Risk assessment of eighteen elements leaching from ceramic tableware in China

ABSTRACT Ceramic products are among the most frequently used food contact materials. Health risks associated with ceramic tableware usually arise from the migration of heavy metals. In this study, 767 pieces of ceramic tableware of different shapes and types were collected across China, and the migration levels of 18 elements were determined using inductively coupled plasma mass spectrometry. Migration tests were conducted according to the Chinese National Food Safety Standard – Ceramic Ware (GB 4806.4) with microwaveable and non-microwavable samples under different conditions. The food consumption of consumers via different shapes of ceramic tableware was obtained through a self-reported web-based survey, and the estimated dietary intakes of the studied elements were calculated accordingly. The exposure assessment showed that certain metals leached from the ceramic tableware at levels of concern. In addition, the applicability of the migration experiment conditions for microwaveable ceramic ware in GB 4806.4 needs to be further investigated.


Introduction
Ceramics are generally made using mixtures of clay, earthen elements, powders and water, which are then shaped into desired forms and hardened them by heating to high temperatures. Ceramic dinnerware and cookware are normally glazed to produce a nonporous, waterproof and glass-like surface (González-Soto et al. 2000). Therefore, they have many merits, including corrosion resistance, high-temperature insulation and ease of cleaning (Aderemi et al. 2017).
In general, the glazing process involves materials containing heavy metals such as lead and cadmium (Bulut et al. 2021). Lead (Pb) is added to glazing materials for multiple reasons. First, lead can enhance the durability of chemical components to help them withstand the attack of detergents. Second, lead plays an important role in forming a smooth and durable hygienic surface that resists scratching damage. Third, lead compounds strengthen the connection between the glaze and ceramic body (Abou-Arab 2001;Ahmad et al. 2017). Cadmium (Cd) is a critical ingredient in some glaze compositions to produce a bright red colour and it can also improve the lustre and brilliance of other colours present in the glaze (Beale et al. 1991;Ajmal et al. 1997;Dong et al. 2015). Moreover, inorganic salts of barium (Ba), cobalt (Co), chromium (Cr), copper (Cu) and nickel (Ni) are also used as pigments to create colourful decorative glazes that coat the surfaces of ceramic bodies (Gazulla et al. 2007;Lin et al. 2014).
Among these elements, lead and cadmium have been recognised as toxic and non-essential to humans for many decades. Lead causes long-term harm in adults, including an increased risk of high blood pressure and kidney damage. Exposure of pregnant women to high levels of lead can cause miscarriage, stillbirth, premature birth and low birth weights. Young children are particularly vulnerable to the toxic effects of lead and can suffer profound and permanent adverse health impacts, particularly on the development of the brain and nervous system (Lanphear 2015). Cadmium primarily exerts toxic effects on the kidneys (Qing et al. 2021). However, it can also cause skeletal damage (Engström et al. 2012). Human exposure to cadmium has been statistically associated with an increased risk of cancer in the lungs, endometrium, bladder and breast (Larsson and Wolk 2015). In 1993, the International Agency for Research on Cancer (IARC) classified cadmium and cadmium compounds as human carcinogens (Group 1). Accumulating evidence indicates that Cd is also a neurotoxin that may lead to Alzheimer's diseaselike neurodegenerative disorders ). Elements such as cobalt, manganese, chromium, nickel, aluminium, barium, copper, antimony, vanadium and zinc, although some are essential to humans, can also adversely affect health when excessive exposure occurs (Zoroddu et al. 2019).
Metals from glazed ceramic food contact materials (FCMs) can leach into food and are potentially significant contributors to the dietary metal intake of humans (Wallace et al. 1985;Assimon et al. 1997;Sheets and Turpen 1997;Aderemi et al. 2017;Ahmad et al. 2017;Mandal and Das 2018). The leaching or migration process has been theoretically explained in the literature (Demont et al. 2012;Partyka and Lis 2012). Briefly, migration starts from an ion-exchange reaction and a state of equilibrium is formed when hydrolysis gradually becomes the active process. This migration process can be affected by multiple parameters. For instance, if glazed ceramics are not properly heated in terms of temperature and time, toxic metal elements from the materials may be released in nonnegligible amounts into foods and pose a health threat to consumers.
As a safeguard against the health risk of heavy metal migration from ceramic FCMs, many administrative regulators worldwide have set specific migration limit (SML) and migration test requirements for lead and cadmium. For example, in 1984, the European Union issued Directive 84/500/EEC, stipulating that the quantities of lead and cadmium transferred from ceramic articles shall not exceed their respective limits. The EU is currently considering lowering the limits for lead and cadmium in the directive and adding limits for other metal elements (European Commission 2019).
Chinese people have a long history of using ceramics as dishware and ceramic FCMs continue to play a very important role in Chinese culinary culture today because of their practical and aesthetic functions. The Chinese National Food Safety Standard -Ceramic Ware (GB 4806.4-2016) of the China National Health and Family Planning Commission (2016a) was implemented to ensure that ceramic FCMs sold and used in China do not adversely affect consumer health. Considering the changing dietary patterns of Chinese people, the development of ceramicware production processes and the increasing research into the toxicity of heavy metals, whether the permissible limits for Pb and Cd in GB 4806.4-2016 are still appropriate and whether it is necessary to set limits for other metal elements in GB 4806.4-2016 are topics that are worth investigation using various scientific measures, including risk assessment.
The present work analysed the migration of 18 elements leached from different shapes and types of ceramic tableware. An exposure assessment was conducted using preliminary food consumption survey data focusing on ceramic usage. To the best of our knowledge, risk assessments focused on Chinese consumers using ceramic tableware have rarely been reported to date.

Reagents
Unless otherwise stated, all of the reagents used were of analytical grade. Concentrated nitric acid (65% v/v) and acetic acid were purchased from Sinopharm Chemical Reagent Co., Ltd. (Shanghai, China). Ultrapure water (18 MΩ cm) was prepared using a Milli-Q water purification system (Millipore, Milford, MA, USA). Food simulant 4% (v/v) acetic acid was freshly prepared by adding 40 mL acetic acid to a 1000 mL volumetric flask and filling to the mark with Milli-Q water. A multielement standard solution (100 mg/L) and rhodium (Rh) as the internal standard (1000 mg/L) were obtained from Guobiao Testing & Certification Co., Ltd. (Beijing, China).

Ceramics
A total of 767 brand-new ceramic tableware samples were randomly collected from retail markets, online shops and manufacturers in seven provinces across China, regardless of the glaze and pigment type (Table 1). Each sample contained at least four identical articles from the same batch. According to GB 4806.4-2016 National Food Safety Standard -Ceramic Products, the samples were categorised into four groups based on their shape and size: flatware (178), small hollowware (204), large hollowware (144) and cups and mugs (241). Samples labelled as microwaveable (391) were subjected to migration tests under different conditions from those used for samples without labels identifying them as microwaveable (376).

Migration test
Migration tests were carried out using the "filled method" based on National Standard of the People's Republic of China GB 5009.156-2016 -General Principle of Migration Test Pre-treatment Method of Food Contact Materials and Their Products (China National Health and Family Planning Commission 2016b). After removing any visible contaminants attached to the surfaces, the ceramic pieces were washed with distilled water, rinsed with ultrapure water and air-dried at room temperature. According to GB 4806.4-2016, non-microwavable samples were filled with 4% acetic acid solution at room temperature (22°C) for 24 h, whereas microwaveable tableware samples containing the food simulant were heated in a (non-microwave) laboratory oven at 100°C for 15 min. For clarification, the microwaveable samples were preheated to 100°C in a (non-microwave) oven for an hour, after which a 4% acetic acid solution (also brought to 100°C in advance) was used to fill the samples. Two replicates of each sample were used for testing and each was covered with a glass plate to prevent evaporation of the food simulant.
Given that the ceramics are intended for repeated use, each specimen was subjected to three consecutive migration tests under the same conditions with the same amount of food simulant. Between the three tests, the samples were washed with ultrapure water and air-dried before the addition of the food simulant for the next experiment. The test solutions from the three consecutive migration tests were directly transferred for inductively coupled plasmamass spectrometry (ICP-MS) analysis. If precipitated matter was present in the test solution, a PTFE filter was used to remove it before analysis. The results of the third experiment were registered as the final results.
The calibration standard solutions of Li, Ti, V, Cr, Mn, Co, Ni, Cu, As, Mo, Ag, Sb, Ba, Cd and Pb ranged from 0.5 µg/L to 20 µg/L and those of Al, Fe and Zn ranged from 10 µg/L to 200 µg/L. Calibration curves were prepared in 4% acetic acid by stepwise dilution of multi-element standard solutions. The concentration of the internal standard (Rh) was kept constant independent of the dilution applied (20 μg/L). Further dilution was applied if the concentrations of the target elements were outside the calibrated range. Prior to each batch of samples pure solvent was injected as to avoid cross contamination. The stability of the ICP-MS system was checked by running a calibration standard every ten to twelve samples.

Risk assessment
The target hazard quotient (THQ), which is the ratio of the potential human exposure to a substance to the level at which no adverse effects are expected (the oral reference dose of the substance, RfD) or the benchmark dose lower confidence limit (BMDL) of the substance (no threshold for the toxicity of the substance), was employed to assess the health risk resulting from the intake of metal elements through migration from ceramic tableware. Generally, when THQ ≥ 1, the adverse effects caused by the intake of metal are considered nonnegligible, whereas when THQ < 1, the risk from the target metal elements can be ignored. As for food contact material, such as ceramic articles, an allocation factor of 20% is set in the exposure assumption since the primary source of metal elements exposure is considered dietary other than FCMs (European Union 2018).
In the present study, when the intake level of a heavy metal through migration of ceramics exceeds 20% (THQ > 0.2) of the health-based guideline values (HBGV) or BMDL, the risk of adverse effects from prolonged use of these FCMs cannot be neglected. The estimated daily intake was obtained by multiplying the daily food consumption via ceramic tableware by the mean concentration of metal elements in the migrating solutions, divided by the average body weight. The corresponding equations are THQ where EDI x is the total estimated daily intake of a metal element migrating from ceramic tableware (µg/kg bw/d); Rf D x is the oral reference dose of the element (µg/kg bw/d); BMDL x is the benchmark dose lower confidence limit which is used for metals whose toxicity has no threshold, for example, Pb; M t is the mean migrant concentration of a specific metal (x) from ceramic tableware type (t) (mg/L), C t is the amount of daily food consumption via a ceramic tableware type (t) (g) and BW is the mean body weight (kg). The ceramic tableware types include flatware (FW), small hollowware (SH), large hollowware (LH) and cups and mugs (CM).
The HBGV or BMDL of each element is listed in Table S1 of the supplementary material, except for titanium. Only health-based guidance values issued by authoritative agencies, such as the WHO, JECFA, EFSA, or USEPA, were used for the assessment. The hazard index (HI) was calculated by adding the THQ values of each element: HI ¼ P THQ: When HI < 1, the health risk of migration of the studied elements from the ceramic tableware is considered negligible, while HI ≥ 1 indicates a potential health concern.

Food consumption and ceramic usage data
The food consumption of Chinese consumers using ceramic dinnerware is not available in the literature or in industry, especially the respective amounts from different ware types. Therefore, a preliminary 24 h recall survey was conducted from 1 January 2022, to 31 May 2022. The questionnaire entries were imported to the professional online survey platform Wenjuanxing (www.wjx.cn) and distributed via WeChat, a social media app, within China. The questionnaire in the present study was designed and revised by experienced researchers who participated in the China National Nutrition and Health Survey (CNNHS) and were personally involved in the face-to-face interviews.
The questionnaire consisted of three parts. The study objectives and instructions for the questionnaire were provided in the first part. The second part collected demographic information, including the gender, age, body weight and residence (urban/rural) of the respondent. Food consumption data were collected in the third part of the questionnaire. The types of ceramic ware were also described and illustrated in this section. The participants were asked to fill in the amount (e.g. 150 g) in the corresponding boxes and only those consumed via a ceramic article were documented. Participation was voluntary, the responses were anonymous and no identifiable information was collected. Submission of the questionnaire was regarded as consent to participate.

Statistical analysis
A normal distribution was not observed in any of the element migration data groups and a two-sided Wilcoxon rank-sum test was performed using R (R Core Team, R Foundation for Statistical Computing, Vienna, Austria. URL http://www.R-project.org/ver sion4.0.2). Statistical significance was set at p < 0.05.

Method performance
The instrumental method used in this study was validated based on multiple parameters, including the recovery, precision, linearity and limit of quantification (LOQ). The recoveries obtained from spiking samples at the three concentrations (low, medium and high) were 76.0%-117%. The precision was expressed as a percentage relative standard deviation (RSD) ranging from 0.5% to 19.9% for the studied elements. The linear correlation coefficient (R 2 ) obtained for the calibration curve for each element was greater than 0.996. To determine the LOQ, 11 blank samples containing 4% (v/v) acetic acid were measured. The LOQ of each element was calculated as ten times the standard deviation (SD) of the blank results. The spiked levels, recoveries, precisions and LOQs of the 18 elements are listed in Table 3.

Migrant concentrations
The migration levels of 18 different elements from 767 ceramic pieces were analysed. In brief, the frequency of quantification (≥LOQ) for all 18 elements was less than 60%. Ba (51%) had the highest quantification frequency, followed by Co (49.5%), Pb (43.8%), Mn (41.9%), Al , Cd (11.9%) and As (4.0%). The lowest quantifiable rate was obtained for Sb (0.3%). The concentrations of leached metals from flatware articles are shown as μg/dm 2 and the others are expressed as μg/L. For the calculation, the left-censored results (data below the LOQ) were replaced by a value equal to half the LOQ.
Of these 18 elements, Pb and Cd are of the greatest concern because of their toxicity. The average concentrations of Pb leached from the four types of ceramic foodware were 7.64 μg/dm 2 for FW, 8.86 μg/L for SH, 30.8 μg/L for LH and 0.32 μg/L for CM. Cd was found in the leachates of FW, SH, LH and CM with mean values of 0.007 μg/dm 2 , 0.71 μg/L, 0.10 μg/L and 0.042 μg/L, respectively. The mean Cd concentration was significantly lower than that of Pb. The same trend was observed by Li and Zhang (2021) in kitchenware from Belgian markets. The mean concentrations of Pb were 0.019 mg/L for hollowware and 0.016 mg/dm 2 for flatware, whereas the mean concentrations of Cd were 0.003 mg/L for hollowware and 0.0007 mg/dm 2 for flatware.
Many countries and international organisations have set limits for the migration of Pb and Cd from different types of ceramics. In the present study, only one large hollowware product (1083 μg/L) exceeded the SML for Pb of 1000 μg/L for LH, as stipulated in China National Standard GB 4806.4-2016. Similar incidents were found by Yaman and Buyukaslan (2013), where some Turkish porcelain plates (flatware) released Pb up to 4.140 mg/ dm 2 and Cd up to 0.084 mg/dm 2 with 4% acetic acid, which is higher than the FDA limits of 0.8 mg/dm 2 for Pb and 0.07 mg/dm 2 for Cd.
More than half of the studied samples released Ba into the 4% acetic acid solutions, with concentrations from <LOQ to 44.4 μg/dm 2 for FW, <LOQ to 188.2 μg/L for SH, <LOQ to 456.7 μg/L for LH and <LOQ to 109.2 μg/L for CM. Barium was suggested to replace lead in frit formulations for glazes and barium salts act as colourants of the glazes, both of which could result in a relatively high the frequency of quantification.
The highest average value for Zn was 13.4 μg/dm 2 for FW, 37.0 μg/L for SH, 46.5 μg/L for LH and 17.3 μg/L for CM. This was followed by Al, with 3.32 μg/dm 2 for FW, 15.9 μg/L for SH, 26.9 μg/L for LH and 13.6 μg/L for CM. Zinc oxide and aluminium oxide are two important components in forming ceramic glaze and account for up to 6% of the total weight. Therefore, the high leaching levels are understandable.
Arsenic is a metalloid element with a dual identity: it is present in clay materials as a contaminant and arsenic trioxide, an opacifying oxide in glazes. In the literature, the leaching of As from ceramic articles has exhibited discrepancies. Henden et al. (2011) and Çiftçi and Henden (2016) noted that the amounts of As in leachates obtained from glazed ceramics were lower than those from non-glazed ceramics and glazing had a positive effect on avoiding the leaching of As in 4% acetic acid. Aderemi et al. (2017) found that arsenic leached from ceramic foodware in the range of 1.927-15.001 mg/L, which is considerably higher than the results obtained by Henden et al. (2011) and Çiftçi and Henden (2016), who reported all their findings were below 1 mg/L. The positive findings of arsenic in this study were mainly obtained from large hollowware, with a mean value of 4.97 μg/L and the maximum detected result was from a small hollowware specimen (888.7 μg/L). The presence of other elements in food simulants at various concentrations may be attributed to the different uses of these elements in making ceramic products. For example, oxides or salts of V, Ni, Co, Cu, Fe, Cr and Mn are commonly used in glazes as pigments, lithium, in the form of Li 2 O or Li 2 CO 3 , is used both for fluxing and for encouraging crystal growth in crystalline glazes. Oxides of titanium and antimony usually serve as opacifiers to give the ceramic surface whiteness. Further, titanium oxide provides bactericidal properties to some extent. Similarly, silver has been identified in the literature as an excellent antimicrobial agent (Yoshida et al. 2010).

Microwaveable wares
Among the 767 ceramic products collected in this study, 391 had labels claiming that they could be used in a microwave oven, including 114 flatware, 75 large hollowware, 126 small hollowware and 76 cups and mug items.
It is stipulated in the Chinese National Standard GB 4806.4-2016 that in migration tests for microwaveable ceramic products, the samples should be filled with 4% acetic acid solution for 15 min at 100°C. The simulants were heated to 100°C and duplicate samples were preheated at the same temperature in an oven for one hour.
The average values of the elements migrated from microwaveable samples of the four types of ceramic foodware and their corresponding values from nonmicrowaveable samples are listed in Table 4. The twosided Wilcoxon rank-sum test was applied to compare the mean leaching levels of each element from both groups in the four categories. It was found that elements including Li, Al, Ti, Cr, Mn, Fe, Co, Ni, Zn, Ba and Pb from microwaveable small hollowware and Al, Zn, Ba and Pb from microwaveable large hollowware had higher release amounts than those from nonmicrowavable articles (Z < 0, P < 0.05). Microwave heating is generally considered to be a harsher migration test condition. Li (2020) and Mandal and Das (2018) reported that the quantities of metals extracted by microwave heating were greater than those extracted by conventional thermal heating. However, this was not confirmed by the results of the present study. This could be attributed to several factors. First, the third results of the three consecutive migration tests were used in this study instead of the first results. As the migration tests proceed, or as the duration of the migration tests increases, the leachable metal elements reach a certain equilibrium. Second, heating in a nonmicrowave oven cannot simulate the process of metal migration from ceramics through microwave heating. In other words, 15 min at 100°C is not as harsh as microwaving at 140°C for 2 min and 30 s in convection mode, as proposed by Mandal and Das (2018), or microwaving at 400 W for 1 h or 600 W for 30 min, as proposed by Li (2020). Owing to restrictions in size and shape, not all of the microwaveable articles could be placed in microwave heating equipment for migration tests. Therefore, it is feasible and desirable to find migration test conditions (heating time and temperature) for conventional heating in a non-microwave oven to properly simulate the microwave heating process. Additional investigation is required to revise and update GB 4806.4.

Risk assessment
Data on food consumption through ceramic usage were obtained via the preliminary survey. A total of 152 valid questionnaires were collected. On average, 1568 g of food and beverage were consumed daily using ceramic tableware, including 231 g from flatware, 865 g from small hollowware, 80 g from large hollowware and 392 g from cups and mugs. All of the food and beverage densities were set to 1 kg/L for calculation purposes. The mean body weight of the respondents was 61.5 kg. The consumption data for flatware was subsequently multiplied by the average surface-to-volume ratio (S/V) of the flatware samples (15.8 dm 2 /L) to obtain a daily intake estimation of 3.65 dm 2 . The estimated daily intake (EDI, expressed in μg/ kg bw/day) of the elements from the four types of articles and from the total ceramic usage presented in Table 5 were calculated along with the THQ.
The decreasing order of the metals leached from ceramic specimens according to their THQ values was Co > Pb > V > As >Mo > Li > Cd > Sb > Ag > Al > Zn > Ni = Ba > Fe > Mn > Cu > Cr. None of the EDIs of the studied elements exceeded RfD or BMDLx. The THQ values for Mo, Li, Cd, Sb, Ag, Al, Zn, Ni, Ba, Fe, Mn, Cu and Cr were below 0.01, suggesting that the migration of these elements did not pose a significant health risk to ceramic tableware consumers. The THQ values for V and As were below 0.05, at 0.0352 and 0.0289, respectively. Arsenic is highly toxic in its inorganic form and is widely distributed throughout the environment in water and land. Manufacturers should use raw materials and water with low arsenic levels whenever possible to reduce the risk of arsenic migration from ceramic tableware.
However, for Co and Pb, the THQ values were 0.7088 and 0.4716, respectively. In general, diet and water are the two main sources of oral exposure. When the intake of a heavy metal through migration of FCMs exceeds 20% of the HBGV (THQ > 0.2), the risk of adverse effects from prolonged use of ceramics cannot be neglected (European Union 2018). The toxic properties of Pb are cumulative. Even trace amounts of Pb in continued daily intake will cause damage to multiple body systems. Cases of lead poisoning resulting from the consumption of food and beverages stored in glazed ceramics have been reported previously (Mohamed et al. 1995;Pantic et al. 2018). Li and Zhang (2021) revealed that the risk of exposure to lead by Belgian consumers needs to be considered, particularly for vulnerable children. The THQ of Pb in this study suggests that the lead intake through the use of ceramic tableware is close to 50% of the BMDL 01 of Pb, indicating the potential health risk could not be completely ruled out. Cobalt is an essential trace element in the human body and is generally thought to be relatively non-toxic. However, Z < 0 (appear in bold type) means that the amount of the element migrated from microwaveable foodwares is greater than from non-microwaveable ones. P < 0.05 (followed by an asterisk) means the result of the Wilcoxon test is statistically significant. long-term excessive intake could endanger human health (Leyssens et al. 2017). As elucidated by the US EPA (2008), the RfD used in this study is of low confidence. Nevertheless, a THQ of 0.71 suggests that regulatory authorities should collect sufficient toxicological and risk assessment data to consider whether to include the migration limit of cobalt in GB 4806.4. A similar previous study has been reported by the Norwegian Scientific Committee for Food Safety, which assessed the metal leakage from handcrafted ceramic articles produced by individual potters sold in the Norwegian market. The results suggested that the risk of exposure to Co from ceramic ware could not be completely excluded for the subgroup using ceramic ware with a large release of Co (EFSA Publication 2007). The Hazard Index (HI) exceeded one in this study, indicating that the collective effect of all heavy metals was not negligible, with cobalt and lead contributing the most to the potential health risk. A more refined cumulative risk assessment approach, such as the maximum cumulative ratio (MCR) method, should be employed to evaluate the combined effect of metal migration from ceramics.

Limitation of this study
This study had three main limitations. First, the food consumption of consumers using ceramic tableware was obtained by completing an online questionnaire survey. While this approach is easier to organise and conduct and more convenient to obtain data, a threeday, 24-h retrospective household survey is still a more acceptable means to obtain food consumption data. The method used in this study was a preliminary attempt. Second, for ceramic samples labelled as microwaveable, only one migration experimental condition was used in this study, according to the regulations of the Chinese national standard. To further investigate whether the migration experimental conditions in the national standard for microwaveable ceramic tableware are sufficiently stringent, experiments should be conducted using different migration conditions at different temperatures and times. Third, confidence in some RfDs (e.g. silver, vanadium and cobalt) was considered low to medium. Thus, moderate uncertainty exists in the risk assessment results.

Conclusions
In this study, migration tests were conducted on 767 pieces of ceramic tableware and the migration levels of 18 elements were measured using an ICP-MS. Among them, only 1 large hollow product showed noncompliance with the SML of 1000 μg/L for Pb from the LH stipulated in the Chinese national standard. A risk assessment conducted through a preliminary survey of ceramic tableware use revealed that health risks associated with the migration levels of lead and cobalt in ceramic tableware were of concern to Chinese consumers. This study also concluded that the experimental conditions of the migration tests used in the national standard for microwaveable samples require additional investigation to ensure their applicability.

Disclosure statement
No potential conflict of interest was reported by the author(s).