Why EPA tested for it
EPA added lithium to its national testing (UCMR 5) to learn how common it is before deciding whether to regulate it. About a quarter of systems reported results above 10 µg/L. Of the cities we cover, the highest were Phoenix (160 µg/L), El Paso (133 µg/L), Los Angeles (100 µg/L), Albuquerque (71 µg/L).
- Strong evidenceFederal agency technical fact sheet
EPA’s 10 µg/L lithium screening level is not a legal limit; it was derived from side effects seen in patients taking lithium as medicine, divided by a large safety factor, assuming water supplies 20% of lithium intake.
Limits: EPA says it 'cannot confidently estimate the risk' between 10 µg/L and therapeutic-equivalent levels; based on a limited PPRTV assessment (2008); no NOAEL established.
US EPA Office of Water. (2023). Technical Fact Sheet: Lithium in Drinking Water
- Strong evidenceFederal nationwide monitoring dataset summary
In EPA’s nationwide sampling, about a quarter of public water systems reported lithium above EPA’s 10 µg/L screening level.
Limits: Single results above the level, not averages; results above 15 µg/L are counted as above the reference (rounding rule).
US EPA. (2026). The Fifth Unregulated Contaminant Monitoring Rule (UCMR 5) Data
- Strong evidenceNational groundwater survey
Natural lithium is common in US groundwater, highest in arid Western aquifers and lowest in places like south Florida.
Limits: Abstract read. Untreated groundwater, not finished tap water.
Mental health and other studies
Lithium is a psychiatric medicine at doses far above what’s in tap water, which is why researchers have asked whether natural lithium in water affects mood. Regional comparisons say maybe; the largest individual-level study says no. Studies on dementia and pregnancy are single national datasets with mixed or early findings.
What this doesn’t show: that lithium in water prevents suicide, or that levels above 10 µg/L are harmful.
- Weak evidenceSystematic review and meta-analysis of ecological studies
Studies comparing regions have found lower suicide rates where natural lithium in drinking water is higher, but these comparisons cannot show that lithium is the cause.
Limits: Abstract only (paywalled). Ecological studies are subject to ecological fallacy (area averages, not individuals); high heterogeneity. Authors call for randomized community trials.
- Weak evidenceSystematic review and meta-analysis
A second meta-analysis also found an association between more lithium in water and fewer suicides, but it did not account for confounding factors.
Limits: Abstract only (paywalled). Authors state the analysis used crude relationships 'without considering the role of confounders’. Mostly ecological.
- Weak evidenceSystematic review and meta-analysis
A third meta-analysis covering 113 million people found a modest association with lower suicide rates, along with strong heterogeneity and signs of publication bias, and said more research is needed before any public health recommendation.
Limits: Abstract only (paywalled). Mostly ecological; publication bias; heterogeneity.
- Mixed evidenceNationwide individual-level register cohort
A study following all 3.7 million Danish adults for 22 years found no link between lithium levels in their drinking water and suicide.
Limits: Abstract read (open access, PMC5486313). Narrow lithium range (under 31 µg/L); exposure estimated from waterworks data, not personal intake.
- Weak evidenceNationwide nested case-control study
A Danish nationwide study found a nonlinear association: the highest lithium exposure was linked to less dementia, but a middle range was linked to more.
Limits: Abstract read (PMC5710473). Authors note confounding by municipality cannot be excluded; inconsistent dose pattern.
- Weak evidenceNationwide population-based case-control study
A Danish study found that higher natural lithium in a mother’s drinking water during pregnancy was associated with a higher chance of autism in her child; the authors say this needs further study.
Limits: Abstract read (PMC10071398). Exposure estimated by interpolating 151 waterworks measurements to home addresses; observational; drew published commentary and letters.
How we gathered these
Every source was opened and read in November 2026, from EPA’s technical fact sheet and data summary, the U.S. Geological Survey (USGS), and PubMed records.
Questions
Is lithium in tap water harmful?
There’s no legal limit and no proof either way at the levels found in tap water. The U.S. Environmental Protection Agency’s (EPA’s) 10 micrograms per liter (µg/L) screening level is a cautious benchmark, and EPA says it "cannot confidently estimate" risk above it.
Does lithium in drinking water lower suicide rates?
Three meta-analyses of regional studies found lower suicide rates where natural lithium is higher, but these comparisons can’t show cause and the results vary widely. A study following all 3.7 million Danish adults for 22 years found no link. The researchers call for trials before any recommendation.