Lithium and the Aging Brain: The 2025 Deficiency Discovery and Why the Human Trials Still Say No
A 2025 Nature paper found that brain lithium falls early in Alzheimer's disease and that lithium orotate reversed pathology in mice. A late-2025 meta-analysis of six human trials and a 2026 Pittsburgh pilot trial of low-dose lithium carbonate both came back negative or inconclusive in people. Here is what each study actually found.
Research context: This article discusses peer-reviewed preclinical research, a meta-analysis, and a pilot clinical trial. It is intended for informational and educational purposes only and does not constitute medical advice. Do not begin lithium supplementation, in any form, without medical supervision. Consult a qualified healthcare professional before making changes to diet, lifestyle, or supplementation.
A mouse discovery outran the human evidence, and the gap matters
An August 2025 paper from Bruce Yankner's lab at Harvard Medical School, published in Nature, reported that lithium is naturally present in the brain and falls early in the course of Alzheimer's disease, and that restoring it with a specific salt, lithium orotate, reversed pathology and memory loss in mouse models (Aron L, Ngian ZK, Qiu C, et al., "Lithium Deficiency and the Onset of Alzheimer's Disease," Nature 2025;645(8081):712-721, PMID: 40770094). Coverage of the paper was extensive, and lithium orotate is now sold as an over-the-counter supplement marketed on the strength of that single mouse study. Two other papers, less widely covered, tell the part of the story that matters most for anyone actually deciding whether to take it: a meta-analysis of six human trials of conventional lithium salts published two months later, and a dedicated pilot trial of low-dose lithium carbonate in people with mild cognitive impairment (MCI) published in 2026. Both come from methodologically careful human research, and neither supports treating the mouse finding as a demonstrated human benefit.
What the Nature paper actually found
The Yankner lab's starting observation was that of several metals measured in postmortem human brain tissue, lithium was the one consistently reduced in people with MCI, a precursor stage to Alzheimer's disease, and reduced further in Alzheimer's disease itself. The paper proposes a specific mechanism for that decline: amyloid-beta, as it accumulates, binds and sequesters lithium, lowering its bioavailability in brain tissue. To test whether that depletion matters functionally, the researchers depleted dietary lithium in wild-type and Alzheimer's-model mice. Cutting cortical lithium by roughly half increased amyloid-beta deposition and phospho-tau accumulation, activated pro-inflammatory microglia, and drove loss of synapses, axons, and myelin, with accelerated cognitive decline, an effect the authors tie at least partly to activation of the kinase GSK3β. Replacing lithium with lithium orotate, a salt the authors report binds amyloid less than standard lithium salts, prevented these pathological changes and memory loss in both the Alzheimer's-model and aging wild-type mice.
That is a genuinely novel mechanistic finding, and it is entirely a mouse and postmortem-tissue result. The paper does not report a human trial of lithium orotate, dosing data in humans, or any clinical outcome in living people. Its own framing is explicit: altered lithium homeostasis "may be an early event" in Alzheimer's pathogenesis, and lithium replacement with amyloid-evading salts is "a potential approach" to prevention and treatment, which is the language of a hypothesis to test next, not a treatment already shown to work. Senior author Yankner was widely quoted making the same point to journalists: nobody should take anything based on mouse studies alone.
What happened when conventional lithium salts were actually tested in people
Lithium is not a new idea in dementia research. Lithium carbonate, the form used to treat bipolar disorder, has been tested in small human trials for Alzheimer's disease and MCI for close to two decades, well before the 2025 mechanistic paper. A meta-analysis published in November 2025 pooled six randomized, placebo-controlled trials of lithium, totaling 435 participants with MCI or Alzheimer's disease, spanning study durations from about 10 weeks to two years and testing lithium carbonate, lithium gluconate, and lithium sulfate (Kishi T, Matsunaga S, Saito Y, Iwata N, "Lithium for Alzheimer's Disease: Insights from a Meta-Analysis," Neuroscience and Biobehavioral Reviews 2026;180:106458, PMID: 41205642). Pooling across trials, lithium produced no statistically significant improvement in cognitive outcomes, measured mainly with the Alzheimer's Disease Assessment Scale-Cognitive Subscale and the Mini-Mental State Examination, compared with placebo. Behavioral symptoms such as agitation and aggression also showed no significant benefit, and adverse events and discontinuation rates were comparable between lithium and placebo groups. Subgroup analyses by diagnosis, trial length, dose, and salt formulation found no advantage for lithium in any category the authors examined.
The authors' own explanation for the mismatch between encouraging mechanistic biology and flat clinical results centers on bioavailability: conventional lithium salts like carbonate are more readily bound by amyloid and may not reach effective concentrations in the brain at tolerable doses, which is the same bioavailability problem the 2025 Nature paper proposes lithium orotate might sidestep. That proposal, as of this meta-analysis, remains untested in a controlled human trial; none of the six pooled trials used orotate.
The 2026 pilot trial: feasibility and safety, not yet efficacy
The most directly relevant human data come from a pilot randomized trial run at the University of Pittsburgh, published in 2026 in JAMA Neurology, which tested low-dose lithium carbonate specifically in older adults with MCI (Gildengers AG, Ibrahim TS, Anderson SJ, et al., "Low-Dose Lithium for Mild Cognitive Impairment: A Pilot Randomized Clinical Trial," JAMA Neurology 2026;83(4), DOI: 10.1001/jamaneurol.2026.0072, ClinicalTrials.gov NCT03185208). The trial ran from February 2018 to August 2024 with two years of follow-up, randomizing 83 adults aged 60 and older with MCI to daily low-dose lithium carbonate or placebo, with 80 starting treatment. It was explicitly designed and sized as a feasibility and safety pilot, with six prespecified coprimary outcomes covering cognitive tests, hippocampal volume, cortical gray matter volume, and brain-derived neurotrophic factor (BDNF), the growth factor central to the neuroplasticity mechanisms covered in our BDNF and neuroplasticity review.
None of the six coprimary outcomes met the trial's prespecified significance threshold. The closest result was on a memory test, the California Verbal Learning Test-II delayed recall: scores declined 1.42 points per year in the placebo group versus 0.73 points per year in the lithium group, a difference of 0.69 points per year that landed right at the edge of significance (95% CI 0.01 to 1.37, p = 0.05). Hippocampal and cortical gray matter volumes declined over time in both groups with no significant treatment-by-time interaction. On safety, serious adverse events occurred in 12 of 41 (29%) on lithium versus 9 of 39 (23%) on placebo, none judged definitely treatment-related, and more common adverse events including elevated creatinine, diarrhea, tiredness, and tremor occurred somewhat more often on lithium than placebo, consistent with lithium's known tolerability profile at low doses.
The authors describe their own result accurately: the trial established feasibility, confirmed a workable safety and tolerability profile for low-dose lithium carbonate over two years in this population, and generated effect-size estimates to power a future, larger efficacy trial. It did not demonstrate that lithium slows cognitive decline in MCI. A single borderline result on one of six coprimary outcomes, in a trial explicitly not powered to detect efficacy, is a reason to run a bigger trial, not a reason to conclude the question is answered.
Reading the three studies together
Lined up, these three papers tell a coherent and honest story rather than a contradictory one. The Nature paper identifies a plausible, mechanistically detailed reason why the aging and Alzheimer's brain might become lithium-deficient, and shows that correcting that deficiency with a specific, less amyloid-reactive salt reverses pathology in mice. The meta-analysis shows that decades of small human trials using the conventional, more amyloid-reactive salts already on the market have not produced a measurable cognitive benefit, which is consistent with, rather than a refutation of, the bioavailability problem the mouse paper describes. The pilot trial shows that low-dose lithium carbonate is feasible and reasonably well tolerated over two years in people with MCI, generates one borderline-significant memory signal worth following up, and still does not constitute evidence of benefit. What is genuinely missing from the literature, and what none of these three papers provide, is a controlled human trial of lithium orotate itself. Readers interested in how this fits the broader landscape of tested and untested Alzheimer's prevention strategies can see our Alzheimer's prevention research review for other compounds facing the same gap between mechanistic promise and clinical proof.
Why the orotate marketing is ahead of the science
Lithium orotate supplements sold for cognitive or mood support predate the 2025 Nature paper, but the paper has visibly accelerated interest in the specific claim that orotate is safer or more effective than standard lithium salts for brain health. That claim currently rests entirely on rodent and cell-culture data. No published, controlled human trial has tested lithium orotate against placebo for cognition at any dose, and the two groups that have run the relevant human trials, the Kishi meta-analysis authors and the Pittsburgh pilot trial team, both tested conventional salts, not orotate. Low-dose lithium, regardless of salt form, still carries lithium's known risks, including effects on kidney function, thyroid function, and a narrow margin between a dose that does something and a dose that causes toxicity, which is why both human trials discussed here were conducted under medical supervision with monitoring built into the protocol.
Key takeaways
A 2025 Nature paper found that brain lithium is depleted early in Alzheimer's disease, partly through amyloid sequestration, and that restoring it with lithium orotate reversed pathology and memory loss in mouse models, a genuinely novel mechanistic finding with no human trial data behind it yet. A November 2025 meta-analysis of six randomized trials in 435 people with MCI or Alzheimer's disease found that conventional lithium salts, tested over nearly two decades of research, produced no significant cognitive or behavioral benefit compared with placebo. A 2026 pilot trial of low-dose lithium carbonate in 80 adults with MCI established feasibility and an acceptable safety profile over two years but did not meet its prespecified efficacy thresholds, with one memory measure landing at the border of statistical significance. No controlled human trial of lithium orotate specifically, the form the mouse study tested and the form now marketed on the strength of that study, has been published.
FAQ
Does the 2025 Nature study prove lithium orotate helps the human brain?
No. It is a mouse and postmortem human tissue study. It identifies a real and detailed mechanism, amyloid-driven lithium depletion, and shows that lithium orotate corrects it in mice, but it includes no human trial of orotate, no human dosing data, and no human cognitive outcome. The paper's own authors describe their proposed treatment approach as a "potential approach," not a demonstrated one, and the senior author has publicly cautioned against self-treating based on mouse data alone.
Did any human trial test lithium orotate specifically?
Not that has been published through 2026. The two human studies discussed here, the November 2025 meta-analysis and the 2026 Pittsburgh pilot trial, both tested conventional lithium salts: carbonate, gluconate, and sulfate. Lithium orotate's proposed advantage, lower amyloid binding, is a hypothesis drawn from the mouse data, and testing it directly in people would require a dedicated trial that, as of this writing, does not exist in the published literature.
Does the negative 2025 meta-analysis mean lithium definitely doesn't work for Alzheimer's?
It means that pooling the trials run so far, conventional lithium salts have not produced a statistically significant cognitive or behavioral benefit over placebo in people with MCI or Alzheimer's disease. The meta-analysis authors themselves propose that bioavailability limits, including amyloid binding of the salts tested, may explain the gap between preclinical promise and clinical results, which is the same limitation the 2025 mouse study was designed to address with a different salt. The negative result applies to the forms and doses tested, not necessarily to every possible lithium formulation.
What did the 2026 pilot trial actually show?
It showed that low-dose lithium carbonate can be given safely and tolerably to older adults with MCI over two years, with adverse event rates broadly similar to placebo, and that the trial design is feasible for running in a larger study. None of its six prespecified outcomes, covering cognition, hippocampal and cortical volume, and BDNF, reached statistical significance, though one memory test showed a difference that sat exactly at the conventional significance cutoff (p = 0.05). The authors describe this as generating effect-size estimates for a future, adequately powered efficacy trial, not as evidence that lithium works.
Primary sources: Aron L, Ngian ZK, Qiu C, Choi J, Liang M, Drake DM, Hamplova SE, Lacey EK, Roche P, Yuan M, Hazaveh SS, Lee EA, Bennett DA, Yankner BA. "Lithium Deficiency and the Onset of Alzheimer's Disease." Nature 2025;645(8081):712-721. DOI: 10.1038/s41586-025-09335-x. PMID: 40770094. Open-access full text: PMC12443616. Kishi T, Matsunaga S, Saito Y, Iwata N. "Lithium for Alzheimer's Disease: Insights from a Meta-Analysis." Neuroscience and Biobehavioral Reviews 2026;180:106458. DOI: 10.1016/j.neubiorev.2025.106458. PMID: 41205642. Gildengers AG, Ibrahim TS, Anderson SJ, Emanuel JE, Santini T, Diaz JL, Lopresti BJ, Royse SK, Lopez OL, Zeng X, de Almeida B, Alkhateeb SK, Chu C, Karikari TK, Lee L, Weinstein AM, Butters MA. "Low-Dose Lithium for Mild Cognitive Impairment: A Pilot Randomized Clinical Trial." JAMA Neurology 2026;83(4). DOI: 10.1001/jamaneurol.2026.0072. PMID: 41770546. Open-access full text: PMC12954601. Trial registration: ClinicalTrials.gov NCT03185208. This article summarizes published, peer-reviewed research and is not a substitute for individualized medical advice.
Disclaimer
This article is for educational and research-context purposes only. It does not constitute medical advice and is not a substitute for consultation with a qualified healthcare practitioner. Lithium, in any salt form, requires medical supervision and monitoring of kidney and thyroid function, and it has a narrow margin between an active dose and a toxic one. Nothing in this article should be read as a recommendation to begin lithium supplementation, including lithium orotate, without medical guidance. Always consult a doctor before beginning, modifying, or stopping any supplement, particularly if you have a diagnosed cognitive, kidney, thyroid, or psychiatric condition, are pregnant or breastfeeding, or take prescription medications.