Public health risk due to aflatoxin and fumonisin contamination in rice in the Mekong Delta, Vietnam

Mycotoxin contamination in rice can lead to a health risk for consumers. In this study, the health risk among different age groups of Vietnamese population in the Mekong Delta, Vietnam was evaluated through rice consumption. Total aflatoxins (AFs) and fumonisins (FBs) in raw rice samples (n = 50) were analyzed using an ELISA method. A survey (n = 155) was used to collect data on rice consumption and consumer practices for the evaluation of mycotoxin exposure. Results showed that the frequency of AFs and FBs contamination was 60 and 74% with the average concentrations in raw rice ranging from 1.88–4.00 ng/g and 227–290 ng/g from the lower bound (LB) to the upper bound (UB), respectively. The average AFs exposure due to rice consumption was estimated from 0.81 to 2.44 ng/kg bw/day at scenarios LB – UB with the medium bound (MB) of 2.10, 1.60, 1.92 and 1.23 ng/kg bw/day for children, adolescents, adults and elderly, respectively. These values ranged from 343 to 724 ng/kg bw/day with respect to FBs (scenarios LB - UB), which are below the provisional maximum tolerable daily intakes (PMTDI) value (2000 ng/kg bw/day). The margin of exposure (MoE) to AFs ranged from 160 to 1585,179-2669,149–2175 and 206–3480 for children, adolescent, adults and elderly, respectively from UB - LB, indicating a high health risk for this carcinogenic hazard since the values are so lower than 10,000 (safe limit). However, for FBs, MoE value ranged from 105 to 575 (UB-LB) for all groups, which are higher compared to 100 (safe limit), indicating no risk for public health. The mean cancer risk due to estimated AFs exposure at LB - UB was 0.05–0.13 cases/year/100,000 individuals with MB of 0.08–0.13 cases/year/100,000 people for all four age groups. This study provides new insights into probabilistic risk assessment and potential health impact of mycotoxins in rice in the Mekong Delta, Vietnam.


Introduction
Mycotoxins are toxic secondary metabolites produced primarily by fungi. Of those, aflatoxins (AFs) that are produced mainly by Aspergillus flavus and A. parasiticus were frequently found in agricultural products in pre-and post-harvest (Gonçalves et al. 2019). AFs are considered as Group 1 Carcinogens by the International Agency for Research on Cancer (IARC 2002), causing 5-28% of all global hepatocellular carcinoma (HCC) cases (Liu and Wu 2010). More than 80% of the HCC occurred in poor countries, where people have a high-risk source of dietary exposure to AFs, chronic hepatitis B and hepatitis C viral infection (HBV and HCV) (Majeed et al. 2018). Fumonisins (FBs) produced by Fusarium proliferatum and F. vertillioides were also prevalently detected in farming products (Gonçalves et al. 2019). FBs are categorized to group 2B, possibly carcinogenic to humans (IARC 2002).
Rice (Oryza sativa L.) is the staple food for 50% of the Asian population (Devi and Ponnarasi 2009). Also, rice is a main staple food in Vietnam. The average annual rice consumption was 218 kg/capita in 2017 surveyed by Food and Agriculture Organization Statistics in 2020 (FAOSTAT 2020). Rice is grown in the Summer-Autumn and Autumn-Winter crop seasons (General Statistics Office of Vietnam 2020, Phan et al. 2021a, b), in which frequent and heavy rainfalls occur, especially during harvest, leading to the rice crop prone to fungal infection (Reddy et al. 2008). Several reports have indicated the presence of aflatoxin B1 (AFB1) and fumonisin B1 (FB1) in rice or cereals in the Mekong Delta, northern and central regions of Vietnam , Huong et al. 2016bNguyen et al. 2007;Thieu et al. 2008;Trung et al. 2001). However, these studies did not contain enough data to indicate a comprehensive understanding of mycotoxin presence in the Mekong Delta rice linked to public health risk, where more than 50% of Vietnam's rice production was produced.
Risk assessment involves hazard identification, hazard characterization, exposure assessment and risk characterization (Borchers et al. 2010). Regarding carcinogenic mycotoxins (e.g., AFs), it is overall assumed that there is no cut off dose below leading to no cancer induction. That means, there is no tolerable daily intake for AFs since such toxins are genotoxic and carcinogenic contributing to tumor growth while regarding FBs, PMTDI has been established at 2000 ng/kg bw/day (EFSA 2014). If daily consumption is below the proposed PMTDI value, no adverse human health impacts would appear over a lifetime. Mycotoxin exposure assessment depends on the mycotoxin levels in food as well as food consumption. The risk is usually evaluated based on the MoE for both AFs and FBs (EFSA 2005). In addition to MoE, in terms of AFs, risk is also assessed based on HCC risk (JECFA 1999).
In recent years, the application of probabilistic methods to estimate exposure to mycotoxins in food has been increased (Panrapee et al. 2016;Meerpoel et al. 2021;Udovicki et al. 2021). Based on the approaches, proper statistical descriptions were employed to assess the data and to describe the range of consumer exposures (EFSA 2011). Thus, the aim of this study was (i) to estimate total aflatoxins and fumonisins contamination in rice, (ii) to estimate dietary exposure via rice consumption and (iii) to assess human health risks for different age group categories of Vietnamese population in the Mekong Delta (i.e., Can Tho, Dong Thap and An Giang provinces) using the probabilistic Monte Carlo simulation as potential input for further risk management.

Sampling of raw rice
Fifty raw rice samples were collected from farmers' households in July 2019, in Can Tho (n = 17), Dong Thap (n = 16) and An Giang (n = 17) provinces of the Mekong Delta ( Fig. 1). Each sample was collected about 0.5-1.0 kg from farmers'containers or bags which contained 20-50 kg raw rice. The collected samples were kept in plastic bags, transported to Ho Chi Minh city University of Food Industry's laboratory and stored at -20 °C before mycotoxins analysis.

Rice consumption data
The consumption survey dedicated to rice was performed in July 2019 in three provinces, namely Can Tho, Dong Thap and An Giang in the Mekong Delta, Vietnam (Fig. 1). One hundred and fifty-five participants from 50 families (1-6 persons/family) were interviewed face to face to avoid uncertain answers. During the survey, also raw rice samples from interviewed families were collected (as mentioned in section 2.1). The participants consisted of children (6-9.9 years, 3%), adolescents (10-17.9 years, 14%), adults (18-64.9 years, 77%) and elderly (> 65 years, 6%). The survey was carried out in participants' houses at lunch-time or dinner-time to verify participants' answers. The respondents in this survey were farmers, housewives, workers and their children who always consumed rice prepared by their families. Moreover, children did not consume rice in their schools or school canteens. In terms of rice consumption, the number of rice bowls consumed by each individual per day was recorded (range: 1-12 bowls/day). Besides, grams of rice in each bowl (range: 105-218 g) used by respondents were measured using the study organizer's scale (Kitchen scale, Max 5000 g, CH 303A-1, China). Afterward, the total grams of rice consumed (g) per day per person was calculated based on the following equation (Eq. 1).
Additionally, participants' body-weight were measured by the study organizer's scale (Nhon Hoa, Vietnam). Other information was also collected including washing, (1) Rice consumption g∕day∕person = the total number of rice bowls x gram of rice in each bowl cooking method, etc. The survey data were analyzed by @RISK (version 8.1, Palisade Corporation, USA).

Determination of aflatoxins and fumonisins in raw rice by enzyme-link immunosorbent assay (ELISA)
Fifty samples (range: 0.5-1.0 kg) were individually ground using a laboratory blender (Phuong Thanh, Vietnam) for 30 seconds and stored at -20 °C before mycotoxin determination. The total aflatoxins and fumonisins were analyzed by ELISA (AgraQuant ® , Romer, USA). The limit of detection (LOD) and limit of quantification (LOQ) were 3 ng/g and 4 ng/g, respectively for AFs and these values were 200 ng/g and 250 ng/g, respectively in regard to FBs. The method was conducted according to the manufacturer's instructions. Briefly, the samples (25 g raw rice powder) were homogenized and mixed with 100 mL methanol (Merck, German)/water (70/30 v/v) for 2 min to extract mycotoxins. The mixture was centrifuged for 10 min at 4000 rpm. The supernatant was collected for toxin detection. Optical density was measured using a microtiter plate reader at 405 nm (Chromatic Reader, USA). The concentration of AFs and FBs were calculated on a dry weight basis according to the specifications of the manufacturer.

Mycotoxin reduction rates in rice during processing based on literature
Raw rice is usually processed before consumption. Many processing methods are applied such as normal cooking in boiling water, pressure cooking, etc.. Washing and cooking methods, leading to a reduction in mycotoxin levels, were mentioned by several studies and screened in a literature search (Table 1). Key words 'effect of cooking or washing' and 'reduction of mycotoxins in rice or cereals' were applied to search literatures published on Web of Science, PubMed and Google Scholar from 2000 to 2021. The reported mycotoxin reduction ranged from 7 to 88%, depending on cooking methods ( Table 1). The highest AFs reduction was found when rice was cooked by pressure cooking (78-88%) followed by normal cooking (7-84%) while 23-48% of FBs were removed by a combination of washing and cooking approaches. AFs reduction through washing methods was lower (14-24%), compared to a cooking process. Based upon our questionnaire on the Vietnamese households' practices, raw rice was washed (procedure is similar to methods reported by Majeed et al. (2018), Park et al. (2005)) before cooking, and most respondents cooked rice using normal cooking procedure (96%) the same as methods of Park and Kim (2006), Sani et al. (2014), Sakuma et al. (2013) and Hussain and Luttfullah (2009). Few families used the pressure cooking approach (4%), which was more popular in families having high incomes or/ and living in big cities (i.e., Ho Chi Minh, Ha Noi, etc.). Moreover, Nguyen et al. (2007) analyzed mycotoxin in raw rice and estimated them in rice based on the reduction rate during processing mentioned in previous researches and then calculated exposure based on rice consumption data. Thus, based on this information, a range of reduction rates from 7 to 88% (normal cooking -pressure cooking) and 14-24% (washing) with respect to AFs and 23-48% (combination of washing and cooking) for FBs was applied.

Dietary exposure assessment related to aflatoxins and fumonisins due to rice consumption
A probabilistic approach was used to evaluate AFs and FBs exposure due to rice consumption in the Mekong Delta based on raw rice-contamination data, reduction factors to rice preparation, cooking and rice consumption data. The mycotoxins concentration in rice was calculated by multiplying the best fit distribution of mycotoxin contamination in raw rice (ng/g) with the best fit distribution of the mycotoxin reduction (%) in washing and cooking processes (Table 1). Thus, dietary exposure was calculated as follows: However, mycotoxins levels in some samples were low; therefore, no probability method could be performed. Hence, in the current work, AFs and FBs exposure was evaluated integrated three scenarios analysis (LB, MB and UB) with the calculation of such toxins contamination data of the non-detects (NDs) and below the limit of quantification (<LOQ) (Vinci et al. 2012). Management of the leftcensored contamination data is considered as the main factor of uncertain exposure. Replacement of the ND values by zero (0) and by the limit of detection (LOD) value for the LB and UB, respectively was the most frequent approach in mycotoxins risk assessment studies (EFSA, European Food Safety Authority 2010). Thus, ND values were replaced by zero, half of the LOD (1/2 LOD) and LOD, while <LOQ values were replaced by half of the LOD (1/2 LOD), LOD and LOQ for the LB, MB and UB, respectively. (2) Dietary exposure ng∕kg bw∕day = Distribution of mycotoxin in raw rice ng∕g x distribution of mycotoxin reduction (%) x distribution of rice consumption g∕kg bw∕day The probabilistic exposure assessment was also used to consider the variances and uncertainties related to the mycotoxins intake determinants. The fractions of the ND, <LOQ and > LOQ values were calculated. The fractions of >LOQ, <LOD and > LOQ, <LOQ and > LOQ were calculated by an "if " function of the Microsoft excel at LB, MB and UB, respectively with regard to the risk output calculations (Yogendrarajah et al. 2014). Best fit distributions were observed at the three scenarios of the AFs and FBs content data based on the Chi square statistics. Also, the probability/probability and quantile/quantile plots were evaluated to identify the best fit distribution for both cooked rice consumption and mycotoxins concentration data. For consumption data, best fit distribution functions such as Log-logistic were used to calculate raw rice consumption distribution. Regarding reduction rates during processing, a uniform distribution function (minmax) was used ranging from 14 to 24% (washing), 7 to 88% (cooking) for AFs and 23-48% (washing and cooking) for FBs (based on the data in Table 1). Monte Carlo simulations were performed at 100,000 iterations with the add-in @RISK (version 8.1, Palisade Corporation, USA). Based on these input data, a best fit distribution was performed with @RISK (Table 2).

Risk characterization
The calculated exposure values were compared with the proposed PMTDI value to assess the risk of the exposure for FBs. AFs are genotoxic carcinogens, leading to unsafe at any level of exposure. Risk characterization was performed by two approaches namely MoE (EFSA 2005) and HCC associated risk (JECFA 1999) for AFs while in terms of FBs, such evaluation was based on MoE. Regarding the MoE, this parameter is calculated as the ratio between a toxicological reference point (a dose, causing a low measurable response) and the estimated exposure. EFSA recommends to use the benchmark dose lower confidence limit 10% (BMDL 10 ) (the lowest dose, causing no more than a 10% of cancer incidence in rodents or human) (EFSA 2007). In this study, a BMDL 10 of 870 ng/kg bw/day for AFs (EFSA 2018) and 150,000 ng/kg bw/day for FBs (Bondy et al. 2012) was used. If the MoE value is less than 10,000 (EFSA 2005) and 100 (ChemSafetyPro 2022), the exposure is considered as a public health concern with respect to AFs and FBs, respectively. In this study, it is assumed that AFs and FBs contained 100% of the AFB1 (Manizan et al. 2018;Majeed et al. 2018) and FB1, respectively.

Mycotoxin concentration (ng/g) in raw rice
AFs (

Distribution of aflatoxins and fumonisins in raw rice and rice
Best fit distribution functions used in this study are related to mycotoxins contents in raw rice, IF-function for fitting of AFs and FBs levels at three scenarios (LB, MB and UB) and mycotoxin reduction during processing (Table 2). These data were fitted to best fit distributions according to the @RISK (version 8.1, Palisade Corporation, USA). The estimated distribution of AFs and FBs contamination in raw rice and rice is presented in Table 3. In terms of mycotoxins contamination in raw rice, the prevalence of contaminated samples by AFs was 60% (n = 30/50) and the estimated average levels of AFs concentration in raw rice ranged from 1.88 to 4.0 ng/g from LB-UB with the mean MB of 2.98 ng/g (Table 3). AFs contents at all levels were lower than the Vietnamese regulation limit for food (15 ng/g) (Vietnamese Ministry of Health 2011). Moreover, AFs concentrations from percentile 75 (P 75 ) to percentile 95 (P 95 ) at three scenarios were also lower than the European maximum limit for unprocessed rice (10 ng/g) (EC 2006). The prevalence of contaminated samples with FBs was 74% (n = 37/50), with the average content at LB, MB and UB of 277, 261 and 290 ng/g, respectively in raw rice. These mycotoxins contents were lower than Vietnamese regulation limit for corn (1000 ng/g) (Vietnamese Ministry of Health 2007). In general, both AFs and FBs contents were lower than Vietnam and European maximum limit for food and unprocessed rice, respectively even at P 95 of MB or/and UB (5.64 ng/g for AFs and 433 ng/g for FBs). However, rice is an important staple food and a key exported product in Vietnam, hence contents of these toxins should be as low as possible, especially AFs.
Furthermore, raw rice is washed and cooked during processing before consumption by Vietnamese people. Table 3 shows that the mean contents of AFs were 0.18 ng/g, 0.27 ng/g and 0.35 ng/g and these FBs values were 81 ng/g, 93 ng/g and 103 ng/g at LB, MB and UB, respectively when the processing factors were used to shift from raw rice to rice.
The result of this study is in good agreement with previous reports in Vietnam and Malaysia. For instance, the Table 3 Distribution of aflatoxins and fumonisins contamination in raw rice and rice with applied reduction of washing 14-24% and cooking 7-88% of the raw rice contamination in the Mekong Delta (ng/g) The distribution mycotoxins in rice = the distribution of mycotoxins raw rice x distribution of washing reduction x distribution of cooking reduction at LB, MB and UB; LB Lower Bound, MB Medium Bound, and UB Upper Bound based on the best fit distributions presented in Table 3; P50 Percentile 50%, P75 Percentile 75%, P90 Percentile 90%, P95 Percentile 95%; n = 20 (<LOD); n = 16 (>LOD and < LOQ) and n = 14 (>LOQ) for AFs, for FBs n = 13 (<LOD); n = 4 (>LOD and < LOQ) and n = 33 (>LOQ);

Mycotoxins
Scenarios Rice (raw rice) (ng/g) mean of AFB1 was 3.31 ng/g in raw rice of five provinces in central Vietnam (Nguyen et al. 2007), 2.99 ng/g (MB) and 2.4-3.0 ng/g (LB-UB) in Lao Cai rice (Huong et al. 2020) and 0.68-3.79 ng/g (AFs) in Malaysia (Reddy et al. 2011). However, the levels of AFs in rice in this study were higher than those of AFB1 reported in Thailand (range: 0.05-1.66 ng/g) (Panrapee et al. 2016). In contrast, higher values were detected in rice in Philippines (range:1-2546 ng/g) (

Consumption of rice in the Mekong Delta
The information of 155 participants belonging to 50 families interviewed in this survey is presented in Table 4. According to the survey, the average of rice consumption for the overall population was estimated at 6.20 g/kg bw/ day. The mean of rice intake was highest (7.00 g/kg bw/ day) for children and then adults (6.43 g/kg bw/day), adolescents (5.27 g/kg bw/day) and elderly (4.25 g/kg bw/day). This seems to accord with general findings on the global consumption of rice for different population groups (Udovicki et al. 2021;Majeed et al. 2018). Typically, children and adults ate more rice per unit bodyweight than adolescents and elderly (Udovicki et al. 2021;Majeed et al. 2018). Because adults and children usually consumed rice at breakfast, lunch and dinner, prepared by their family, and they rarely ate other food outside. Moreover, adults must work hard on their farm requiring more energy; thus, they consumed a lot of rice. In contrast, adolescents who are secondary and high school students ate variety of food replacing for rice such as milk, milk-tea, soup, corn, sweet potato, bread, instant noodle, etc. at school canteens or street food markets. Similarly, the elderly usually consumed soup, rice noodle, sweet potato, etc. instead of rice prepared by their family. Moreover, elderly worked less than adults; therefore, they did not consume a lot of food. Thus, adolescents and elderly consumed less rice than children and adults due to a different consumption pattern.
As regard rice consumption, the mean of daily rice intake per capita in this study (341 g/day/capita) calculated by multiplying daily intake (g) per body weight (kg) (6.20 g/kg bw/day) with the mean body weight of 55 kg is quite higher than the results surveyed in northern Vietnam namely Ha Giang, Ha Noi and Thanh Hoa provinces in Vietnam, which ranged from 244 to 301 g/day/adult (Do et al. 2020), China (183 g/day/person), Japan (157 g/day/ capita) (Abdullah et al. 2006), Pakistan (108 g/day/people) (Majeed et al. 2018) and Iran (107 g/day/individual) (Yazdanpanah et al. 2012). However, rice consumption in this work is lower than that reported in other Asian countries namely Thailand (377 g/day/capita). The difference could be related to the characteristics of participants (ages, profession etc.) and regions surveyed. Indeed, most participants in this survey were farmers (adults, 77%) who consumed rice higher than the people working in the offices, adolescents (14%) and elderly (6%) ( Table 4), which maybe result in higher consumption data than other studies. Also, rice is the main staple food in Vietnam; therefore, Vietnamese farmers have consumed a large amount of rice per day, leading to higher rice intake data in the current study, compared to that in other countries.

Estimated aflatoxins and fumonisins exposure related to rice consumption
The calculated exposure related to AFs and FBs in rice is presented in Table 5. The mean levels of AFs intake based on rice consumption were estimated to be 1.30 to 2.44 ng/kg bw/day for children and adults, whereas for elderly and adolescents, these values were slightly lower of 0.81 to 2.04 ng/kg bw/day from LB to UB. Regarding FBs, the values for the LB-UB were estimated from 517 to 724 ng/kg bw/day for children and adults, while these Table 4 Distribution of rice household consumption in different age groups in Mekong Delta (g/kg bw/day) obtained from the survey in 2019 using best distribution functions presented in Table 2 Statistical description Daily intake (g)/kg body weight in age group (n = 155) Daily intake(g)/kg body weight overall population Children (n = 5) Adolescents (n = 21) Adults (n = 120) Elderly (n = 9) values were lower ranging from 343 to 546 ng/kg bw/day for adolescents and elderly. The results indicated that although AFs intake were low due to rice consumption, this may be harmful to human health because these toxins are genotoxic and carcinogenic compounds (JECFA 1999), which exposure above zero level is harmful. Moreover, the Mekong Delta population consumed a big amount of rice, leading to a risk for human health (Table 3).
For FBs, the estimated mean of FBs exposure values due to rice consumption in this work is slight lower than that in other studies in Vietnam. For instance, a lower average of FB1 exposure (536 ng/kg bw/day and 1019 ng/kg bw/day) was found in Northern Vietnamese adults and children in Vietnam, respectively (Huong et al. 2016b). Our results are also lower than those observed in Ha Giang children (851-1199 ng/kg bw/day) and adults (1106-1325 ng/kg bw/day) (Do et al. 2020). In contrast, the results in the present work are higher than Thanh Hoa, Ha Noi children (6.5-256 ng/kg bw/day), adults (127-209 ng/kg bw/day) (Do et al. 2020), Nigerian population (19.13 ng/kg bw/day) (Abdus-Salaam et al. 2016), Pakistani children (31.2-64.2 ng/kg bw/day (LB-UB)) and adults (30.8-61.9 ng/kg bw/day (LB-UB)) (Majeed et al. 2018). This difference may be associated with sampling, levels of mycotoxins contamination in rice, processing method, consumption data, etc..

Risk characterization of aflatoxins and fumonisins related to rice consumption Increase risk associated with the MoE
Concerning MoE associated with AFs exposure, the mean of MoE ranged from 160 to 1585, 179-2669, 149-2175 and 206-3480 for children, adolescents, adults and elderly, respectively (Fig. 2). These MoE values were lower when compared to 10,000, leading to a public health concern. However, the average of MoE related to FBs was higher than 100, ranging from 105 to 575 for all groups. Moreover, FBs exposure values were lower than tolerable daily intake and PMTDI established by the European Union Scientific Committee for Food and JECFA (2000 ng/kg of body weight) (International agency for research on cancer (IARC) and world health organization (WHO) 2012; EFSA 2014). Therefore, it is safe for human health due to rice intake regarding FBs.
Based on these results, the intake of contaminated rice was considered as a great public health concern with respect to AFs for the Vietnamese population. The  (Ding et al. 2012), Northern Vietnam (Huong et al. 2020;Do et al. 2020), Pakistan (Majeed et al. 2018), Gambia (Shephard 2008) and Japan (Sakuma et al. 2013) was lower than that in the present study. However, the MoE through AFB1 intake in Serbian rice (Udovicki et al. 2021) was higher than this value in our study. This could be related to mycotoxin contamination contents in rice, consumption levels, type of BMDL 10 (for human or rodent), etc.. as mentioned above.

Prevalence of liver cancer risk or HCC associated with AFs contamination in rice
AFB1 is considered the most biologically active and abundant in the AFs. The risk of liver cancer in people exposed to both chronic HBV infection and AFs was 30 times higher than the risk in people exposed to AFs only (Groopman et al. 2014). The burden of AFs induced HCC has been recently evaluated in different countries (Liu and Wu 2010).
The risk characterization based on AFs exposure through rice consumption was estimated using the liver cancer risk approach ( Table 6). The mean of HCC risk at three scenarios (LB, MB and UB) ranged from 0.05 to  0.13 cases/year/100,000 individuals for children, adolescents, adults and elderly.
The average liver cancer risk due to AFs in this study was slight higher than that due to AFB1 intake through colored and brown rice in all age groups of Thailand (0.010-0.039 cases/year/100,000 individuals) (Panrapee et al. 2016 In order to reduce the risk of cancer, improvements in HBV vaccination and methods to eliminate AF contamination of rice (i.e., remove off residue crops and spray bio-decomposer to decompose residue crops, use disease-resistant paddy varieties, inorganic fertilizers, biocontrol in pre-harvest, and hygienic storage containers, essential oil, milling and polishing in post-harvest (Phan et al. 2021a, b)) should be promptly implemented as risk mitigation strategies. The risk characterization of AF and FB from rice consumption (Majeed et al. 2018;Park et al. 2005 &Kim 2006) is helpful in establishing priority control approaches for mycotoxins.

Uncertainty analysis
Regarding rice consumption surveys, under or over (−/+) reporting of consumption data, misreporting of consumed rice and the erroneous estimation of consumed quantities (based on portion sizes) could contribute to an underestimation or overestimation of rice consumption (−/+), affecting the exposure assessment (Vinci et al. 2012). Moreover, consumption data that were used in this study were surveyed by 155 people in three provinces of Mekong Delta, in which only 3 and 6% of respondents were children and elderly, respectively, and consuming habits may have been different from other provinces in this Delta, leading to lower or higher rice consumption (−/+). This made overestimation or underestimation (+/−). In terms of sampling, collecting samples and/or the number of samples (50 samples) in this study could result in underestimation or overestimation (−/+) in mycotoxin intake evaluation. For mycotoxins analysis method, in this study using an ELISA method to analyze mycotoxins in raw rice may have a limitation because of the possibility of matrix effect and cross-reactivity, leading to a certain level of uncertainty in exposure assessment (−/+) (Udovicki et al. 2021).
In addition to problems such as consumption surveys, sampling and mycotoxins analysis method as mentioned above, the distribution fitting to literature input data for reduction of mycotoxins during processing as cooking and washing could lead to high mean reductions of mycotoxins in comparison to actual practices (+). Because washing and cooking were only performed in small sizes (100 g rice and 200 mL drinking water in laboratory for washing and cooking). Moreover, AFs and FBs are stable and hard to destroy them during cooking (normal cooking) or/ and mycotoxins could transfer to other structures which could be toxic or not, and they were not identified during mycotoxins analysis or mycotoxins reduction could be related to other factors. Moreover, we do not have clear and original data from literatures as a result in under or over estimation (−/+) in exposure assessment. Also, this study assumed that AFs and FBs containing 100% of AFB1 and FB1, respectively used to evaluate MoE and HCC may result in overestimation (+).
The bioavailability of mycotoxin in target organs in humans depends on factors as bio-accessibility, percentage of mycotoxin released (partially or totally) from the matrix during digestion in the gastrointestinal tract, bio-accessible fraction transported across the intestinal epithelium as well as metabolism (Meca et al. 2012;González-Arias et al. 2013;Bordin et al. 2017; Van et al. 2020. In our study, these factors are not included; thus, this study could make overestimation in exposure and risk assessment (+).

Conclusions
This study is the first report on probabilistic risk assessment related to AFs and FBs exposure in rice for the Mekong Delta, Vietnam population. The dietary exposure to aflatoxins and fumonisins through rice consumption was higher for children and adults than adolescents and elderly, due to the high consumption of adults (farmers needed high energy input) and children (high consumption compared to lower body weight). The MoE related to aflatoxins exposure was remarkably lower than the recommended safe limit, leading to a health concern when consuming the Mekong Delta rice. In addition, there is a potential risk due to rice consumption, associated with aflatoxininduced HCC. Thus, AF contamination in such commodity should be decreased to ensure food safety. This work highlights the need to establish risk management strategies and set (regulatory) guidelines for the rice cultivation (pre and post-harvest) in the Mekong Delta rice to prevent contamination, followed by a regular monitoring of highly consumed foodstuff, especially rice. Also, a cumulative risk assessment from the exposure of multi-mycotoxins, especially in the HBV and HCV-positive population should be studied in the future to estimate the full burden on human health due to mycotoxins.