Arsenic in drinking water: what the evidence says about children and health risk

A headline says arsenic and most readers picture poison. That instinct is not wrong, but it skips the two questions that actually decide how worried to be: which form of arsenic, and at what dose. Get those wrong and the same word can describe a harmless trace in a piece of fish or a serious long-term hazard in well water.

Arsenic is a useful case for thinking clearly about an environmental exposure, in the same careful spirit as the work on autism and the environment, where measuring an exposure correctly is half the battle. It rounds out a look at the toxic metals that reach children mainly through what they ingest, alongside earlier explainers on mercury in fish and lead. This is an explainer, not medical guidance. The goal is to separate what the evidence firmly establishes from what it only suggests.

The question behind the question: which arsenic, and from where

Arsenic comes in two broad chemical families, and the distinction is the spine of everything that follows. Organic arsenic, the form bound to carbon-containing molecules, is common in seafood and is largely non-toxic to humans; the body absorbs little of it and clears most of it quickly. Inorganic arsenic, the form bound to elements like oxygen, sulfur, or chlorine rather than carbon, is the one that drives the health concern, and it is the form found in groundwater.

This is why a fish can contain measurable "arsenic" and still be safe to eat, while water with a much smaller number on a lab report can be the real problem. The chemistry, not just the quantity, determines the risk. When a claim about arsenic does not say which form it means, it has skipped the first thing a careful reader should ask.

Where drinking-water arsenic comes from

The mental image of arsenic is usually industrial: a factory, a spill, a deliberate poison. In drinking water, the more common story is geological. Inorganic arsenic is naturally present in many rocks and soils, and over long periods it leaches into the groundwater that aquifers and wells draw from. In many regions, the arsenic in tap water was never put there by anyone; it came from the ground the water moved through.

That natural origin has a practical consequence. Public water systems in the United States are regulated and routinely tested, but private wells are not covered by federal drinking-water standards, which places the responsibility for testing on the well owner. This is a factual point about how the rules are structured, not a recommendation about what any individual should do. It matters here only because it explains why two homes in the same area can face very different exposures depending on their water source.

Why the limit is 10 parts per billion

Regulators express the arsenic limit as a Maximum Contaminant Level, or MCL, the highest concentration of a contaminant legally allowed in public drinking water. For arsenic, that level is 10 parts per billion, written ppb, which is the same as 10 micrograms per liter, a microgram being one millionth of a gram. Ten ppb is a very small amount, which underscores how seriously the contaminant is treated.

The number has a history worth knowing plainly. The United States Environmental Protection Agency (EPA) lowered the arsenic standard from 50 ppb to 10 ppb; the rule was promulgated in January 2001, and public water systems had to comply by January 23, 2006, according to the EPA's technical fact sheet on the final rule. The World Health Organization (WHO) uses the same figure as a provisional guideline value of 10 micrograms per liter, as set out in its arsenic fact sheet.

There is a detail in the rule that often confuses readers. The EPA set a Maximum Contaminant Level Goal of zero for arsenic, while the enforceable limit is 10 ppb. These are not contradictory. The goal of zero reflects a scientific judgment that for a carcinogen, no level of exposure is assumed to be entirely risk-free. The enforceable limit reflects what is achievable and measurable in practice. The gap between the two is a window into how risk-based regulation actually works.

What the cancer evidence firmly establishes

Here the evidence is strong, and the language can be strong with it. Long-term ingestion of inorganic arsenic is an established cause of cancer in humans, specifically of the bladder, lung, and skin. This is why both the EPA and the WHO classify arsenic as a known human carcinogen rather than a suspected one. The WHO arsenic fact sheet lays out this classification and the cancers most clearly linked to chronic exposure.

It is worth pausing on the word "established," because it is used carefully in this field. Much of environmental epidemiology deals in associations that fall short of proven cause. Arsenic and these cancers is one of the cases where the body of evidence, built over decades and across populations with naturally high exposures, supports causal language. When the science earns that word, it should use it. The point of distinguishing strong from weak evidence is not to downplay either, but to match the claim to what the data can bear.

What the child-development evidence can and cannot show

The neurodevelopment question is genuinely different, and honesty requires treating it differently. A 2025 systematic review of 24 studies reported a consistent inverse association between arsenic exposure and cognitive performance in children, with higher exposure linked to lower IQ scores. That is a meaningful and concerning signal, and it points in one direction across most of the included studies.

It is also, by its nature, weaker evidence than the cancer finding, and the review itself notes the limits. A minority of the studies found no effect. More importantly, these are observational studies, meaning researchers measured exposure and outcomes in real populations rather than assigning exposure in a controlled trial, which would be neither possible nor ethical. Observational designs struggle to fully separate one exposure from everything that travels with it. Communities with higher arsenic in their water often differ in income, nutrition, and other co-exposures, and any of those can influence a child's cognitive development.

For that reason, the responsible phrasing is that arsenic exposure is "associated with" lower cognitive scores in children, not that it "causes" them, even though the cancer evidence supports the stronger word. Researchers watch prenatal and early-childhood exposure most closely, because those are the windows when the developing brain is thought to be most sensitive. The science here is active and suggestive, not closed.

Reading arsenic claims with a clear head

The useful habit is not to memorize numbers but to ask a few questions of any arsenic claim you encounter. Is it talking about inorganic arsenic, the groundwater form that drives risk, or organic arsenic from seafood, which is largely benign? At what dose, and over what length of exposure? And does the language say "associated with," which signals an observational finding still subject to confounding, or "causes," which should be reserved for established effects like the cancer link?

It also helps to know how exposure is actually measured. Biomarkers such as urinary arsenic can indicate recent exposure, but they have limits: they reflect a short recent window rather than a lifetime, and without care they can be inflated by the harmless organic arsenic from a recent seafood meal. This is the same theme that runs through every careful exposure study, that measuring the thing well is as important as the thing itself.

This article is educational and is not a substitute for personal medical advice. Its aim is narrower and, in a noisy information environment, more useful: to show how the evidence on arsenic is built, where it is firm, and where it is still being worked out. For readers who want to go to the primary literature, the peer-reviewed work on environmental exposures and child health is the place to continue.