Mini-review and systematic assessment published in the Journal of the Endocrine Society in 2026 by Wang, Gerassimidou, Geueke, and colleagues, examining (1) antimony migration into food and food simulants from the full range of food contact material (FCM) types using the FCCmigex database, (2) population-level urinary and blood antimony biomonitoring from the United States, Canada, and China, and (3) a systematic review of 85 studies linking antimony exposure to cardiometabolic outcomes. The paper does not generate new migration measurements; it synthesizes published FCCmigex entries and biomonitoring surveys. Funding was provided by the National Institutes of Health and the Food Packaging Forum Foundation.
Study design
The paper has two analytical components, described in the methods section (p. 2).
The FCCmigex component queried the FCCmigex database using the search terms “antimony” and “Sb” on May 12, 2025. This returned 99 entries for antimony as a food contact chemical, yielding 102 database entries across 54 references after filtering for studies that analyzed migration from all types of FCMs (single-use, unclear/unknown, or repeat-use). The 102 entries covered 78 plastics, 14 paper/board, 3 glass/ceramic, and 3 other FCM records (Figure 1 of the paper). The database provides structured migration data under various experimental conditions; the paper summarizes range, fraction below the limit of detection, and representative high and low values by FCM category (Table 2 of the paper).
The biomonitoring component used publicly available data from four national surveys: NHANES (United States, four cycles 2011-2018), the Canadian Health Measures Survey (CHMS), and the China National Human Biomonitoring Survey (CNHBS). For the NHANES analysis, the paper reports that 11,457 participants were eligible for urinary antimony measurements and that the analytical sample for the geometric mean included 10,901 without missing measurements. NHANES urine samples were measured by mass spectrometry. The lower LOD for NHANES urine antimony using mass spectrometry was 0.022 micrograms/L; 23.92 percent of samples were below the LOD and were assigned a concentration of LOD divided by the square root of 2 (p. 3 of the paper).
The cardiometabolic systematic review searched PubMed from 1954 to May 12, 2025 using 13 MeSH terms combined with keywords covering antimony and diabetes, glucose, insulin, obesity, dyslipidemia, fatty liver, NAFLD, steatohepatitis, hypertension, atherosclerosis, health disparities, and health risks. After deduplication and abstract review, 85 studies met inclusion criteria.
Key numbers
Antimony migration by FCM type (Table 2 of the paper; food simulant concentrations in micrograms/L)
| FCM type | Entries (n) | Fraction below LOD | Reported migration range (micrograms/L) |
|---|---|---|---|
| Plastics | 78 | 71% (55/78) | 0.001 to 0.01 |
| Paper and board | 14 | 0% | 0.001 to 0.03 |
| Metals | 3 | 0% | not tabulated in paper |
| Glass and ceramic | 3 | 30% (1/3) | 0.00004 to 15 |
| Other FCMs | 3 | 0% | not tabulated in paper |
Notes: (1) Ceramics produced the highest reported maximum (15 micrograms/L); the paper notes that crystalware values ranged from 0.016 to 0.024 micrograms/L (p. 2 and Fig. 1). (2) The 71 percent below-LOD fraction for plastics means most plastic FCM entries in the database did not detect antimony in the food simulant under the conditions tested. (3) Migration units are food simulant concentrations (micrograms/L); the EU specific migration limit for antimony from plastic FCMs is expressed in mass per food mass (0.04 mg/kg, equivalent to 40 micrograms/kg food, per EU Regulation 10/2011, cited in the paper).
The paper states that antimony migration from PET bottles is typically below 2.5 micrograms/L under ambient conditions and can exceed 5 micrograms/L when temperature is elevated, matrix is acidic, or contact time is prolonged (p. 2). The paper further states that migration increases with the use of antimony-containing recycled plastic waste in FCAs (citing references 78 and 79 of the paper). Higher baseline antimony levels in FCAs raise concern for increased leaching during storage of acidic foodstuffs such as soft drinks and juices (citing references 61, 74, 77 of the paper).
Human biomonitoring for antimony (Table 1 of the paper)
| Population | Survey | Limit of detection (micrograms/L) | Urinary geometric mean (95% CI) (micrograms/L) | Blood geometric mean (micrograms/L) |
|---|---|---|---|---|
| United States | NHANES 2011-2018 | 0.022 | 0.0473 (0.0462, 0.0483) | not available |
| Canada | CHMS | 0.02 | 0.048 (0.046, 0.050) | not available |
| China | CNHBS 2017-2018 | 0.045 (urine) | 0.05 (95% CI: 0.02, 0.09) | 2.35 (LOD 0.03 micrograms/L) |
Additional NHANES detail from p. 3 of the paper: geometric mean for urine antimony was 0.0473 micrograms/L (95% CI: 0.0462, 0.0483) among 10,901 participants weighted to represent the general US population. Among participants aged 3 to 5 years and a random one-third subsample aged 6 and older, 23.92 percent of samples were below the LOD of 0.022 micrograms/L.
The paper notes that for the CNHBS population, the analytical sample was 6,311 individuals aged 3 to 79 years, with 20 percent of values below the LOD of 0.02 micrograms/L. Blood measurements from the CNHBS population showed an average concentration of 2.35 micrograms/L (LOD 0.03 micrograms/L for blood). No population-level biomonitoring data for Japan could be identified by the authors.
A recent analysis of global antimony biomonitoring (cited as reference 59 in the paper) found a consistent positive association between antimony levels and DNA methylation in umbilical cord blood, indicating prenatal exposure.
Cardiometabolic systematic review
The systematic review analyzed 85 studies: 3 in vivo, 4 case studies/series, 11 case-control studies, 28 cross-sectional studies, and 39 cohort studies (Figure 2 of the paper, p. 4). Key associations reported:
Diabetes and glucose metabolism: several cohort studies found that higher urinary antimony was associated with increased risk of gestational diabetes. A cross-sectional NHANES analysis found no significant association between antimony and blood glucose at the aggregate level, though some studies indicated associations with insulin resistance, impaired glucose homeostasis, and hemoglobin A1c levels in adolescents (p. 5 of the paper). The paper characterizes the evidence as “inconclusive.”
Obesity: a 1999-2011 NHANES analysis found no significant association between antimony and body mass index. Data from the Guangdong Provincial Residents’ Chronic Disease and Nutrition Survey (n not specified in the review) also found no association. The paper classifies the relationship between antimony and obesity as “inconclusive” but notes that metal mixtures studies suggest antimony contributes to this outcome (p. 5 of the paper).
Liver disease: epidemiological studies linked elevated antimony to steatotic liver disease, described as “nonalcoholic fatty liver disease” in older literature. An animal model study in rats given antimony at 5, 20, and 60 mg/kg/day demonstrated liver enzyme abnormalities and irregular liver-to-body weight ratios (p. 6, Table 2 not reproduced here). The paper identifies hepatotoxicity as a mechanistic pathway through mitochondrial dysfunction and oxidative stress.
Hypertension: a Wuhan-Zhuhai Adult Cohort study positively associated antimony with mean platelet volume; a Polish cohort found antimony associated with atherosclerotic plaque severity in the left main coronary artery; Aragon Workers Health Study data identified antimony as a potential epigenetic modifier and atherosclerotic risk factor (p. 6 of the paper).
Atherosclerotic cardiovascular disease (ASCVD): multiple cohort studies, including NHANES analyses (1999-2006 and 2003-2016), found that antimony was one of the strongest contributors to cardiovascular disease mortality risk. A Spanish Hortega Study and NHANES 1999-2006 showed associations between antimony exposure and cardiovascular disease incidence (p. 6 of the paper).
Methods (brief)
This is a narrative mini-review and systematic assessment, not an original analytical study. No new laboratory measurements were generated.
FCCmigex component: structured database search on May 12, 2025 using “antimony” and “Sb” as food contact chemical search terms. Database entries cover all FCM types, single-use and repeat-use. Entries were filtered by detection type (detected, not detected, or unclear). Migration experiment types included extraction into food simulants and migration into actual food products; extraction-only experiments were excluded. The number of FCCmigex entries analyzed is 102 across 54 references.
Biomonitoring component: secondary analysis of publicly available NHANES data (cycles 2011-2018, four cycles, complex sampling with survey weighting per NHANES guidance). NHANES urine antimony measured by mass spectrometry; LOD 0.022 micrograms/L; values below LOD assigned LOD/sqrt(2). Geometric means estimated with SAS software version 9.4.
Systematic review component: PubMed search from 1954 to May 12, 2025 using 13 MeSH terms and 12 keyword pairs. Initial yield: 694 results. After abstract review, 85 studies met inclusion criteria covering cardiometabolic outcomes. Studies focused on environmental outcomes, toys, cosmetics, or parasitic infections (Leishmaniasis, Schistosomiasis) were excluded.
Analytical instruments: not applicable (review). LOD/LOQ: not applicable for the review itself; NHANES urine LOD is 0.022 micrograms/L as reported by the authors (p. 3). CRM and recovery data: not reported in this paper.
Evidence fitness
This source provides context-level evidence and secondary synthesis. Its principal value is the aggregated FCCmigex database summary characterizing the range and distribution of antimony migration across FCM types, which no single primary migration paper covers in full. The biomonitoring table (Table 1) provides internally consistent population-level geometric means from three national surveys under a single analytical frame, useful for cross-country comparison.
The cardiometabolic systematic review synthesizes 85 studies but does not perform meta-analysis; associations cited are drawn from primary studies that vary in design, population, and antimony measurement method. Most cited associations are from cross-sectional data, which cannot establish causal direction. The paper explicitly notes that existing literature does not fully account for the rising tide of cardiometabolic disorders attributable to toxic metals acting as endocrine-disrupting chemicals.
This source does not support: (1) a certified migration limit for any specific food contact article; (2) an independent derivation of an oral reference dose (the 0.35 micrograms/kg bw/day figure cited in the paper traces to an EPA document that was not opened for this source page); (3) speciated antimony migration data (the FCCmigex entries do not uniformly report antimony(III) vs antimony(V)); (4) causal risk quantification for cardiometabolic outcomes.
Limitations
The FCCmigex database entries underlying Table 2 reflect diverse laboratory methodologies, food simulants, temperatures, and contact durations. Migration ranges across studies are not strictly comparable. The ceramics maximum of 15 micrograms/L is a single outlier value within a dataset where 30 percent of entries were below LOD; it should not be treated as a typical ceramics migration value.
Unit inconsistency in the paper: the EU SML for antimony from plastic FCMs is 0.04 mg/kg (food mass basis), whereas migration in Table 2 is expressed in micrograms/L (simulant volume basis). These units are not interchangeable without knowledge of food density and food-to-contact-material ratio.
The biomonitoring section notes that blood antimony may be a more accurate systemic measure than urine because of antimony’s long biological half-life and differential partitioning into tissues (p. 9 of the paper), but population-level blood data are limited to China in the survey table.
The cardiometabolic systematic review covers endpoints and exposure metrics that vary substantially across the 85 included studies; the paper’s conclusions about associations are therefore heterogeneous and descriptive rather than quantitative.
Related evidence
- Antimony
- Antimony in baby products
- Plastic tubs
- Cans/lids
- Antimony leaching from polyethylene terephthalate (PET) plastic used for bottled drinking water
- Antimony in Polyethylene Terephthalate-Bottled Beverages: The Migration Puzzle
- Contamination of Canadian and European bottled waters with antimony from PET containers
Update history
No substantive edit history is available in this build. The full commit record is available in git.
