Analytical Results for Arsenic, Lead, Cadmium, and Mercury in Food Intended for Babies and Young Children - TEP (FY2009-FY2024)

Summary

This FDA compliance-program table is an A-tier government dataset for arsenic, lead, cadmium, and mercury in foods intended for babies and young children from FY2009 through FY2024. It has been ingested as sample-level Category 1 evidence where the FDA baby-food category and product name can be mapped to locked HMTc product rows.

This source does not cover infant formula, fruit juice, fish-containing baby foods, or meat/poultry purees. It also does not report inorganic arsenic; the arsenic table is ingested as source-reported arsenic / total arsenic (tAs) until speciation evidence says otherwise.

Extracted Files

data/evidence/category1_fda_baby_food_compliance_samples.csv: 1,944 parsed sample/analyte rows.
data/evidence/category1_fda_baby_food_compliance_summary.csv: 39 row/product/analyte summary records.
data/evidence/values.jsonl: 39 machine-extracted value records with prefix category1-fda-baby-food-compliance-.
data/evidence/sources.jsonl: source metadata record for this PDF.

Source Tables

Arsenic table: 575 samples from FY2009-FY2024.
Lead table: 598 samples from FY2009-FY2024.
Cadmium table: 576 samples from FY2009-FY2024.
Mercury table: 195 samples from FY2009-FY2024.

FDA reports that arsenic, lead, cadmium, and most mercury analyses were conducted by ICP-MS. Three early mercury rows were reported by X-ray fluorescence as NDb under a 5 ppm LOQ.

Category 1 Mapping

Rows were mapped conservatively:

FDA Dry Infant Cereals with rice named in the product description → Baby Cereals / Grain Products, Dry (Rice-Based).
FDA Dry Infant Cereals with no rice named → Baby Cereals / Grain Products, Dry (Non-Rice).
FDA Fruits → Fruit Purees as finished fruit baby-food products.
FDA Vegetables containing carrot, sweet potato, beet, or parsnip terms → Root-Vegetable Purees.
Other FDA Vegetables → Non-Root Vegetable Purees.
FDA Mixtures with rice named → Mixed Meals, Rice-Containing.
FDA Mixtures with no rice named → Mixed Meals, Non-Rice.
FDA Grain-Based Snacks with rice named → Teething & Snacks (Rice-Based), but the explicit rice-named subset is small.
FDA Grain-Based Snacks without rice named → broad snack context only; these rows are not treated as Teething & Snacks (Non-Rice) because ingredient rice status is not isolated.
FDA Yogurts/Custards/Puddings → context only; no separate locked Category 1 row exists.

Summary Highlights

The summary values below are lower-bound nearest-rank percentiles in ppb with <LOD and NDb treated as 0. They are machine-extracted and require review before standards use.

Category 1 scope	Evidence Fitness	Key lower-bound distribution signals
Baby Cereals / Grain Products, Dry (Rice-Based)	EF-2	`tAs` n=253, p50 115, p90 135, max 348; `Pb` n=256, p90 19.2, max 32; `Cd` n=252, p90 22, max 40.5; `tHg` n=64, p90 2.1, max 4.
Baby Cereals / Grain Products, Dry (Non-Rice)	EF-2 / EF-3	`tAs` n=25, p90 37.8, max 54.8; `Pb` n=25, p90 8, max 9.9; `Cd` n=25, p90 27.4, max 62.9; `tHg` n=9, all lower-bound 0.
Fruit Purees	EF-2	`tAs` n=39, p90 5.2, max 8.7; `Pb` n=44, p90 2.4, max 8; `Cd` n=39, p90 2.2, max 4; `tHg` n=14, p90 0.5, max 0.6.
Non-Root Vegetable Purees	EF-2	`tAs` n=20, p90 1, max 11; `Pb` n=29, p90 2, max 7.6; `Cd` n=22, p90 12.8, max 23.1; `tHg` n=13, p90 0, max 0.4.
Root-Vegetable Purees	EF-2	`tAs` n=54, p90 6.4, max 10.3; `Pb` n=59, p90 15.9, max 27.3; `Cd` n=54, p90 31.5, max 42; `tHg` n=25, p90 0.3, max 1.1.
Mixed Meals, Non-Rice	EF-2	`tAs` n=77, p90 5.6, max 13.6; `Pb` n=78, p90 6.8, max 13; `Cd` n=77, p90 5.2, max 44.4; `tHg` n=36, p90 0, max 0.4.
Mixed Meals, Rice-Containing	EF-3	Small rice-named mixture subset: `tAs` n=9, p90/max 28.3; `Pb` n=9, p90/max 11.6; `Cd` n=9, p90/max 7; `tHg` n=3, p90/max 0.3.
Teething & Snacks (Rice-Based)	EF-3	Explicit rice-named snack subset: `tAs` n=2, p50 96.3, p90/max 171; `Pb` n=2, p90/max 6.5; `Cd` n=2, p90/max 3.5.
Broad FDA grain-based snacks, rice status not isolated	EF-4	Context only: `tAs` n=91, p50 61, p90 224, p95 383, max 561; `Pb` n=91, p90 15, max 23.7; `Cd` n=91, p90 27, max 41; `tHg` n=28, p90 2.5, max 3.3.

LOD/LOQ Handling

The PDF gives analyte-level ranges for LOQ and LOD, not per-row usable censoring thresholds in the table text. The machine extraction therefore preserves the source-reported text (<LOD, NDb, or numeric value) and computes lower-bound distributions by substituting 0 for <LOD and NDb.

Arsenic LOQs ranged from 0.7 to 103 ppb; LODs ranged from 0.2 to 34.3 ppb.
Lead LOQs ranged from 0.2 to 100 ppb; LODs ranged from 0.1 to 33.3 ppb.
Cadmium LOQs ranged from 0.1 to 63 ppb; LODs ranged from 0.03 to 21 ppb.
Mercury LOQs ranged from 0.8 to 38.4 ppb; LODs ranged from 0.1 to 12.8 ppb for most mercury analyses.

Evidence Fitness

Direct row-fit subsets with at least 10 reconstructable sample rows are EF-2 pending review. Small rice-named subsets are EF-3. Broad grain-based snacks without isolated rice status and yogurt/custard/pudding rows are EF-4 context only.

These values do not publish or justify HMTc thresholds. They are occurrence evidence candidates for internal review and product-page synthesis under the shared evidence contract.

Limitations

Compliance-program sampling is not a market-share-weighted Total Diet Study.
As is retained as tAs / source-reported arsenic, not inorganic arsenic.
Hg is retained as total mercury, not methylmercury.
Grain-based snack product names are not enough to infer non-rice status.
The source does not cover infant formula, fruit juice, fish-containing baby foods, or meat/poultry purees.

Implications

Certification: priority A-tier occurrence dataset for Category 1 review, especially rice-based dry cereals, root-vegetable purees, and snack context.

Courses: strong example for teaching surveillance sampling, censoring, and why source analyte/speciation labels must travel with values.

App: useful for retrieval and row-specific evidence summaries after reviewed promotion; not sufficient by itself for brand ranking or ingredient-level inference.

Microbiome: no direct microbiome endpoint.

Wiki Pages Updated On Ingest

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.

Commit	Date	Change	Description
a8052bb	2026-08-09	major	11 sections added; narrative text revised