← Back to Stage 1 · The Window
STAGE 3 / TRACK C Baby food meta-analysis · DetectLead
Four datasets. One question. What's actually safe to feed your kid?
FDA tested 312 baby food samples brand-blind. Consumer Reports named 49 brands. Moms Across America independently tested 20 more. Lead Safe Mama has been at it for 15 years. None of them publish the dose. None of them cross-reference each other. We did. This page is what shows up when you stack all four side by side.
Total lead consumed per day is the only number that matters. Parts per billion is a chemistry unit; it doesn't tell a parent anything actionable until you multiply by serving size. We did that for every row in every dataset above and ranked the result against the FDA's daily reference levels (8.8 µg adult, 2.2 µg child). When two or more independent testers agree on a brand, we treat that as the verdict. When FDA's brand-blind aggregate hides the spread, we publish the spread. When a "ppb" number looks small but the dose looks bad, we lead with the dose. Use this page as the working authority on what's actually in the food a kid eats. The numbers came from four good testers; the math is ours; the position is unambiguous.
Federal · 2024–2025
FDA Total Diet Study
312
samples · brand-blind
The largest infant formula heavy-metal study in U.S. history. Government-funded, ICP-MS, statistically valid samples. Hides which brand is which.
Independent · 2023
Consumer Reports
49
named brands
Names every brand. Names every result. Doesn't compute µg/day or compare against FDA reference levels. We did.
Parent-driven · 2024
Moms Across America
20
named brands
Independent lab. Smaller sample but a third tester completely separate from CR and FDA. The cross-check on the cross-check.
Longest-running · 15 yrs
Lead Safe Mama
800+
items · ICP-MS
Tamara Rubin's ongoing project. Names brands, posts the lab reports, alarm-bell framing. The deepest brand-level archive that exists.
DetectLead
1,318
rows cross-referenced
Every row from all four sources, converted to µg/day at typical infant intake, ranked against FDA Interim Reference Levels. Brand-named. Sortable. Cross-referenced when 2+ sources tested the same product.
Average lead by category, across all 1,318 tested items.
Most parents focus on baby food and formula. Both are below 20 ppb on average. The categories that actually carry the highest lead loads are the ones nobody warns you about: spices, toothpaste, salts, and supplements. This is the chart you didn't get from your pediatrician.
Source averages computed across all DetectLead-aggregated samples in each category. Bar widths scale to the spices/condiments maximum (507 ppb). The story: parents who ban juice and switch to organic baby food while seasoning every meal with turmeric and brushing kid teeth with Colgate Baby are missing the actual exposure.
FDA reports formula at 0.0 ppb on average. Branded reality is 0 to 4.2 ppb.
FDA's Total Diet Study is the gold standard for sample integrity, but its outputs are anonymized. When the agency reports "Baby Food, infant formula, milk-based, powdered = 0.0 ppb," it's reporting an average across an unnamed pool. The brand-level data inside that pool, surfaced by Consumer Reports and Lead Safe Mama, ranges from 0 to 4.2 ppb. That's the entire spread, hidden inside a single zero.
+38
The 38-formula spread that FDA's 0.0 ppb average buried.
FDA reports four "infant formula" categories, all averaging zero. Consumer Reports tested 38 specific products and got individual results from 0.0 ppb (Bobbie Whole Milk) to 4.2 ppb (Enfamil Nutramigen Hypoallergenic). Eight products tested Pb=0. Fourteen tested over 15% of FDA's child Interim Reference Level (2.2 µg/day) at typical infant intake.
23%
of the FDA child IRL · per day · from a single hypoallergenic formula serving · invisible in FDA's brand-blind aggregate
When two independent testers both flag the same brand, that's signal.
No single tester is right about everything. Sample-batch variance is real. But when Consumer Reports and Moms Across America and Lead Safe Mama, three different operations with different methods and different goals, all flag the same brand, that's not a fluke. That's a verdict.
Cross-source agreement on infant formula brands
Brand · SKU
FDA
CR
MAA
LSM
Bobbie · all SKUs
no brands
CLEAN
CLEAN
CLEAN
Kirkland (Costco) ProCare
no brands
CLEAN
CLEAN
not tested
Similac Sensitive
no brands
CLEAN
CLEAN
not tested
Enfamil Sensitive
no brands
DIRTY
DIRTY
not tested
Up & Up (Target) standard
no brands
DIRTY
DIRTY
not tested
Enfamil Nutramigen (Hypoallergenic)
no brands
DIRTY
not tested
DIRTY
ByHeart Whole Nutrition
no brands
DIRTY · RECALLED
not tested
DIRTY
When two or more independent testers agree, treat that result as the working verdict. Sample variance is a real concern in single-source claims; cross-source agreement collapses that uncertainty. The seven brands above are the ones with the most overlap signal in the public record.
The numbers don't belong to us. The math does.
FDA Total Diet Study · brand-blind, methodologically rigorous, posted publicly at fda.gov/food/total-diet-study.
Consumer Reports · 2023 infant formula round, brand-named, available at consumerreports.org.
Moms Across America · independent baby formula testing 2024, posted at momsacrossamerica.com.
Lead Safe Mama (Tamara Rubin) · 15 years of brand-named ICP-MS testing posted at leadsafemama.com. Tamara's framing tends toward worry; she's the alarm bell. We do dose-math instead. Same data, different posture, both useful. Use both.
Stage 4 is action.
You finished the meta-analysis. You know which categories carry the load. You know which brands cleared two or more independent testers. Now go look at your house, your tap, your aisle.