Flawed Study Asks Important Questions on Lithium

Not wanting to waste anyone’s time, I hesitated in posting this. I suspect we will be hearing about it.

Low-Dose Lithium for Mild Cognitive ImpairmentA Pilot Randomized Clinical Trial

Findings In this pilot randomized clinical trial of 80 participants, none of the 6 coprimary outcomes reached the prespecified significance threshold; for verbal memory, scores declined by 1.42 points annually in the placebo group vs 0.73 points in the lithium group, which did not meet the prespecified threshold for multiple comparisons. Exploratory analyses suggested possible larger effects among amyloid-positive participants.

https://jamanetwork.com/journals/jamaneurology/fullarticle/2845746

1 Like

They used lithium carbonate at 150mg or 300mg. I’m not sure how that dose compares with lithium orotate 5mg, but have seen where orotate works better.

Looks like it didn’t do much.

2 Likes

Sorry but that was a waste of time

2 Likes

So if you’re here, and you are reading about this lithium study, you might be interested in this n=1 experience. I had been taking 5 gm of lithium orotate every night. Titrated up from 1 gm to alternating 1 and 5 gm to 5gm every night. Past few months have had swelling (edema) in ankles and feet. Might be from hours standing at an easel (at age 76). But, I looked up lasix, which I have no intention of taking. I encountered an interaction between lasix and lithium that causes water retention. Then looked into lithium orotate and found a connection with water retention and weight gain. Backed off the lithium. Too early to tell, and also reducing the standing/taking breaks and walking more. We’ll see. . . .

2 Likes

Do you mean grams or milligrams @Deborah_Hall?

3 Likes

Dang, I hope you meant mg.
Yours is the first one I have noticed that low-dose lithium orotate has this effect. While I don’t dispute your results, they are unlikely to be from the lithium orotate you are taking.

Since I have not personally experienced this or seen anyone else posting this side effect, I believe it is a coincidence.

I asked Claude Opus 4.7

The short answer: “At therapeutic (prescription) lithium doses this is real and well-documented, but at your dose range (5–10 mg elemental lithium from orotate) the claim has essentially no direct evidence — it’s a plausibility argument extrapolated from high-dose pharmacology.”

Therapeutic serum lithium sits in the 0.6–1.2 mmol/L range. A 5–10 mg elemental dose from orotate produces serum levels roughly two to three orders of magnitude below that — typically in the low micromolar to sub-micromolar range.

Individual sensitivity is also a real thing — there’s a sodium/aldosterone axis involvement where some people may respond at lower thresholds than the population mean, though I’m not aware of documented cases at the 5–10 mg range.

REF1, REF2

2 Likes

Thanks Rob Tuck and Charles, yes, mg. not gm. oof!
I hope you’re right as I would like to go back to the 5 mg at least a few times per week.

I believe you are right @Deborah_Hall in noting that the potential for significant individual differences. This is typifies a common problem I see when we reason from population to individual data. I think I will write a post on that topic one of these days as I run into it quite a bit. For what it is worth, I experience no effects of any kind that I can attribute to taking 1.0 mg/day of orotate (the Life Extension formulation). It has been a part of my PM supplement stack for a few years. A dose in this range is concordant with the intersection of a variety of source of evidence – no combination of them being determinative. A long term dose of 1.0 mg/day confers a potential upside (if you are at risk and if the various snippets of research and conjectures turn out to be correct) with virtually zero downside risk . . . unless you turn out to be someone sensitive to the mineral.

3 Likes

Those of us taking telmisartan might want to modulate the lithium dose downward or at least monitor the relevant kidney metrics. Also, older people clear lithium more slowly. I’m going to move my current 1.0 mg/day down to 0.5 mg/day.

1 Like

1. Core Efficacy Data: None Met Pre-specified Statistical Endpoints

The study pre-specified 6 coprimary outcome measures (3 cognitive tests, 2 neuroimaging measures, and 1 blood biomarker). To control for false positives, the study set a strict statistical significance threshold of P < .01. The objective result is: none of the 6 primary outcomes met this pre-specified threshold.

  • Cognitive Function Measures:
    • CVLT-II (California Verbal Learning Test-II) : Scores declined by 1.42 points annually in the placebo group versus 0.73 points in the lithium group (a difference of 0.69 points; P = .05). Because it did not meet the pre-specified threshold of P < .01, in a strict statistical sense, this cannot be characterized as an effective treatment and is merely a nominal difference.
    • BVMT-R and PACC : For these two tests concerning visual memory and comprehensive cognition, no significant treatment × time interactions were observed between the two groups.
  • Structural Brain Imaging:
    • Over the two years, both the hippocampal volume and cortical gray matter volume of participants in both groups showed a decline (atrophy) over time.
    • There was no statistical difference in the rate of decline between the two treatment groups (hippocampus P = .09; cortical gray matter volume P = .78).
  • Biomarker (BDNF):
    • The trajectory of brain-derived neurotrophic factor (BDNF) showed no significant treatment × time interaction between the two groups (P = .93).

2. Safety and Tolerability Data: Significant Side Effects and High Dropout Rates

The data indicate that long-term use of low-dose lithium still presents tolerability thresholds and adherence challenges in the elderly MCI population.

  • High Discontinuation Rate : During the 2-year trial period, the overall study medication discontinuation rate was as high as 36% (34% in the lithium group and 38% in the placebo group). This suggests that maintaining long-term medication adherence is highly challenging for the elderly patient population, even with placebo or very low doses of lithium.
  • Doubled Typical Adverse Events : Although the incidence of serious adverse events (SAEs) was similar between the two groups (29% with lithium vs. 23% with placebo), the lithium group showed a distinct tendency towards intolerance for certain specific side effects. Specifically, the incidence of diarrhea was 29% in the lithium group (vs. 15% with placebo), tiredness was 29% (vs. 15% with placebo), and tremor was 24% (vs. 15% with placebo).
  • Dose Ceiling : Based on actual operational data, the authors acknowledged in the discussion that older adults have substantial difficulty tolerating doses greater than 300 mg daily.

3. Objective Limitations in Sample and Design: Weakened Study Power

From the raw baseline data, it is evident that this trial had significant flaws in its participant screening mechanism, which directly impacted the validity of the data:

  • Proportion of Alzheimer’s Disease (AD) Pathology was Too Low : Lithium was hypothesized to intervene against amyloid plaques and AD-related neurodegeneration. However, baseline data showed that only 27% (11/41) of the lithium group were amyloid-beta positive (Aβ+), and only 26% (10/39) in the placebo group. The vast majority of participants (66%-69%) were amyloid-negative (Aβ-). This means that the MCI in most enrolled participants was likely not driven by AD pathology, leaving the drug without its presumed “pathological target” in the majority of subjects.
  • Lack of Deterioration in the Control Group : The study data revealed that the placebo group actually did not show significant decline over the two-year period on the BVMT-R and PACC cognitive measures. If there is no significant decline during the natural progression of the condition (in the control group), the trial fundamentally loses its baseline for measuring whether the drug can “delay decline”.

Objective Summary

Stripping away the embellishing conclusion that “the trial established feasibility” , and looking purely at the data metrics and pre-specified statistical thresholds: this is a negative clinical trial. The existing data failed to prove that low-dose lithium can significantly delay cognitive decline in patients with mild cognitive impairment, nor did it prove that it can slow brain atrophy. The trial’s main data contribution is ruling out the feasibility of dosing regimens exceeding 300 mg daily in the elderly population , and pointing out that future similar studies must confirm participants have amyloid-positive (Aβ+) characteristics prior to enrollment

wasn’t aware of Lithium having these potential “major” side effects. I’ll be simply throwing it in garbage. why risk a lot for something that might (might being key word) help a little.

I find the generalization sound that this study failed to find impact under the study parameters but I do not think the study ruled much out. The problem in this instance lies not in the findings but in the study’s design. As a rule, I hesitate to be critical of design; when you look into the context of bad designs, you often find that the researchers were doing the best they could within the context presented to them. In this case, they might have improved a parameter or two.

Research designs need to fit their context. Somewhere in contemporary research training students were educated to value the universal application of the randomized full experimental design because of the precision of the inferences it can generate. (Thinking this way is bad science. Perhaps a post on this some day.)

Given the decades long course of development of AD, plus what is known about the lack of concordance between physical damage and cognitive attributes, it would be enormously expensive to construct a valid RC design that would capture enough variance to be useful. On the other hand, we have decently robust retrospective data, primarily emerging from high-quality national registries in Denmark and the UK, as well as ecological studies in the United States and Japan, that suggest a non-linear, protective association between trace lithium exposure and AD risk.

The 2017 Danish study is a statistically rigorous longitudinal case-control design covering over 800,000 individuals. The study found that long-term exposure to higher levels of naturally occurring lithium in drinking water was associated with a reduced rate of dementia. Individuals exposed to lithium levels above 15.0 mug/L (micrograms per liter) showed a 17% reduction in dementia incidence compared to those exposed to 2.0 mug/L.

The data also suggested a non-linear relationship. Moderate levels (5.1 mug/L) actually showed a slight increase in risk compared to the lowest levels, whereas the highest quartiles showed a significant decrease. This suggests a potential threshold effect where a minimum concentration is required to engage neuroprotective mechanisms like GSK-3β inhibition.

The theoretical logic is interesting as well but that is a different conversation.

Wow, your choice of words once again reflects the unwritten rule of this forum that I’ve pointed out many times: at moderate concentrations of 5.1–10.0 µg/L, the incidence rate ratio (IRR) for dementia is 1.22. That means a 22% increase in the risk of developing dementia compared to the reference group. Yet you describe that as “a slight increase.” At the highest concentration above 15.0 µg/L (i.e., 15.1–27.0 µg/L), the IRR is 0.83, indicating a 17% decrease in risk—and for that, you choose the phrase “a significant decrease.”

(It’s Sunday, and you know I have time to (love to) rant.) :stuck_out_tongue_closed_eyes:)

Frankly, I’m getting tired of reading responses that parrot a misinterpreted study that is often not meaningful, relevant, or just plain clickbait. Anecdotal reports of negative effects are just that, not a scientific fact. It is easy to see from the many negative reports that the particular supplement use reported in the forum can be attributed to coincidence or other confounding factors. Every time you read an anecdotal report, good or bad, including one from me, it is basically just a personal observation, not a scientific fact, not unlike telling you that a certain food tastes good or bad or that it gives you an upset stomach.

As someone who has researched the effects of low-dose lithium supplements at doses below 20 mg of elemental lithium for decades, regarding their life extension and mental health,

For God’s sake, get it through your heads, PhD, MD, or layman. Less than 20 mg of elemental lithium daily is safe except for a very small percentage of outliers. This is far below the therapeutic doses of lithium prescribed by doctors.

"It is most commonly prescribed as lithium carbonate

The therapeutic range for most indications is 0.4–0.8 mmol/L (12-h post-dose sampling), typically requiring a dose for adults between 400 and 1200 mg daily, individualised according to age, sex, weight, and, importantly, renal function."

Personally, I have taken lithium orotate at 5 to 20 mg of elemental lithium orotate daily for close to 30 years, with nothing but positive effects.

If you throw away your supplement or drug every time you read something negative about it from a forum member, you are not thinking for yourself or doing any of your own research.

Is there any primary evidence of harm below 20 mg of elemental lithium that supplementation causes harm in a statistically meaningful way?

Bottom line first:

No. Across:

A formal preclinical toxicology battery (Murbach 2021) — no harm

A systematic review of 16 interventional studies (Strawbridge 2023) — unanimous safety

The only controlled human dosing study at 5 mg/day (Smith 2024) — no AEs

Decades of post-market surveillance — no published safety signal

Population-level natural-exposure studies — protective associations, not harm

There is no primary evidence of statistically meaningful harm from daily lithium orotate or aspartate supplementation delivering under ~20 mg of elemental lithium. Every controlled study, every preclinical toxicology study, and decades of post-market surveillance have failed to produce a clear safety signal at this dose range. The one published case of “toxicity” at low-ish doses involved someone who deliberately took 18 capsules at once. The evidence base is thin, mostly short-term, and almost entirely uncontrolled. A recent self-selected user survey found subjectively reported side effects were “more prevalent than anticipated,” but without a control arm those numbers have no anchor.

Primary evidence on safety <20 mg elemental Li

1. Strawbridge et al. 2023 — Systematic review of low-dose lithium interventional studies

This is the most rigorous synthesis available. It pooled 16 interventional studies of low-dose lithium across neuropsychiatric outcomes. Of the 16 included studies, reports were unanimous on the safety of low-dose lithium across all studied populations PubMed Central, and the authors concluded that low-dose lithium is safe and may improve cognition as well as affective symptoms ScienceDirect. The accompanying narrative review noted there is a paucity of research into adverse effects of low-dose lithium, although one as-yet unpublished study found increased (though clearly non-significant) CKD rates in people with serum lithium levels of 0.3–0.59 PubMed Central — well above what supplements produce.

2. Murbach et al. 2021 — Formal toxicological evaluation of lithium orotate

The most direct safety study ever done on the compound. In a 28-day, repeated-dose oral toxicity study, rats were administered 0, 100, 200, or 400 mg/kg body weight/day of lithium orotate by gavage. No toxicity or target organs were identified; therefore, a no observed adverse effect level was determined as 400 mg/kg body weight/day. These results are supportive of the lack of a postmarket safety signal from several decades of human consumption. ScienceDirect The compound also was not mutagenic or clastogenic in bacterial reverse mutation and in vitro mammalian chromosomal aberration tests, and did not exhibit in vivo genotoxicity in a micronucleus test in mice ScienceDirect.

That 400 mg/kg/day NOAEL in rats translates to roughly 27,000 mg of lithium orotate daily for a 70 kg human — about 225× the typical 120 mg supplemental dose.

URL: https://www.sciencedirect.com/science/article/pii/S0273230021001136

3. Pacholko & Bekar 2022 — Direct head-to-head animal comparison

Re-ran the historically-cited Smith & Schou (1979) toxicity experiment that scared researchers away from orotate for 40 years. Concluded that lithium orotate has a safer kidney profile than lithium carbonate, and that both increased TSH only in females, but the increase was lower in the orotate group Wikipedia. The original 1979 study used what the original authors later acknowledged were toxic-range doses.

URL: https://pmc.ncbi.nlm.nih.gov/articles/PMC8413749/

3 Likes

This is not what most people would call a “low dose” study…

They used lithium carbonate at 150mg or 300mg. vs. 5mg Lithium rotate. They really are not in the same class of dosing. I think this study (the one at the top of the thread) is irrelevant for this forum.

Related reading: Lithium Supplementation

2 Likes

My best take is that this side-effect is definitely minor and then only in cases where clearance issues are at issue. The concerns emerged with psychiatric doses of lithium carbonate (several hundred mg/day). I posted here out of an abundance of caution and should probably have suggested that anyone with diminished kidney function would want to take a look at the research.

1 Like

Indeed. These can be meaningful. Many moons ago, as they say, I taught a very bright and specialized group of medical students how to read and interpret scientific research in their disciplines (at the time, one that was in its infant stages). They were good students but stumbled frequently in my classes. That was when I first saw how few people understand how to interpret, assess sources of error, and evaluate the generalizability of scientific research. Someone looking for a public figure who does it well would want to study Attia; he makes fewer mistakes than anyone I follow and, before Epstein, often gently corrected the errors committed by his frequently prestigious guests. As for the goo greasing your soapbox today, I would add to your list the importance of reading carefully – rereading if necessary – before deciding and reacting to what someone said.

=====
I should add that the leading edge of my own understanding of design and methodology is blunted by some of the clever adaptations that have developed since I was more active. Whenever I find that comprehending some of these innovations would require more effort than I have put in, I look to the interpretations of others.

1 Like

Yeah, I hear you but Lithium’s main benefit seems to be against dementia, and I am in a very low risk (never had anyone in my bloodline on either side with it) so I see little benefit with some risk. Not worth for me. My rule of thumb for supplements is if there are known and somewhat potentially severe side effect I just skip it all together. Not worth it, because the benefits we get from supplements are minor at best (regardless of what some people say) than I just scrap that supplement all together, why risk it at all. It is different with drugs because the benefits (at times) can be significant so it makes more sense to me to take them even if I knew there might be some side effect. As an example, if I take statins and lowered lipids, apoB etc from mid 100’s to say into 30’s or 40’s that is a HUGE benefit so even if I knew I might be prone to some side effect I still think it is worth taking it. . As for supplements is a totally different story so if my chance of getting dementia is 4% (it is about 10% I think in general but I’m probably way lower than that) and risking kidney injury and other sides to get to 3% chance it nuts in my opinion. So, it all comes down to risk vs reward. I already through my lithium in garbage this morning. Selegeline will do much better job at no known side effect (especially in my case)

You are right as far as I know. Moreover, the only robust evidence suggests benefit when lithium is elevated in drinking water, presumably over a lifetime or at least for a long period. The evidence based on studies of short term supplementation is weak.

Your comment causes me to examine my own motivations. I’m (e3,e3) and have no AD on either side of my family until you get to a second cousin. Taking the other side, there have been periods in my life where my diet was not great and my ApoB was elevated before I intervened to lower it. Additionally, I want to slow as much as possible the diminution in cognitive functions that everyone experiences as they age. Will lithium contribute to that slowing? Who knows. It seems like a low risk/low cost/low reward coefficient/high reward impact kind of equation.