Why Did Lithium in Tap Water Become an Autism Story?

In 2023, researchers published a large Danish study examining whether naturally occurring lithium in drinking water during pregnancy was associated with autism spectrum disorder in children. The study attracted attention because lithium is biologically active, can cross the placenta, and is used in much higher doses as a prescription mood stabilizer.

The researchers identified 8,842 children diagnosed with autism and compared them with 43,864 children without a recorded diagnosis. They estimated lithium concentrations in the water serving each mother’s residential area during pregnancy. Higher estimated lithium levels were associated with higher odds that the child would later receive an autism diagnosis.

The result was statistically significant, and the association appeared to become stronger across higher exposure categories. Predictably, the internet translated this into something resembling, “Tap water causes autism.” The actual scientific conclusion was much narrower: estimated exposure was associated with diagnosis, and the possibility deserved further investigation.

That wording difference is not decorative academic frosting. It is the difference between finding a clue and solving the mystery.

Association Does Not Prove Causation

An observational study can reveal that two things occur together. It cannot automatically show that one produced the other. Umbrella sales and wet sidewalks rise on the same days, but umbrellas are not secretly operating a sidewalk-moistening conspiracy. Rain explains both.

The same logic applies to environmental health research. Areas with different lithium concentrations may also differ in population density, geology, income, diet, health care access, diagnostic practices, air pollution, parental age, water sources, and other factors. Researchers can statistically adjust for known differences, but no model can guarantee that every relevant variable has been measured perfectly.

What the Danish study did well

The study had genuine strengths. Denmark maintains unusually detailed national health and residential records, allowing researchers to examine a large population and track addresses during pregnancy. Autism diagnoses came from a national psychiatric registry rather than social-media surveys or parental recollection. The investigators also adjusted their models for several socioeconomic characteristics, smoking during pregnancy, urbanization, maternal age, birth year, sex, and air pollution.

Those features make the findings worthy of attention. They do not, however, transform an observational association into proof that lithium caused neurodevelopmental changes.

What the study could not establish

The researchers did not measure lithium in each participant’s blood, urine, food, or personally consumed water. Exposure was assigned primarily from residential location and modeled water concentrations. That approach assumes that people living in the same water-service area receive similar exposure, which may not be true.

One pregnant person may drink two liters of unfiltered tap water daily. Another may drink bottled water, use a workplace water supply, rely on a private well, install a specialized filter, or survive mainly on iced tea and determination. Their residential estimates could look similar even when their actual lithium intake is not.

The water measurements were also collected during limited periods and geographically modeled across communities. Although naturally occurring lithium tends to be relatively stable, estimated concentration at a waterworks is not identical to a verified dose inside an individual human body.

The Difference Between Trace Lithium and Prescription Lithium

The word “lithium” can sound alarming because people associate it with batteries or psychiatric medication. Chemically, lithium is a naturally occurring element found in rocks, groundwater, some foods, and subsurface brines. Its presence in water does not mean someone dropped a laptop battery into the reservoir.

The Danish study examined water concentrations ranging from approximately 0.6 to 30.7 micrograms per liter. A microgram is one-thousandth of a milligram. Prescription lithium compounds, by contrast, are administered in carefully managed doses measured in hundreds of milligrams, with blood tests used to monitor treatment.

The comparison is not perfectly direct because pharmaceutical labels may refer to the mass of a lithium compound rather than elemental lithium. Nevertheless, exposure from typical drinking water is orders of magnitude below therapeutic medication exposure.

Dose matters in toxicology. Oxygen keeps us alive, yet too much pressurized oxygen can cause injury. Water is essential, yet drinking a ridiculous amount in a short period can be dangerous. Finding that a substance has biological effects at a medical dose does not prove that trace environmental exposure produces the same effects.

Prescription lithium does require medical supervision, especially during pregnancy. That issue should not be confused with the unproven claim that ordinary tap-water exposure causes autism. People taking prescribed lithium should never stop it suddenly because of a drinking-water article. Untreated bipolar disorder can carry serious risks, and pregnancy treatment decisions require an individualized discussion of benefits and potential harms.

Why Autism Research Requires Extra Care

Autism is a neurodevelopmental condition with substantial biological and behavioral diversity. There is no single autism cause. Research indicates that genetic factors play a major role, while biological conditions, pregnancy-related factors, environmental exposures, and interactions among these influences may alter probability in some people.

That complexity makes simple explanations emotionally appealing and scientifically suspicious. A headline naming one chemical, food, medicine, or household product offers the comforting illusion that a complicated condition has been neatly solved. Unfortunately, human development did not consult the marketing department.

Autism diagnosis is also influenced by screening, clinical awareness, access to specialists, diagnostic criteria, educational services, and cultural attitudes. A region with more diagnostic resources may record more cases even if the underlying biological prevalence is similar.

Experts reviewing the Danish research specifically questioned whether geographic differences in diagnostic access could have affected the association. The investigators attempted to adjust for urban and rural residence, but broad categories may not capture every difference between neighborhoods or health systems.

Other Reasons Scientists Remain Skeptical

The study needs independent replication

A first study is a starting point. Strong causal conclusions generally require findings to be reproduced in different populations, with different water systems, exposure measurements, research teams, and analytical methods.

Ideally, future studies would measure lithium in individual biological samples, document drinking-water behavior, account for food and bottled-water sources, and track pregnancy exposure prospectively. Researchers would also examine whether the association persists in regions where lithium levels do not overlap so closely with urbanization and diagnostic access.

Unmeasured confounding remains possible

Several critiques pointed to factors that were missing or difficult to model, including paternal age, gestational characteristics, birth weight, dietary intake, and individual water consumption. The study authors responded that some variables could be mediators rather than confounders and that including them might introduce a different kind of statistical bias.

This debate is normal. It does not mean the study was worthless or fraudulent. It means epidemiology is hard, particularly when exposure cannot be randomly assigned. No ethical committee is going to approve a trial in which pregnant volunteers are deliberately given different concentrations of a suspected contaminant to see what happens.

A diagnosis is not the same as an exposure effect

The study’s outcome was a recorded autism diagnosis. A diagnostic registry is valuable, but it cannot detect autistic people who were never evaluated or whose traits were recognized later. Geographic differences in recognition and referral could therefore influence the numbers.

Researchers must distinguish between factors that change the probability of being autistic and factors that change the probability of receiving a diagnosis. Those are related questions, but they are not interchangeable.

What the EPA Says About Lithium in Drinking Water

The U.S. Environmental Protection Agency included lithium in its Fifth Unregulated Contaminant Monitoring Rule. From 2023 through 2025, participating public water systems monitored lithium along with other unregulated contaminants so the agency could better understand where it occurs and at what concentrations.

“Unregulated” does not mean the EPA has declared lithium harmless. It means there is currently no enforceable federal maximum contaminant level for it. The agency has also not established a formal drinking-water health advisory for lithium.

The EPA uses a screening health reference level of 10 micrograms per liter to help evaluate occurrence data. That number is not a regulatory safety boundary and was derived using information from adverse effects observed at therapeutic exposure levels, combined with conservative assumptions and uncertainty factors.

The agency acknowledges that evidence about long-term health effects at ordinary drinking-water concentrations is limited. It continues to collect data and evaluate whether further studies or regulation may be needed. That is a responsible position: monitor the substance, investigate uncertainty, and avoid pretending that missing evidence proves either perfect safety or certain danger.

Can boiling remove lithium?

No. Lithium is a dissolved element, so boiling, disinfecting, or leaving water dramatically on the counter will not remove it. Boiling can actually concentrate dissolved minerals slightly as water evaporates.

Some ion-exchange and specialized adsorption systems may reduce lithium, while common pitcher filters are not necessarily designed or certified for that purpose. Bottled water is not automatically lithium-free, and the Food and Drug Administration has not established a specific bottled-water lithium standard.

Should Pregnant People Stop Drinking Tap Water?

Current evidence does not support a general recommendation that pregnant people abandon tap water because of autism concerns. Dehydration is not a clever public-health strategy, and switching blindly to bottled water may introduce expense, plastic waste, and uncertainty about the water’s mineral content.

People with a specific concern can review their water utility’s annual Consumer Confidence Report and any available unregulated-contaminant results. Private-well users may need to contact a certified laboratory or local health department because private wells are not monitored like public systems.

A clinician or environmental-health specialist can help interpret a confirmed result, especially during pregnancy or when prescription lithium is involved. Decisions should be based on measured local conditions and personal medical circumstances, not a viral post featuring a red warning emoji and seventeen exclamation points.

It also helps to keep relative and absolute risk separate. The Danish study reported higher odds across exposure categories. A percentage increase in relative odds does not mean the same percentage of all exposed children will develop autism. Individual probability depends on the baseline rate and many other factors.

What Would Convincing Evidence Look Like?

To determine whether trace lithium contributes to autism likelihood, researchers would need a broader body of consistent evidence. Useful studies could include:

• Prospective pregnancy cohorts with repeated blood or urine lithium measurements
• Detailed records of tap, filtered, bottled, workplace, and private-well water consumption
• Measurement of lithium from food and supplements
• Replication across several countries and geological regions
• Careful adjustment for diagnostic access, parental characteristics, and other environmental exposures
• Evidence of a credible dose-response pattern using individually measured exposure
• Laboratory research demonstrating effects at concentrations comparable to real-world drinking-water exposure

If these methods repeatedly pointed in the same direction, confidence in a causal relationship would increase. If better exposure measurements weakened or eliminated the association, the original finding would more likely reflect confounding or classification error.

Real-World Experiences: Responding Without Panicking

The following are composite scenarios based on common questions and practical situations. They are not descriptions of specific patients.

The expectant parent who sees a frightening headline

Imagine an expectant parent scrolling through the news while waiting for a prenatal appointment. A headline announces that lithium in tap water is “linked to autism.” Within minutes, the kitchen faucet looks less like plumbing and more like a tiny chrome supervillain.

A calmer review changes the picture. The report concerns one observational study in Denmark. It estimated exposure from residential water systems rather than measuring the individual’s actual dose. It found an association with later diagnosis but could not prove causation. The practical next step is not to stop drinking water. It is to check the local water report, write down questions, and bring them to the prenatal clinician.

That experience illustrates a useful rule: the scarier the headline, the more important it is to read the study design.

The family using a private well

A family in the western United States may reasonably pay closer attention because lithium concentrations can be higher in some arid, groundwater-dependent regions. Their private well does not receive the same routine oversight as a community water system.

Instead of purchasing the first filter advertised by an influencer wearing a lab coat, the family contacts a state-certified laboratory. Testing reveals what is actually in the water. If lithium is elevated, a water-treatment professional can recommend an appropriate technology and verify performance with follow-up testing.

The value of this experience is measurement. Fear says, “Something might be present.” Good environmental health practice asks, “What was measured, by whom, using which method, and at what concentration?”

The patient taking prescription lithium

Another reader takes lithium for bipolar disorder and becomes worried after encountering the tap-water story. The person considers skipping medication before speaking with a psychiatrist.

That would confuse two very different exposure questions and could create immediate medical risk. Prescription lithium has known benefits, known side effects, and established monitoring procedures. Pregnancy may change the treatment plan, but stopping suddenly can increase the risk of mood relapse.

The safer experience begins with a call to the prescribing clinician. The patient discusses pregnancy plans, medication dose, blood monitoring, hydration, kidney and thyroid testing, and the benefits and risks of available options. A population-level water study cannot replace that personalized conversation.

The parent searching for someone to blame

Parents of autistic children are often targeted by claims implying that one drink, medicine, parenting decision, or household exposure caused their child’s neurodevelopment. These messages can produce guilt without producing useful knowledge.

The current science does not justify blaming a parent for drinking tap water during pregnancy. Autism arises through complex developmental pathways, and a single observational association cannot reconstruct one child’s personal cause. Families are better served by supportive health care, educational resources, accommodations, and respect for autistic people than by an endless hunt for a household culprit.

These experiences lead to the same conclusion: lithium in drinking water deserves continued monitoring and better research, but the available evidence does not demonstrate that ordinary tap-water exposure causes autism. Curiosity is appropriate. Testing may be appropriate in specific locations. Panic, guilt, and unilateral medication changes are not.

Conclusion

The Danish study raised a scientifically legitimate question, but it did not answer that question conclusively. Higher estimated lithium concentrations were associated with higher odds of an autism diagnosis, yet actual individual exposure was not measured, residual confounding could not be eliminated, and diagnostic geography may have influenced the results.

Based on the available evidence, saying that lithium in tap water causes autism goes far beyond what the research supports. A more accurate conclusion is that trace lithium is an emerging environmental-health topic worth studying carefully. Until stronger, independently replicated evidence appears, your faucet has not been convicted.