FDA expands infant formula contaminant testing datasets

The US Food and Drug Administration has expanded its published infant formula contaminant monitoring results with new survey data and a public chemical contaminants dashboard. The 29 October 2026 update covers a supplemental survey of 104 powdered formula samples and a separate mini market basket of 134 powdered products tested for metals and aluminium. Those additions extend a larger 312-sample market basket whose principal findings were already released in April.

The 104-sample supplemental survey was initiated in 2025 and analysed in 2026. It added powdered products from domestic and overseas brands newer to the US market, including cow milk, soy, and goat milk formulations. Both routine formulas and specialist extensively hydrolysed or amino acid-based products were represented. This survey complements the earlier market basket, which included powder, ready-to-feed, and concentrated liquids, so the two sets do not describe identical product populations.

For the supplemental survey, the FDA reports that measured arsenic, lead, cadmium, and mercury levels were generally very low or undetectable. Pesticides were not found in any of the 104 samples, and 29 of the 30 PFAS compounds included in the screen were absent from every sample. Detectable chemical results are reported by compound and concentration; an absence of a detected result is a laboratory finding against the method’s detection capability, not proof that a substance cannot occur under other conditions.

Phthalates were undetectable in 72 of the 104 supplemental samples, equivalent to 69%. The non-phthalate plasticiser DEHA was detected in 20 samples, while DINP appeared in 12, BBzP in two, and DBzP in one. The FDA reports occurrences by substance and sample rather than treating every detection as a failure of a product specification. Concentration, analytical detection limits, and estimated exposure are separate inputs to any assessment of health risk.

The same 104-sample survey tested for process-related contaminants associated with certain fat and oil refining conditions. The substance 3-MCPD ester was detected in 102 samples, while glycidyl esters were detected in 34. The FDA says levels of these compounds in its infant formula monitoring have declined since 2013, reflecting mitigation measures and industry work. Those detections should be read with the agency’s wider assessments and cannot be treated as equivalent to a product recall or an individual manufacturing violation.

A second 2026 mini market basket tested 134 powdered formula samples for arsenic, lead, cadmium, mercury, and aluminium. The FDA reports that heavy metal levels were generally low or undetectable, consistent with its earlier market basket. Aluminium was included to establish baseline concentrations in formula, rather than as the basis for a newly announced compositional limit. The public results cover formula types using milk, soy, and goat protein, including specialist products.

The 2023–2025 market basket remains a separate dataset. It tested 312 samples across powdered formula, ready-to-feed liquids, and liquid concentrates and was the subject of the FDA’s April 2026 results publication. That survey found no detectable pesticides in 309 samples and no detections of 25 of the 30 PFAS compounds across the group. Reintroducing those figures without their original publication date would obscure what the October update actually adds.

Alongside the new surveys, the FDA corrected a subset of historical perfluorobutanoic acid, or PFBA, entries in its earlier market basket. It says the correction does not change the conclusions drawn from those results. The agency has also made arsenic, cadmium, lead, and mercury findings available through a public chemical contaminants dashboard, providing a more accessible record of its heavy-metal monitoring as further datasets are added.

Because the surveys draw products from retail channels, they document concentrations in sampled formulas rather than testing every production lot or pinpointing an individual contamination source. Ingredient composition, product form, analytical method, and batch history may differ across the market. A detected contaminant can originate through raw materials or processing inputs, but the published results alone do not trace it to a particular manufacturing step. That distinction is important when using the data to review supplier specifications or investigate an individual result.

Infant formula plants control ingredient selection, production hygiene, and finished-product specifications through separate checks, while regulators monitor contaminants across products already on the market. The 104-sample survey includes process-related contaminants associated with oil refining as well as substances linked to environmental exposure. Those data can inform assessment of raw-material sources and process controls; they do not establish a new microbiological contamination incident. Cronobacter and Salmonella remain subjects of separate surveillance and control requirements.

The FDA has now added the 104-sample supplemental survey, the 134-sample powdered-formula metal survey, a public dashboard, and corrected PFBA entries to its monitoring resources. The 312-sample main market basket was already public in April. Some sampled products appear in more than one 2026 survey, so the sample counts should not be added together as a total of distinct products. Additional published surveys will extend the evidence available to evaluate contaminant levels across different formulas and production periods.