Skip to content
Heavy Metal Index

From Farm to Retail: Decoding the Elemental Landscape of Milk and Dairy Products Across Organic and Conventional Production Systems Using ICP-MS

Overview

Marcus and colleagues analyzed 307 milk, yogurt, cheese and mozzarella samples from organic farms, conventional farms and retail outlets in north-western Romania by ICP-MS for lead, cadmium, mercury, arsenic, chromium, nickel, aluminum, tin, copper and zinc. Toxic metals were low throughout, mercury was below detection in every dataset, and cadmium was below detection in all yogurts and cheeses. Retail cow products carried higher cadmium, arsenic, chromium, nickel and aluminum than organic farm products.

Key numbers

  • Organic cow milk lead 0.00809 ± 0.00182 mg/kg (n = 30) vs retail cow milk 0.00667 ± 0.00200 mg/kg (n = 9); cadmium 0.00106 vs 0.00186 mg/kg; arsenic 0.00106 vs 0.00425 mg/kg (Table 2).
  • Cow yogurt lead 0.00550 (organic) vs 0.00846 mg/kg (retail); arsenic below detection in organic yogurt, 0.00360 mg/kg in retail yogurt (Table 3).
  • Cow cheese lead 0.00727 (organic) vs 0.00946 mg/kg (retail); cadmium and mercury below detection in both (Table 4).
  • Buffalo mozzarella lead 0.008 mg/kg, cadmium and mercury below detection (Table 1); donkey milk arsenic 0.06641 mg/kg, partially censored (Section 3.1 text).

Methods (brief)

Samples were digested with nitric acid and hydrogen peroxide in a closed microwave system and analyzed by ICP-MS with helium collision mode. Accuracy was checked with NIST SRM 1549 (recoveries 96 to 104%) and a FAPAS whole-milk proficiency material. Results are on a wet-weight basis; means exclude values below the LOD, and numeric LODs are only in the Supplementary Material.

Evidence Fitness

A large, recent, well-dated sample set with certified reference material checks. The tables have internal conflicts: a Table 1 row labelled buffalo milk cheese carries organic cow cheese codes, the goat milk row in Table 1 largely repeats the organic cow milk row, and several text means differ from the tables, so Tables 2 to 4 are the most reliable figures. Means of partially censored groups use only quantified values and run high. For the thin pages it was pulled for (reduced-fat milk, fromage frais, custard, cheese pizza, low-calorie ranch dressing), this is a dairy ingredient proxy only: it measures milk, yogurt and cheeses of unstated fat level, not those foods.