Lithium Was the One Metal Missing From the Brains of People With Early Memory Loss. A Microdose Restored Memory in Mice. The Human Evidence Is Thinner Than the Headlines.
For seventy years lithium has lived in one box: a psychiatric drug with a narrow safety margin, prescribed for bipolar disorder and monitored with blood tests. In August 2025 a Harvard team published a paper in Nature that tried to open a second box. Lithium, they reported, was the only metal among dozens measured that was significantly lower in the brains of people in the earliest stage of memory decline, and restoring it in mice with a tiny dose reversed the damage. The story spread fast, and lithium orotate capsules sold out in several online shops within weeks. The mouse work is striking. The human evidence behind it is real but small, and it comes mostly from drinking water records and trials with fewer than 120 people. This article sets out what each piece of evidence shows, how the doses compare, and where the safety line sits for men over 40.
Key Takeaways: What the Research Shows About Low-Dose Lithium and Dementia
A 2025 Nature study by Aron and colleagues, using donated brain tissue from the Rush Memory and Aging Project, found lithium was the only metal measured that was significantly lower in the prefrontal cortex of people with mild cognitive impairment than in people with normal memory. In mice, a lithium-restricted diet that roughly halved brain lithium accelerated amyloid plaques, tau pathology and memory loss, while lithium orotate at about one thousandth of a psychiatric dose reversed those changes. A Danish registry study of 73,731 dementia patients and 733,653 controls, published by Kessing and colleagues in JAMA Psychiatry in 2017, found that long-term drinking water lithium above 15 micrograms per litre was associated with a 17 percent lower dementia rate, with an incidence rate ratio of 0.83. A 2022 PLOS Medicine study of 29,618 NHS mental health patients aged 50 and over found prescribed lithium was associated with a hazard ratio for dementia of 0.56. The largest randomized trial of lithium for memory enrolled 113 people with Alzheimer's disease, and no trial has yet shown that low-dose lithium prevents dementia in healthy men.
Two Human Signals That Point the Same Way
Ratios for dementia associated with lithium exposure. The Danish drinking water registry reported an incidence rate ratio of 0.83, 95 percent confidence interval 0.81 to 0.85, for water above 15 micrograms per litre. The 2022 PLOS Medicine study of NHS mental health records reported a hazard ratio of 0.56, 95 percent confidence interval 0.40 to 0.78, for people prescribed lithium. Both are observational, and neither shows that adding lithium prevents dementia.
- Ratio (below 1 = fewer dementia cases)
What did the 2025 Nature study find about lithium and Alzheimer's disease?
The 2025 Nature study found that lithium levels fall in the human brain at the earliest stage of memory decline and that lithium shortage alone was enough to drive Alzheimer's-like changes in mice. Bruce Yankner's group at Harvard Medical School measured 27 metals in brain and blood samples from the Rush Memory and Aging Project, a long-running study in which older adults agree to cognitive testing during life and brain donation after death. Lithium stood out as the one metal significantly reduced in the prefrontal cortex of people with mild cognitive impairment. The team also found that amyloid plaques bind lithium, which suggests the plaques themselves pull lithium out of circulation in the surrounding tissue. In mice fed a diet that cut cortical lithium by about half, amyloid and tau pathology increased, the brain's immune cells lost their ability to clear amyloid, and memory declined faster. Lithium orotate, a form the plaques bind less tightly, reversed those changes in Alzheimer's model mice and in normal aged mice at a dose roughly a thousand times lower than the lithium carbonate doses used in psychiatry.
How could such a small amount of lithium affect the brain?
Trace lithium appears to act on the same enzyme that psychiatric doses inhibit, an enzyme called glycogen synthase kinase 3 beta, or GSK3 beta, which adds phosphate groups to tau protein. Hyperphosphorylated tau is the main component of the neurofibrillary tangles that spread through the Alzheimer's brain, so restraining GSK3 beta is a plausible route to slower tangle formation. In the Nature mice, lithium deficiency raised GSK3 beta activity, and blocking that enzyme recreated much of the protective effect of lithium replacement. Lithium also influences microglia, the brain's resident immune cells, and in the 2025 study lithium-deficient microglia were worse at engulfing amyloid. None of this proves that microgram lithium does the same in a human brain. The mechanism is coherent and testable, which is a better starting point than most longevity candidates have, and it explains why researchers take the drinking water data seriously rather than dismissing it as coincidence.
Does lithium in drinking water lower dementia risk?
Higher natural lithium in drinking water was associated with lower dementia rates in the largest study of the question, though the relationship was not a straight line. Kessing and colleagues linked Danish national health registers to lithium measurements from 151 waterworks and estimated each person's long-term exposure from 1986 onward. The study compared 73,731 people diagnosed with dementia between 1970 and 2013 against 733,653 matched controls, all aged 50 to 90. Against a reference band of 2.0 to 5.0 micrograms per litre, exposure above 15.0 micrograms per litre carried an incidence rate ratio of 0.83, a 17 percent lower rate, and 10.1 to 15.0 micrograms per litre sat at 0.98, essentially no difference. The middle band of 5.1 to 10.0 micrograms per litre was associated with a 22 percent higher rate, an incidence rate ratio of 1.22, which nobody has convincingly explained. A Texas county analysis by Fajardo and colleagues in 2018 also reported lower Alzheimer's mortality where water lithium was higher. Water studies cannot rule out confounding by geography, income or other minerals, so the Danish result is a signal worth testing and not a demonstrated cause.
Dementia Rates by Drinking Water Lithium in Denmark: Lower at the Top, Not a Straight Line
Incidence rate ratios for dementia by estimated long-term lithium in drinking water, from Kessing and colleagues in JAMA Psychiatry in 2017, comparing 73,731 people with dementia against 733,653 matched controls aged 50 to 90. Against the 2.0 to 5.0 microgram per litre reference band, exposure above 15.0 micrograms per litre carried a ratio of 0.83, while the 5.1 to 10.0 band carried a ratio of 1.22. Values below 1 indicate lower dementia rates.
Does lithium in water affect how long people live?
Two lines of evidence tie trace lithium to overall survival, but both are observational in humans. Zarse and colleagues reported in the European Journal of Nutrition in 2011 that across 18 neighbouring municipalities in Oita prefecture, Japan, higher tap water lithium was associated with lower all-cause mortality. In the same paper, adding lithium to the growth medium of the nematode worm C. elegans at low concentrations extended its lifespan. Separately, water lithium has been studied as a suicide factor: Ohgami and colleagues found an inverse association between tap water lithium and suicide rates in the same Oita region in 2009, and a 2020 meta-analysis by Memon and colleagues in the British Journal of Psychiatry pooled studies from several countries and found the same direction. These findings make low-level lithium look biologically active at environmental doses. They do not show that raising intake in someone who already drinks low-lithium water will add years of life.
What do human trials of lithium for memory show?
Three small randomized trials have tested lithium for memory decline, and all three leaned positive, but together they enrolled fewer than 220 people. Forlenza and colleagues gave lithium carbonate to 45 older adults with amnestic mild cognitive impairment for 12 months in a 2011 British Journal of Psychiatry trial, targeting a low serum level of 0.25 to 0.5 millimoles per litre, and reported lower cerebrospinal fluid phosphorylated tau and better performance on attention tasks than placebo. The same group extended the design to 61 participants over two years, reporting in 2019 that the lithium group had more stable cognitive and functional scores. Nunes, Viel and Forlenza tested a true microdose in 113 people with Alzheimer's disease in Current Alzheimer Research in 2013: 300 micrograms of lithium a day for 15 months. Mini-Mental State Examination scores held steady in the lithium group and declined in the placebo group. No trial has tested low-dose lithium for dementia prevention in cognitively healthy adults, and a Columbia University trial called LATTE, testing low-dose lithium carbonate in mild cognitive impairment, has not yet reported a definitive result.
The Big Numbers Are Observational. The Randomized Trials Are Tiny.
People included in each major human study of lithium and dementia or memory decline, on a logarithmic scale. Blue bars are observational studies: the Danish drinking water registry of 807,384 people and 29,618 NHS mental health patients. Violet bars are randomized trials in people who already had memory decline, with 113, 61 and 45 participants. No randomized trial has tested lithium for dementia prevention in healthy adults.
Is prescribed lithium linked to lower dementia rates?
People prescribed lithium have had lower dementia rates in several record-based studies, the most detailed of which found a hazard ratio of 0.56. Chen and colleagues published the analysis in PLOS Medicine in 2022, using electronic records from Cambridgeshire and Peterborough NHS Foundation Trust for 29,618 patients aged 50 and over. Of those, 548 had been prescribed lithium, and 9.7 percent of them developed dementia during follow-up against 11.2 percent of people never given lithium. After adjustment for age, sex, other diagnoses and other medicines, the hazard ratio was 0.56, with a 95 percent confidence interval of 0.40 to 0.78. The weakness is that people are prescribed lithium for bipolar disorder or severe depression, conditions that themselves raise dementia risk, and the comparison group differs in ways adjustment may not fully capture. The study also involved only 53 dementia cases among lithium users, which is why its confidence interval is wide.
How does a microdose of lithium compare with a psychiatric dose?
A lithium microdose is thousands of times smaller than a psychiatric dose, and the gap matters because the side effects of lithium scale with blood level. A typical bipolar regimen of 900 milligrams of lithium carbonate a day delivers about 169 milligrams of elemental lithium, since lithium makes up about 18.8 percent of the carbonate by weight. The low-dose memory trials by Forlenza started at 150 milligrams of carbonate, or about 28 milligrams of elemental lithium. The Nunes microdose trial used 0.3 milligrams, or 300 micrograms. A person drinking 2 litres of Danish water at 15 micrograms per litre takes in about 0.03 milligrams a day from water. Over-the-counter lithium orotate capsules are commonly sold at 1 to 5 milligrams of elemental lithium, which places a single 5 milligram capsule about 17 times above the Nunes trial dose and more than 150 times above the Danish high-water intake. Dietary lithium from grains, vegetables and some mineral waters varies widely by region and is poorly characterized, so no official daily requirement exists.
From Tap Water to Psychiatric Dose: Four Orders of Magnitude of Lithium
Approximate daily elemental lithium in milligrams, on a logarithmic scale. Two litres of water at 15 micrograms per litre supplies about 0.03 mg. The Nunes 2013 microdose trial used 0.3 mg. A common lithium orotate capsule carries 5 mg. The Forlenza low-dose memory trials started at 150 mg of lithium carbonate, about 28 mg of lithium, and a standard 900 mg carbonate psychiatric regimen supplies about 169 mg. Lithium is about 18.8 percent of lithium carbonate by weight.
Is low-dose lithium safe?
Low-dose lithium has a clean safety record in the small trials that tested it, but the known risks of lithium at psychiatric doses are serious and well documented. At therapeutic serum levels of 0.6 to 1.0 millimoles per litre, long-term lithium use is linked to hypothyroidism, raised calcium from overactive parathyroid glands, a reduced ability of the kidneys to concentrate urine, and in some people a gradual decline in kidney function. Levels above about 1.5 millimoles per litre can cause toxicity, with tremor, confusion, vomiting and in severe cases seizures. The Forlenza trials held serum lithium to 0.25 to 0.5 millimoles per litre and reported no excess of serious adverse events, and the 300 microgram Nunes trial produced serum levels too low to affect routine monitoring. Lithium is cleared by the kidneys, so anything that reduces kidney clearance raises blood levels. Thiazide diuretics, ACE inhibitors, angiotensin receptor blockers and regular NSAID use such as ibuprofen or naproxen all do this, and those drugs are common in men over 40 with high blood pressure or joint pain. Dehydration from heat, heavy training or illness has the same effect.
Where the Blood Levels Sit: Low-Dose Trials, Psychiatric Treatment and Toxicity
Serum lithium ranges in millimoles per litre. The Forlenza memory trials targeted 0.25 to 0.5. Standard maintenance treatment for bipolar disorder targets 0.6 to 1.0, the range linked to long-term thyroid and kidney effects. Toxicity becomes likely above about 1.5. The 300 microgram Nunes microdose produces levels far below the left edge of this chart. Diuretics, ACE inhibitors, angiotensin receptor blockers, NSAIDs and dehydration all push levels higher.
What about lithium orotate supplements?
Lithium orotate is sold as a dietary supplement in the United States, is not an approved drug, and has almost no human trial data of its own. The 2025 Nature paper renewed interest in orotate because the salt bound less tightly to amyloid in mice, but that finding comes from animal brain tissue and has not been reproduced in people. Older rat work from the late 1970s by Smith and Schou reported higher kidney lithium concentrations after orotate than after carbonate, a question that was never settled in humans. Supplement labels are not checked for accuracy before sale, and the amount of elemental lithium per capsule is not always stated clearly. Case reports of lithium toxicity from orotate exist, including a documented overdose in a young adult who took a large number of capsules at once. Researchers involved in the Nature study have publicly cautioned against self-dosing until human trials report, and that caution is the reasonable reading of the data today.
What should a man over 40 take from the lithium evidence?
A man over 40 should treat low-dose lithium as a promising research question, not a supplement protocol, and should raise it with a doctor before taking any. The strongest human data are observational, the randomized trials are small and enrolled people who already had memory problems, and the mechanistic case comes mostly from mice. For men already prescribed lithium, the dementia data are reassuring and support continuing monitored treatment. For men thinking about orotate capsules, the practical questions are kidney function, thyroid function and the medicines already in the cabinet, because diuretics, blood pressure drugs and anti-inflammatories change how lithium is handled. The things with firmer dementia-risk evidence behind them remain blood pressure control, hearing correction, regular exercise, not smoking and treating diabetes. If a trial like LATTE shows that microgram lithium slows decline in people with early memory loss, the calculation will change, and it will change on the back of human data rather than a mouse result.
Frequently Asked Questions
Have more questions? Contact us or explore our full peptide intelligence platform.
Related Articles
Does Telomere Length Determine Lifespan? Short Telomeres Track With Earlier Death. The Genetics Say Longer Ones Raise Cancer Risk.
Does telomere length determine lifespan? Short telomeres track with earlier death, yet genetic studies show longer ones raise the risk of several cancers.
Low-Dose Aspirin Added 8 Percent to Male Mouse Lifespan. In 19,114 Healthy Older Adults, It Added No Healthy Years and Deaths Ran Higher.
Does low-dose aspirin help longevity? It extended male mouse lifespan 8%, but the ASPREE trial in 19,114 healthy older adults found no gain and more bleeding.
Is NAD+ a Peptide? No. It Is a Coenzyme Built From Two Nucleotides, and the Peptide Label Comes From the Clinic, Not the Chemistry
Is NAD+ a peptide? No. NAD+ is a coenzyme built from two nucleotides. What NAD+ is used for, what IV NAD+ studies found, and how it differs from SS-31.