Cadmium exposure in children: the sources, and what the evidence shows

Lead, mercury, arsenic, and manganese are the metals most families have heard of. Cadmium usually is not, even though for children in the United States it arrives by the most ordinary route there is: the food on the table.

That makes cadmium a good case for a habit worth practicing on any contaminant. What is established, what is associated, and what is still being worked out are three different categories of knowledge, and most confusion about environmental health comes from collapsing them into one. This explainer keeps them apart.

The same care applies here as in the work on autism and the environment, where how well an exposure is measured usually decides how much a study can actually claim.

Where cadmium in a child's diet comes from

Cadmium is a naturally occurring metal that plants take up from soil. According to the U.S. Food and Drug Administration (FDA), cadmium levels in soil are higher "in areas where some phosphate fertilizers are used, and where industrial processes such as smelting, mining, and burning of fossil fuels occur," as described on its page on cadmium in food and foodwares. Because the metal enters through the roots, it ends up distributed across ordinary crops rather than concentrated in a single suspect food.

An FDA analysis of Total Diet Study data from 2018 to 2020 estimated where that exposure actually comes from by age. For non-breastfed infants, the processed baby food and infant formula group was the largest contributor. For young children, grains and baking plus vegetables were the top cadmium sources (Hoffman-Pennesi et al., Food Additives and Contaminants Part A, 2024).

Chocolate comes up often in coverage of cadmium, and it is worth putting in proportion. The FDA cites a 2021 World Health Organization assessment finding that the contribution of cocoa products to dietary cadmium exposure was minor, between 0.1 and 9.4 percent, even in countries with high cocoa consumption.

For adults, the picture is different. The Agency for Toxic Substances and Disease Registry (ATSDR) lists smoking cigarettes or breathing cigarette smoke alongside food and contaminated water as main sources for the general population, in its ToxFAQs for cadmium. Tobacco is a major adult route that does not apply to most children directly, which is one reason childhood exposure is discussed as a dietary question.

How is cadmium exposure measured?

Two biomarkers do most of the work, and they answer different questions. A biomarker is a measurable signal in the body that stands in for an exposure that cannot be observed directly.

• Blood cadmium reflects recent exposure.
• Urinary cadmium reflects both recent and past exposure, which makes it the usual choice when a study wants an indicator of accumulated body burden.

ATSDR states the distinction plainly: the amount in blood shows recent exposure, while the amount in urine shows both recent and past exposure. That difference matters when reading any study, because a single measurement of the wrong biomarker can badly misrepresent a child's actual history with a metal that has a long biological half life.

This is the same measurement problem covered in more depth in the explainer on how scientists measure heavy metal exposure in children.

What cadmium does in the body: the established part

The best established effects of long-term cadmium exposure are not in the brain. They are in the kidney and the skeleton.

ATSDR describes long-term exposure at lower levels as leading to a buildup of cadmium in the kidneys and possible kidney disease, and lists fragile bones among the long-term effects. It also notes that in animal studies the young are more susceptible than adults to loss of bone and decreased bone strength. The U.S. Department of Health and Human Services has determined that cadmium and cadmium compounds are known human carcinogens.

Two honest caveats belong with that paragraph. Much of this evidence comes from adult and occupational exposure, at levels well above what a child encounters in a normal diet. And moving from that evidence to a statement about ordinary dietary exposure in children is an inference, not an observation. Reference values exist precisely to make that inference in a structured, conservative way.

What the evidence says about the developing brain

This is the part where proportion matters most, because it is the part most likely to be overstated.

A cross-sectional analysis of NHANES data from 1999 to 2004, covering children aged 6 to 15, compared those in the highest urinary cadmium quartile with those in the lowest. It reported higher odds of a learning disability, with an odds ratio of 3.21 (95 percent confidence interval 1.43 to 7.17), and of special education use, odds ratio 3.00 (95 percent confidence interval 1.12 to 8.01). It found no association with attention deficit hyperactivity disorder, odds ratio 0.67 (95 percent confidence interval 0.28 to 1.61) (Ciesielski et al., Environmental Health Perspectives, 2012).

Those odds ratios look large, and the study design is the reason to hold them loosely. Cross-sectional means exposure and outcome were measured at the same moment, so the design cannot establish which came first. A single urine measurement taken years after the developmentally relevant window is a rough proxy for the exposure that would actually matter.

A 2025 systematic review and meta-analysis of 17 observational studies looked at prenatal exposure specifically. It found that each 50 percent increase in cadmium levels in body fluids was associated with a 0.44 point lower full scale IQ in children aged 5 to 9 (95 percent confidence interval negative 0.67 to negative 0.21). The authors are explicit about a limitation: each component of the analysis included fewer than ten studies, so publication bias was not assessed (Su et al., Medicina, 2025).

Read that effect size carefully. Less than half an IQ point per 50 percent increase in exposure is small for any individual child and would be invisible in a clinic. Across a whole population it is not nothing, because shifting the average of a distribution moves the number of children at both tails. Both statements are true at once, and neither is a claim about causation. These are observational findings, and confounding by co-exposures, nutrition, and socioeconomic circumstances is genuinely difficult to remove.

How to read a reference value without misreading it

The FDA has adopted a toxicological reference value for cadmium of 0.21 to 0.36 micrograms per kilogram of body weight per day. The same FDA Total Diet Study analysis estimated mean dietary cadmium exposures for non-breastfed infants aged 0 to 11 months and children aged 1 to 6 years at 0.18 to 0.47 micrograms per kilogram of body weight per day, depending on the age group and on how non-detected samples were handled.

Those ranges overlap, and the upper end of the exposure estimate sits above the reference range. That comparison is worth understanding rather than reacting to.

A toxicological reference value is an estimate of daily exposure over a lifetime that is unlikely to cause appreciable harm, built with deliberate safety factors. It is not a line separating harm from no harm, and a single day above it is not an event. When a population mean approaches or exceeds such a value, the correct reading is a statement about the food supply, and a signal to monitor and reduce exposure at the source. It is not a diagnosis of any individual child.

What is changing

Two developments are worth watching.

The FDA has listed action levels for cadmium and inorganic arsenic in baby and toddler foods among its 2026 priority deliverables under the Closer to Zero initiative, stating that it "will advance this work in 2026 through activities to establish action levels for cadmium and inorganic arsenic in baby and toddler foods." That is a stated priority for the year, not a rule already in force, and the distinction matters when reading news coverage of it.

The second is upstream of any single country's food supply. An analysis of 796,084 soil samples drawn from 1,493 regional studies, published in Science in April 2025, estimated that 14 to 17 percent of global cropland is contaminated by at least one toxic metal, with cadmium the most widespread, and put the number of people living in high-risk areas at 0.9 to 1.4 billion (announcement of the study). Cadmium in a child's cereal is, in that sense, a soil question before it is a grocery question.

What the evidence can and cannot say

The kidney and bone effects of sustained cadmium exposure are established, mostly from adults. The neurodevelopmental findings are associations, modest in size, drawn from designs that cannot settle causation. Dietary exposure in young children is close enough to the reference value that regulators are acting on monitoring data rather than waiting for stronger evidence, which is a reasonable response to a metal that accumulates slowly and leaves the body slowly.

What would sharpen the picture is better exposure measurement in the window that matters, prospective cohorts followed long enough to observe outcomes rather than infer them, and analyses that can separate cadmium from the other exposures that travel with it. That is unglamorous work, and it is the work that turns an association into something a policy can rest on.

Readers who want to see how these methods are applied to environmental exposures and child neurodevelopment can explore the peer-reviewed publications behind this work.