Monitoring arsenic species concentration in rice-based processed products distributed in South Korean markets and related risk assessment

Abstract

Rice is an important grain as a major source of carbohydrates in Asia but contains more arsenic (As) than other grains. A total of 239 rice-based processed foods (rice, n = 30; rice cake, n = 30; porridge, n = 39; noodles, n = 33; bread, n = 20; snack, n = 59; powder, n = 28) were purchased in 2019 from domestic markets to measure total As (tAs) and As species. The average tAs and inorganic As (iAs) in each sample group ranged from 20 to 180 μg/kg (porridge for baby to noodle) and 4.4–85 μg/kg (porridge for baby to powder), respectively. The correlation between the iAs and tAs was affected by the variety of ingredients, such as the presence of seaweed (tAs) and the milling type of rice (iAs). Although rice cakes and baby rice-based powders are a source of concern for both adults and children, respectively, risk assessments indicate that most rice-based foods are generally safe to consume in South Korea.

Graphical abstract

Supplementary Information

The online version contains supplementary material available at 10.1007/s10068-023-01270-9.

Introduction

Rice is one of the most commonly consumed foods in South Korea, which is comparable to other Asian countries (e.g., China, India, Malaysia, and Thailand) in terms of relatively high yearly consumption (59.2 kg per person) (KOSIS, 2020). Rice can be consumed in processed forms, such as rice cakes, noodles, porridges, and snacks (Han and Gouk, 2014). As the demand for gluten-free foods has increased over time, instead of wheat, rice has become one of the most common carbohydrate sources for individuals with gluten sensitivity or celiac disease (Bascuñán et al., 2017). Rice-based meals are a substantial source of carbohydrates for babies owing to their nutritional value, low allergy risk, lack of gluten, ease of preparation, and good gourmet cultural settings (Jung, 2018). Although these rice-based processed products are becoming more popular, they pose health risks owing to their higher arsenic (As) content relative to that of other grains (Signes-Pastor et al., 2016; Williams et al., 2007), this is because rice has a high propensity to accumulate As from its anaerobic flooded growth environment (Jackson et al., 2012; Norton et al., 2009).

As is the top compound on the Agency for Toxic Contaminants and Disease Registry's priority list of substances that pose the highest risk to human health (ASTDR, 2019). In previous studies, inorganic As (iAs; arsenite (AsIII) and arsenate (AsV)) was linked to bladder, skin, and lung cancers (Hindmarsh et al., 1986). Further, the International Agency for Research on Cancer (IARC) has classified iAs as a group 1 carcinogen (IARC, 1987; Rousseau et al., 2005). The IARC also categorized organic As (oAs; monomethylarsonic acid (MMA), and dimethylarsinic acid (DMA)) as Group 2B, which indicates possibly carcinogenic to humans (IARC, 2012). Previously, the Joint FAO/WHO Expert Committee on Food Additives (JECFA) suggested benchmark dose lower confidence limit (BMDL) values for iAs. The BMDL₀₁ range for iAs is 0.3–8 μg/kg body weight (bw)/day (Alexander et al., 2009) while its BMDL_0.5 range is 3 μg/kg bw/day (WHO, 2006). The BMDL values of oAs remain unknown. Rice consumption may be the main source of iAs exposure. In our previous study, iAs comprised 43–91% of total As (tAs) in polished rice: tAs = 0.088 ± 0.021 mg/kg and iAs = 0.060 ± 0.013 mg/kg (Lee et al., 2018).

It is necessary to monitor the iAs concentration in rice and investigate its impact on the health of South Korean consumers. The ingestion rate, frequency of intake, length of food consumption, age, and weight of consumers are factors that contribute to risk assessment (Sharafi et al., 2019). As rice is the main staple in South Korea, the above characteristics have a considerable influence on the rate of intake and risk to human health posed by iAs. Using inductively coupled plasma mass spectrometry (ICP-MS) and HPLC-ICP-MS, the present study monitored the content of tAs and four different As species (AsIII, AsV, MMA, and DMA) in 239 rice-based processed foods from ten categories distributed in domestic markets in South Korea. The current study was also performed to estimate the carcinogenic risk of iAs in rice-based processed products using the margin of exposure (MOE) approach. The findings of this study could serve as a basis for the development of useful dietary guidelines for controlling the quality of rice-based processed foods in terms of iAs and determining the effect of their intake on the health of South Koreans.

Materials and methods

Instrumentation

All As analyses were carried out using ICP-MS (Agilent 7700x, Agilent Technologies, Santa Clara, CA, USA), and the experimental conditions used in previous studies that employed the same instrument for As evaluation was slightly modified and applied in this study (Lee et al., 2018). The ICP-MS working conditions are presented as follows: RF Matching, 27.12 MHz; RF Power, 1550 W; Auxiliary gas flow, 1.0 L/min; Nebulize (carrier) gas flow, 1.12 L/min; Collision cell, He flow 4.0 mL/min; single-ion monitoring at m/z 75; Lens voltage, 10 V; sampling depth, 8.0 mm.

An HPLC system (Agilent 1200 series, Agilent Technologies, Santa Clara, CA, USA) equipped with a CAPCELL PAK C18 MG column (4.6 × 250 mm, 5 μm, Shiseido, Tokyo, Japan) coupled with ICP-MS was used to determine the As species. The operational settings for HPLC-ICP-MS were optimized based on previous studies (Kim et al., 2020). The HPLC working conditions are presented as follows: Column temperature, 25℃; Mobile phase [0.05% (v/v) methanol, 10 mM sodium 1-butane sulfonate, 4 mM malonic acid, 4 mM tetramethyl ammonium hydroxide (TMAH) (pH 2.7, nitric acid)]; flow rate, 0.75 mL/min; injection volume, 15 μL. The detailed HPLC working conditions for As specification analysis are presented in Table S1. The ICP-MS instrument consisted of a double-pass spray chamber and a micro-mist concentric nebulizer. A PEEK capillary was used to transport the mobile phase from the separation column to the nebulizer.

Reagents and solutions

Ultrapure deionized water (YL WPS System, Young Lin Instrument Co., Gyeonggi-do, South Korea) was used in the present study. Arsenic standard solutions for ICP (Fluka, Buchs, Switzerland), hydrogen peroxide, and nitric acid (Chemitop, Chungcheongbuk-do, South Korea) were used for tAs analyses. Arsenic oxide hydrate (Inorganic Ventures, Christiansburg, VA, USA), arsenic trioxide (Inorganic Ventures), disodium methyl arsonate hexahydrate (Chemservice, Pittsburgh, PA, USA), and sodium cacodylate trihydrate (Sigma-Aldrich, St. Louis, MO, USA) were used as speciation analysis standards. For the chromatographic mobile phase, malonic acid (Wako, Osaka, Japan), HPLC-grade methanol (Burdick & Jackson, Ulsan, South Korea), sodium 1-butanesulfonate (Sigma-Aldrich), and tetramethylammonium hydroxide (Sigma-Aldrich) were used. The analytical method was validated using certified reference material (CRM) NIST SRM 1568b rice flour (National Metrology Institute of Japan, Ibaraki, Japan). All chemicals used in this study were of analytical grade and were used without additional purification.

Samples

A total of 239 rice-based processed food samples were purchased from domestic markets from February to August 2019. Based on the number of search results for the item in the online mall, the top ten categories were selected from the report's categorized products (Lee and Eun, 2008) and baby products, including home meal replacement (HMR) rice (n = 30), rice cake (n = 30), porridge (n = 25), noodles (n = 33), bread (n = 20), snack (n = 45), powder (n = 19), snack for baby (n = 14), porridge for baby (n = 14), and baby rice-based powder (n = 9). The number of samples in each category was determined according to the category's sales volume and purchasing preferences (Han and Gouk, 2014; Kim et al., 2018).

tAs analysis

The tAs concentration was analyzed whole As (including As species) exiting in a sample using the Microwave method and ICP-MS analysis. Rice-based processed product samples were digested according to the Korea Food Code (KFC) method 8.9.1.4 (KFC, 2020a). Briefly, 0.5 g of each dried and powdered food sample was added to 4 mL of nitric acid (70%, v/v) in polytetrafluoroethylene (PTFE) tube. Thereafter, the sample mixtures were heated on a hot plate (140 °C) for 90 min. After cooling to 25 °C, 1 mL of hydrogen peroxide (30%, v/v) and 3 mL of nitric acid (70%, v/v) were added. The PTFE tubes were sealed and placed in a microwave digestion apparatus (ETHOS EASY, Milestone, Italy). Digestion was carried out in five steps: (1) 5 min at 1600 W to increase the temperature to 80 °C, (2) 5 min at the same voltage to lower the sample temperature to 50 °C, (3) 15 min at 1600 W to increase the temperature to 190 °C, (4) 20 min at the same conditions, and (5) a cooling interval to lower the sample temperature to 60 °C or lower. The sample mass was adjusted to 20.0 g before ICP-MS analysis by adding distilled water after digestion. The tAs was quantified using an 8-point external calibration ranging from 0.5 to 50 µg/kg. Standard samples were prepared in 2.5% (v/v) nitric acid.

As speciation analysis

The pretreatment procedure for As speciation analysis was conducted according to the iAs analysis method in the KFC method. 8.9.1.5 (KFC, 2020b). Briefly, 1.0 g of each rice-based food sample was placed in a polypropylene tube, which was then filled with 5 mL of 1% (v/v) nitric acid and heated in a water bath at 90 °C for 90 min. To enable improved mixing of the sample and nitric acid during the first 30 min, the samples were vigorously shaken for 30 s at 5-min intervals. The extract was cooled to 25 °C and distilled water was added to adjust the volume to 25 mL. Diluted samples were centrifuged for 10 min at 3000×g using an Allegra X-15R centrifuge (Beckman Coulter, USA), and 10 mL of the supernatant was collected. After the supernatants were centrifuged for 10 min at 3000×g, 3 mL was filtered through a 0.45 µm nylon membrane filter and analyzed as samples. Two iAs and two oAs were separated using a C18 reversed-phase column and HPLC-ICP-MS. The quantification of As species was measured using 7-point external calibration ranging from 0.5 to 50 µg/kg. The chemical standards were dissolved in ultrapure deionized water at a concentration of 400 mg/kg to prepare a stock solution of each of the four As species. Stock solutions of standards were stored at 4 ℃ until analysis. Stock solutions of the As species standard were combined in equal amounts and diluted with 0.2% (v/v) nitric acid to prepare fresh working solutions for each analytical batch.

Validation

Linearity, precision, and accuracy were used to validate the method according to International Conference for Harmonization (ICH) guidelines Q2R1 (2005). A representative sample was collected per each group in equal amounts to reflect the characteristics of the matrix (Kim et al., 2020). The representative sample was spiked by adding the arsenic standards to it. A standard curve was plotted using five different concentrations of the element to span the range of spiked sample concentrations (2.5, 5, 10, 25, and 50 μg/kg). Method detection limit (MDL) and method quantification limit (MQL) were obtained through seven calibration curves by performing seven repeats of the spiked sample. MDL and MQL were calculated by 3.14 σ/S (σ is the standard deviation of the y-intercept, and S is the mean of the slope) and 10 σ/S, respectively. The certified reference material, CRM NIST SRM 1568b rice flour (tAs, 285 μg/kg; iAs, 92 μg/kg; MMA, 11.6 μg/kg; DMA, 180 μg/kg), was analyzed to derive the accuracy (Eq. 1) and precision (Eq. 2). An equal quantity of reference material was pretreated, and seven replicates were analyzed in an intraday test. This intraday test was repeated three times for the subsequent three days to produce an inter-day test. For quality control, CRM NIST SRM 1568b rice flour was evaluated for both tAs and As species measurements using each batch of samples. The external calibration standards for quantification were prepared daily from standard stock solutions. The calibration ranged from 0.5 to 50 μg/kg, and tAs and each As species were analyzed in all batches. The formulas used for the accuracy and precision are as follows:

$$\text{Accuracy}\left(\text{\%}\right)=\frac{\text{Mean}\text{measured}\text{concentration}\left(\mathrm{\mu }\text{g/kg}\right)}{\text{Nominal}\text{concentration}\left(\mathrm{\mu }\text{g/kg}\right)}\times 100$$ 1

$$\text{Precision}\left(\%\text{RSD}\right)=\frac{\text{Standard}\text{deviation}\text{of}\text{measured}\text{concentration}\left(\mathrm{\mu }\text{g/kg}\right)}{\text{Mean}\text{measured}\text{concentration}\left(\mathrm{\mu }\text{g/kg}\right)}\times 100$$ 2

where RSD represents relative standard deviation.

Risk assessment

A risk assessment was performed for ten categories based on our monitoring data and the rice-based food category intake by age. To collect the consumption data of processed foods to conduct the risk assessment of iAs, data from the 7th edition (2016–2018) of the Korea National Health and Nutrition Examination Survey (KNHNES), published by the Korea Center for Disease Control and Prevention (KCDC), were employed (Tak et al., 2018). Data for estimated daily consumption were generated by separating the data into five categories based on age (Alexander et al., 2009). The iAs intake through the consumption of rice-based foods was estimated by multiplying category-specific food intake data by each food group's mean iAs concentration, including the value of not detected (ND) set to be equal to MDL level (0.73 μg/kg) of iAs (Eq. 3).

$$\text{Daily}\text{iAs}\text{intake}=\frac{\text{Daily}\text{consumption}\left(\text{g/d}\right)\times \text{iAs}\text{concentration}\left(\mu \text{g/kg}\right)}{\text{Average}\text{body}\text{weight}\left(\text{kg}\right)}$$ 3

The risk of iAs was assessed using the MOE. Based on previous epidemiologic studies, the MOE (Eq. 4) was estimated using the benchmark dosage (BMDL₀₁; 0.3–8 µg/kg bw/day), which was assessed by the EFSA and found to cause a 1% increased risk of skin, lung, and bladder cancers (Alexander et al., 2009). A BMDL₀₁ intake value of 0.3 µg/kg bw/day was selected for the risk assessment, thereby aligning with the lower end of the BMDL₀₁ value range. The average body weight of each group (age 1–2: 12.0 kg; age 3–5: 17.6 kg; age 6–18: 45.5 kg; age 19–64: 64.5 kg; age over 65: 60.6 kg) was used to calculate daily iAs intake (Tak et al., 2018). The formula used for the risk assessment is as follows:

$$\text{MOE}=\frac{\text{BMDL}}{\text{EDI}}=\frac{0\text{.3}\left(\mu \text{g/kg/d}\right)}{\text{Daily}\text{iAs}\text{intake}\left(\mu \text{g/kg/d}\right)}$$ 4

where EDI represents estimated daily intake.

Statistical analysis

All statistical analyses were performed using R statistical software (http://www.r-project.org/). The analyzed data were processed to compare the levels among food groups. Significant differences were defined using the t-test and one-way ANOVA. Statistical significance was set at p < 0.05.

The correlation factor (r) was used to determine the correlation between tAs and iAs for each dietary group, and a correlation analysis was performed using the p-value at a 95% confidence level. The reason of selected the Noodles and snacks was that were the top 2 groups with the largest number of samples and were able to classify the based types of rice or the contained raw material such as seaweeds. The raw material was used to compare the tAs and iAs concentrations in noodles, whereas the type of rice was used to compare the concentration in snacks. The following equation was used to calculate the correlation factor (Eq. 5) and t-factor (Eq. 6):

$$\text{Correlation}\text{factor (r)}=\frac{\text{cov(x,y)}}{\sqrt{\text{var}\left(\text{x}\right)\text{var(y)}}}=\frac{\sum \left({\text{x}}_{\text{i}}-\overline{x}\right)\left({\text{y}}_{\text{i}}-\overline{y}\right)}{\sqrt{\sum {\left({\text{x}}_{\text{i}}-\overline{x}\right)}^2{\left({\text{y}}_{\text{i}}-\overline{y}\right)}^2}}$$ 5

$$t\text{factor}=\frac{r\sqrt{\text{n}-2}}{\sqrt{1-r^2}}$$ 6

where x and y denote tAs and iAs concentrations, respectively. The tAs and iAs values of each sample are indicated by xi and yi, respectively. The degrees of freedom are denoted by n-2, and the correlation factor is denoted by r.

To calculate the variety of food ingredients by category, the following equation was used:

$$\text{Diversity}\text{index}\left(H^{\prime }\right)=-\sum _{i=1}^n{p}_i\text{ln}{p}_i$$ 7

where n is the number of all ingredients used within the category, except rice, water, and seasoning. p_i denotes the prevalence of the ith component. The results are presented as median ± standard deviation (SD) (*p < 0.05).

Results and discussion

Validation of the analytical methods

As shown in Table 1, the coefficient of determination (R²) was used to determine linearity in the calibration curve ranging from 0.5 to 50 g/kg for tAs and each of the four As species. The R² values of the tAs and four As species were greater than 0.99, the MDL for tAs was 0.32 µg/kg, and the MDL for As speciation ranged from 0.26 to 0.37 µg/kg. For tAs, the accuracy and precision were 98.2% and 1.2%, respectively. Because the AsV and AsIII contents were not independently confirmed in the CRM NIST SRM 1568b, the sum of AsV and AsIII was employed as an iAs value; the CRM NIST SRM 1568b manufacturer provided the combined value of AsV and AsIII. The accuracies for iAs, MMA, and DMA ranged from 94.9 to 103%, while the precisions ranged from 0.94 to 4.0%. Based on these findings, the analytical procedures utilized in this study were deemed adequate for analyzing tAs and As species in rice-based processed food products.

Table 1.

Analytical performance of total arsenic (tAs) and arsenic species, and the validation data

Calibration curve
Analyte	tAs	AsV	AsIII	MMA	DMA
Regression equation	y = (4940 ± 1413)x + (5109 ± 509)	y = (4169 ± 126)x + (1075 ± 484)	y = (1910 ± 512)x − (3308 ± 223)	y = (4084 ± 197)x + (55 ± 335)	y = (3150 ± 88)x + (430 ± 307)
R2	0.9961	0.9998	0.9997	0.9998	0.9997
MDLa (μg kg−1)	0.32	0.36	0.37	0.26	0.31
MQLb (μg kg−1)	1.03	1.16	1.17	0.82	0.97

Reproducibility
Analyte	tAs	iAs (AsV + AsIII)	MMA	DMA
Intraday1
Accuracy (%)c	104.4	84.6	95.8	94
Precision (% RSD)d	3.5	2.9	11.3	2.1
Intraday2
Accuracy (%)	106.8	84.3	103	92.1
Precision (% RSD)	2.7	4.4	12.7	2.8
Intraday3
Accuracy (%)	82.9	115	102	111
Precision (% RSD)	2.2	6.2	12	2.1
Interday
Accuracy (%)	98.2	94.9	103	98.9
Precision (% RSD)	1.2	4	8.2	0.94

RSD = relative standard deviation; Arsenate (AsV); Arsenite (AsIII); Monomethylarsonic acid (MMA); Dimethylarsinic acid (DMA)

^aMethod detection limit, S/σ = 3.14

^bMethod quantifiation limit, S/σ = 10

^c(Mean measured concentration / nominal concentration) × 100

^d(Standard deviation/mean) × 100

Quantitative analysis of tAs and As species in rice-based processed foods

The concentrations of tAs and the four As species in rice-based processed food samples are presented in Table 2. The concentrations of tAs and the four As species varied depending on the ten sample categories. The average tAs concentrations in 239 rice-based processed products ranged from 40 ± 30 μg/kg in bread to 180 ± 190 μg/kg in noodles.

Table 2.

Concentrations and detection frequencies of arsenic species in rice-based processed foods ordered by food category

Food categories (n)a	tAs (μg/kg)	Concentration of each arsenic species (μg/kg)
		AsV	AsIII	iAs	MMA	DMA
HMR rice	110 ± 150b	3.4 ± 9.1	27 ± 14	31 ± 18	1.3 ± 3.7	8.8 ± 14
(n = 30)	(97)c	(23)	(97)	(97)	(13)	(37)
Rice cake	69 ± 33	1.0 ± 5.5	36 ± 16	37 ± 18	1.5 ± 3.6	14 ± 15
(n = 30)	(97)	(3)	(97)	(97)	(17)	(60)
Porridge	49 ± 43	3.6 ± 8.2	15 ± 15	19 ± 22	0.80 ± 3.9	2.9 ± 6.4
(n = 25)	(72)	(24)	(72)	(72)	(8)	(24)
Noodle	180 ± 190	9.5 ± 13	39 ± 26	48 ± 34	4.8 ± 8.6	22 ± 30
(n = 33)	(94)	(39)	(93)	(93)	(25)	(64)
Bread	40 ± 30	3.0 ± 5	20 ± 15	23 ± 16	0.30 ± 1.3	5.7 ± 12
(n = 20)	(90)	(20)	(85)	(90)	(20)	(28)
Snack	120 ± 150	4.6 ± 8	39 ± 31	43 ± 36	3.6 ± 7.8	14 ± 18
(n = 45)	(84)	(27)	(82)	(82)	(59)	(51)
Powder	160 ± 170	6.6 ± 11	78 ± 48	85 ± 56	8.2 ± 12	12 ± 14
(n = 19)	(100)	(37)	(100)	(100)	(53)	(53)
Snack, for baby	110 ± 80	12 ± 8	49 ± 52	62 ± 55	0.65 ± 2.4	19 ± 11
(n = 14)	(100)	(79)	(100)	(100)	(8)	(86)
Porridge, for baby	22 ± 40	NDd	4.4 ± 8	4.4 ± 8	2.7 ± 8.3	3.0 ± 11
(n = 14)	(29)	(0)	(29)	(29)	(14)	(7)
Powder, for baby	160 ± 70	6.3 ± 9.9	67 ± 28	74 ± 31	17 ± 11	24 ± 20
(n = 9)	(100)	(33)	(100)	(100)	(100)	(78)

^an is the number of samples

^bAs species and total As concentration (μg/kg)

^cFrequency of detection (%)

^dResults < MDL were set to be equal to 0, and not detected (ND); inorganic As (iAs)

As was detected in most samples, except some baby porridge samples. The lowest amounts of tAs were found in bread samples, with a range of not detected (ND) to 118.0 g/kg. However, tAs was found in 90% of the bread samples assessed. Of the 186 samples, MMA was not detected in 125 samples, whereas DMA was not detected in any of the samples. According to several research, rice not only contains iAs, but also traces of MMA and DMA, whereas other cereals, such as wheat, only have trace amounts of iAs (Llorente-Mirandes et al., 2014; Rasheed et al., 2018).

To compare the ranges for iAs measurements, the present and prior studies (Gavilanes-Terán et al., 2019; Islam et al., 2017; Kollander et al., 2019; Munera-Picazo et al., 2014) on iAs concentrations in rice-based processed products were summarized (Table 3). Notably, the most common chemical forms of As are oAs and iAs (Devesa et al., 2001). According to several research, iAs, including AsIII and AsV, were suggested to be substantially more hazardous than oAs, with AsIII being markedly more toxic than AsV (Christensen and Luginbyhl, 1975; Penrose and Woolson, 1974). However, oAs species were determined to be reduced in the metabolic process in vivo, and toxicity studies on the reduced species in human cell lines revealed that they are among the most toxic arsenic species (Cubadda et al., 2017; Kim et al., 2016; Luvonga et al., 2020; Styblo et al., 2000). Although the current risk assessment for dietary As exposure mainly focused on iAs, to more accurately estimate As risk, the toxicological characteristics must be properly defined, and more research should be performed on oAs species that may be harmful to human health.

Table 3.

Comparison of the results of this study with those of published studies

Sample type	Country	iAs Range (μg/kg)	Mean (μg/kg)	iAs/tAs ratio (%)	References
HMR rice	South Korea	ND–90.5	30.9	29	This study
Rice cake	South Korea	ND–87.2	36.6	53	This study
Porridge	South Korea	ND–73.0	18.8	39	This study
Noodle	South Korea	ND–150	48	26	This study
Bread	South Korea	ND–46.4	22.7	57	This study
Snack	South Korea	ND–144	43.2	36	This study
Powder	South Korea	ND–283	84.7	52	This study
Baking flour	Spain	52.8–179	93.2	81	Sandra et al. (2014)
Pasta	Spain	73.7–90.5	82.1	69	Sandra et al. (2014)
Pastries	Spain	ND–198	133	90	Sandra et al. (2014)
Crackers	Austrailia	59–270	126	53	Islam et al. (2017)
Crackers	Thailand	58–63	61	36	Islam et al. (2017)
Rice cakes	Austrailia	27–158	105	51	Islam et al. (2017)
Rice cakes	Netherlands	45–141	96	63	Islam et al. (2017)
Rice cakes	Ecuador	114–132	123	NAa	Gavilanes et al. (2019)
Breakfast cereals	Ecuador	48–165	90.5	NA	Gavilanes et al. (2019)
Cereal bars	Ecuador	51–197	103	NA	Gavilanes et al. (2019)
Biscuit	Ecuador	67–122	94.5	NA	Gavilanes et al. (2019)
Crackers	Sweden	86–322	152	62	Kollander et al. (2019)
Breakfast cereals	Sweden	24–91	52	65	Kollander et al. (2019)
Porridge, ready to eat	Sweden	10–17	12	75	Kollander et al. (2019)

^aNon applicable

Compared with previously published data, our monitoring results revealed comparable mean iAs values within the relevant category. Further, the total iAs/tAs ratio appeared to decline when the amount of arsenic-containing supplementary items, besides rice, increased. Bread made with ingredients known to be free of As, such as eggs and butter, has a high iAs/tAs ratio (Martorell et al., 2011). However, noodles and porridge supplemented with shrimp, and seaweed that exclusively contains oAs, have a low iAs content (Choi et al., 2011; Khan et al., 2015; Wang et al., 2019). Brown seaweeds are frequently utilized in seafood-based foods owing to their umami flavor (Mouritsen et al., 2019). Seaweed supplementation in HMR rice, porridge, noodles, bread, and snack samples may have contributed to the comparably low iAs/tAs ratio obtained herein.

Although gluten-free rice-based processed products are increasingly available in the markets, such products are nutritionally unbalanced, and are associated with a low intake of fiber, vitamins (B group and D), and minerals (Ca, Fe, Mg, and Zn) (Vici et al., 2016). Seaweed is a natural source of nutrients for humans, including fatty acids, minerals, vitamins, and pigments (Fradinho et al., 2019). In the current study, because the concentration of tAs was significantly (p < 0.05) higher in the seaweed-containing group than the control group, seaweed-containing rice noodles and frequently-served snacks were examined to compare the concentrations of tAs and iAs in both categories by regrouping the material used (Fig. 1A–D). Husked rice had more tAs and iAs than polished rice, implying that although the milling type of rice had a relatively higher impact on the tAs/iAs concentration of processed rice meals, the incorporated seaweed components had a greater impact on the tAs concentration.

Fig. 1.

Comparison of the total arsenic (tAs) and inorganic arsenic (iAs) concentrations in rice-based processed foods. A–B Comparison of tAs and iAs concentrations in polished rice noodles (n = 26) and husked rice noodles (n = 7), respectively; C–D Comparison of tAs and iAs concentrations in rice noodles without seaweeds (n = 14) and with seaweeds (n = 19), respectively; E–F Comparison of the tAs and iAs concentration in polished rice snacks (n = 35) and husked rice snacks (n = 10), respectively; The results are presented as median ± standard deviation (*p < 0.05)

Because only one snack product that contained seaweed and one containing cod extract were employed in this study, and the remaining products were mainly rice with cereals, such as wheat and corn, the levels of tAs and iAs in snacks were evaluated based on the type of rice used. As shown in Fig. 1E, F, the iAs content of husked rice snacks was significantly (p < 0.05) higher than that of polished rice snacks. These results suggest that seaweed affects tAs concentrations, while husked rice affects iAs concentrations in rice-based products.

Correlation analysis between tAs and iAs

A positive correlation was found between tAs and iAs concentrations in rice samples (Lee et al., 2018). If a correlation exists between tAs and iAs levels in rice-based products, this correlation could be used to estimate iAs concentration using tAs measured value; this is because tAs levels are considerably simpler than iAs levels in samples. To estimate the impact of iAs relative to the tAs level in rice-based processed products, the correlation between the two was examined (Table 4). The correlation factor was employed to assess the relationship between tAs and iAs in each food group, while the p-value was used to evaluate the significance. The presence of a correlation was determined if the p-value of the t-test was < 0.05. The snack for the infant group was the only one product with a correlation value greater than 0.99, whereas the HMR rice category had a correlation factor of less than 0.07. These findings may be linked to the diversity of components in each food group, which were assessed by identifying and listing substances, besides rice, water, and seasoning for each individual sample in the category. The larger the variety of elements, the higher the diversity index (H′). The degree of association between iAs and tAs is inversely proportional to material diversity. Most samples in the snack for the infant group that had a very high correlation coefficient were solely comprised of rice; however, the HMR rice and noodle groups had a low correlation value owing to samples including additional items, such as fish, meat, other crops, and vegetables. DMA accounts for most of the As in sea mustard, whereas arsenobetaine accounts for most of the As in shrimp (Choi et al., 2011; Khan et al., 2015; Wang et al., 2019). Consequently, the r values between tAs and iAs may be limited in products containing a large proportion of subsidiary components carrying the most As in the organic form. There was no correlation among HMR rice, snacks, or noodles; however, a correlation was found in all other categories after confirming the significance of the correlation factor between tAs and iAs for each category using a t-test. Therefore, estimating the concentration of iAs in rice-based foods using tAs is not a suitable alternative, and As species analysis is necessary to more accurately predict As levels.

Table 4.

Correlation coefficient between tAs concentration and iAs concentrations by category

Category	Correlation factor (r)	p value	Diversity index (H′)
HMR rice	0.0657	7.30 × 10–1	3.45
Snack	0.280	6.23 × 10–2	3.23
Noodle	0.311	7.81 × 10–2	2.87
Porridge	0.565	3.23 × 10–3	2.95
Rice cake	0.781	3.56 × 10–7	2.88
Bread	0.823	8.47 × 10–6	2.29
Porridge for baby	0.829	2.51 × 10–4	2.85
Powder for baby	0.896	1.08 × 10–3	2.03
Powder	0.925	1.54 × 10–8	1.97
Snack for baby	0.990	1.20 × 10–11	1.84

If the p value was less than 5 × 10^–2 based on the t-test, it was judged that a correlation exists

Risk assessment

Table 5 shows the daily food intake and dietary exposure to iAs among South Koreans. The estimated daily consumption data for each age group were derived from the 7th edition (2016–2018) of the KNHNES (Tak et al., 2018). The MOE method was used to analyze the carcinogenic risks associated with rice-based processed food intake. Further, a conservative approach was adopted for the risk assessment using the lower end of the BMDL₀₁ value (Alexander et al., 2009). As shown in Table 5, the MOE calculations yielded values greater than 10⁵ for porridge intake and lower than 10² for rice cake intake. If the estimated value of MOE exceeds 100, it is regarded as safe for carcinogenic effects; if this value is less than 100, there is a considerable carcinogenic risk (Bessems et al., 2017; Demaegdt et al., 2021; Lachenmeier and Rehm, 2015). Rice cake and snack intake is a concern for individuals older than six and six–eighteen, respectively, whereas powdered baby food intake is a concern for children aged one–two. The ingestion of HMR rice, rice porridge, rice noodles, and baby rice-based foods, except baby rice-based powder, is safe based on the value of iAs. Considering the intake of all categories, the MOE estimations for those over the age consuming six rice-based processed products in our investigation yielded values more than 10³.

Table 5.

Estimated dietary exposure of South Koreans to arsenic and margin of exposure to iAs in food categories

Food category	HMR rice	Rice cake	Porridge	Noodle	Bread	Snack	Powder	All categories
iAs concentration
Meana	31.2	36.9	19.2	48.2	23.1	43.6	84.9	41.0
Age 6–18
Consumptionb	0.0517	15.5	0	0.502	0.306	8.45	0.303	25
EDIc	0.0355	12.6	0	0.532	0.155	8.10	0.565	22.5
MOEd	8.46 × 103	2.40 × 10	-	5.64 × 102	1.93 × 103	3.71 × 10	5.31 × 102	1.92 × 103
Age 19–64
Consumption	0.0387	17.3	0.006	0.878	0.176	2.74	1.01	22.1
EDI	0.0187	9.90	0.0018	0.656	0.0630	1.85	1.33	14.1
MOE	1.60 × 104	3.03 × 10	1.68 × 105	4.57 × 102	4.76 × 103	1.62 × 102	2.26 × 102	2.71 × 104
Age ≥ 65
Consumption	0.0062	15.8	0.0011	0.36	0.115	1.61	0.85	18.8
EDI	0.0032	9.62	0.0003	0.286	0.0438	1.16	1.19	12.7
MOE	9.40 × 104	3.12 × 10	8.61 × 105	1.05 × 103	6.84 × 103	2.59 × 102	2.52 × 102	1.37 × 105

Food category	Snack for baby	Porridge for baby	Powder for baby	All categories
iAs concentration
Mean	61.6	5.04	73.7	46.8
Age 1–2
Consumption	0.0009	0.02	1.67	1.69
EDI	0.0046	0.0084	10.3	6.59
MOE	6.49 × 104	3.57 × 104	2.93 × 10	3.36 × 104
Age 3–5
Consumption	0	0.19	0.05	0.24
EDI	0	0.0544	0.209	0.638
MOE	-	5.51 × 103	1.43 × 103	3.47 × 103

^aUnit: μg/kg

^bUnit: g/day

^cEstimated daily intake, Unit: ng/kg bw/day

^dMargin of exposure = BMDL/Estimated exposure dose, MOE < 100: concern (Lachenmeier and Rehm, 2015; Bessems et al., 2017; Demaegdt et al., 2020)

The MOE values for iAs in rice have previously been reported to be less than 1 for BMDL₀₁ (Rintala et al., 2014) and 5.1–59.8 BMDL_0.5 in multiple countries (Brandon et al., 2014; Huang et al., 2015). Compared to prior findings, our results do not indicate a particularly high risk. Of note, our findings did not contain data from individuals who made and consumed rice on their own, and the worst-case scenario was estimated for the MOE using the lowest BMDL₀₁ of 0.3 g/kg bw/day. The MOE based on BMDL₀₁ is the most conservative evaluation; however, selecting an upper limit of 8 g/kg bw/day increases the MOE of all categories to a number greater than 100. As a result, the intake of As from rice-based processed foods is of little concern for South Koreans; however, the Low As Practically Possible risk management approach necessitates increasing South Koreans’ awareness of risks associated with processed rice product consumption.

In conclusion, because the ratio of iAs to tAs varies among the food categories of rice-based processed foods, an analysis of iAs is required to enable an appropriate risk assessment of these products. The tAs concentration is affected by the presence of seaweed, while the iAs concentration in rice-based products is affected by the milling type of rice. The variability in food subgroup ingredients influences the correlation between iAs and tAs. According to the risk assessments, most rice-based foods are safe to consume, on average, in South Korea, although rice cakes and baby rice-based powder forms are a source of concern for adults of all ages and children, respectively. The study provides a foundation for analyzing the dietary risk of As in other rice-based processed food-consuming regions of the world, given the current trend of increasing rice consumption worldwide.

Supplementary Information

Below is the link to the electronic supplementary material.

Declarations

Conflict of interest

The authors declare no conflict of interest.