Status Report: Review of Metals in the Toy Safety Standard, ASTM F 963

Overview

This U.S. Consumer Product Safety Commission (CPSC) staff status report, dated 14 March 2012 and prepared by Dominique W. Johnson, MPH (CPSC Directorate for Health Sciences, Division of Health Sciences), is the federal regulator’s formal Section 106(d)(1) Consumer Product Safety Improvement Act (CPSIA) assessment of whether the migration limits for seven heavy metals in ASTM F 963-07 §4.3 (antimony, arsenic, barium, cadmium, chromium, mercury, selenium; lead is regulated separately under CPSIA §101) remain adequate to protect children. The 172-page bundle contains (a) the CPSC staff memorandum and conclusion, (b) Tab A: the Versar Inc. / SRC Inc. final contractor report “Review of Toxicity Data and Assessment of Seven Metals” dated 16 July 2010 under Contract CPSC-D-06-0006 Task Order 008, covering arsenic, chromium, mercury, and selenium toxicity reviews based on a 2000-2010 literature search of ~13,000 references, (c) Tab B: CPSC staff-prepared toxicity assessments for antimony, barium, and cadmium (the three metals for which Versar found insufficient new data or that CPSC handled in-house), (d) Tab C: peer-review comments from five reviewers (two Versar-selected, three CPSC-selected from U.S. federal agencies) and CPSC’s responses (61 comments total), and (e) Appendix A: tables of U.S. federal-agency standards (US EPA RfD/MCL/RSD/RSC, NIOSH, OSHA, FDA) and a comparison to the EU 2009 Toy Safety Directive intake limits (effective 2013). The bundle re-derives acceptable daily intakes (ADIs) for all seven metals and compares the resulting toy-intake limits against the existing ASTM F 963-07 standard. Notwithstanding that the staff-derived ADIs imply intake limits that are 5- to 500-fold lower than the existing standard for arsenic, cadmium, chromium, mercury, and selenium (i.e., the existing limits are weaker than recent toxicity data would support), and 5- to 35-fold higher than the existing standard for antimony and barium (i.e., the existing limits are tighter than recent toxicity data require), CPSC’s formal staff recommendation is that no change be made to the migration limits in ASTM F 963-07 §4.3, on the grounds that the existing limits are sufficiently protective for children and that the F 963-11 amendment already expanded the standard’s scope from paints/surface coatings to all substrate materials (plastics, metal, glass, ceramic toys, and parts of toys), which CPSC supports.

Key numbers

Existing ASTM F 963-07 §4.3 migration limits (Table 2-1; mg/kg in soluble migrated element from toy paint and surface-coating material)

The ASTM F 963-07 limits, identical to the EU EN 71-3 (1994) values that they were derived from, applied at the time of this report only to paints and surface coatings. F 963-11 (approved by ASTM 1 December 2011, published shortly thereafter) expanded the scope to substrate materials (plastics, metal, glass, ceramic toys) and parts of toys intended to be mouthed, intended for children under 6 years, or that are small parts per 16 CFR Part 1501.

Antimony (Sb) 60 mg/kg
Arsenic (As) 25 mg/kg
Barium (Ba) 1000 mg/kg
Cadmium (Cd) 75 mg/kg
Chromium (Cr) 60 mg/kg
Mercury (Hg) 60 mg/kg
Selenium (Se) 500 mg/kg

EU EN 71-3 separately specifies tighter limits for modeling clay and finger paint: As 25 mg/kg, Cd 50 mg/kg, Cr 25 mg/kg, Hg 25 mg/kg, others unchanged (Appendix A Table 1). Lead is regulated by CPSIA §101 at 100 mg/kg total content for children’s products ≤12 years of age (content limit, not migration limit; Appendix A Table 1).

Derivation assumption used by ASTM F 963-07 and EN 71-3 (Section 2.1, Versar report)

The legacy limits assume a child ingests 8 mg/day of toy material (mean value), uses a 12 kg child body weight (about 26.5 lb, approximately 1 year old per Ogden et al. 2004), and assumes adult weekly dietary intake reduced by 50 % as the child estimate; toys are then allowed to contribute between 0.1 % and 10 % of the resulting child dietary intake (the per-metal allowable relative contribution depends on the metal’s toxicity — 0.1 % for known carcinogens like arsenic and chromium, 5–10 % for less toxic metals like barium and selenium).

Existing F 963-07 daily intake limits (µg/day from toys, 12-kg child) — main memo Table 1

Antimony (Sb) 0.2
Arsenic (As) 0.1
Barium (Ba) 25
Cadmium (Cd) 0.6
Chromium (Cr) 0.3
Mercury (Hg) 0.5
Selenium (Se) 5

These map back to the migration limits via the 8 mg/day toy-ingestion assumption.

CPSC-derived ADIs (Versar 2010 contractor + CPSC staff)

For each metal, the report identifies a point of departure (POD; NOAEL or LOAEL) from the most sensitive critical-effect study identified in the 2000-2010 literature, then applies an uncertainty factor (UF) per established U.S. EPA / ATSDR practice.

Antimony (Sb) — POD: 90-day rat drinking-water NOAEL of 6 mg Sb/kg/day for reduced body weight (NTP 1992, dose group 50 ppm). UF = 1000 (10 inter-species × 10 intra-species × 10 subchronic-to-chronic). ADI = 6 µg Sb/kg/day. (Tab B, Antimony Toxicity Assessment.)
Arsenic (As, inorganic) — POD: chronic human LOAEL of 0.001 mg inorganic arsenic (iAs)/kg/day for increased risk of skin lesions in adults in the Ahsan et al. 2006 Bangladesh drinking-water cross-sectional study (n = 11,746). UF = 30 (10 LOAEL→NOAEL × 3 human variability; the partial factor of 3 rather than 10 for sensitive individuals is consistent with U.S. EPA 1993 and ATSDR 2007 treatment of the older arsenic dataset). ADI = 3 × 10⁻⁵ mg iAs/kg/day (0.03 µg/kg/day). (Versar Section 3.4.)
Barium (Ba) — POD: 2-year mouse drinking-water NOAEL of 60 mg Ba/kg/day for nephropathy in male mice (NTP 1994). ATSDR (2007) intermediate and chronic minimal risk level (MRL) = 0.2 mg Ba/kg/day, derived using UF = 100 (10 × 10) plus a modifying factor of 3 for inadequate developmental toxicity data. CPSC main memo Table 1 adopts the ATSDR MRL of 0.2 mg/kg/day as the ADI; CPSC Tab B reverts the modifying-factor-of-3 adjustment to derive a “proposed ADI” of 0.6 mg Ba/kg/day. See verification notes for the internal inconsistency between the two parts of the bundle. (Tab B, Barium Toxicity Assessment.)
Cadmium (Cd) — POD: lower confidence-limit benchmark dose (BMDL₀.₀₅) of 0.5 µg Cd/g urinary creatinine for kidney effects in Swedish women (Suwazono et al. 2006); translated to an intake of 0.33 µg Cd/kg/day. UF = 3 for sensitive populations. ADI = 0.1 µg Cd/kg/day. (Tab B, Cadmium Toxicity Assessment, summarizing Williams 2010 CPSC review.)
Chromium, trivalent (Cr-III) — POD: 52-week Sprague-Dawley rat dietary NOAEL of 1.6 to 3.0 mg Cr/kg/day for niacin-bound chromium (III) (Shara et al. 2007). UF = 100 (10 × 10). Tentative ADI = 0.016 mg Cr-III/kg/day (1.6 × 10⁻²). Versar notes the tentative status reflects derivation from a free-standing NOAEL without supporting reproductive-toxicity data. (Versar Section 4.6.1.)
Chromium, hexavalent hexavalent chromium (Cr-VI) — POD: chronic 2-year B6C3F1 mouse drinking-water LOAEL of 0.4 mg Cr/kg/day for epithelial hyperplasia in the duodenum, histiocytic infiltration of the liver and mesenteric lymph node, and cytoplasmic alterations in the pancreatic acini (Stout et al. 2009 / NTP 2008b TR-546). UF = 1000 (10 inter-species × 10 intra-species × 10 LOAEL→NOAEL). ADI = 4 × 10⁻⁴ mg Cr-VI/kg/day. (Versar Section 4.6.2.)
Mercury (Hg, inorganic) — POD: 3-month male rat drinking-water LOAEL of 0.001 mg Hg/kg/day for degenerative histological changes in seminiferous tubules and epididymis (Penna et al. 2009; mercuric chloride). UF = 1000 (10 inter-species × 10 intra-species × 10 LOAEL→NOAEL). ADI = 1 × 10⁻⁶ mg iHg/kg/day. Versar notes additional UFs for subchronic-to-chronic and developmental-toxicity database gaps could be considered. (Versar Section 5.4.)
Selenium (Se) — POD: chronic human LOAEL of 200 µg Se/day (0.003 mg/kg/day) for increased risk of type 2 diabetes from a 7.7-year secondary analysis of the Nutritional Prevention of Cancer trial (Stranges et al. 2007; n = 1,202; HR = 1.55, 95 % CI 1.03–2.33). UF = 3 (LOAEL is close to the U.S. RDA of 55 µg/day and within the typical 30–220 µg/day U.S. dietary + supplement intake range, limiting the defensible UF). ADI = 1 × 10⁻³ mg Se/kg/day (1 µg Se/kg/day). (Versar Section 6.4.)

Main memo Table 1 — comparison of existing F 963-07 limits to derived intake limits using ASTM/EN 71-3 modifying factors (µg/day from toys, 12-kg child)

The comparison applies the same allowable-relative-contribution percentages and 12-kg body weight that were used to derive the existing F 963-07 limits, so the “Intake Limit” column is directly comparable to the existing standard.

Metal	Existing F 963 (µg/day)	CPSC ADI (µg/kg/day)	Intake Limit with F 963 modifying factors (µg/day)
Sb	0.2	6	7.2
arsenic (As)	0.1	0.03	0.00036
Ba	25	200	120
cadmium (Cd)	0.6	0.1	0.06
chromium (Cr)	0.3	0.4	0.05
mercury (Hg)	0.5	0.001	0.001
Se	5	1	1

The chromium ADI used in Table 1 is the Cr-VI ADI of 0.4 µg Cr-VI/kg/day (4 × 10⁻⁴ mg/kg/day), on the basis that chromium in toys is most likely to be hexavalent. (Versar Section 4.6.3.)

Main memo Table 2 — CPSC’s “preferred” intake limits using staff’s standard approach (ADI × 12 kg body weight, no relative-contribution allocation)

CPSC notes that staff’s normal practice when setting a daily intake limit is to use the ADI multiplied by body weight, without further dietary-allocation adjustment. The resulting limits are several orders of magnitude higher than the Table 1 modified-comparison values.

Metal	Existing F 963 (µg/day)	CPSC ADI (µg/kg/day)	Intake Limit by CPSC staff approach (µg/day)
Sb	0.2	6	72
arsenic (As)	0.1	0.03	0.36
Ba	25	200	2400
cadmium (Cd)	0.6	0.1	1.2
chromium (Cr)	0.3	0.4	4.8
mercury (Hg)	0.5	0.001	0.012
Se	5	1	12

Appendix A Table 8 — EU 2009/48/EC Toy Safety Directive intake limits (effective 19 July 2013) for comparison (µg/day)

The 2009 EU Directive’s updated migration limits converted to intake limits using the same 8 mg/day toy ingestion and 12-kg body weight assumption.

Metal	Current ASTM F 963 (µg/day)	EU 2009 Directive (µg/day, effective 2013)
Sb	0.2	0.37
arsenic (As)	0.1	0.03
Ba	25	37
cadmium (Cd)	0.6	0.02
chromium (Cr)	0.3	0.31 (Cr-III); 0.0001 (Cr-VI)
mercury (Hg)	0.5	0.06
Se	5	0.31

The 2009 Directive includes significantly tightened intake limits for cadmium, mercury, selenium, and a separate dedicated limit for hexavalent chromium not present in F 963-07.

Recent migration-test data identified by Versar (Section 2.2)

Versar reported that the 2000-2010 literature search yielded only three studies measuring metal migration from toys; “minimal data” were found, paralleling the same observation by RIVM (2006). Each is from outside the United States.

Kawamura et al. 2006 (Food Hyg. Saf. Sci. 47:51-57; Japan, summary from English abstract because article text is Japanese-only): 45 PVC baby toys + 10 paints, screened for eight harmful elements. Substrate concentrations: barium 0.3–3700 mg/kg (all samples), cadmium 0.2–26 mg/kg (several samples), chromium 0.5–280 mg/kg (several samples), lead 1.5–1300 mg/kg, antimony 5.3 mg/kg (1 sample). Migration testing per ISO 8124-3 (Migration of Certain Elements) gave migration values lower than the ISO 8124-3 limits for Ba, Cd, Cr, and lead (Pb).
Kawamura et al. 2009 (Food Hyg. Saf. Sci. 50:93-96; Japan, English abstract): vinyl chloride resin enamel and acrylic resin enamel containing 1000 mg/kg Cd and Pb on a dried basis, painted on glass plates and soaked in water or migration solvent. In water, both Cd and Pb were <0.1 µg/mL (limit of determination). In 4 % acetic acid or 0.07 mol/L HCl, 0.3–2.3 µg/mL Cd and Pb migrated from the acrylic enamel but no migration was observed from the vinyl chloride enamel. With the ISO scratched-paint test method (paint scratched from glass, soaked in 0.07 mol/L HCl at 37 °C for 1 hour ± shaking), Cd and Pb migration reached 310 to 910 mg/kg, i.e., 3.5- to 12-fold above the migration limits. Cadmium migrated more extensively than lead, and both migrated more readily from acrylic enamel than from vinyl chloride enamel.
Kumar & Pastore 2006 (Toxics Link, India): 111 non-branded PVC and soft toys randomly purchased in Delhi, Mumbai, and Chennai. Of these, 77 PVC and 11 non-PVC samples were analyzed for cadmium and lead at the National Accreditation Board for Testing and Calibration Laboratories. Both metals were present in all analyzed samples. Cadmium average concentrations ranged 15.71–26.53 ppm across product groups; maximum range 11.6–188 ppm; minimum range 0.03–0.16 ppm.

Comparison of derived intake limits to existing F 963-07 limits (Section IV and Conclusion)

Tighter than existing (toxicity data support lower limits): arsenic (≈ 300×), mercury (≈ 500×), cadmium (10×), chromium (≈ 6×), selenium (≈ 5×).
Looser than existing (toxicity data permit higher limits): antimony (≈ 35×), barium (≈ 5–14× depending on whether the ATSDR MRL or the CPSC-staff-derived 0.6 mg/kg/day ADI is used).

Staff recommendation (Section IV)

CPSC staff finds that “the existing intake limits in ASTM F 963-07 and EN 71-3 are sufficiently protective of children who use toys that conform to the current standard” and recommends no change to the migration limits in ASTM F 963-07 §4.3 at this time. The staff position rests on three rationales:

The F 963-07 limits are lower (more protective) than the ADI-derived intake limits when staff’s standard ADI × body-weight approach is used (Main memo Table 2 comparison).
CPSC’s limited testing data show that mercury or mercury compounds are not currently being used in the manufacture of children’s toys, so tightening the Hg limit would not increase protection in practice.
The F 963-11 amendment (December 2011) already expanded the scope of §4.3 from paints/surface coatings to all toy substrate materials (plastics, metal, glass, ceramic toys, parts of toys intended to be mouthed, parts intended for children under 6, and small parts per 16 CFR Part 1501), aligning more closely with EN 71-3’s broader scope. Staff actively encouraged this expansion through the ASTM F15.22 toy safety subcommittee working group.

Staff recommends continued cooperation with ASTM on improvements to the toy safety standard and continued monitoring of changes in international standards.

Methods (brief)

Literature search (Versar contractor). PubMed, ATSDR toxicological profiles and chemical evaluation documents, and TOXLINE were searched for English-language data from 2000 to 2010 on acute and chronic (repeated-exposure oral) toxicity, migration from products, toxicological effects, and suggested maximum daily intakes for inorganic forms of antimony, arsenic, barium, cadmium, chromium, mercury, and selenium (lead was excluded because of separate CPSIA §101 treatment). Results were grouped in metal-specific EndNote databases as “All References” (n ≈ 13,000 total; per-metal counts in Table 1-1), “May Impact” (potentially affecting ADI; per-metal counts also in Table 1-1), and “Key Studies” (selected for use in ADI derivation; per-metal counts also in Table 1-1).

ADI derivation approach. For each metal, the Versar contractor (As, Cr, Hg, Se) or CPSC staff (Sb, Ba, Cd) identified a point of departure (NOAEL or LOAEL from the most sensitive critical-effect study) and applied uncertainty factors per established U.S. EPA / ATSDR practice: 10× for inter-species extrapolation, 10× for intra-species (human variability), 10× for LOAEL→NOAEL extrapolation when no NOAEL was identified, 10× for subchronic-to-chronic extrapolation when only subchronic data were available, and additional 3–10× modifying factors for database deficiencies. The arsenic, mercury, and chromium-VI ADIs were derived from LOAELs (NOAEL not identified), so each carries a LOAEL→NOAEL factor; the arsenic ADI uses a partial factor of 3 (rather than 10) for human variability, consistent with U.S. EPA 1993 IRIS and ATSDR 2007 treatment of the older arsenic dataset.

Comparison-to-existing-standard calculations. Two parallel calculations were performed for each metal. The “Table 1” comparison applies the same modifying factors (allowable-relative-contribution percentage and 12-kg body weight) that the EU 1985 Scientific Advisory Committee used when deriving the EN 71-3 limits, so the resulting “intake limit” is directly comparable to the existing F 963-07 standard. The “Table 2” comparison uses staff’s standard approach (ADI × 12-kg body weight, no allocation), yielding intake limits that are 100-1000× higher than the Table 1 values for most metals. The discussion explicitly acknowledges that the dietary-allocation approach and the risk-based approach produce different values and that the goal of the comparison is to facilitate direct comparison with the existing standard, not to endorse one approach over the other.

Migration-test data review. The Versar literature search also looked for migration-test studies (i.e., laboratory measurement of metal release from toys under simulated mouthing or chewing conditions), per Section 2.2. Three studies were located; methodological detail is summarized in Key numbers above. CPSC staff explicitly notes that the lack of migration data is itself a finding: “Staff believes that additional testing will be required to identify the use of metals in children’s toys and to quantify the potential migration of the metals from children’s toys” (Section II).

Peer review. CPSC staff initiated an external scientific peer review in February 2011 covering both the CPSC staff document and the Versar contractor report. Two reviewers were selected by Versar; three reviewers were selected by CPSC staff from other U.S. federal agencies. 61 comments were received across reviewers, organized in Tab C by topic area (general information and understanding, key studies and endpoints used, comparison of dietary and risk-based approaches, use of uncertainty factors, relating findings to recommended dietary amounts, and other current heavy-metals standards). Staff revised the CPSC staff metals document (Tab B) in response to peer-review comments; the Versar contractor report did not require revisions.

Toy material categories covered. ASTM F 963-07 §4.3 at the time of report applied only to paints and surface coatings on toys. F 963-11 expanded the scope to plastics, metal, glass, and ceramic toys and parts of toys; the report’s analysis covers both the §4.3 paints/coatings scope and the F 963-11 expanded substrate scope. The report does not separately address specific toy categories (teethers, pacifiers, rattles, soft toys, painted wood, metal jewelry, ceramic tea sets, etc.); the analysis is at the toy-materials level rather than the toy-product-category level.

Implications

Regulatory baseline (heavy metals in U.S. toy materials, post-CPSIA). This report is the federal regulator’s formal Section 106(d)(1) CPSIA evaluation of the ASTM F 963 metals limits. The status of the U.S. toy-metals regulatory floor immediately after the report’s publication is: lead at 100 mg/kg total content for any accessible substrate of a children’s product ≤12 years old (CPSIA §101); the other seven metals (Sb, As, Ba, Cd, Cr, Hg, Se) at the EN 71-3-equivalent migration limits in F 963-11 (effective then) for paints/surface coatings and substrate materials of toys intended to be mouthed, toys for children under 6, and small parts per 16 CFR Part 1501. CPSC explicitly declined to lower the metals migration limits despite finding that the recent toxicity literature would support 5–500× lower limits for five of the seven metals; the gap between literature-supported limits and the regulatory limits is the precise pattern that motivates third-party certification programs that ratchet below the federal floor.
Speciation in toy-metals regulation. The F 963-07 migration limits are written as total elemental concentrations (Cr, Hg, As) and do not differentiate species (Cr-III vs Cr-VI; iHg vs methylmercury (MeHg); iAs vs tAs). The Versar ADI derivations do differentiate species — chromium uses the Cr-VI ADI as the regulatory basis on the rationale that “chromium in toys is most likely to be hexavalent,” mercury uses an inorganic-Hg ADI on the rationale that “only inorganic forms are considered relevant for toy-related exposures,” and arsenic uses an inorganic-As ADI on the rationale that organic forms are very low-toxicity and not present in toys. The 2009 EU Directive (effective 2013) is the first regulatory standard for toys to formalize the Cr-III vs Cr-VI distinction (Appendix A Table 8: 0.31 µg Cr-III/day vs 0.0001 µg Cr-VI/day). Heavy Metal Index pages for toy-metals contamination should follow the Versar / EU 2009 speciation logic rather than the F 963-07 total-element logic.
Migration-data gap. Versar’s 2000-2010 literature search yielded only three migration-from-toys studies (two Japanese, one Indian). The 2006 RIVM Chemicals in Toys advisory report had already noted the same migration-data gap. The implication for any HMTc work on toys is that the literature base for back-calculating substrate content limits from migration limits is thin; substrate-content-vs-migration-rate relationships established for one toy material (e.g., scratched acrylic enamel, Kawamura 2009) are not necessarily generalizable across toy categories.
App. Not directly relevant to ingredient contamination_profile data because no food-matrix occurrence values are reported. Potentially relevant to a future toy-screening surface, where the migration-vs-content distinction (the 2009 Kawamura scratched-paint result of 310–910 mg/kg migration from a 1000 mg/kg painted substrate) is a key consumer-facing fact about how badly damaged toy paint can release metals to a chewing/mouthing child.
Courses. Useful primary regulatory-history document for the toy-metals regulatory-trajectory teaching module, particularly as a worked example of (a) the U.S. dietary-allocation regulatory derivation that produced the ASTM F 963-07 / EN 71-3 limits, (b) the parallel U.S. risk-based ADI × body-weight approach CPSC staff would otherwise apply, and (c) the deliberate policy decision to retain a regulatory floor that is weaker than the literature supports because the regulator judges the gap acceptable given limited migration data and current toy manufacturing practices. The 2009 EU Toy Safety Directive comparison (Appendix A Table 8) is the parallel European trajectory toward tighter, speciation-aware toy-metal limits.

Update history

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