not be taken with the expectation of slowing cognitive decline or altering disease progression. The best available RCT evidence — now 23 trials and over 10,000 patients — does not support that use. There may be a role for adjunctive use to help manage behavioral symptoms such as agitation or anxiety (NPI SMD −0.85), but only under medical supervision and with awareness of the unresolved oxidative-stress signal. The decision belongs to the treating clinician and family, not to supplement marketing claims.
For health-conscious adults focused on Alzheimer's prevention: the observational evidence favors dietary vitamin C at ≥ 75 mg/day from food — not supplements. 3 That threshold is achievable with one medium orange (~70 mg) plus a handful of strawberries (~30 mg), or half a cup of raw red bell pepper (~95 mg). Reaching ≥ 75 mg/day from whole food is straightforward in a diet that includes fruit and vegetables at most meals; isolating vitamin C as a supplement appears to offer weaker and less consistent benefit based on current evidence.
For registered dietitians counseling AD patients or high-risk individuals: communicate clearly that antioxidant supplement use is not supported as a disease-modifying strategy in established Alzheimer's disease — a position now backed by the largest RCT synthesis on the question. For prevention, the vitamin C–food association from Hu et al. 2025 can be cited as observational support for including vitamin-C-rich foods in dietary patterns, while noting that causality has not been established. Monitor for updated guidance as the Li et al. 2026 full text, including funding disclosures and subgroup data, becomes available on the Frontiers in Nutrition platform. 1
Cover image: AI-generated editorial still life depicting antioxidant-rich foods.
References
- 1
- 2
- 3
This story was produced automatically by a channel. One sentence is all it takes for Neodrop to keep producing for you.