Antimony (Sb)

This page draws on evidence from EPA and European Commission, together with 24 peer-reviewed studies and 11 other cited sources. The complete citation list is at the bottom of the page.

This page draws on the ATSDR toxicological profile, IARC Monograph Volume 131, the WHO drinking-water fact sheet and background document, the OECD SIDS profile for diantimony trioxide, the OEHHA public health goal, the NTP inhalation bioassay and Report on Carcinogens monograph, and the regulatory texts transcribed below. Occurrence in infant articles is synthesized product by product at Antimony in baby products. Where a source examined for this page does not print a figure, the gap is named rather than filled.

Overview

Antimony (Sb) is a metalloid, and the sources reviewed here treat its two oxidation states as toxicologically different. Trivalent antimony, Sb(III), includes antimony trioxide and antimony potassium tartrate. Pentavalent antimony, Sb(V), includes compounds used to treat leishmaniasis and the oxo-anion described as the form most likely to leach from antimony-bearing materials into drinking water. The World Health Organization chemical fact sheet, assessment date 2003, states that the form in drinking water is a key determinant of toxicity, that the antimony(V) oxo-anion is the less toxic form, and that the subchronic toxicity of antimony trioxide is lower than that of potassium antimony tartrate, the more soluble form.

The International Agency for Research on Cancer, announcing Monograph Volume 131 on 5 May 2023, classified trivalent antimony as probably carcinogenic to humans (Group 2A), on limited evidence for cancer in humans, sufficient evidence for cancer in experimental animals, and strong mechanistic evidence in human primary cells and in experimental systems. The same announcement classified pentavalent antimony as not classifiable as to its carcinogenicity to humans (Group 3), and stated that pentavalent antimony was evaluated for the first time. The Volume 131 general remarks, as posted on NCBI Bookshelf, record that antimony trioxide had previously been evaluated as possibly carcinogenic to humans (Group 2B) in 1989.

The 2023 classification is a classification of the trivalent class. Antimony trioxide is a trivalent compound, so it sits inside that class, but the announcement is not a sentence that names antimony trioxide alone, and this page does not read it as one.

Antimony is also not one of the eight elemental totals (lead, cadmium, total arsenic, total mercury, nickel, aluminium, total chromium, tin) that dominate food-contaminant panels. A measurement panel that stops at those eight does not report antimony.

At a glance

Three points matter most for anyone trying to interpret an antimony result.

First, the form matters. Trivalent antimony, which includes antimony trioxide, is IARC Group 2A, and pentavalent antimony is Group 3. The animal cancer evidence behind the 2023 classification comes from inhalation of antimony trioxide, while oral carcinogenicity remains weakly quantified, so the classification cannot be read as an oral or dermal potency.

Second, the limits that exist answer different questions. A drinking-water maximum of 0.006 mg/L, a migration limit of 0.04 mg/kg from plastic into food, toy-migration limits in milligrams per kilogram of toy material, and cosmetic impurity guidance of 5 ppm are written on different bases, and none of them is a content limit for a baby wipe.

Third, the evidence on infant articles is a set of screens, not a dose. Surveys have found antimony in wipe components, polyester textiles, and PET plastics, but no study examined here measures how much antimony crosses infant skin from a wipe or a garment.

Chemistry and forms

Elemental antimony and its compounds are not one analyte for hazard classification, and Volume 131 evaluated trivalent and pentavalent antimony separately. Section 4.1.2 of the antimony chapter nevertheless records interconversion between them. About 23.3 percent systemic conversion to trivalent antimony was observed after a single intramuscular dose of 5 mg/kg body weight in five healthy adults. In rhesus monkeys given repeated intramuscular injections, the plasma proportion of trivalent antimony rose from 5 percent on day 1 to 50 percent on day 9. The Working Group treated that interconversion as an evidence gap, not as a reason to collapse the two evaluations.

Antimony trioxide (Sb2O3) is the insoluble trivalent oxide that appears in the inhalation cancer studies and in the flame-retardant and plastics uses named by IARC. Antimony potassium tartrate is the soluble trivalent salt used in the oral studies behind the WHO tolerable daily intake and the ATSDR intermediate oral minimal risk level. The two compounds cannot be averaged into one dose. On the evidence available at its 2003 assessment, WHO states that antimony trioxide, because of low bioavailability, was genotoxic only in some in vitro tests and not in vivo, whereas soluble antimony(III) salts exerted genotoxic effects in vitro and in vivo.

Toxicology

ATSDR’s systematic review names the respiratory tract, heart, gastrointestinal tract, serum glucose, and the developing animal as the sensitive targets. The studies behind that list are predominantly of trivalent compounds, especially antimony trioxide and antimony potassium tartrate. Parenteral therapeutic doses produce a different picture (electrocardiogram changes, vomiting, myalgia), and they are not consumer-product doses.

Acute oral toxicity in the profile is anchored by forestomach ulceration in mice given antimony potassium tartrate (NTP 1992). The acute oral minimal risk level is 1 mg Sb/kg/day, from a NOAEL of 99 mg Sb/kg/day and an uncertainty factor of 100. Intermediate-duration oral exposure uses decreased serum glucose in female rats given antimony potassium tartrate (Poon et al. 1998), and the published intermediate oral minimal risk level is 0.0006 mg Sb/kg/day, with an uncertainty factor of 100. Chronic oral data were inadequate for a minimal risk level, and no chronic oral MRL is stated.

The NOAEL behind the intermediate value needs one note. The profile footnote states it as 0.06 mg Sb/kg/day, and the UK Committee on Toxicity, in its 2024-2025 discussion of a drinking-water guidance value, reads the same ATSDR choice as 60 µg Sb/kg body weight per day. An extracted table cell in the profile text is ambiguous between 0.06 and 0.064 mg Sb/kg/day. This page uses 0.06 mg Sb/kg/day, the footnote and COT reading, and does not adopt 0.064 as a second value.

By inhalation, the acute minimal risk level is 0.001 mg Sb/m3, based on squamous metaplasia of the epiglottis in mice exposed to antimony trioxide (NTP 2016), and the intermediate-duration inhalation value is adopted from that acute value. The chronic inhalation minimal risk level is 0.0003 mg Sb/m3, based on chronic lung inflammation in female rats exposed to antimony trioxide (Newton et al. 1994). Occupational series summarized in the profile report cough, upper-airway inflammation, pneumoconiosis, and bronchitis in workers exposed to antimony oxides, often with co-exposure to arsenic, silica, or other metals. Those series do not describe general-population inhalation risk.

Cardiovascular effects appear in two forms. Electrocardiogram changes are a repeated finding after injected therapeutic antimony. The profile also summarizes observational associations between urinary antimony and high blood pressure in NHANES analyses, alongside associations for several other metals. Those urinary studies do not identify a product source, and they do not establish that antimony caused the blood-pressure difference.

The developing animal is one of ATSDR’s named sensitive targets. Occupational reports of menstrual disturbance and lower infant weight gain are in the profile, and they are not isolated from co-exposures. No developmental study at a dose relevant to wipe or textile contact was examined for this page, so no infant-product developmental dose is stated.

Genotoxicity follows the valence split. On the WHO 2003 reading, antimony trioxide was genotoxic in some in vitro tests and not in vivo, while soluble antimony(III) salts were genotoxic in vitro and in vivo. IARC’s 2023 Group 2A classification for trivalent antimony rests in part on strong mechanistic evidence in human primary cells, including DNA damage, chromosomal aberrations, micronuclei, and sister-chromatid exchange, as described in the Volume 131 general remarks. Pentavalent antimony was not given that classification.

Carcinogenicity classification

The WHO fact sheet still records the 1989 Group 2B conclusion for antimony trioxide, and Group 3 for antimony trisulfide, because its assessment date is 2003. For trivalent antimony as a class, the 2023 Volume 131 classification supersedes that reading. Oral carcinogenicity remains weakly quantified: ATSDR notes no tumor increase in lifetime oral studies of antimony potassium tartrate in rats and mice, and WHO states that there are no data to indicate carcinogenicity by the oral route. The inhalation classification must not be rewritten as an oral cancer potency.

NTP Technical Report 590 (December 2017) is the 2-year inhalation bioassay behind that animal evidence. Under the conditions of those studies there was some evidence of carcinogenic activity in male Wistar Han rats (combined alveolar/bronchiolar adenoma or carcinoma of the lung, and benign pheochromocytoma of the adrenal medulla) and some evidence in female Wistar Han rats (alveolar/bronchiolar adenoma of the lung, and benign or malignant pheochromocytoma of the adrenal medulla).

There was clear evidence in male B6C3F1/N mice (alveolar/bronchiolar carcinoma of the lung) and clear evidence in female mice (alveolar/bronchiolar adenoma and carcinoma, and malignant lymphoma). In the same report, alveolar/bronchiolar tumours from the rats carried Egfr mutations in 13 of 26 and a Kras mutation in 1 of 26. These are inhalation findings in rodents, not an oral potency and not a wipe dose.

Epidemiological health associations

Associations between antimony biomarkers and health outcomes in human populations have been examined in several NHANES analyses and prospective cohort studies. Because most studies use a single urine measurement, which has high within-person variability for antimony, and because they are cross-sectional or of modest size, the associations below are observational leads rather than established causal dose-response relationships. No blood or urinary antimony biomonitoring equivalent has been formally established by a regulatory agency.

A nested case-control study within a birth cohort in Lanzhou, China (N = 291, 97 cases of congenital heart defects and 194 controls), found that maternal blood antimony during pregnancy was significantly higher in mothers of infants with congenital heart defects than in controls (median 14.05 vs. 11.97 microg/L). High maternal blood antimony (above approximately 18.4 microg/L, the cutoff derived from the control distribution, not an established health-based threshold) was associated with 4.5-fold higher odds of any congenital heart defect (OR 4.456, 95% CI 2.829-9.782) and approximately 12-fold higher odds of isolated congenital heart defects (OR 12.648, 95% CI 2.806-56.953, p = 0.003). A restricted cubic spline showed a threshold-like, non-linear association rather than a monotonic dose-response. The study area is a mining-affected region with higher baseline antimony exposure than European or North American populations, and no food-specific antimony source was identified (Shao and colleagues 2026).

An analysis of NHANES 2005-2020 (N = 3,051) found that urinary antimony was significantly associated with higher odds of thyroid problems (OR 5.78, 95% CI 1.19-25.4) and sleep disorders (OR 1.80, 95% CI 1.08-22.1) specifically in females; no significant associations were found in males for any of 16 health outcomes tested (Wu and Yu 2026). A separate NHANES analysis (N = 10,012, 2005-2018) found no significant overall association between urinary antimony and trouble sleeping after adjustment; a marginal inverse association (OR 0.79) appeared at Q3 in males only (Ma and colleagues 2026). The confidence intervals on the thyroid finding are wide (95% CI spans an approximate 20-fold range), and the two sleep analyses used overlapping cycles but produced different results, so neither sex-specific finding should be treated as confirmed.

An analysis of NHANES 2003-2020 (N = 8,814) found a statistically significant association between urinary antimony at the third quartile and depression (adjusted OR 1.34, 95% CI 1.03-1.75, p = 0.029), with no significant association at Q2 or Q4. Restricted cubic spline analysis confirmed a non-linear dose-response (p for nonlinearity less than 0.001). Blood tungsten and blood cadmium were the dominant positive contributors in the weighted quantile sum analysis (Guo and colleagues 2026).

A NHANES 2003-2016 analysis (N approximately 4,363) found that urinary antimony was positively associated with chlamydia prevalence (adjusted prevalence ratio 1.20, 95% CI 1.01-1.44) and HPV 16/18 prevalence (PR 1.07, 95% CI 1.02-1.13); in females the associations were stronger (chlamydia PR 1.30, HPV 16/18 PR 1.27). No significant associations were found in males (Taylor and colleagues 2026).

A prospective cohort study in Mexico City (N = 422 children, blood metals measured at age 4, metabolic outcomes through age 10) found no statistically significant independent antimony effect on fasting glucose, insulin, HOMA-IR, or HOMA-B in single-metal models. In mixture analyses, antimony contributed a small positive weight to fasting glucose and a notable negative weight to beta-cell function (HOMA-B), where it was the second-largest negative contributor after zinc (Gigot and colleagues 2026).

A consistent pattern across multiple NHANES analyses is that antimony health associations are more frequently detected in females than in males, biologically plausible given that antimony activates the estrogen receptor and disrupts thyroid hormone signaling at concentrations found in PET food-contact materials (Wang and colleagues 2026).

Typical exposure routes

The IARC Volume 131 announcement lists uses in flame retardants, lead-acid batteries, lead alloys, plastics, brake pads, clutch discs, glass and ceramics, and ammunition primers, and some pentavalent compounds are used to treat leishmaniasis. The U.S. Environmental Protection Agency’s National Primary Drinking Water Regulations page lists discharge from petroleum refineries, fire retardants, ceramics, electronics, and solder as drinking-water sources of antimony. The WHO fact sheet adds that elemental antimony forms hard alloys with copper, lead, and tin, and that there is little evidence that replacing lead with antimony in solder contributes significantly to drinking-water concentrations.

Diet, drinking water, and bottled water are the oral routes, and they carry the regulatory numbers set out below. Inhalation is the route of the occupational lung findings and of the NTP inhalation bioassay summarized by ATSDR. Dust and brake wear are implied by the IARC use list (brake pads, clutch discs) and by occupational airborne exposure in the ATSDR profile, but no house-dust survey with an antimony concentration was examined for this page, so no dust concentration is stated.

Dermal contact is a plausible route for textiles, wipes, and diapers, and the evidence on it is indirect. ATSDR describes occupational dermatosis in workers exposed to airborne antimony that dissolved in sweat (Stevenson 1965), more often in heat. That is sweat acting on airborne dust, not a study of wipe lotion. The OECD SIDS Initial Assessment Profile for diantimony trioxide (SIAM 27, 14 to 16 October 2008) states that dermal absorption is considered negligible on the basis of an in vitro human-skin study that showed 0.26 percent absorption.

The California Office of Environmental Health Hazard Assessment, in its September 2016 public health goal for antimony in drinking water, names that study as Roper and Stupart 2006, an unpublished contract-laboratory report. It prints two results from six skin samples (one abdominal and five breast): 0.26 percent at 100 µg/cm2 and 0.14 percent at 300 µg/cm2 after 24 hours. The two agencies read the study differently. OECD prints only the 0.26 percent figure and calls the absorption negligible; OEHHA prints both doses and does not, in that paragraph, call the fraction negligible. The primary report is not in this corpus. Neither figure is an absorption fraction for trace antimony on infant skin, and neither comes from a wipe-lotion study.

PET-bottled water and polyester textiles are exposure hypotheses that fit the use pattern. The only bottled-water concentration examined here is the Romanian survey described under Occurrence, and it is not paired with a bottle-wall measurement. The Borca bottle-wall concentrations are totals in the plastic, not concentrations in the water.

Absorption, distribution, and excretion

The ATSDR Toxicological Profile for Antimony and Compounds (October 2019) describes gastrointestinal absorption as poor and form-dependent. Animal estimates for tartrate and trichloride are about 2 to 7 percent, and a dog study of a single gavage dose of pentavalent meglumine antimoniate estimated 10 percent bioavailability. The profile adopts the International Commission on Radiological Protection (1981) reference values for human gastrointestinal absorption: 10 percent for antimony tartrate and 1 percent for all other forms. Those are reference values for dose reconstruction, not measured absorption from a wipe or a meal.

Dermal absorption of a wipe residue or a textile residue has not been measured in the consumer studies examined here. ATSDR reads occluded application of antimony trioxide that killed rabbits as evidence that at least some forms can cross skin (Myers et al. 1978). A repeated dermal study of antimony trioxide in an artificial-sweat paste also killed rabbits, but the application was not occluded, and ATSDR excluded it from the levels-of-significant-exposure table because the animals may have ingested the paste (Fleming 1938). The only numeric percutaneous fractions are the OECD and OEHHA summaries of the unpublished Roper and Stupart 2006 in vitro study: 0.26 percent and, in OEHHA only, 0.14 percent.

OECD also prints oral absorption of diantimony trioxide in rats of 0.3 percent at 100 mg/kg body weight and 0.05 percent at 1,000 mg/kg body weight, and states that 1 percent is assumed at relevant human exposure levels. Those trioxide figures are not the ATSDR and ICRP gastrointestinal values, and they are not averaged with them. OECD prints an oral LD50 above 20,000 mg/kg body weight in rats, a dermal LD50 above 8,300 mg/kg body weight in rabbits, and a 4-hour nose-only LC50 above 5.20 mg/L, and it treats acute toxicity by all three routes as low despite poor study quality for the oral and dermal tests.

Distribution is wide. The profile’s chapter summary states that antimony is distributed throughout the body, with the highest concentrations in the lungs, gastrointestinal tract, red blood cells, liver, kidney, bone, spleen, and thyroid. Antimony is not metabolized in the sense of an organic metabolite, but the profile and the IARC remarks both record interconversion between pentavalent and trivalent forms. Excretion splits by valence: trivalent antimony is predominantly excreted in feces, with smaller amounts in urine, and pentavalent antimony is primarily excreted in urine. Parenteral half-lives from leishmaniasis treatment are not oral half-lives, and this page does not collapse them into one number.

Biomonitoring reference ranges

Population-level biomonitoring studies provide reference distributions for blood, urine, and hair antimony. No health-based blood or urinary antimony biomonitoring equivalent has been formally established; the distributions below are descriptive references, not action thresholds.

Saez and colleagues 2026 measured whole-blood antimony in 541 working adults in Valencia Region, Spain (2021) and found a detection frequency of 99%, a median of 2.36 microg/L (geometric mean 2.18 microg/L, P95 3.36 microg/L, maximum 19.7 microg/L) at a limit of quantification of 1.08 microg/L. Statistically significant exposure correlates in multivariable regression were time spent with windows open and agricultural activity (Saez and colleagues 2026).

Shao and colleagues 2026 measured maternal blood antimony during pregnancy in 194 controls in Lanzhou, China (a mining-affected region) and found a median of 11.97 microg/L (mean 13.05 microg/L, P25 8.80, P75 16.75), approximately five times the Spanish working-adult median (Shao and colleagues 2026). Gigot and colleagues 2026 found a median blood antimony of 4.9 microg/L (IQR 4.3-5.7, 100% detection) in 422 Mexican children at age 4 in Mexico City (Gigot and colleagues 2026).

Whaibeh and colleagues 2026 measured cord blood antimony in 74 Lebanese neonates and found a detection frequency of 98.6% at a LOD of 0.01 microg/L, with a median of 0.04 microg/L (P90 0.07 microg/L, maximum 0.64 microg/L) (Whaibeh and colleagues 2026). Cord blood concentrations are substantially lower than maternal peripheral blood, consistent with limited transplacental antimony transfer.

Multiple NHANES analyses in this deposit report broadly consistent urinary antimony reference distributions in US adults. Creatinine-corrected medians cluster at approximately 0.04-0.05 microg/g creatinine (Ma and colleagues 2026); raw geometric means are 0.04-0.07 microg/L depending on sex and age group, with males consistently above females (Wu and Yu 2026, Taylor and colleagues 2026).

Ortiz-Robles and colleagues 2026 measured scalp hair antimony in 96 mother-newborn pairs in Mexico City at birth: maternal hair median 18.0 ng/g (P25 12.2, P75 29.0, maximum 2870 ng/g dry weight, detection 98.9%), newborn hair median 1.0 ng/g (P75 1.9, detection 96.3%). The mother-newborn correlation for antimony in hair was not statistically significant, unlike the significant correlations for lead, arsenic, and mercury (Ortiz-Robles and colleagues 2026). Pena-Fernandez and colleagues 2026 conclude from a critical review of paediatric hair biomonitoring that scalp hair antimony is a context-specific indicator subject to discrete deposits from product contact, and that urine or plasma provide more interpretable systemic indicators in most settings (Pena-Fernandez and colleagues 2026).

Occurrence in water, plastics, textiles, and children’s products

Water

The WHO fact sheet, assessment date 2003, states concentrations in groundwater less than 0.001 µg/L, in surface water less than 0.2 µg/L, and in drinking water appearing to be less than 5 µg/L. It states that conventional treatment does not remove antimony, that antimony is not normally a raw-water contaminant, and that the most common drinking-water source appears to be dissolution from metal plumbing and fittings. Those are the fact sheet’s own occurrence statements, not a new survey.

The 2003 background document prints a different range: groundwater and surface water normally from 0.1 to 0.2 µg/L, and drinking water appearing to be less than 5 µg/L. Because the two documents do not print the same groundwater figure, this page keeps both.

Ungureanu and colleagues 2022 measured antimony in 50 bottled waters from the Romanian market, from below 0.07 µg/L to a printed maximum of 0.64 ± 6.85 µg/L, with a detection limit for antimony of 0.07 µg/L. The printed uncertainty on the maximum is larger than the maximum itself, and this page preserves that figure as printed rather than correcting it. The result is a Romanian market measurement. It is not a U.S. benchmark, and it is not, by itself, a demonstration that the antimony came from the bottle wall.

PET bottles that are not identified as baby bottles do release antimony into water and beverages in the studies examined here, most clearly at elevated temperature. Carneado and colleagues 2015 found no migration into mineral water at 4 °C and 20 °C, a significant increase at 40 °C that stayed below the EU limit of 5.0 µg/L, and, at 60 °C, migration after 30 days that exceeded that limit, with both Sb(V) and Sb(III) detected. A short-term stress test in the same paper found day-0 totals of 0.3 to 0.7 µg/L and 2.6 to 4.2 µg/L after storage at 60 °C.

Allafi 2020, in commercial bottled water, found no significant change after three months at 25 °C and, after heating to 50 °C, 8.530 ppb at 24 hours and 16.8 ppb at 7 days.

Molaee Aghaee and colleagues 2014 found bottled-water antimony still below a 6 ppb maximum contaminant level at room temperature, outdoors, and 40 °C, and above that level at 65 °C and 80 °C for longer storage. Mehdar 2025 (Water SA) prints bottled-water antimony from 0.02 to 2.14 µg/L across 25 to 60 °C over 200 days, with the maximum of 2.14 µg/L at 60 °C after 200 days, above a Japanese limit of 2.00 µg/L that the paper cites. The same abstract calls 0.844 µg/L a concentration in the PET material of one bottle; that unit does not match a weight percent, and it is not the 2.14 µg/L water result. None of these bottles is identified as a baby bottle.

Carrasedo and colleagues 2023 (Molecules 28(20):7166, DOI 10.3390/molecules28207166) extended this picture to fruit-juice matrices. They measured total antimony and speciated antimony (Sb(III) and Sb(V)) in commercial juices (apple, grape, pear, pineapple, peach, orange, tomato) and mineral waters stored in PET bottles, using four PET container types. Under ambient storage (20 degrees C), migration into all matrices was minimal over 30 days. At 60 degrees C for 30 days, peach juice in PET water bottles reached approximately 3-5 microg/L and pineapple juice reached approximately 7-15 microg/L; juice-specific PET containers produced lower migration than general-purpose PET water bottles for the same beverage. The initial speciation (Sb(V) only) shifted to Sb(III)-dominated at elevated temperatures in juice matrices, Sb(V)-citrate complexes appeared in pineapple juice at approximately 1-1.5 microg/L at 60 degrees C, and standard acetic-acid food simulants used in EU food-contact testing may underestimate actual migration into acidic juice matrices because they do not reproduce this complexing chemistry. The limits of detection were 0.01 microg/L for Sb(V) and 0.03 microg/L for Sb(III). None of the beverage or water samples is identified as a product prepared for infant consumption.

Plastics and textiles

Catalyst residue from polyester and polyethylene terephthalate (PET) is the pathway named for baby wipes on Baby wipes, and total antimony in screened components was measured by Washington Ecology Publication 14-04-014. Borca, Huthwelker, and Filella 2025 (Journal of Environmental Sciences 153:109-117, DOI 10.1016/j.jes.2024.02.006) examined what form that antimony takes. They measured six PET bottles by portable X-ray fluorescence at 351, 331, 361, 322, 370, and 370 mg/kg, and identified the antimony as organically bound, consistent with antimony glycolate or antimony acetate. The same spectra exclude a significant amount of antimony trioxide nanoparticles in those bottles, within the detection limits of the method. The methods paragraph says five PET bottles and the table prints six; this page follows the table.

Three polyester textiles in the same study measured 286, 471, and 452 mg/kg. One textile spectrum resembled senarmontite (an antimony trioxide), and the other two could not be fitted reliably. Electrical objects in the paper were senarmontite at 55,835, 27,812, and 33,122 mg/kg, plus a ring at 28,193 mg/kg. Bromine was measured in the three electrical objects (84,075, 48,712, and 71,949 mg/kg) and not in the ring, and the bromine-bearing objects match a halogenated flame-retardant synergist rather than a catalyst residue. These are consumer plastics and textiles, not samples identified as baby bottles or infant sleepwear.

Biver, Turner, and Filella 2021 (Regulatory Toxicology and Pharmacology 119:104824, DOI 10.1016/j.yrtph.2020.104824) measured both content and release on six polyester textiles bought in the European Union. Portable X-ray fluorescence totals were 125, 185, 286, 357, 452, and 471 µg/g, which the abstract rounds to about 125 to 470 µg/g. Artificial-sweat extraction by ISO 105-E04, three sequential 24-hour extractions at 37 °C, gave 2.572 ± 0.031, 0.904 ± 0.121, 0.480 ± 0.126, 2.477 ± 0.093, 0.402 ± 0.128, and 0.689 ± 0.154 µg/g for those six fabrics, in the same order as the totals.

The abstract summarizes default-parameter mobilization as about 0.05 to 2 percent of total antimony, or 0.1 to 1 µg/g. Two of the table extracts, 2.572 and 2.477 µg/g, sit above that 1 µg/g summary, and this page keeps the table. The result is release into a simulant. It is not a dermal absorption fraction, and the samples are not identified as infant sleepwear.

The paper states that the EU Ecolabel for textile products should not exceed 260 µg/g antimony in polyester fibres, tested on raw fibre before wet processing, with a derogation for polyester from recycled PET bottles, and that the OEKO-TEX STANDARD 100 label requires extractable antimony below 30 µg/g for clothing textiles. Those sentences are Biver’s citations of the 2014 Ecolabel decision and the 2020 label; neither instrument’s own text is relied on in this paragraph.

Washington Ecology Publication 15-03-039 (November 2015) tested children’s clothing, footwear, and accessories. Of 50 component samples analysed for metals, antimony was detected in 72 percent, more often than any other analyte in that panel, mainly in fabric and plastisol prints. Concentrations ranged from below 1 to 2,500 ppm, with a median of 5.9 ppm across sample types and a median of 126 ppm in fabric alone.

The report states that 50 to 300 ppm was typical of polyester clothing and is the range expected from antimony trioxide used as a polyester catalyst. The highest result, 2,500 ppm, was a pair of children’s pajamas labeled flame resistant, with an X-ray fluorescence bromine reading of 4.6 percent. That is one component in a clothing screen, not a study designed around polyester sleepwear as a class.

A Greenpeace Research Laboratories technical report (06/2013; Brigden and colleagues, December 2013) quantified antimony in 36 articles that contained polyester or a polyester blend, and detected it in all 36. The executive summary prints 14 to 293 mg/kg of fabric, median 96 mg/kg.

The results text then gives the fabric-blend concentrations as 14 to 293 mg/kg of polyester, median 96 mg/kg, and the polyester fraction of each fabric as also 14 to 293 mg/kg, with a higher median of 120 mg/kg. A summary row in the same report prints a median of 120. This page keeps both medians and does not choose one. Brand names in the article table are not repeated. The articles are clothing and footwear bought for that survey, not items identified as infant sleepwear.

Children’s products

Washington Ecology Publication 14-04-014 (revised June 2021) is a screening survey, not a random sample of the market: components were prioritized by XRF and then analysed by ICP-MS. In that design, baby-wipe components measured 70.6, 61.3, and 40.6 ppm antimony, and moist baby-wipes packaging measured 95.8 ppm. Baby bibs measured 116 and 99.6 ppm, two clothing samples (jeans) measured 121 and 128 ppm, and two toy components measured 168 and 94 ppm. The agency reported antimony detected in 57 percent of 150 children’s-product components.

The same rows that print antimony in the tens of ppm often print other metals at the reporting floor, and a printed zero without a detection limit is not treated here as a formal non-detect. The study supports the Children’s Safe Products Act reporting threshold of 100 ppm for chemicals of high concern to children. That threshold is a reporting trigger, not a health-based limit.

Choi, Song, and Kim 2016 (Journal of the Korean Oil Chemists’ Society 33(1):195-203, DOI 10.12925/jkocs.2016.33.1.195) is a nine-page paper. The methods print: “Approximately 0.2 g of sample was added into a PTFE vessel containing 7 mL of 70% HNO3 and 1 mL of 50% HF,” and the digest was then “transferred and diluted with 25 mL of ASTM type 1 water.” That sentence is dilution with 25 mL. It is not a statement that the flask was brought to 25 mL, and it does not state a final volume after the 8 mL of acid. The sample mass is printed as approximate. Table 5 reports the digest in µg/L, and antimony was detected in 12 of 24 wet-wipe nonwovens, from 187.86 ± 5.24 to 19,558.35 ± 3,537.3 µg/L. The highest result is the infant sheet labeled rayon alone. Several sheets labeled rayon plus polyester or rayon plus PET had no detected antimony.

The authors do not print a sheet concentration. If the final volume is treated as 25 mL and the mass as 0.200 g, mg/kg is approximately (µg/L × 0.025 L / 0.0002 kg) / 1,000, which is µg/L × 0.125, and the two ends are about 23.48 and 2,444.79 mg/kg. The quotient of micrograms per litre, times litres, divided by kilograms, is micrograms per kilogram, and one milligram is 1,000 micrograms. If the 8 mL of acid remains and the 25 mL of water is added to it, the volume is 33 mL, the factor is 0.165, and the high end is about 3,227.13 mg/kg. Neither figure is printed. The source page keeps both readings beside the twelve printed means and does not use either as the paper’s result. Sample codes are not repeated here.

Mask packs are Table 6, not baby wipes. The source page records one pack at 8,895.28 ± 42.86 µg/L in the nonwoven and 38.95 ± 1.86 µg/L in the liquid, and states that the authors did not publish that split for the wet wipes. Other mask-pack antimony figures are not carried here, because that source page does not adopt them. The authors attribute a high antimony result to antimony trioxide or antimony glycolate used as a polyester catalyst. That is their inference, and the rayon-labeled maximum means it is not a rule for these sheets. Borca’s later X-ray absorption work found organic antimony rather than antimony trioxide in PET bottles, so the catalyst inference and the oxide inference are not the same claim.

A 2025 master’s thesis from Istanbul University-Cerrahpaşa (English abstract in the PDF) reports solution antimony of 695.40 ppb in one product, which the English abstract rounds to 695.4 ppb, and fiber antimony printed as 565.46 ppb and 537.86 ppb, with a further sentence and section 4.1 printing 476.69 ppb for fiber. All three fiber figures are in the thesis, and 476.69 is lower than the other two. Packaging antimony is printed at 0.03 and 0.04 ppb, at or below the quantification limit of 0.10 ppb. The antimony results of several hundred ppb sit above the stated calibration top of 50 ppb. The methods print an approximate sample mass of 0.2 g and the addition of 10 mL of water after digestion, but no made-up final volume, so no milligrams-per-kilogram conversion is made and the unit stays the ppb of the measured solution. Brand names in that thesis are not repeated.

A monitoring page dated 2 April 2026 from the China Association for Quality Promotion of Consumer Products Safety (https://cpqs.org.cn/archives/3229) reports total antimony detected in 20 of 21 baby wet wipes bought online, and not detected in one. The sample-by-sample annex is an image. Preferred reads of total antimony, taken as micrograms per liter, are numeric in 20 of 21 columns. The stable low read is 15.02 µg/L and the stable high preferred read is 450.45 µg/L. The source page’s preferred read for one wrapped cell is 383.7 µg/L, with the last digit low confidence. A second full-cell pass of the same image also produced 490.40 µg/L where the preferred read is 450.45, and 399.43 µg/L where the preferred read is 355.43. Several other wrapped cells have a less certain second digit. The preparation is not described, and the unit is not converted here to milligrams per kilogram of sheet. The narrative names products; this page does not. The document is an association monitoring report, not a peer-reviewed paper.

Toy data are sparse. Turner 2018 measured second-hand plastic toys in the United Kingdom, and under EN 71-3 conditions migratable antimony exceeded 100 µg/g in at least one component; the metals the paper reports as exceeding the EU migration limits, in the cases it names, are cadmium or lead, not antimony. Kawamura and colleagues 2006 found antimony in one of 45 baby toys and 10 paints, at 5.3 mg/kg. The Danish Environmental Protection Agency’s 2005 survey of 15 surface-treated wooden toys found antimony below a 6 mg/kg detection limit in every EN 71-3 extract.

Karaś and Frankowski 2018 print mean antimony after acid mineralisation in the hundreds to thousands of ng/g for several colour groups (green 3,806 ± 981.9 ng/g, black 2,838 ± 846.9 ng/g) and state that the highest mineralisation antimony was in metallic and blue samples. The blue mean in that table is 218.4 ± 422.2 ng/g, and the metallic row is printed in mg/g (antimony 0.004 ± 0.04). This page does not reconcile that prose with that table.

Nyamukamba and colleagues 2023 measured extractable metals in 20 disposable diapers bought in Cape Town, using artificial sweat and artificial urine, and antimony was not an analyte. The panel was lead, copper, arsenic, zinc, cobalt, nickel, chromium, cadmium, manganese, and selenium. Nickel exceeded the 0.1 mg/kg Oeko-Tex value stated by the authors in one sweat extract (0.197 mg/kg) and one urine extract (about 0.224 mg/kg). The paper’s prose gives a sweat lead range of 0.002 to 0.019 mg/kg, while Table 4 in the same HTML lists a lead maximum of 0.094 mg/kg, and this page does not adopt a lead range from that conflict. The diaper finding is that a sweat-and-urine extractable panel built around other metals can miss antimony entirely.

Regulatory limits

The instruments below limit different quantities. A drinking-water concentration, a specific migration limit into food, a toy-migration limit in milligrams per kilogram of toy material, and a cosmetic impurity guidance value are different bases, and none of them is a baby-wipe content limit. Limits are copied from the instruments themselves or from the wiki regulation pages that transcribe them.

Reference values

These values answer different questions, and they are not a range to average.

The WHO tolerable daily intake is 6 µg/kg body weight. It rests on a NOAEL of 6.0 mg/kg body weight per day for decreased body-weight gain and reduced food and water intake in a 90-day rat drinking-water study of potassium antimony tartrate, with an uncertainty factor of 1000 (100 for interspecies and intraspecies variation, and 10 for the short duration). The drinking-water guideline of 0.02 mg/L (20 µg/L, a rounded figure) allocates 10 percent of that TDI to a 60 kg adult drinking 2 litres per day.

The background document WHO/SDE/WSH/03.04/74 prints that derivation and states that the value could be highly conservative, because of the end-points and the size of the uncertainty factor. The same document prints an oral LD50 for antimony trioxide above 20,000 mg/kg, which it attributes to extremely low water solubility, and gastrointestinal absorption of 5 to 20 percent in animals and 5 percent in four people after involuntary acute intoxication with antimony potassium tartrate. Those absorption figures are the background document’s; they are not the ATSDR and ICRP figures above, and they are not averaged with them. Assessment date 2003.

The ATSDR intermediate oral minimal risk level is 0.0006 mg Sb/kg/day, two orders of magnitude below the WHO TDI, because ATSDR and WHO chose different NOAELs from the same Poon et al. 1998 study. The UK Committee on Toxicity, in October 2024, agreed with the WHO and Lynch et al. 1999 reading of a NOAEL of 6,000 µg Sb/kg body weight per day, and recorded that Health Canada and ATSDR had selected 60 µg Sb/kg body weight per day. COT had previously used the WHO TDI of 6 µg/kg body weight per day when it reviewed the 2006 UK Total Diet Study and the 2014 infant-food survey. This page does not average the two NOAELs.

RIVM’s 2012 letter report on environmental risk limits states that EFSA (2004) adopted the WHO TDI of 6 µg antimony/kg body weight per day for antimony trioxide in food-contact materials. The EFSA opinion itself was not examined for this page, so the adoption is reported as RIVM’s statement.

OEHHA’s September 2016 public health goal replaces the agency’s 1997 goal of 20 ppb with a goal of 1 ppb, rounded from a calculated 1.1 µg/L. The point of departure is a BMDL10 of 0.14 mg/kg-day for liver nuclear anisokaryosis in male rats in the Poon et al. 1998 drinking-water study of antimony potassium tartrate.

OEHHA applies an uncertainty factor of 1,000 (10 for interspecies extrapolation, 30 for human variation, and the square root of 10 for subchronic-to-lifetime extrapolation), keeps a relative source contribution of 0.40, and uses a lifetime-average drinking-water intake of 0.053 L/kg-day. The 1997 goal had used a different uncertainty factor, 300, on a LOAEL of 0.43 mg/kg-day from Schroeder et al. 1970. The 1 ppb figure is a California drinking-water goal, not a wipe limit, and the Poon study itself was not re-examined for this page.

Kowalczyk and colleagues 2022, an EFSA Journal EU-FORA report on elements in teas (EFSA Journal 20(S1):e200410, DOI 10.2903/j.efsa.2022.e200410), restate the WHO tolerable daily intake of 6 µg/kg body weight and state that the highest antimony intake they evaluated from tea was 0.03 percent of that TDI. The PDF renders the microgram symbol poorly in that sentence. The number is the WHO TDI already on this page, not a new reference value, and the paper is not the 2004 EFSA food-contact opinion. No tea concentration in micrograms per kilogram is printed in its main text.

Wang and colleagues 2026 (Journal of the Endocrine Society 10(4):bvag011, DOI 10.1210/jendso/bvag011) state that most food-contact migration studies they reviewed fell below the WHO drinking-water figure, which they write as 20 ppb and also as 0.02 mg/kg, and below the EU specific migration limit, which they write as 40 ppb and also as 0.04 mg/kg. Those parentheticals equate a water concentration with a food mass fraction. This page keeps each instrument in its own units: 20 µg/L in water, and 0.04 mg/kg migrating into food. Wang states that leaching below 2.5 µg/L is what many bottle studies report under ordinary storage, and that acidity, heat, and time can push results above 5 µg/L.

Wang also states that the U.S. EPA recommends less than 0.35 µg Sb/kg body weight per day for oral exposure. The IRIS chemical file was not examined for this page, so that sentence is recorded as Wang’s citation and is not adopted here as an oral reference dose, and no RfD is printed. The National Primary Drinking Water Regulations page sets the maximum contaminant level goal and the maximum contaminant level for antimony at 0.006 mg/L, and lists increase in blood cholesterol and decrease in blood sugar as the long-term health effects above the MCL. No chronic oral minimal risk level exists in the 2019 ATSDR profile.

Government limits

The rows in this table are transcribed from the official texts. Each value is the figure printed in that text, with the unit, basis, and scope printed beside it. A drinking-water maximum, a voluntary label criterion, a toy migration limit, and a diaper migration limit are different bases, and they are not averaged.

No US federal limit exists for antimony in diapers or in textiles. The federal texts examined here, 40 CFR 141.51, 40 CFR 141.62, the NPDWR table, and the archived consumer fact sheet, regulate drinking water.

China GB 31701 and GB 18401 were not obtained, so no antimony value from those standards is recorded. GB/T 18885 and FZ/T 73025 were not obtained either, and no value from them is recorded.

Other instruments transcribed in this corpus

The table below indexes further instruments transcribed on the wiki’s regulation pages. Where both tables print a figure, the figure in the government-limits table above is the one read from the official text.

The corpus OEKO-TEX page does not record an extractable antimony limit for STANDARD 100. The 30 µg/g STANDARD 100 figure appears on this page only as Biver’s citation, under Plastics and textiles, and the label’s own text was not examined, so it is not entered in this table. The cosmetic ingredient prohibition is EU Regulation 1223/2009.

Testing

Total antimony by ICP-MS after acid digestion, as in the Washington confirmatory analyses, reports elemental antimony and does not distinguish Sb(III) from Sb(V). EN 71-3 and ASTM F963 report soluble or migrated antimony under acid conditions, in mg/kg of material, which is not a total-content ppm. Artificial-sweat and artificial-urine extractions, as in the diaper study and in Biver 2021, are a third basis, and cosmetic guidance values are finished-product impurity concentrations. A single number cannot be moved across those methods.

Speciation is the missing layer in most of this work. Borca 2025 is the speciation measurement examined here, finding organic antimony in PET bottles, senarmontite in flame-retardant electrical objects, and an unresolved mix in polyester textiles. The wipe and clothing screens cited above did not speciate.

Microbiome context

No study examined here measured the skin microbiome of the diaper area under antimony exposure, so that cell is unknown. The strength of each line of evidence below is stated with it.

Microbial handling of Sb(III) shares hardware with arsenic. Li and colleagues 2017, in Scientific Reports 7:43225 (doi.org/10.1038/srep43225), studied Agrobacterium tumefaciens GW4 in vitro. The paper states that the glycerol facilitator GlpF and its homolog Fps1p take up Sb(III) because Sb(OH)3 resembles glycerol, and that the arsenite efflux pumps ArsB and Acr3 can also efflux Sb(III). A cytoplasmic Sb(III) oxidase, AnoA, a short-chain dehydrogenase, and the periplasmic arsenite oxidase AioAB both oxidize Sb(III) to the less toxic Sb(V).

Knockout of anoA decreased Sb(III) oxidation efficiency by about 30 percent. Deletion of aioA did not decrease log-phase oxidation in GW4 and slightly increased it later, although the paper notes that aioA deletion decreased oxidation in a different strain (5A). At 50 µM Sb(III), growth was unaffected; at 100 and 200 µM, katA, anoA, and aioA were required for resistance. Cellular hydrogen peroxide also oxidizes Sb(III) abiotically (R2 0.93 in vivo and 0.94 in vitro in the experiments reported). The paper states that genes and enzymes for microbial Sb(V) reduction and Sb(III) methylation had not been identified, although the phenomena were described as environmentally widespread, and this page does not name a methylation gene. This is in vitro evidence.

In the animal gut, Wu and colleagues 2025 (International Journal of Molecular Sciences 26(7):3332, DOI 10.3390/ijms26073332) gavaged male C57BL/6 mice, about 20 g and six weeks old, n = 7 per group, every two days for six weeks. The groups were control, antimony at 10 mg/kg as potassium antimony tartrate, copper at 100 mg/kg, and both, so an antimony-only group exists. Alpha-diversity (ACE and richness) was elevated in the antimony group in the results, while the discussion says paired comparisons of alpha-diversity were not significant. Both sentences are in the paper, and this page does not resolve them. Beta-diversity differed across groups (p = 0.001).

Hepatic taurine-beta-muricholic acid was elevated 1.5-fold in the antimony group, and serum total bile acids 1.6-fold. The 1.8-fold serum T-beta-MCA increase is stated for the copper and combined groups, not as an antimony-only serum result. Lactobacillus spp. were reduced 2.6-fold in the wording “antimony- and/or copper-treated groups,” so the fold is not attributed to antimony alone. Liver inflammation on histology was present in all three treated groups and more pronounced when both metals were given. The 10 mg/kg dose is orders of magnitude above dietary or wipe exposure. This is animal evidence at a high dose, not in an infant and not from a wipe.

An Ecotoxicology and Environmental Safety abstract in adult zebrafish (0, 30, 300, and 3000 µg/L for 14 days) reports intestinal-barrier injury and a microbiota shift at 300 µg/L (Cetobacterium down, Vibrio up), and says 30 µg/L was enough to perturb microbiota and metabolites. The full text was not checked line by line. This is fish evidence from an abstract only, not a human infant.

Deng and colleagues 2021 (Frontiers in Microbiology 12:738596, DOI 10.3389/fmicb.2021.738596) review Sb(III)-oxidizing bacteria. They cite a disruption of aioA that decreased Sb(III) oxidation by 25 percent in one strain, and they describe AnoA as a distinct Sb(III) oxidase. On methylation, the review names monomethyl, dimethyl, and trimethyl antimony as the products reported in fungi, archaea, and bacteria, and it says arsenic methylase and S-methionine methyltransferase have been reported to take part in Sb(III) methylation. The same section states that the genes or enzymes required for antimony methylation are still unknown. This is a review of microbial studies, not a new experiment, and not a human gut.

Later and earlier enzyme papers narrow that statement without closing it. Bentley and Chasteen 2002 review microbial methylation of arsenic, antimony, and bismuth. Thomas and colleagues 2011 showed that Methanosarcina mazei methylates antimony in vivo, and that purified methyltransferase MtaA, with cob(I)alamin, methylates antimony in vitro among several metalloids; that is in vitro work and one archaeon, not a human gut. Kambara, Yamamura, and Amachi 2024 identified a DMSO-reductase-family protein, AnrA, as a dissimilatory antimonate reductase in Geobacter sp. SVR, one environmental bacterium rather than a gut commensal.

Shi and colleagues 2018 identified the efflux transporter ArsK as conferring resistance to antimonite as well as to arsenite and two methylated arsenicals, by heterologous expression in vitro. Zhang and colleagues 2022 report that the flavin monooxygenase ArsV oxidizes antimonite as well as methylarsenite, in vitro, in a soil bacterium and a hypersensitive E. coli strain. Terry and colleagues 2015 isolated two mine-sediment bacteria that oxidize Sb(III) with oxygen or nitrate and that also carry an aioA homolog, and Wang and colleagues 2025 demonstrated hydrogen oxidation coupled to antimonate reduction in mine tailings and river sediment; both are environmental isolates, not an infant microbiome. None of these papers measures antimony on a wipe or in formula.

At the level of enzyme structure, Engrola and colleagues 2023 (Journal of Biological Chemistry 299(8):105036, DOI 10.1016/j.jbc.2023.105036) report four crystal structures of arsenite oxidase bound to arsenic or antimony oxyanions, and state that Sb(III) salts are oxidized about 6,500 times more slowly than the corresponding As(III) salt. This in vitro enzymology shows a shared active site. It does not show an effect on a microbiome.

In the human gut, Yan and colleagues 2025 measured 26 metals in serum from 342 infants at Hunan Children’s Hospital, drawn from an initial 459, mean gestational age 34.71 ± 3.36 weeks. The abstract and the Chao1 results text say antimony was negatively associated with the Chao1 index, along with chromium, tungsten, cobalt, copper, lanthanum, praseodymium, and uranium. No single-metal beta for antimony versus Chao1 is printed in the main text, so none is stated. Bayesian kernel machine regression identified antimony, copper, arsenic, and uranium as contributors to the mixture association with Chao1, each with a posterior inclusion probability above 0.7, while the discussion names arsenic and copper as the most influential metals for alpha diversity.

Printed interaction terms that include antimony are Sb-Pr, beta 2.17 (95 percent confidence interval 0.77 to 3.58), and Sb-U, beta 2.14 (0.78 to 3.50). The figure caption assigns full term N = 95, preterm N = 191, and very preterm N = 56, but one results sentence assigns N = 191 to very preterm infants and the next assigns N = 56 to very preterm infants. This page follows the figure caption and records that sentence as inconsistent.

Antimony’s maintained Chao1 association is the one the paper ties to the N = 56 stratum, and in very preterm infants the paper also reports genus-level associations with tin and antimony, without naming the genera in the main text. This is observational human evidence from a hospital sample weighted toward preterm birth, measured in serum rather than as a product dose, and not on skin.

In the mouth, Davis and colleagues 2020 measured 21 metals in saliva from 61 children and adults. After multiple-testing adjustment, bacterial taxa were associated with antimony, arsenic, and mercury, and antimony was associated with decayed teeth. No salivary concentration is extracted on the source page, so none is stated. This is observational human evidence from the oral cavity.

On skin, Zeldin and colleagues 2023 (Environmental Science & Technology Letters 10(5):452-457, DOI 10.1021/acs.estlett.3c00142; author manuscript PMC10485844) reanalyzed U.S. postal-code atopic dermatitis billing against EPA pollutant inventories in a region of low diisocyanate exposure, and identified antimony compounds as the top associated pollutant there. The printed correlation with pediatric atopic dermatitis visits was 0.84 in Montana, Idaho, Wyoming, and the Dakotas, and −0.014 in the remaining states. In vitro, antimony inhibited ceramide-family lipids in the skin commensal Roseomonas mucosa and induced calcium flux in human Schwann cells that was blocked by the TRPA1 antagonist HC030031. The manuscript does not print an antimony concentration for those assays.

The methods, in the author manuscript and in the PMC HTML, name ICD-10 code L40.9 as the billing code. That code is psoriasis, unspecified, while the title and discussion treat the outcome as atopic dermatitis, and this page does not resolve the conflict. The authors state that the work lacks mouse models and direct human challenge. The evidence is an ecological association plus in vitro bacterium and neuron assays. It is not a diaper-area cohort, and it is not a measurement of antimony on infant skin.

Whether microbial oxidation of Sb(III) to Sb(V) lowers toxicity to the host is a mechanistic inference from the in vitro chemistry. No document examined here shows it in an infant gut or on infant skin.

Vulnerable populations

Infants meet antimony, where they meet it, through water used to prepare feeds, through polyester and PET articles against the skin or in the mouth, and through any antimony in the diet. No infant-formula concentration has been extracted from a source page. Source records that list antimony among the analytes in infant food, including the 2003 COT statement on a UK infant-food survey, are indexed on this page and are not quoted as a parts-per-billion result.

Body weight makes the same ingested mass a larger dose in an infant than in a 60 kg adult, which is why the WHO guideline derivation (60 kg, 2 L/day, 10 percent of TDI) does not describe an infant. This page does not rescale the TDI or the MRLs onto an infant body weight, and no dermal dose from a wipe was calculated. The 0.26 percent figure is OECD’s summary of an unpublished in vitro study on diantimony trioxide, not a wipe-lotion measurement, and no wipe-lotion migration result was examined here.

Infant formula itself was not analysed for antimony in any document examined here, and the infant data that exist are urinary. Pikounis and colleagues 2024 measured antimony in urine from infants of about 6 weeks and did not use it as an exposure biomarker, because few published studies assess general-population antimony exposure in urine. Du and colleagues 2025, in 56 infants with reflux or dysphagia, print urinary antimony medians of 0.2 µg/L in the control group, 0.15 µg/L in an added-rice formula group, and 0.68 µg/L in an oatmeal or rice-cereal thickener group (Table 2, p = 0.022).

The same paper’s prose on urinary arsenic (1.78, 0.96, and 0.76 µg/L) does not match Table 2 (0.74, 0.96, and 1.78 µg/L), and this page follows Table 2 for antimony. Those are urine concentrations, not antimony in the formula or in the water used to mix it. The brand name in that table is not repeated.

Open questions

The first open question is dermal absorption. There is no accessible primary dermal absorption study: the 0.26 percent figure, and OEHHA’s additional 0.14 percent, are summaries of the unpublished Roper and Stupart 2006 report, which was not examined, and the EU risk assessment that contains that study was not obtained. Biver 2021 measures release into artificial sweat, which is not absorption.

The second is the infant article itself. No study is specific to PET baby bottles, and none to polyester sleepwear. The PET migration papers examined here use commercial beverage bottles. Washington Ecology 15-03-039 includes one flame-resistant pajama component at 2,500 ppm, which is one clothing component, not a sleepwear class study, and the Greenpeace polyester results are clothing and footwear bought for that survey, not sleepwear. The Washington wipe and textile numbers are an XRF-prioritized screen. The 2026 China monitoring annex prints total antimony in µg/L, from 15.02 to 450.45, and does not print a sheet concentration in milligrams per kilogram.

The third is infant formula water, which is covered only indirectly. Drinking-water limits exist, and Pikounis 2024 and Du 2025 report urinary antimony in infants, but neither paper reports antimony in formula or in the water used to prepare it. The Romanian bottled-water maximum carries a printed uncertainty larger than the value.

The oral reference values have gaps of their own. There is no chronic oral minimal risk level, and no IRIS oral reference dose was re-read for this page. Wang 2026 cites an EPA oral figure of 0.35 µg/kg/day, but that citation is not a substitute for the IRIS file. EFSA’s 2004 food-contact opinion was not examined, and Kowalczyk 2022 is a tea exposure report, not that opinion.

The microbiome questions are the most open. No diaper-area cohort links antimony to the skin microbiome. Zeldin 2023 is an ecological association plus in vitro work on one skin commensal, and its billing code conflicts with its title. Deng 2021 states that the genes required for microbial antimony methylation are still unknown, and Thomas 2011 shows only that MtaA can methylate antimony in vitro, which is not a dedicated antimony methylase demonstrated in an infant gut. The 2024 paper that names ArsM as the antimony biomethylase was not obtained.

A restriction dossier on the oxide as used in electrical and electronic equipment was examined. That dossier sets no baby-product concentration, and none of its occupational or article figures are adopted here.

Competing interests

The Heavy Metal Index reports what the peer-reviewed and regulatory literature supports. It is operated separately from any certification program, does not evaluate or endorse brands, and does not use certification thresholds as evidence for the statements made here.

Frequently asked questions

References

Works cited in this page’s text, in first-appearance order. See Sources for this page’s source inventory. Each title links to a source record with bibliographic details, reported findings, and document provenance.

1. 40 CFR 141.62, maximum contaminant level for antimonyU.S. Environmental Protection Agency · Code of Federal Regulations, Title 40, section 141.62, 1 July 2025 edition · 2025 · www.govinfo.govRegulation
2. Decision (EU) 2023/1809, EU Ecolabel for absorbent hygiene productsEuropean Commission · Official Journal of the European Union, L 234, 22 September 2023, p. 142 · 2023 · data.europa.euRegulation
3. Association between maternal antimony exposure and offspring congenital heart defects: findings from a large birth cohort studyYawen Shao, Lei Cao, Ru Lin, Jie Ran, Qing Liu, and Baohong Mao · Frontiers in Public Health · 2026 · doi.org/10.3389/fpubh.2026.1957018Peer-reviewed
4. Association between urinary trace metals and multiple health outcomesSiyi Wu and Binbin Yu · Science Progress 109(2) · 2026 · doi.org/10.1177/00368504261448301Peer-reviewed
5. Association of eight urinary metals and trouble sleeping: A cross-sectional studyQing Ma, Jing-Zhi Liu, Xiao-Eni Shi, Lin-Lin Li, Xue-Wei Feng, Min Fu, et al. · Medicine 105(36) · 2026 · doi.org/10.1097/md.0000000000050551Peer-reviewed
6. Association between heavy metals exposure and depression: findings of the NHANES from 2003 to 2020Yifei Guo, Yitong Chen, Hao Zhou, Yixing Fan, Tiantian Feng, and Zhuang Ma · European Journal of Medical Research 31:266 · 2026 · doi.org/10.1186/s40001-026-03850-xPeer-reviewed
7. Heavy metals and prevalence of sexually transmitted infections: a cross-sectional study using data from the National Health and Nutrition Examination Survey (NHANES) 2003-2016Brandie D. Taylor, Abigail Noah, Victor Adekanmbi, Yongkang Zhang, Melissa Ingle, and Abbey Berenson · BMC Infectious Diseases 26:810 · 2026 · doi.org/10.1186/s12879-026-13046-7Peer-reviewed
8. Prospective associations of childhood metal mixtures with metabolic health in a Mexico City cohortCorentin Gigot, Pei-Wen Tung, Haotian Wu, Guadalupe Estrada-Gutierrez, Maria-Consuelo Tolentino, Alejandra Cantoral, et al. · Environmental Epidemiology 10:e535 · 2026 · doi.org/10.1097/ee9.0000000000000535Peer-reviewed
9. Antimony: a cryptic metabolism disruptor ubiquitous in food contact materialsLuyu Wang, Spyridoula Gerassimidou, Birgit Geueke, Ksenia J Groh, Eleni Iacovidou, Aaish Majid, et al. · Journal of the Endocrine Society 10(4):bvag011 · 2026 · doi.org/10.1210/jendso/bvag011Peer-reviewed
10. Levels, Exposure and Risk Assessment to Metals in Blood of Adult Working Population in Valencia Region (Spain)Carmen Saez, Alfredo Sanchez, Pablo Dualde, Miguel Angel Aguirre, Antonio Canals, Marta Esteban-Lopez, et al. · Toxics 14(10):874 · 2026 · doi.org/10.3390/toxics14100874Peer-reviewed
11. Correlates of prenatal heavy metal contamination in Lebanese infants using cord blood analysisElie Whaibeh, Joy Abi Hanna, Jan Kuta, Nathalie Aouad, Nada Yaghi, Andrea Abi Nader, et al. · Scientific Reports 16:19491 · 2026 · doi.org/10.1038/s41598-026-46981-1Peer-reviewed
12. Hair metal concentrations in a mother-newborn population from a polluted megacity: an indicator of prenatal metal exposureCitlali Denisse Ortiz-Robles, Mariana Paz-Sabillon, Angel Barrera-Hernandez, Luz del Carmen Sanchez-Pena, Monserrat Nereida Rangel-Calvillo, Luz Maria de la Razo, et al. · Journal of Exposure Science and Environmental Epidemiology 36:788-795 · 2026 · doi.org/10.1038/s41370-026-00869-4Peer-reviewed
13. Technology-Related Trace Elements in Paediatric Scalp-Hair Biomonitoring: Biomarker Validity, Environmental Relevance, and Implications for Risk AssessmentAntonio Pena-Fernandez, Rafael Moreno-Gomez-Toledano, Beatriz Martinez-Alonso, and Maria Angeles Pena Fernandez · Toxics 14:813 · 2026 · doi.org/10.3390/toxics14090813Peer-reviewed
14. Occurrence of Potentially Toxic Elements in Bottled Drinking Water-Carcinogenic and Non-Carcinogenic Risks Assessment in Adults via IngestionUngureanu EL, Soare AD, Mocanu AL, Iorga SC, Mustatea G, and Popa ME · Foods · 2022 · doi.org/10.3390/foods11101407Peer-reviewed
15. Antimony in Polyethylene Terephthalate-Bottled Beverages: The Migration PuzzleSandra Carrasedo, Esther Hernandez-Nataren, Jose F. Lopez-Sanchez, and Angeles Sahuquillo · Molecules 28(20):7166 · 2023 · doi.org/10.3390/molecules28207166Peer-reviewed
16. Metals in Children’s and Consumer Products and PackagingStone A · Washington State Department of Ecology, Hazardous Waste and Toxics Reduction Program, Publication 14-04-014 (Revised June 2021) · 2021 · apps.ecology.wa.govRegulation
17. A step towards understanding plastic complexity: Antimony speciation in consumer plastics and synthetic textiles revealed by XASCamelia N. Borca, Thomas Huthwelker, and Montserrat Filella · Journal of Environmental Sciences 153:109-117 · 2025 · doi.org/10.1016/j.jes.2024.02.006Peer-reviewed
18. Antimony release from polyester textiles by artificial sweat solutions: A call for a standardized procedureBiver M, Turner A, and Filella M · Regulatory Toxicology and Pharmacology 119:104824 · 2021 · doi.org/10.1016/j.yrtph.2020.104824Peer-reviewed
19. Determination of heavy metals on the non-woven in wet wipes using ICP-MSChoi SM, Song JK, and Kim SJ · Journal of the Korean Oil Chemists’ Society 33(1):195-203 · 2016 · doi.org/10.12925/jkocs.2016.33.1.195Peer-reviewed
20. Determination of Selected Harmful Substances in Baby Diapers Available on the South African MarketNyamukamba P, Mququ Z, Nkosi S, and Isaacs S · International Journal of Environmental Research and Public Health 20(2): 1023 · 2023 · doi.org/10.3390/ijerph20021023Peer-reviewed
21. Risk assessment of rare earth elements, antimony, barium, boron, lithium, tellurium, thallium and vanadium in teasEwelina Kowalczyk, Lucas Givelet, Heidi Amlund, Jens Jørgen Sloth, and Max Hansen · EFSA Journal 2022;20(S1):e200410 · 2022 · doi.org/10.2903/j.efsa.2022.e200410Peer-reviewed
22. OEKO-TEX® Limit values: New regulations 2024OEKO-TEX® Association · OEKO-TEX® Service GmbH customer information notice, 09 January 2024, Zurich, Switzerland. Updates applicable test criteria, limit values, and requirements for the OEKO-TEX® STANDARD 100, ORGANIC COTTON, LEATHER STANDARD, ECO PASSPORT, and STeP certification labels; new values binding from 01 April 2024 after a three-month transition period · 2024 · www.oeko-tex.comGuidance
23. Regulation (EC) No 1223/2009 of the European Parliament and of the Council of 30 November 2009 on cosmetic products (recast)European Parliament and Council of the European Union · Official Journal of the European Union, L 342, 22.12.2009, p. 59-209 · 2009 · eur-lex.europa.euRegulation
24. Abiotic and biotic factors responsible for antimonite oxidation in Agrobacterium tumefaciens GW4Li J, Yang B, Shi M, Yuan K, Guo W, Wang Q, et al. · Scientific Reports 7:43225 · 2017 · doi.org/10.1038/srep43225Peer-reviewed
25. Gut Microbiota and Its Metabolite Taurine-beta-Muricholic Acid Contribute to Antimony- and/or Copper-Induced Liver InflammationDandan Wu, Qiwen Lin, Senao Hou, Xiaorui Cui, Na Shou, Xuefeng Yuan, et al. · International Journal of Molecular Sciences 26(7):3332 · 2025 · doi.org/10.3390/ijms26073332Peer-reviewed
26. A Critical Review of Resistance and Oxidation Mechanisms of Sb-Oxidizing Bacteria for the Bioremediation of Sb(III) PollutionRenjian Deng, Yilin Chen, Xinpin Deng, Zhongjie Huang, Saijun Zhou, Bozhi Ren, et al. · Frontiers in Microbiology 12:738596 · 2021 · doi.org/10.3389/fmicb.2021.738596Review
27. Arsenite oxidase in complex with antimonite and arsenite oxyanions: Insights into the catalytic mechanismFilipa Engrola, Marcia A. S. Correia, Cameron Watson, Carlos C. Romao, Luis F. Veiros, Maria Joao Romao, et al. · Journal of Biological Chemistry 299(8):105036 · 2023 · doi.org/10.1016/j.jbc.2023.105036Peer-reviewed
28. Association between infants’ serum levels of 26 metals and gut microbiota: a hospital-based cross-sectional study in ChinaXing Yan, Jun Qiu, Ruiwen Huang, Xiaoming Peng, Shi-ting Xiang, Kunyan Zhao, et al. · Frontiers in Microbiology 16:1669475 · 2025 · doi.org/10.3389/fmicb.2025.1669475Peer-reviewed
29. Low levels of salivary metals, oral microbiome composition and dental decayDavis E, Bakulski KM, Goodrich JM, Peterson KE, Marazita ML, and Foxman B · Scientific Reports · 2020 · doi.org/10.1038/s41598-020-71495-9Peer-reviewed
30. Antimony Compounds Associate with Atopic Dermatitis and Influence Models of Itch and DysbiosisJordan Zeldin, Tan T. Tran, Manoj Yadav, Prem Prashant Chaudhary, Brandon N. D’Souza, Grace Ratley, et al. · Environmental Science & Technology Letters 10(5):452-457 · 2023 · doi.org/10.1021/acs.estlett.3c00142Peer-reviewed
31. COT statement on a survey of metals in infant foodCommittee on Toxicity of Chemicals in Food and Consumer Products and the Environment · Committee on Toxicity statement · 2003Government
32. RoHS Annex II Dossier for Diantimony trioxide (flame retardant)Oeko-Institut · RoHS Annex II Dossier, Version 3, 30 March 2020 · 2020 · rohs.exemptions.oeko.infoRegulation

Sources

Source records associated with this topic. Inclusion does not establish that every finding applies to this product or ingredient. References above identify works cited in the text.

Recent source records are shown below. Browse the complete evidence table for all source records and their reported findings.

Update history

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