Karen Harris

Karen Harris is a professional writer and editor with more than two decades of experience in research-driven nonfiction writing. With a background in journalism and a Master's degree in English, she brings a meticulous approach to fact-finding, source evaluation, and clear communication of complex topics. Her work spans health, lifestyle, and consumer education, always with a focus on accuracy and readability.

Context

Human milk and infant formula are the two main foods for infants in the United States, so any environmental contaminants they carry matter for early development. Metals such as arsenic, cadmium, lead, and mercury, and the large family of per- and polyfluoroalkyl substances (PFAS, often called “forever chemicals”), can enter the food supply through air, water, soil, ingredients, or packaging. Yet the United States has no dedicated program for monitoring these contaminants in human milk.


This systematic review, part of a set commissioned by the US Food and Drug Administration, set out to describe measured concentrations of arsenic, cadmium, lead, mercury, and selected PFAS in human milk and infant formula in the United States, and to map the evidence available from other countries that could fill the gaps.

Study Overview

Design: Systematic review following PRISMA and meeting AMSTAR 2 criteria, registered a priori on PROSPERO (CRD42024528756) as one of six reviews commissioned by the FDA's Human Foods Program. Five databases (MEDLINE, Embase, CAB Abstracts, CINAHL, and CENTRAL) were searched through 2 April 2025 with no date restrictions. Risk of bias was assessed with the RoB 2, ROBINS-I, and ROBINS-E tools for human milk and a purpose-built tool for infant formula, and the certainty of evidence was rated with GRADE. A meta-analysis was planned but judged inappropriate, so the synthesis was narrative.


Included studies: From the United States, 14 human milk studies (reported in 15 articles) and 16 infant formula studies, drawn from 6255 unique records screened. The review additionally identified 317 human milk and 108 infant formula articles from other countries, presented as heat maps rather than pooled.


Comparators: Not an intervention comparison. Contaminant concentrations were synthesized narratively by matrix (human milk or infant formula) and by contaminant, and where possible compared against regulatory reference values.


Outcomes assessed: Measured concentrations of arsenic (total and inorganic), cadmium, lead, mercury (total and methylmercury), and the PFAS compounds PFOA, PFOS, PFNA, and PFHxS.


Funding: Funded by the US Department of Health and Human Services, Food and Drug Administration (contract #75F40123P00255). The authors state the FDA commissioned the review but that it was conducted independently by the research team to minimize bias.


Evidence Certainty (GRADE): Rated per conclusion. Moderate certainty for PFOA and PFOS in human milk (downgraded for risk of bias) and moderate certainty for PFAS in infant formula (downgraded for publication bias). For arsenic, cadmium, lead, and mercury, the evidence did not support any conclusion, so no GRADE rating was assigned.

Key Findings

PFAS in human milk

PFOA was detected in 439 of 559 samples and PFOS in 485 of 559 samples, across three of the four studies. Median PFOA concentrations ranged from undetected to 36.1 pg/mL and median PFOS from undetected to 106 pg/mL, both at moderate certainty. PFNA and PFHxS were detected in only two of the four studies (79 of 559 and 68 of 559 samples), and the inconsistent detection meant the evidence did not support a conclusion for them.


PFAS in infant formula

PFAS were largely not detected. Across 38 infant formula samples (at least 17 unique products) tested between 2008 and 2024, PFOA and PFNA were not detected in any sample, while PFOS and PFHxS were detected in no more than 10% of samples in one study and no more than 30% in the other. The certainty was moderate, downgraded for publication bias.


Arsenic, cadmium, lead, and mercury

The evidence did not support conclusions for these metals in either human milk or infant formula. Studies were few, samples were small and drawn from limited regions with little demographic detail, and most data were collected before 2000. For lead in human milk, for example, only two of eight studies had collected data within the previous 25 years. In infant formula, mercury was not detected in the single study that measured it, and reported arsenic concentrations ranged between 2.6 and 20 ng/g, at or below the European Union maximum guideline for powdered formula (0.02 mg/kg, or 20 ng/g).


Evidence from other countries

Across countries rated high or very high on the Human Development Index, lead was the most frequently assessed contaminant in human milk (171 articles), followed by cadmium, mercury, arsenic, and PFAS. The authors present these as heat maps to flag priorities for future reviews.

Limitations

The central limitation is the thinness and age of the United States evidence, especially for the four metals, where small, unrepresentative samples and pre-2000 data prevented any conclusion. Because so few contemporary studies existed and many lacked detail on populations or samples, the planned meta-analysis was not feasible and only a narrative synthesis was possible. By design the review included only published, peer-reviewed literature, with no gray literature, no trial registries, and no contact with authors for unpublished data, which the authors note limits scope and is why the infant formula PFAS conclusion was downgraded for publication bias. Inconsistent reporting of sample preparation, particularly dilution methods for older formula studies, made it hard to judge whether the products tested represented the current United States market. The review also did not assess health effects, infant exposure, or bioavailability, which were outside its scope.

Neutral Interpretation

This review maps what the peer-reviewed United States evidence can and cannot show. Its clearest signals are that PFOA and PFOS are detectable in United States human milk and that PFAS are largely undetected in infant formula, both at moderate certainty. For arsenic, cadmium, lead, and mercury, the evidence is too sparse and dated to support conclusions, a gap the authors emphasize and propose addressing through programs such as the FDA's Total Diet Study. The findings describe concentrations, not health risk or how much an infant actually absorbs, and the authors reiterate that public health agencies still recommend human milk because its benefits are considered to outweigh the potential risks of contaminant exposure. Since powdered formula is mixed with water at home, formula results reflect the product rather than the final prepared feed. The work was funded by an FDA contract and states it was conducted independently, and one author serves as an associate editor of the publishing journal.

Full Citation

Thoerig RC, O'Connor LE, Spill MK, Balalian AA, Trivedi R, Advani SM, Uffelman CN, Bosse T, Foster MJ, Holland KM, Dewey KG, Fisher MM, Galusha AL, Huset CA, MacFarlane AJ. Assessment of arsenic, cadmium, lead, mercury, and per- and polyfluoroalkyl substances concentrations in human milk and infant formula in the United States: a systematic review. Am J Clin Nutr. 2025;122(4):1006-1026. doi:10.1016/j.ajcnut.2025.07.039. PMID: 41043870.

Available at: https://pubmed.ncbi.nlm.nih.gov/41043870/

Disclosure

This summary is based on the published systematic review. It is provided for informational purposes only and does not constitute medical advice.