Soy-based infant formula

Overview

This is a structural ingredient/profile node for soy-based infant formula routing. Finished formula occurrence values belong on the relevant formula product pages unless a source reports ingredient-only values.

References

Works cited in this page’s text, in first-appearance order. This is not the full corpus for this page; it is only what the prose above draws on. The complete set of sources is listed under Sources below. Each title links to its source record, which carries the ingest receipt, the extracted values, and the file hash of the document it was built from.

1. There is (still) too much aluminium in infant formulasShelle-Ann M. Burrell and Christopher Exley · BMC Pediatrics · 2010 · doi.org/10.1186/1471-2431-10-63Review
2. Assessment of arsenic, cadmium, lead, mercury, and per- and polyfluoroalkyl substances concentrations in human milk and infant formula in the United States: a systematic reviewRachel C. Thoerig, Lauren E. O’Connor, Maureen K. Spill, Arin A. Balalian, Rupal Trivedi, Shailesh M. Advani, et al. · American Journal of Clinical Nutrition, Vol. 122, pp. 1006-1026 · 2025 · doi.org/10.1016/j.ajcnut.2025.07.039Review
3. Estimated Exposure to Arsenic in Breastfed and Formula-Fed Infants in a United States CohortCourtney C. Carignan, Kathryn L. Cottingham, Brian P. Jackson, Shohreh F. Farzan, A. Jay Gandolfi, Tracy Punshon, et al. · Environmental Health Perspectives, Vol. 123, No. 5, pp. 500-506 · 2015 · doi.org/10.1289/ehp.1408789Review
4. Arsenic concentration and speciation in infant formulas and first foodsBrian P. Jackson, Vivien F. Taylor, Tracy Punshon, and Kathryn L. Cottingham · Pure and Applied Chemistry, Vol. 84, No. 2, pp. 215-223 · 2012 · doi.org/10.1351/PAC-CON-11-09-17Review
5. Simultaneous Determination of Arsenic, Cadmium, Mercury, and Lead in Raw Ingredients, Nutritional Products, and Infant Formula by Inductively Coupled Plasma Mass Spectrometry: Single-Laboratory ValidationLawrence H. Pacquette and Anumolu Anumula · Journal of AOAC International, Vol. 99, No. 3, pp. 766-779 · 2016 · doi.org/10.5740/jaoacint.15-0304Review
6. Final Opinion on tolerable intake of aluminium with regards to adapting the migration limits for aluminium in toysSCHEER · 2017 · doi.org/10.2875/264211Government

Sources

Auto-generated from source-page frontmatter. The "Used on this page for" column is populated by the orchestrator's POPULATE-SOURCE-LEGEND action; pending entries appear as *[awaiting synthesis]*.

Why this commodity accumulates heavy metals

Soy-based infant formula uses soy protein isolate as the protein base rather than cow-milk protein. The soy protein isolate carries the source-soybean Al, Ni, and Cd profile at concentrated per-mass levels because protein extraction concentrates these metals; see Soy for the soybean-stage rationale. Soy-based formula consequently carries elevated Al, Ni, and Cd relative to cow-milk-based formula even when the rest of the formulation (carbohydrate source, fat blend, vitamin-mineral premix) is identical. Multiple infant-formula surveys document this consistently: Burrell 2010 UK soy powder formula at 4.3 µg/g aluminum, which translates to roughly 629 µg/L prepared-for-feeding under standard 1:7 reconstitution.

Soy-based formula is medically indicated for infants with cow-milk-protein allergy or galactosemia, and is otherwise consumed by infants of vegan families or by parents of preference. The HMTc Cat 1 Step 0 lock splits soy-based formula from non-soy into separate product rows (Infant Formula, Powder (Soy-Based) and Infant Formula, RTF Liquid (Soy-Based)) because the Al/Ni/Cd-elevated profile warrants separate certification analysis. The HMTc panel concerns for soy-based infant formula are dominantly Al, Ni, and Cd (the soy-protein-isolate inheritance pathway), plus the standard Pb concerns common to all infant formulas. Arsenic occurrence in soy-based formula sits at 7-12 ppb tAs/iAs medium-confidence based on the U.S. systematic review by Thoerig 2025, the U.S. cohort biomarker work by Carignan 2015, and the speciation survey by Jackson 2012.

Heavy metal contamination profile

The body-level analyte snapshot for soy-based formula follows the per-format pages: see Infant Formula, Powder (Soy-Based) and Infant Formula, RTF Liquid (Soy-Based) for the format-specific concentration tables. The dominant analytes of concern (Al, Ni, Cd) reflect the soy-protein-isolate inheritance described above; the contributing sources documented in the Source legend collectively establish the multi-fold Al elevation versus cow-milk formula.

Ranges by source, region, and variety

Variance within soy-based infant formula tracks four dimensions: source-soybean origin (US Midwest vs Brazilian vs Argentine vs Chinese vs European soy production), the protein-isolation process used by the soy-isolate supplier, the manufacturer’s formulation choices (vitamin-mineral premix supplier, water source), and historical generation (post-2020 vs pre-2010 manufacturer responses to the Al-in-soy-formula debate documented in Burrell 2010 and earlier work).

US-market soy-based formula, EU-market soy-based formula, and emerging-market soy-based formula carry different baseline profiles reflecting these per-region soy-supplier and per-manufacturer differences. The Al elevation in soy formula relative to cow-milk formula is consistent across markets at multiple-fold; absolute levels vary by manufacturer. Arsenic occurrence in U.S.-market soy-based formula sits in the 7-12 ppb range across the Jackson 2012 and Carignan 2015 cohort work, with the systematic review by Thoerig 2025 confirming the range across the broader U.S. literature.

Processing effects

Soy-based infant formula manufacturing follows the same processing chain as cow-milk-based formula (see Infant Formula Powder and Infant Formula Rtf Liquid) with soy protein isolate substituted for the milk-protein-and-lactose ingredients. The protein-isolation step at the upstream soy-supplier stage is the dominant metal-concentration-from-raw-soybean event; once the isolate is incorporated into formula, downstream processing follows the standard infant-formula pathway without changing the Al/Ni/Cd-elevated profile. ICP-MS analytical methods for soy-formula matrices are validated in Pacquette 2016.

Ingredient-derivative risk

Soy-based formula derivatives are the same format set as other infant formulas: powder, ready-to-feed (RTF) liquid, and concentrated liquid. The Cat 1 Step 0 lock variants (Infant Formula, Powder (Soy-Based), Infant Formula, RTF Liquid (Soy-Based)) carry the same soy-protein-isolate inheritance; format affects per-volume concentration via reconstitution ratio but not per-gram-protein concentration. Older soy-formula generations carry higher Al than current production because manufacturer responses to Burrell 2010 and subsequent regulatory attention have driven reductions; historical stock and emerging-market supply may still carry higher Al concentrations.

Mitigation options

Sourcing levers (Supply-chain screening) for soy-based formula are limited because the soy-isolate ingredient itself carries the elevated Al/Ni profile by species characteristic. The dominant brand-side decisions are: soy-supplier specification (testing of incoming soy protein isolate for Al, Ni, Cd against the per-mass-in-finished-formula compliance target); soy-origin selection within low-Al soybean production regions (limited efficacy because species-level Al accumulation dominates); and hydrolysate or amino-acid formula substitution for infants with cow-milk-protein allergy as an alternative to soy-based formula where medically appropriate.

Agronomic levers (Agronomic mitigation) operate at the soybean-production stage; see Soy for soil-Cd remediation and Al-uptake-management interventions.

Processing levers (Processing mitigation) are limited at the formula-manufacturing stage; the Al/Ni profile is in the soy protein isolate already. Upstream soy-isolate processing improvements (acid-base extraction modifications, ion-exchange polishing) carry potential but are not standard in commodity-grade isolate production.

Formulation levers (Formulation mitigation) include the alternative-base-protein substitution (hydrolyzed cow-milk protein, amino-acid-based formula) for medical-indication infants where soy is not the only option.

Testing and QC levers (Testing and quality-control mitigation) are mature: lot-level Al, Ni, Cd, Pb testing on finished soy-based formula against the EU 10 ppb Pb prepared-for-feeding ML and applicable Al/Ni reference levels (EFSA TWI for Al). The ICP-MS method validated by Pacquette 2016 is the operative analytical platform.

Packaging and storage levers (Packaging and storage mitigation) include can-lining specification (BPA-NI epoxy or food-grade alternative) and aluminum-foil-lined packaging considerations to avoid Al migration from packaging into prepared formula.

Regulatory limits that apply

• eu-2023-915 — EU Reg. 2023/915 sets binding maximum levels for infant formula: Pb 10 ppb prepared-for-feeding (≈70 ppb powder basis), Cd 5 ppb, iAs 20 ppb, Hg 20 ppb. These apply to soy-based formula identically to cow-milk-based.
• fda2020-inorganic-arsenic-infant-rice-cereal — FDA Closer to Zero iAs framework covers rice-containing formula.
• Codex Alimentarius CXS 72-1981 (infant formula) and CXS 156-1987 (follow-up formula) establish composition standards.
• JECFA Al PTWI of 2 mg/kg b.w./week and EFSA TWI of 1 mg/kg b.w./week are the relevant dose-response references per SCHEER 2017; soy-based formula at the upper end of the Al-concentration range documented by Burrell 2010 and subsequent literature can deliver Al intakes that approach or exceed these reference levels for exclusive-soy-formula-fed infants.
• California Prop 65 (california-prop65) Pb MADL applied to infant formula yields stringent serving-based screen.

Update history

The five most recent substantive edits to this page, classified major (evidence or structure moved), correction (a published value or statement was wrong and has been fixed), or minor (narrative rewritten without changing the underlying evidence). Each description is derived from what the edit did to this page; the linked commit is the authoritative record, routine regeneration passes are excluded, and the full version history lives in git. When DOI minting comes online (see schema docs), each entry below will also link to a version-pinned DataCite DOI.