In this post from a promising-but-niche substack I came across, author Jon Brudvig lays out a compelling case for the high expected-value of taking low-dose lithium supplements for Alzheimer's' prevention.  Recent studies of many different kinds -- mouse models, observational studies of areas with high lithium in drinking water, studies of psychiatric patients, and recent preclinical work identifying plausible mechanisms^[1] -- all suggest that dramatic reductions in Alzheimer's rates (20% - 50%) could be achieved with lithium supplementation of around 0.02 - 0.3 mg / day. These doses are hundreds to thousands of times lower than common psychiatric doses, putting them well below any known risk threshold.

For decades, evidence has been building that extremely low doses of this trace element can prevent the development of Alzheimer’s disease. The first hints came from observational studies in psychiatry, where patients treated with clinical (high) doses of lithium exhibited reduced rates of dementia. Next came interventional studies in Alzheimer’s patients, where the drug improved cognitive performance and slowed cognitive decline. Around the same time, epidemiological studies demonstrated that healthy individuals exposed to low levels of naturally occurring lithium in their water supply exhibit decreased rates of Alzheimer’s.

Most recently, we’ve seen several strong preclinical studies that have uncovered the pathways through which lithium exerts its effects. Apparently, there are several relevant mechanisms. Lithium inhibits an enzyme called glycogen synthase kinase-3β (GSK-3β), which drives tau hyperphosphorylation and other brain changes in Alzheimer’s disease. It also shifts amyloid-β processing away from plaque-forming pathways while promoting clearance of existing deposits. Together, these activities prevent the early brain changes that set the disease cascade into motion.

Beyond the huge-if-true object-level benefits of lithium for Alzheimer's prevention, Brudvig's post strikes me as a great example of reasoning under uncertainty in terms of expected value -- he critiques the "wait for RCT data in humans" stance espoused by others, like Eric Topol, who have also remarked on lithium's promise.

[After citing Topol's commentary:] Positions like this are typical of many academic scientists, who rightly value rigorous, controlled studies, but who also inadvertently promote an unfortunate bias towards inaction, even when the best read of the current evidence is that an intervention would be smart to implement. In high-stakes scenarios like this one, where a safe, low-cost measure could prevent an irreversible disease, waiting for confirmatory evidence could mean missing the window for prevention. Acting now is wise.

And with regard to dose, yes, human dose-response studies will be important to determine the optimal dose for Alzheimer’s treatment and prevention. But we don’t need to wait for those to act. The human drinking water studies and early interventional trials already provide us with an effective dosing window, and we can always adjust that dose if and when definitive studies dial in on a different level.

In the full post, Brudvig suggests practical supplementation via liquid drops (~0.05 mg/drop) or by taking a 1 mg pill every 2-3 days.

If you receive benefits similar to those in people who receive the same dose in their drinking water (a very likely outcome), you’ll reduce your risk of developing Alzheimer’s disease by as much as 22%! Combined with protective lifestyle factors like exercise, a healthy diet, and a robust social and mental life, you could be buying yourself a whole bonus chapter of healthy life.

To add some of my own speculation -- Claude and I tried to do some back-of-the-envelope calculations quantifying the benefit for a person like me (in my 30s, with no special risk factors like APOE-4 genes), which I've copy-pasted here.  We figure that lithium's Alzheimer's-preventative effect, if real^[2], probably scales in tune with Alzheimer's exponential increase in prevalence as people age.  So, taking low-dose lithium for a decade starting in your thirties would provide a real but relatively marginal benefit (10 years of pill-taking --> perhaps equivalent to 1-2 weeks of added life expectancy).  But the benefit (again, assuming it's real) might be several times stronger later in life --months instead of weeks.  In addition to considering supplementation yourself, consider forwarding Brudvig's well-written article to some of the older people in your life!

1. ^{^}

   Namely, the idea that lithium inhibits GSK-3β (which drives tau hyperphosphorylation) and shifts amyloid-β processing away from plaque formation.

2. ^{^}

   And if lithium's protective effect indeed comes from the causal pathways that early studies are focusing on.  As Claude says, "Lithium's proposed mechanisms are more about modulating ongoing enzymatic activity and clearance rates. If lithium at 0.05 mg/day can inhibit GSK-3β and promote amyloid clearance when you're 55 and those processes actually matter, starting at 30 probably doesn't give you much additional benefit — you're inhibiting an enzyme that isn't yet doing much harm, and clearing amyloid deposits that don't yet exist in meaningful quantities."