Lithium has not been shown to prevent or slow Alzheimer's in humans. The evidence we have comes in three pieces. In 2025, a Harvard team reported in Nature that lithium is depleted in brains with Alzheimer's and that low-dose lithium orotate prevented the damage in mice. Population studies have found that dementia is slightly less common in areas with more lithium in the drinking water. The randomized trials that tested low-dose lithium carbonate in mild cognitive impairment, meanwhile, have been small and their results mixed: the most recent, LATTICE (JAMA Neurology, 2026), failed to cross its prespecified significance threshold on any of its six primary outcomes.
For lithium orotate, the form sold as a supplement, the picture is even clearer: as of November 2026, there is not a single randomized trial with published results that has tested it against Alzheimer's in humans. The rest of this article works through these pieces one by one and ends by explaining why taking lithium on your own is not a good idea.
Liviu Aron and Bruce Yankner's team measured 27 metals in postmortem brain tissue and blood from older adults with no cognitive impairment, from people with mild cognitive impairment (the stage considered a precursor to Alzheimer's) and from people with Alzheimer's disease. In the prefrontal cortex, one of the regions hit hardest by the disease, lithium was the only metal that was significantly reduced in both mild cognitive impairment and Alzheimer's. The reduction in Alzheimer's was also seen in a second, independent set of tissue samples.
Two details matter. First, blood lithium levels did not differ between the patient groups and the healthy group; in other words, this is not the kind of deficiency a blood test can pick up. Second, lithium was concentrated inside the amyloid plaques. According to the researchers, the plaques bind lithium, leaving less of it for the surrounding brain tissue. The lower the lithium in plaque-free tissue, the lower the scores people had achieved on memory tests while they were alive.
This part is a comparison made in postmortem tissue. It tells us that the loss of lithium and the disease occur together; on its own, it does not prove which came first. For the question of cause and effect, the team turned to mice.
The researchers cut the lithium in the mice's feed by 92%; lithium in the brain cortex fell by about half. In Alzheimer's-model mice, amyloid deposition and phospho-tau increased, microglia shifted into a state that fuels inflammation, and the loss of synapses and myelin and the decline in memory accelerated. At least part of the effect ran through the activation of an enzyme called GSK3β.
Next, 16 different lithium salts were compared. Lithium carbonate, the standard in psychiatry, bound to amyloid more strongly; lithium orotate bound more weakly. In mice given a very low dose of orotate in their drinking water (4.3 microequivalents of lithium per liter), plaque and tau accumulation decreased and memory loss was largely reversed; lithium carbonate at the same dose showed no significant effect. In an aged mouse model, plaque burden fell by about 70%. With long-term dosing, the mice's kidney and thyroid markers (blood urea nitrogen, creatinine, TSH) did not change. To give a sense of the dose: according to the authors, the lithium used in bipolar disorder raises the blood level to roughly 1,000 times the natural level, whereas the amount given to the mice stayed within the natural range.
All of this is mouse data. Genetically engineered mouse models are not an exact copy of the human disease, and these results have not yet been replicated in another laboratory. The claim that the orotate salt is superior is also contested. In 2026, Hajek and colleagues wrote in the British Journal of Psychiatry that lithium orotate would dissociate in the stomach and be absorbed as the ordinary lithium ion, that in experimental studies the two salts are distributed similarly when taken by mouth, and that focusing on a single salt could be problematic from a safety standpoint. A review published the same year in JAMA Psychiatry likewise noted that the deficiency-replacement hypothesis still awaits confirmation by independent teams. As we saw in the case of taurine, there is no guarantee that a brilliant mouse study will produce the same result in humans.
Lithium occurs in nature in trace amounts and also finds its way into drinking water. In 2017, Kessing and colleagues used Denmark's national registries in a study published in JAMA Psychiatry: 73,731 people with dementia were compared with 733,653 people matched for age and sex. Among people living in areas with more than 15 micrograms of lithium per liter of water, the incidence of dementia was 17% lower than among those in the 2–5 microgram range (incidence rate ratio 0.83). The relationship did not follow a straight line: in the 5.1–10 microgram range, the risk was 22% higher. The authors stated explicitly that other factors linked to the municipality of residence may have confounded the result.
A US study published in 2026 asked the same question at the county level, using data from about 62 million Medicare beneficiaries. In high-lithium counties, dementia incidence was slightly lower (ratio 0.958), while mortality among people with dementia was slightly higher (1.017). The differences are small and the data are not at the level of individuals: it is not known who drank which water, or how much. An analysis covering 808 cities and wards in Japan found the association only in women.
The 2026 meta-analysis by Campedelli and colleagues pulled this picture together: across four cohort studies that followed people exposed to lithium, the risk of dementia was about 37% lower (hazard ratio 0.633). In the authors' own words, the observational evidence does not allow a firm causal inference.
The lithium that has been tested in humans is not the orotate salt but prescription lithium carbonate. Forlenza and colleagues at the University of São Paulo (British Journal of Psychiatry, 2019) assigned 61 older adults with amnestic mild cognitive impairment to lithium or placebo for two years. The dose was adjusted so that the blood level stayed below the range used in psychiatric treatment (0.25–0.5 mEq/L). The placebo group showed cognitive and functional decline; the lithium group held steady over the two years.
LATTICE, the trial that put this encouraging result to the test with more comprehensive measurements, was published in JAMA Neurology in 2026. At the University of Pittsburgh, 83 people aged 60 and over with mild cognitive impairment were randomized; 80 started treatment and were followed for two years. Among those who completed the trial, the mean daily dose was 195 mg and the mean blood level was 0.17 mEq/L. The results were as follows:
One weakness of the trial was that most participants had no amyloid buildup in the brain; only 21 of the 80 were amyloid positive. In this small subgroup the effect looked larger, but that is an exploratory analysis. A Bayesian reanalysis of the same data put the probability of a benefit in verbal memory at 96.5% and still described the overall picture as "inconclusive." A meta-analysis published in 2025, pooling six randomized trials and 394 participants, likewise concluded that lithium did not provide a consistent benefit in cognition or daily functioning.
No. In November 2026 we searched PubMed for published studies on lithium orotate; not a single completed randomized trial in humans aimed at preventing Alzheimer's or dementia turned up. The human data we could find are these: a methods study in which nine healthy men were given 5 mg of lithium orotate a day for up to 28 days, and which tested only whether lithium could be imaged in the brain, and a survey of 211 people who use lithium supplements. The survey's authors reported that side effects, and symptoms that appeared after stopping, were more common than they had expected.
The first direct test is on its way. According to its ClinicalTrials.gov record, Johns Hopkins University is planning a 40-person phase 1/2 trial (LiO-AD, NCT07459959) that will compare lithium orotate with placebo in people with Alzheimer's disease. When this article was published, the record's status was "not yet recruiting"; the start is projected for December 2026 and primary completion for June 2029. A trial of this size gives an idea of feasibility and safety; on its own, it cannot answer the question of whether lithium orotate prevents the disease.
This article is not a recommendation for you to take lithium orotate or any other lithium product. Lithium is a substance with a narrow margin between the level at which it is effective and the level at which it is harmful. Nor has it been shown in humans that the orotate salt is safer than carbonate.
If you have kidney or thyroid disease, or if you take blood pressure medication or regular painkillers, do not start any product that contains lithium without consulting your doctor. If you take prescription lithium for bipolar disorder or for any other reason, nothing in this article is a reason to change your dose or to switch salts; that decision is made with your doctor.
Among old drugs being reassessed for Alzheimer's, a similar wait continues with metformin, and among new targeted approaches, with senolytics. Lithium has joined this list as one of the candidates with the best-described mechanism; the human evidence, however, is still at the very start of the road.
There is no evidence showing this in humans. Harvard's 2025 Nature study showed that lithium is depleted in brains with Alzheimer's and that low-dose lithium orotate prevented the damage in mice. The randomized trials in humans, however, are small and their results mixed: in the 80-person LATTICE trial published in 2026, none of the six primary outcomes crossed the prespecified significance threshold.
There is no human evidence to support it. As of November 2026, no published randomized trial shows that lithium orotate prevents or slows Alzheimer's in humans; the positive results come from mouse experiments. Johns Hopkins University's first placebo-controlled trial, with 40 people, has not yet started enrolling participants. Lithium is a substance that affects the kidneys and thyroid and interacts with other medications; it should not be used without consulting a doctor.
No. In the Harvard study, blood lithium levels in people with mild cognitive impairment and Alzheimer's were no different from those of healthy older adults; the difference was seen only in brain tissue examined after death. There is currently no test in clinical use that shows a lithium deficiency in the brain.
Some population studies have found such an association. In the Danish study covering 73,731 people with dementia, dementia incidence was 17% lower in areas with more than 15 micrograms of lithium per liter of water; in areas with intermediate lithium levels, however, the risk was higher. These studies are observational, cannot separate out other differences between regions and do not show cause and effect.
In the LATTICE trial, conducted under medical supervision and with blood monitoring, low-dose lithium carbonate was generally tolerated, but diarrhea, fatigue and tremor were more common in the lithium group, and older participants had difficulty tolerating doses above 300 milligrams a day. This result does not show that unsupervised supplement use is safe; the margin between lithium's effective level and its harmful level is narrow.
"The finding that lithium is depleted in the brain deserves to be taken seriously. That lithium orotate prevents Alzheimer's, however, has so far been shown only in mice; the best human trial was done with the carbonate salt and did not meet its primary endpoints."
— Vivelong Longevity Science Team