Soares et al. 2000 - Cr(VI) in powdered milk formulas
Summary
This peer-reviewed AOAC International paper adapts an ion-exchange / electrothermal atomic absorption spectrometry method for selective hexavalent chromium measurement in powdered milk formulas. The paper is useful because it reports Cr(VI), not total chromium, in 20 commercial powdered milk products from Portugal. Because the source repeatedly describes the matrix as powdered milk / follow-up milk / dietetic milk and does not mention soy or plant protein, the occurrence rows route to the HMTc non-soy powdered formula row as dairy/milk-formula context. The occurrence data remain small, market-specific, and summary-only.
Key Numbers
| Scope | N | Basis | Cr(VI) range | Mean | Table/section | HMTc handling |
|---|---|---|---|---|---|---|
| Infant formulas for newborn infants | 7 | powder as sold | <10 to 75 ng/g | 24 ng/g | Table 4 | Direct non-soy/dairy formula context; source summary only. |
| Follow-up milks for infants older than 4 months | 5 | powder as sold | <10 to 26 ng/g | 12 ng/g | Table 4 | Direct non-soy/dairy follow-up formula context; source summary only. |
| Dietetic milks for infants in pathological situations | 8 | powder as sold | <10 to 75 ng/g | 33 ng/g | Table 4 | Direct non-soy/dairy specialty formula context; source summary only. |
Because 1 ng/g equals 1 ppb on a mass basis, the reported means correspond to 24 ppb, 12 ppb, and 33 ppb Cr(VI), with source-scope maxima up to 75 ppb. The lower end of the range is censored at <10 ng/g. The paper reports grouped means/ranges rather than the underlying sample-level values needed for later HMTc benchmark-pool percentile math.
Methods Brief
Powdered milk was reconstituted in doubly deionized water at 1 g in 15 mL, proteins were precipitated, and Cr(VI) was retained on a Chromabond NH2 anion-exchange column before nitric-acid elution. Quantification used electrothermal atomization atomic absorption spectrometry at 357.9 nm.
The method validation reported a 1.8 ug/L detection limit in reconstituted milk, a 1.8-50.0 ug/L calibration range, analytical precision of 4.1 percent, overall-procedure precision of 6.5 percent, and standard-addition recoveries above 93 percent.
Evidence Fitness
Evidence Fitness: EF-3, modeled or limited evidence. The source is species-specific for Cr(VI), directly relevant to dairy/non-soy powdered formula context, and reports product-group means and ranges. It is not a reconstructable sample-level distribution and should remain summary/context evidence unless another traceable sample pool is added.
Limitations
- The survey covers 20 Portuguese-market products and does not report brand-level values.
- The product groups are milk-formula groups and should not be routed to the soy-based formula row unless another source supplies soy-specific evidence.
- The lower range is censored at <10 ng/g.
- The paper does not publish the individual product values behind the grouped means/ranges, so the products cannot be admitted one by one into benchmark-pool calculations from this table alone.
- The method measures Cr(VI) after reconstitution and sample preparation; comparability to newer Cr speciation methods should be reviewed before cross-source pooling.
Implications
Certification: Useful Cr(VI) species-specific warning context for non-soy/dairy powdered formula. It is not sufficient by itself for HMTc benchmark-pool math because it provides grouped means/ranges from a small Portugal-market survey rather than individual sample values; HMTc percentile calculations belong to the admitted benchmark pool, not to this paper.
Courses: Good teaching example for why total chromium and Cr(VI) must be kept separate, and why method validation does not automatically make a small occurrence survey standards-ready.
App: Supports keeping Cr(VI) as a distinct analyte for non-soy formula-powder evidence retrieval while excluding total chromium substitutes.
Microbiome: No direct microbiome endpoint.
Provenance Notes
The local PDF was supplied in raw/Working Together with ChatGPT/. DOI and bibliographic metadata were checked against the Oxford Academic article record and PubMed record on 2026-05-08. The source is public-reference-only; cite the article record rather than redistributing the local PDF.