Lithium (Low-Dose)

Verdict: Suggestive (population-level, observational) / Mostly hype (specific microdosing claims) Last reviewed: 2026-04-25 Triangulated against anchor: NMN (Suggestive)

TL;DR

Drinking water lithium concentrations associate with reduced suicide rates, dementia incidence, and possibly all-cause mortality in observational data. The mechanism (likely GSK-3β inhibition) has cellular plausibility for aging effects. C. elegans lifespan extension by lithium is replicated. Specific microdosing supplementation claims are weakly supported; therapeutic-dose lithium has substantial side effect / monitoring requirements. Verdict: Suggestive for the observational signal; Mostly hype for specific microdosing protocols beyond what dietary water-source variation already provides.

What it is

An alkali metal; lithium carbonate / orotate at clinical doses (600-1800 mg/day) is first-line for bipolar disorder. "Microdosing" supplements typically deliver 1-5 mg elemental lithium (orotate form is popular but pharmacologically not clearly superior). Drinking water concentrations vary widely; some Texas/Japan regions have notably high naturally-occurring water lithium.

Proposed mechanism

• GSK-3β inhibition — relevant to neurodegeneration, autophagy, mood
• Inositol monophosphatase inhibition — the bipolar mechanism
• Effects on circadian gene expression
• Possible effects on tau phosphorylation (Alzheimer's relevance)
• Neuroprotective signaling (BDNF upregulation)

Confidence: Established for the bipolar-relevant mechanisms at therapeutic doses; Plausible for aging-relevant mechanisms at low doses.

Evidence ladder

Invertebrate (T5)

C. elegans lifespan extension by lithium replicated across multiple labs. Effect modest but consistent.

Mouse / rat (T4)

• Healthspan / behavioral / neurodegenerative model studies.
• ITP not tested.
• Lifespan effects single-lab.

Human

Observational (T2):

• Drinking water lithium and suicide rates — multiple ecological studies (Japan, Greece, Texas, Austria) associate higher water lithium with lower suicide rates. Effect sizes meaningful in pooled analyses.
• Dementia incidence — observational links between water lithium and reduced dementia incidence (Kessing 2017 JAMA Psychiatry, similar studies).
• All-cause mortality — weaker but suggestive observational signals.

Interventional (T2):

• Therapeutic doses for bipolar — clear efficacy; not relevant to "microdosing" claims.
• Low-dose lithium for cognitive decline — small RCTs (Forlenza 2011 in MCI, Forlenza 2019) suggest modest cognitive benefit. Replication and scale are limited.
• No aging-endpoint mortality RCT.

Confounds

• Ecological observational data — confounded by everything that varies geographically (climate, diet, demographics, healthcare access).
• Therapeutic vs trace dose — extrapolation from drinking-water trace exposure to supplement dosing is non-trivial.
• Side effects scale with dose — therapeutic lithium has serious renal, thyroid, parathyroid concerns; low doses minimize but don't eliminate.
• Lithium orotate vs carbonate — orotate marketing claims of better bioavailability are not robustly supported.

Conflict of interest scan

• Therapeutic lithium is generic; little industry incentive.
• Microdosing supplement market is small; claims are not rigorously trial-tested.

Human translation

Honest decomposition:

1. For bipolar disorder: lithium at therapeutic dose is foundational treatment with extensive evidence. Strong (in indication).
2. For dementia / suicide prevention at population level: drinking water lithium signals are real but observational; no large RCT.
3. For "longevity microdosing" via supplements: weak; the dose response between drinking-water-trace and supplement-mg is poorly mapped, and outcomes data is essentially absent.
4. For MCI / early cognitive decline: small RCT signals justify a Suggestive verdict in this narrow indication.

Calibrated verdict

Suggestive for the population/observational signal and the MCI signal. Mostly hype for the broader "lithium microdosing for general longevity" framing.

Compared to NMN (Suggestive), lithium has better observational human data on hard endpoints (suicide, possibly dementia) but weaker mouse lifespan / mechanism translation.

Compared to resveratrol (Mostly hype), lithium has cleaner human observational signal and is "above" resveratrol; not yet at NMN-level evidence quality for general use.

Confidence interval on verdict

• Could move to Probable if a large MCI / dementia-prevention RCT reads positive; some smaller trials in progress.
• Stable for the broader longevity claim absent outcomes evidence.

Open questions

• Q: At what dose does low-dose lithium engage GSK-3β meaningfully in human brain — and is the supplement dose above or below that threshold?
• Q: Will the dementia-prevention signal from observational data hold up in pre-registered RCTs?
• Q: Is there a chronic-low-dose safety profile that's well-characterized at modern microdosing levels (1-5 mg/day)?

Sources

• Kessing et al. 2017, JAMA Psychiatry — Lithium in drinking water and dementia incidence
• Forlenza et al. 2011 — Disease-modifying properties of long-term lithium treatment for amnestic MCI
• Drinking water lithium and suicide rate ecological studies (multiple)

Produced under methodology locked 2026-04-24. Triangulated against NMN anchor.