Heavy Metal Index
Ask the IndexAboutIndex Pro

A Feasibility Study Investigating Action Limits for Certain Heavy Metal Impurities in Cosmetic Products

Cite this pageVersion history

Overview

This report is a UK government feasibility study commissioned by the Office for Product Safety and Standards (OPSS) and executed by LGC under Contract FM21190. It combines (1) a comprehensive international review of regulatory limits and non-regulatory guidance values for technically-unavoidable heavy-metal impurities (Sb, As, Cd, Pb, Hg) in cosmetic products, (2) identification and verification of a fit-for-purpose analytical method based on ISO 21392:2021 (microwave-assisted HNO₃/HCl ± HF digestion + ICP-MS), and (3) a small UK-market survey of 91 cosmetic products covering 10 product types across high-street and online retailers. The report concludes that the German BVL guidance values (Sb <0.5, As <0.5, Cd <0.1, Pb <2.0 general / <0.5 toothpaste / <5.0 theatre-make-up, Hg <0.1 mg/kg) can be applied to the UK market with minimal impact at the 90 % compliance threshold the BVL itself uses, with eye shadow as the principal exception (3 of 9 eye-shadow samples exceeded the BVL 2 mg/kg Pb general-cosmetics limit; max Pb 8.8 mg/kg). Regulatory framing rests on UK EU-retained law Regulation (EC) No 1223/2009 Article 17 (technically-unavoidable trace impurities permitted in good manufacturing practice) read with Article 3 (safety) and Article 8 (GMP).

Key numbers

Regulatory / guidance limits compiled by the literature review (mg/kg unless stated)

  • China — Technical Safety Standard for Cosmetics 2015 (Table 2, p. 7): As ≤2, Cd ≤5, Pb ≤10, Hg ≤1; general (whole-product limits, not impurity-only).
  • Thailand — single regulatory cadmium limit Cd 3 mg/kg in cosmetic products (Section 2.2 and Table 10, p. 21).
  • USA — 21 CFR §700.13 mercury preservative: Hg ≤1 mg/kg general cosmetics, ≤65 mg/kg eye-area preservative; FDA 2016 recommended industry maximum Pb ≤10 mg/kg in cosmetic lip products and externally-applied cosmetics; FDA colour-additive heavy-metal limits (typical: As ≤3 ppm, Pb ≤20 ppm, Hg ≤1 ppm on colorant-as-ingredient basis) (p. 7-8, p. 11).
  • Germany BVL — non-regulatory guidance, technically-avoidable impurity limits in cosmetics (Table 3, p. 9; Table 16f, p. 40):
    • General cosmetics: Sb <0.5, As <0.5, Cd <0.1, Pb <2.0, Hg <0.1 mg/kg
    • Toothpaste: Pb <0.5 mg/kg
    • Theatre/carnival make-up: As <2.5, Pb <5.0 mg/kg
  • ASEAN — guidance limits for heavy-metal impurities in cosmetics (Table 4, p. 10): As <5, Cd <5 (Thailand regulatory 3 mg/kg overrides), Pb <20, Hg <1 mg/kg.
  • Health Canada — guidance on heavy-metal impurities in cosmetics (Table 5, p. 10): action thresholds at Sb >5, As >3, Cd >3 mg/kg.
  • ICCR (International Cooperation on Cosmetics Regulation) — recommended target levels: Pb ≤10 mg/kg in finished cosmetic products excluding oral-cavity products (ICCR 2013); Hg ≤1 mg/kg as total mercury in finished cosmetic products (ICCR 2016) (p. 7-8).
  • KvW (Netherlands) — 1980 guidance values for technically-unavoidable heavy-metal impurities (Table 8, p. 16): Sb <5, As <5, Cd <5, Pb <20, Hg <1 mg/kg.

Method LODs / LOQs

  • Hepp et al. (FDA, 2014) — microwave HNO₃/HF digestion + ICP-MS (Table 6, p. 13): LOD As 0.048, Cd 0.018, Pb 0.0084, Hg 0.0010 mg/kg; LOQ As 0.16, Cd 0.058, Pb 0.028, Hg 0.0032 mg/kg.
  • Perkin Elmer application note — microwave HNO₃/HF digestion + ICP-MS, estimated detection levels in solid cosmetics (Table 7, p. 15): As 0.014, Cd 0.0069, Pb 0.0096, Hg 0.037 mg/kg.
  • ASEAN method (GF-AAS / FIAS-AAS) reported recoveries from spiked cosmetic creams: As 84-86 %, Cd 66-71 %, Pb 85-99 %, Hg 95-108 % (p. 12). Determination limits As <2.5, Cd <1, Pb <10, Hg <0.5 µg/g.
  • ISO 21392:2021 inter-laboratory trial (10 labs, 7 cosmetic samples, Pb): relative repeatability 7.5-15.5 %, relative reproducibility 15-80 % (Section 2.4, p. 15).

Reference data from prior published surveys cited by OPSS (selected, all in mg/kg unless stated)

  • JRC EUR 24886 EN (2010-2011 European market): 113 lip products (81 lipsticks + 32 lip glosses) from 34 brands across 12 EU Member States; 30 of 81 lipsticks (37 %) >1 mg/kg Pb; 1 of 32 lip glosses (3 %) >1 mg/kg Pb; max 3.75 mg/kg Pb in a red lipstick (p. 16-17).
  • BVL 2019 monitoring — lipsticks WITH glitter (Table 9a, p. 20; n analysed varies by metal): Sb n=139, mean 0.096, median 0.050, max 0.410, BVL guidance 0.5, 0 samples exceeded; As n=135, mean 0.124, median 0.063, max 0.507, BVL guidance 0.5, 1 sample exceeded; Cd n=168, mean 0.022, median 0.025, max 0.030, BVL guidance 0.1, 0 exceeded; Pb n=169, mean 0.401, median 0.250, max 6.130, BVL guidance 2.0, 3 exceeded; Hg n=152, mean 0.028, median 0.013, max 0.428, BVL guidance 0.1, 2 exceeded.
  • BVL 2019 monitoring — lipsticks WITHOUT glitter (Table 9b, p. 20): Sb n=103, mean 0.587, median 0.125, max 24.2, BVL 0.5, 8 exceeded; As n=100, mean 0.070, median 0.019, max 0.997, BVL 0.5, 1 exceeded; Cd n=103, mean 0.013, median 0.005, max 0.014, BVL 0.1, 0 exceeded; Pb n=103, mean 0.330, median 0.190, max 7.610, BVL 2.0, 1 exceeded; Hg n=108, mean 0.033, median 0.033, max 0.431, BVL 0.1, 6 exceeded. The mean (0.587)/median (0.125) divergence and max (24.2 mg/kg) for Sb in lipsticks-without-glitter indicate a heavily right-skewed distribution driven by a small number of high outliers.

LGC UK-market survey (n = 91, 9 product groups analysed; makeup powders verified but not analysed) — analytical detail

Method: ISO 21392:2021-derived microwave acid digestion (Multiwave Go, Anton Paar; mouthwash on Discover SP-D, CEM); two digestion programmes (Short 30-min ramp + 30-min hold at 200 °C; Long ramp/hold at 160/180/200 °C totalling 125 min) selected by matrix; acid mix HNO₃ 8 mL + HCl 1 mL + HF 0.5 mL on Long for lipstick/lip gloss/lip liner/mascara/eye powders/eye liners/face powders/liquid foundation, HNO₃ 5 mL + HCl 1 mL + H₂O₂ 1 mL on Short for mouthwash, HNO₃ 8 mL + HCl 1 mL + HF 0.5 mL on Short for toothpaste and sun protection (Table 13, p. 29). Sample mass 0.1 ± 0.01 g, final volume 50 mL water (mouthwash 20 mL), analytical dilution 2-fold (mouthwash 5-fold), total dilution factor 1000-fold for all matrices (Table 14, p. 30). Detection: Agilent 7700x / 7900 ICP-MS, He collision mode for ⁷⁵As, no-gas mode for ¹¹¹Cd, ¹²¹Sb, ²⁰²Hg, ²⁰⁸Pb; internal standards ¹¹⁵In, ¹²⁸Te, ¹⁵⁹Tb, ¹⁷⁵Lu; ⁹⁵Mo monitored to correct ¹¹¹Cd interferences (Table 15, p. 30). All samples in duplicate; for 12 samples where one replicate was <LOQ and the other above, the worst-case (above-LOQ) value is reported.

LGC UK-market survey — per-product-group summary vs BVL guidance (Tables 16a-16e, p. 38-39; mg/kg)

  • Sb (Table 16a, p. 38; BVL general guidance 0.5 mg/kg): lipstick LOQ 0.096, 7/9 <LOQ, 0 >BVL, max 0.292. Lip gloss LOQ 0.050, 9/9 <LOQ, 0 >BVL, max 0.050. Toothpaste LOQ 0.031, 6/9 <LOQ, 0 >BVL, max 0.040. Mouthwash LOQ 0.029, 6/6 <LOQ, 0 >BVL, max 0.029. Lip liner LOQ 0.018, 5/10 <LOQ, 2 >BVL, max 0.665. Sun protection LOQ 0.047, 9/9 <LOQ, 0 >BVL, max 0.047. Eye shadow LOQ 0.028, 1/9 <LOQ, 2 >BVL, max 1.053. Mascara LOQ 0.015, 8/9 <LOQ, 0 >BVL, max 0.074. Eye liner LOQ 0.011, 3/9 <LOQ, 0 >BVL, max 0.194. Foundation LOQ 0.015, 4/9 <LOQ, 1 >BVL, max 0.585.
  • As (Table 16b, p. 38; BVL general 0.5 mg/kg): lipstick LOQ 0.023, 1/9 <LOQ, 0 >BVL, max 0.283. Lip gloss LOQ 0.039, 6/9 <LOQ, 0 >BVL, max 0.090. Toothpaste LOQ 0.101, 8/9 <LOQ, 0 >BVL, max 0.140. Mouthwash LOQ 0.007, 6/6 <LOQ, 0 >BVL, max 0.007. Lip liner LOQ 0.060, 1/10 <LOQ, 1 >BVL, max 0.505. Sun protection LOQ 0.087, 9/9 <LOQ, 0 >BVL, max 0.087. Eye shadow LOQ 0.060, 0/9 <LOQ, 3 >BVL, max 1.620. Mascara LOQ 0.048, 2/9 <LOQ, 0 >BVL, max 0.255. Eye liner LOQ 0.060, 1/9 <LOQ, 0 >BVL, max 0.418. Foundation LOQ 0.045, 0/9 <LOQ, 0 >BVL, max 0.144.
  • Cd (Table 16c, p. 39; BVL general 0.1 mg/kg): no samples exceeded the BVL limit in any product group. Lipstick LOQ 0.003, max 0.015. Lip gloss LOQ 0.001, max 0.012. Toothpaste LOQ 0.005, max 0.046. Mouthwash LOQ 0.001, max 0.001. Lip liner LOQ 0.002, max 0.026. Sun protection LOQ 0.002, max 0.014. Eye shadow LOQ 0.004, max 0.045. Mascara LOQ 0.002, max 0.006. Eye liner LOQ 0.001, max 0.031. Foundation LOQ 0.001, max 0.018.
  • Pb (Table 16d, p. 39; BVL general 2.0 mg/kg, toothpaste 0.5 mg/kg, theatre 5.0 mg/kg): lipstick LOQ 0.012, 0 >BVL (general), max 1.447. Lip gloss LOQ 0.009, 0 >BVL, max 0.267. Toothpaste LOQ 0.026, 3/9 >BVL (toothpaste 0.5 mg/kg threshold), max 0.639. Mouthwash LOQ 0.007, 0 >BVL, max 0.050. Lip liner LOQ 0.009, 3/10 >BVL, max 2.545. Sun protection LOQ 0.168, 0 >BVL, max 0.248. Eye shadow LOQ 0.129, 6/9 >BVL, max 8.769. Mascara LOQ 0.011, 0 >BVL, max 0.859. Eye liner LOQ 0.014, 1/9 >BVL, max 3.780. Foundation LOQ 0.008, 0 >BVL, max 0.764. Across the full 91-sample set: 13/91 samples exceeded Pb 2 mg/kg, 4/91 exceeded Pb 5 mg/kg, 0/91 exceeded Pb 10 mg/kg.
  • Hg (Table 16e, p. 39 — note: the source mis-titles Table 16e as “Pb” but its column structure (LOQ values <0.002-<0.073 mg/kg vs the actual Pb LOQs <0.007-<0.168 mg/kg in Table 16d) and the BVL 0.1 mg/kg threshold for Hg general cosmetics, together with the discussion’s Hg-eye-shadow finding, confirm Table 16e is Hg; see Verification notes): lipstick LOQ 0.010, 0 >BVL, max 0.058. Lip gloss LOQ 0.073, 1/9 >BVL, max 0.101. Toothpaste LOQ 0.014, 0 >BVL, max 0.025. Mouthwash LOQ 0.002, 0 >BVL, max 0.002. Lip liner LOQ 0.003, 0 >BVL, max 0.071. Sun protection LOQ 0.008, 0 >BVL, max 0.008. Eye shadow LOQ 0.005, 3/9 >BVL, max 0.519. Mascara LOQ 0.005, 0 >BVL, max 0.009. Eye liner LOQ 0.006, 0 >BVL, max 0.050. Foundation LOQ 0.015, 1/9 >BVL, max 0.585 — but this entry is identical to the Sb-Foundation row in Table 16a (LOQ 0.015, 1 >BVL, max 0.585) and contradicts the Section 4.5 Hg-by-product chart (p. 37), which shows all 9 Foundation Hg values well below 0.1 mg/kg. The Foundation Hg row in Table 16e appears to be a copy-paste error from Table 16a Sb in the published report; see Verification notes.

LGC UK-market survey — headline category-level finding

Eye shadow drove the exceedances for every metal except cadmium: 1.053 mg/kg Sb (red eye shadow), 1.620 mg/kg As (purple eye shadow), 8.769 mg/kg Pb (mid-price red-purple eye shadow), 0.519 mg/kg Hg (one of two eye-shadow samples with detectable Hg, low-price red-purple). Lip liner was the second-most-contaminated category for Sb, As, Pb. Toothpaste exceeded the BVL toothpaste-specific Pb threshold (0.5 mg/kg) in 3 of 9 samples while remaining well below the general-cosmetics 2 mg/kg threshold. Highly-coloured products (orange-red, brown, red eye shadows and lip liners) were associated with the largest Pb and Sb signals, with the commonly-listed reddish pigment being CI 77491 (iron oxide) (Discussion, p. 41-42). Price and place-of-purchase (high-street vs online) did not appear to correlate with metal levels.

Discussion conclusion (Section 2.6 and Section 5)

OPSS concludes that the BVL guidance values, applied with the BVL’s own 90 % compliance criterion, could be applied to the UK market with minimal impact on the basis of this 91-sample exploratory survey (Section 5, p. 42). The report flags that the small sample size and exploratory scope require further confirmatory work with larger market coverage, and recommends extending the scope to chromium (Cr) and other elements detected in the semi-quantitative screen.

Evidence Fitness

This is a UK government feasibility study with a primary 91-sample analytical survey, full method validation against ISO 21392:2021, and an A-tier comprehensive international regulatory review. The analytical lab (LGC) operates under ISO 17034 and ISO 17043 accreditations for the relevant work. Evidence the report supports:

  • Reconstructable per-product-group occurrence dataset for Sb, tAs, Cd, Pb, tHg in 9 of 10 declared cosmetic groups on the UK market in 2022 (per-sample numerical results in Appendix 2 with the printed report; this page captures the group-level summary tables 16a-16e and the discussion findings).
  • Methods-fitness evidence: ISO 21392:2021 with HF acid adaptation is fit-for-purpose for the UK cosmetic matrix, including refractory SiO₂/TiO₂-containing samples.
  • Regulatory-context evidence: comprehensive 2023 snapshot of cosmetic heavy-metal regulations and non-regulatory guidance across UK/EU, US, China, Thailand, Japan, Canada, Germany, ASEAN, and ICCR.
  • Eye-shadow Pb finding (6/9 > BVL 2 mg/kg; max 8.8 mg/kg; no sample > 10 mg/kg FDA recommended industry max) is a direct UK-market observation pertinent to any UK or downstream HMT&C threshold discussion on adult eye-makeup category.

The report does not support exposure modelling of children’s cosmetic use; sample selection covered 10 adult-cosmetic categories not children’s products. The 91-sample size is too small to support percentile-distribution claims by product group (OPSS itself flags this as exploratory). The single-replicate worst-case reporting for 12 marginal cases means the per-product values for those entries are slight over-estimates rather than true duplicate means.

Methods (brief)

Sample selection by OPSS-defined purposive sampling: low, mid and high price ranges, ~70/30 high-street/online split, range of colours/shades, 10 cosmetic-product groups risk-rated by SiO₂/TiO₂ content and ingestion/eye/inhalation pathway (Table 1, p. 3); ratings 1 (lip and oral products) and 2 (eye and sun-protection products) were prioritised, makeup powders verified but not analysed due to time/budget constraints.

Sample preparation: matrix-specific homogenisation (lipstick heated 85 °C/1 h; lip gloss stirred; lip liner cut open with stainless-steel scalpel + dry-ice pestle-and-mortar pulverisation; mascara/eye liner disassembled with sharp tweezers; face powders crushed with pestle and mortar). Sample mass 0.1 ± 0.01 g.

Digestion: microwave acid digestion per ISO 21392:2021, two programmes selected by matrix (Tables 12-13, p. 29). Multiwave Go (Anton Paar, UK) Teflon vessels for all matrices except mouthwash (Discover SP-D, CEM Corporation, UK, with quartz vessels). Acids trace-element grade (SpA, Romil, UK).

Method validation: 8 reagent blanks + 1 duplicate spiked sample + QC material (lipstick + lip-gloss proficiency-testing materials from LGC Standards under ISO 17043, used at consensus values) or NIST SRM 2709a San Joaquin Soil (selected for its high SiO₂/TiO₂ content). Mouthwash validated by triplicate spiked samples only (no fit-for-purpose QC available). Acceptance criteria: linearity >0.99, internal-standard recovery 75-125 %, IQC standard recovery 80-120 %, triplicate-spike recovery 70-150 %, QC/CRM recovery 75-125 % (excluding mouthwash), duplicate RSD ≤20 %. Verification passed for every matrix.

Analytical determination: Agilent 7700x or 7900 collision/reaction-cell ICP-MS, helium collision mode for ⁷⁵As, no-gas mode for ¹¹¹Cd, ¹²¹Sb, ²⁰²Hg, ²⁰⁸Pb; Pb and Hg reported as sum of isotopes; ⁹⁵Mo correction for ¹¹¹Cd; ¹¹⁵In/¹²⁸Te/¹⁵⁹Tb/¹⁷⁵Lu internal standards (Table 15, p. 30). External calibration with ISO 17034-accredited standards (Romil, UK); IQC from independent supplier batch. Semi-quantitative scan identified Ba, Co, Cr, Cu, Li, Mo, Ni, Sn, Tl, V, Zr in Appendix 3 (p. 53); Zn additionally mentioned in Section 3.5.4 (p. 31) as an interference-source element observed during method verification.

Statistical handling: duplicate measurements averaged for graphical presentation; for the 12 instances where one replicate was <LOQ and the other above, the above-LOQ value reported as worst-case.

Speciation status: total elemental concentrations only. Arsenic measured as ⁷⁵As (total As, recorded as tAs). Mercury measured as ²⁰²Hg by ICP-MS (total Hg, recorded as tHg). No inorganic/methyl speciation. Chromium-hexavalent vs total chromium is NOT distinguished in the semi-quantitative Cr data (Cr, not Cr-VI).

Limitations OPSS acknowledges: small sample size (9-10 per product group), no analysis of makeup powders, exploratory scope, single-laboratory measurements, reliance on proficiency-testing-derived QC consensus values rather than certified reference materials (no matrix-matched cosmetic CRMs exist), reproducibility 15-80 % for Pb across labs in the ISO inter-laboratory trial (i.e., the method works for fit-for-purpose decisions but inter-lab comparisons require care).

Implications

  • Certification: this is the highest-quality publicly-available UK-government regulatory feasibility evidence on heavy-metal occurrence in adult-market cosmetics as of 2023. Use as the regulatory-context anchor for any Cat 13 (Cosmetics — Leave-on) row that maps to the categories surveyed: eye-makeup, makeup-foundation-powders-blush, sun-suntan-products. Use as the regulatory-context anchor for Cat 20 (Oral Care) rows mouthwash-oral-rinse and any future adult-toothpaste row. The eye-shadow Pb finding (6/9 > 2 mg/kg, max 8.8 mg/kg) is the strongest direct-evidence signal in the corpus for Cat 13 Row 7 (eye-makeup).
  • Courses: regulatory-literacy module reference. The 20× range between EU Pb 0.5 ppm (colour-additive level) and Health Canada Pb 10 mg/kg (impurity-action level), and the ICCR 1 ppm Hg / 10 ppm Pb policy-driven target levels, illustrate that “cosmetic heavy-metal limits” are not a single number but a layered set of impurity, colour-additive, ingredient, and finished-product caps that resolve differently by jurisdiction.
  • App: not directly relevant to the food-app contamination-profile workflow; this is a cosmetics paper.

Update history

No substantive edit history is available in this build. The full commit record is available in git.

Referenced by