Estimated long-term dietary exposure to lead, cadmium, and ... - Nature

European Journal of Clinical Nutrition (2014) 68, 1322–1326 © 2014 Macmillan Publishers Limited All rights reserved 0954-3007/14 www.nature.com/ejcn

ORIGINAL ARTICLE

Estimated long-term dietary exposure to lead, cadmium, and mercury in young Korean children DW Kim1, HD Woo1, J Joo2, KS Park3, SY Oh4, HJ Kwon5, JD Park6, YS Hong7, SJ Sohn8, HJ Yoon9, MS Hwang9 and J Kim1 BACKGROUND: Controlling for day-to-day variation is a key issue in estimating long-term dietary exposure to heavy metals using 24-hour recall (24HR) data from a relatively small number of days. OBJECTIVES: This study was conducted to estimate long-term dietary exposure to lead, cadmium and mercury among Korean children using the Iowa State University (ISU) method and to assess the contributions of different food groups to heavy metal intake. METHODS: We analyzed 2 days of 24HR data from 457 children between 0 and 6 years of age in 2010. Using bootstrapped concentration data for 118 representative foods, 93.5% of total intake was included in the exposure estimates in this study. Using the 2-day exposure data, we estimated long-term exposure by controlling for within-individual variation using the ISU method. RESULTS: The long-term dietary exposure estimates (mean ± standard deviation) for lead, cadmium, and mercury were 0.47 ± 0.14, 0.38 ± 0.20, and 0.22 ± 0.08 μg/kg bw/day, respectively. For lead and cadmium, the percentages of children whose exposure was greater than the reference value were 35 and 42%, respectively. Fruits were an important source of lead exposure, and cereal and fish and shellfish made the greatest contributions to the total cadmium and mercury exposure. CONCLUSIONS: Our findings also suggest that the long-term exposure to lead and cadmium was somewhat greater than the reference values, whereas mercury exposure was well below than the reference value in this population. Further studies may be necessary to evaluate the food items contributing to heavy metal exposure, and continuous monitoring is needed to ensure the safety of food intake and dietary patterns among vulnerable groups in Korea. European Journal of Clinical Nutrition (2014) 68, 1322–1326; doi:10.1038/ejcn.2014.116; published online 25 June 2014

INTRODUCTION Heavy metals in the environment can appear as residues in foodstuffs.1 They can be harmful for humans because they are not metabolized, and thus, they accumulate in the human body.2 Recently, studies have reported that the renal and dopaminergic systems in infants and children are particularly vulnerable to metal poisoning.3 In addition, the organ systems of infants and children are still developing, and they consume more food in relation to their body weight than adults consume. However, relatively little research has been conducted on individual-level dietary exposure to heavy metals in children.4–6 Quantitative evaluations of dietary heavy metal exposure can utilize direct or indirect forms of assessment. In a direct assessment, exposure is measured directly by monitoring individuals, whereas an indirect assessment combines relevant input data using a model or a deterministic approach.7 Therefore, an indirect assessment is less accurate than a direct assessment for determining exposure to food contaminants. However, direct methods are burdensome for subjects as well as costly and timeconsuming. Thus, indirect methods combining food consumption data as collected via 24-hour recalls (24HR) and concentration

data as collected in total diet studies (TDS) have been used in many countries to evaluate exposure to heavy metals.8–14 Conventional approaches to exposure assessment are deterministic, using single fixed values (point estimates) for concentration and consumption data. However, both concentration and consumption data are variable, and there may be significant uncertainty in the analytical measurement of concentration levels.15 When data are sufficient, probabilistic approaches could quantify variation and uncertainty by using probability distributions rather than point estimates. It could give more reliable results in both the exposure and effects in risk assessment.16 Furthermore, if data are only available for a relatively few number of days, the estimates of heavy metal exposure are biased toward the typical heavy metal exposure associated with normal eating habits. This bias is caused by the substantial day-to-day variations in the eating habits of free-living humans.17 Therefore, controlling for day-to-day variation is a key issue when food consumption data are available for a relatively small number of days. Fortunately, several statistical approaches have recently been developed to eliminate day-to-day variation. The Iowa State University (ISU) method18 has been widely used since 1996 and

1 Molecular Epidemiology Branch, National Cancer Center, Goyang-si, Gyeonggi-do, Republic of Korea; 2Biometric Research Branch, National Cancer Center, Goyang-si, Gyeonggido, Republic of Korea; 3Advanced Analysis Center, Research Planning & Coordination Division, Korea Institute of Science and Technology, Seoul, Republic of Korea; 4Department of Food & Nutrition, College of Human Ecology, Kyung Hee University, Seoul, Republic of Korea; 5Department of Preventive Medicine, College of Medicine, Dankook University, Cheonan-si, Chungcheongnam-do, Republic of Korea; 6Department of Preventive Medicine, College of Medicine, Chung-Ang University, Seoul, Republic of Korea; 7Dong-A University, Regional Cardiocerebrovascular Center, Busan, Republic of Korea; 8Department of Preventive Medicine, Chonnam National University Medical School, Gwangju, Republic of Korea and 9Risk Analysis & Research Division, National Institute of Food and Drug Safety Evaluation, Cheongwon-gun, Chungcheongbuk-do, Republic of Korea. Correspondence: Dr J Kim, Cancer Epidemiology Branch, National Cancer Center, 111 Jungbalsan-ro, Ilsandong-gu, 323 Ilsan-ro, Ilsandong-gu, Goyang-si, Gyeonggi-do 410-769, Republic of Korea. E-mail: [email protected] Received 7 May 2013; revised 10 March 2014; accepted 13 April 2014; published online 25 June 2014

Estimated long-term dietary metal exposure DW Kim et al

1323 the estimation of proportion of the population above or below a specific reference values is possible for ISU method. In this study, short-term dietary exposure to three heavy metals was estimated in 457 Korean children using 2-days of 24HR consumption data and bootstrapped concentration data for 118 core foods. These estimates were then used to determine longterm heavy metal exposure using the ISU method. Next, the estimated long-term heavy metal exposures were compared with reference values to assess the percentage of children with an exposure exceeding these values. MATERIALS AND METHODS Selection of food items for the chemical analysis The core food items that contribute substantially to heavy metal exposure were selected using individual food intake data from the 2007/2008 Korea National Health and Nutrition Examination Survey (KNAHNES).19 The KNHANES is a cross-sectional survey conducted in a nationally representative sample, and the nutrition part of the survey used the 24-hour recall (24HR) method to determine the food intake of each individual. Trained dietitians visited each household and administered a single 24HR assessment using two-dimensional food models and various sized food containers to improve the estimates of portion sizes. We selected 118 core food items eaten by Koreans in order of frequency and amount consumed according to the KNHANES. The amounts of lead, cadmium, and mercury concentrated in the 118 core food items were obtained from the database of the Korean Research Project on the Integrated Exposure Assessment to Hazardous Materials for Food Safety, which was conducted in 2010. In that study, more than 3 brands per foodstuffs were purchased at markets in seven cities in South Korea (Seoul, Daejeon, Daegu, Busan, Incheon, Gwangju, and Gangreung). The collected food samples were analyzed using heavy metal analysis methods specified in the Korean Food Standard Codex.20 For lead and cadmium, the samples were digested using a microwave digestion method in the pre-treatment process, and the digested samples were then analyzed with an inductively coupled plasma-mass spectrometer (ICP-MS). For mercury, the gold-amalgamation method was used with the raw food samples.

Food consumption data The consumption data for children aged 0 to 6 were obtained from kindergartens, health centers, nurseries, and hospitals in five regions throughout the country. A total of 481 children agreed to participate, and 457 children (234 boys and 223 girls) completed non-consecutive 24HRs for 2 days between June and August 2010. Additional demographic information, such as sex, age, and weight, was collected for use as covariates. To calculate food intake, a trained interviewer facilitated the 24HR in face-to-face interviews with children and their mothers in addition, another non-consecutive 24HR was conducted by telephone within one week to control for within-individual variation. After collecting the dietary data, we calculated individual food intake using CAN-PRO 3.0 (Computer Aided Nutritional Analysis Program, The Korean Nutrition Society, Seoul, Korea). This study was conducted according to the guidelines in the Declaration of Helsinki, and all of the procedures involving human subjects/patients were approved by the Institutional Review Board (IRB approval number: DKUHIRB2010-04-0093). Written informed consent was obtained from the parents of all the children.

Linking the concentration data and the food consumption data In total, 3823 concentrations in 118 core foods were suitable for use in the exposure assessment. The samples with no detected levels were assigned concentrations equal to half the limit of detection (LOD = 0.20 μg/kg) because 1/2 LOD is often seen as the most optimal scenario possible when addressing contaminants, resulting in the most likely exposure estimates. The 2-day consumption data included 871 food items for 457 children, and we linked these food items with 118 core food items to calculate daily heavy metal consumption. During this procedure, we matched the consumption data and the concentration data for 613 food items, so 258 foodstuffs could not be linked to a concentration value and were not included in the assessment. Ultimately, 93.5% of the total consumption (762.9 kg from 816.3 kg) was included in our estimates of exposure in this study (Table 1). © 2014 Macmillan Publishers Limited

Daily estimates of exposure to lead, cadmium, and mercury The dietary exposure to lead, cadmium, and mercury was estimated by multiplying the food consumption (g/day) by the heavy metal concentration (μg/g) for each day. The multiplication was performed using bootstrapping (semi-parametric Monte Carlo simulation), where the food consumption data were treated as constant.21 The consumption of a foodstuffs was multiplied with a randomly selected concentration in corresponding foodstuffs from the concentration data set. The number of iterations used for the calculation was 10 000, and this procedure yields a daily estimate of exposure distribution. We used the bootstrapping method because bootstrapping is used to determine the uncertainty in percentile estimates. The medians produced from 10 000 point estimates and confidence intervals were calculated directly from the bootstrapped data. The estimated exposure per foodstuff was summed per day and individual, and the sum was subsequently divided by the participant’s body weight. Finally, daily dietary heavy metal exposure was calculated.

Estimates of long-term exposure to lead, cadmium, and mercury To estimate long-term dietary heavy metal exposure, we assessed the exposure to heavy metals using the 2-day data by controlling for withinindividual variation using a statistical approach. This adjustment was applied after calculating the bootstrapping exposure estimate. The ISU method developed by Nusser and co-workers18 was used to estimate individual heavy metal exposure. The bootstrapped exposure data for lead, cadmium, and mercury (μg/kg bw/day) were used to estimate individual exposure using the ISU method, after the effects of other sociodemographic variables such as age and sex were partially removed. For modeling positive amounts, the ISU method first transforms the positive daily exposure distribution into a more normal distribution using a power transformation, and an additional spline transformation is used to improve normality. After removing within-individual variation, a backtransformation was performed to return to the original scale. In this paper, we calculated the long-term heavy metal exposure distribution for selected percentiles (50th, 75th, 90th, 95th, and 99th) and a 95% confidence interval for each percentile. The X-Windows-based version of the Software for Intake Distribution Estimation (C-SIDE) (version 1.02, 1996; available from the Center for Survey Statistics and Methodology, Iowa State University, Ames, IA, USA) was used to estimate long-term heavy metal exposure with the ISU method. To assess the health risk associated with the intake of heavy metals, we compared the estimated usual heavy metal exposure distribution with reference values. Tolerable weekly intake (TWI) values of 2.5 ug/kg bw/week for cadmium and 4.0 ug/kg bw/week for mercury were derived by the EFSA Panel on Contaminants in the Food Chain and by the Joint FAO/WHO Expert Committee on Food Additives (JECFA). For lead, 0.50 μg/kg bw/day was suggested by the EFSA CONTAM panel as the level

Table 1.

Assignment of food consumption data to concentration data Food type Total Assigned

Total 868 Cereals 208 Potatoes 20 Sugars 36 Beans 23 Seeds 22 Vegetables 109 Mushrooms 10 Fruits 70 Meats 86 Eggs 11 Fish and shellfish 97 Seaweeds 19 Milk, dairy products 70 Fat and oils 8 Beverages 29 Spices and Condiments 50

613 166 16 8 21 6 64 6 44 67 8 85 14 59 5 12 32

Food amount (kg) % 70.4 79.8 80.0 22.2 91.3 27.3 58.7 60.0 62.9 77.9 72.7 87.6 73.7 84.3 62.5 41.4 64.0

Total Assigned 816.3 160.4 18.0 5.6 22.3 2.1 78.6 1.5 130.7 39.1 19.8 23.5 1.6 228.4 4.7 68.8 11.4

763.0 141.8 17.9 2.8 22.1 0.2 74.2 1.4 121.6 35.9 18.2 23.0 1.5 223.9 4.6 63.6 10.2

% 93.5 88.4 99.4 50.0 99.1 9.5 94.4 93.3 93.0 91.8 91.9 97.9 93.8 98.0 97.9 92.4 89.5

European Journal of Clinical Nutrition (2014) 1322 – 1326


1324 Table 2.

Long-term dietary exposure to lead, cadmium, and mercury (μg/kg bw/day (μg/day)) in Korean children (0–6 years)

Lead Cadmium Mercury

Mean

SD

Skewness

P50

P75

P90

P95

P99

0.47 0.38 0.22

0.14 0.20 0.08

0.690 2.214 1.379

0.46 (0.43–0.48) 0.34 (0.32–0.36) 0.20 (0.19–0.21)

0.56 (0.52–0.59) 0.46 (0.42–0.49) 0.26 (0.24–0.27)

0.66 (0.61–0.71) 0.62 (0.54–0.70) 0.32 (0.29–0.35)

0.73 (0.66–0.80) 0.75 (0.63–0.88) 0.37 (0.32–0.42)

0.88 (0.78–0.98) 1.11 (0.86–1.36) 0.49 (0.41–0.57)

Notes: The 95% confidence intervals around the percentiles of exposure are reported in parentheses. Long-term dietary exposure was estimated using the Iowa State University method and two-day dietary recall data. Bootstrapping was performed with 10,000 rounds for the concentration data.

of dietary lead exposure that may be associated with developmental neurotoxicity for children.

RESULTS In this study, 457 healthy children (234 boys and 223 girls) completed 2-day 24HRs. The subjects had the mean age of 3.5 years and a mean weight of 16.2 kilograms. Table 2 shows the descriptive statistics for dietary heavy metal exposure among the Korean children calculated with the ISU method. The means, standard deviations and selected percentiles are presented. The cadmium exposure distribution was the widest of the three heavy metals, whereas mercury had the narrowest distribution. The median long-term dietary exposures to lead, cadmium and mercury were 0.46, 0.34 and 0.20 ug/kg bw/day, respectively. The long-term exposure distribution had a similar mean value to the 2-d mean exposure, but as expected, the long-term exposure was less dispersed than the 2-d mean exposure (data not shown). To assess the health risk associated with the accumulation of heavy metals in the human body, we compared the estimated long-term exposure distributions with reference cut-off values. Mercury exposure over the TWI level is not a cause for concern because the P99 value was below the reference line (TWI/7 for mercury = 0.57 μg/kg bw/day). However, the distribution of longterm lead exposure showed that 35% of children had dietary lead exposure that exceeded the reference value (0.50 ug/kg bw/day). With regard to cadmium, 42% of the children exceeded the TWI/7 value (0.36 ug/kg bw/day). Table 3 shows the sources of exposure to lead, cadmium and mercury for Korean children. For lead, fruit (23.4%) and milk and dairy products (21.3%) were the most important sources of exposure. Other important contributors of lead were vegetables (15.4%), fish and shellfish (11.3%), and cereals (11.1%). For cadmium and mercury, the top three food groups contributed 77.6 and 75.0% of the total dietary exposure, respectively. For cadmium, cereals (37.9%), fish and shellfish (22.1%), and seaweeds (17.6%) were the three food groups that contributed most to exposure. For mercury, fish and shellfish (31.3%), cereals (25.4%), and milk and dairy products (18.3%) were the top food groups that contributed most to exposure. DISCUSSION In this study, we assessed the long-term dietary intake of three heavy metals and the most important food sources of these metals for Korean children. Due to the relatively fewer number of days (2) to represent the long-term intake of children, we estimated long-term heavy metal exposure distribution using the ISU method to control within-individual variation. The longterm exposure to lead and cadmium among Korean children was exceeded the relevant reference values, whereas the exposure to mercury was below this value. The most important food group contributing to lead intake was fruit. For cadmium and mercury, the top three food groups (of 16 total food groups) accounted for almost 75% of the total dietary exposure of Korean children. European Journal of Clinical Nutrition (2014) 1322 – 1326

Table 3. Contribution (%) of various food groups to the total dietary heavy metal exposure of Korean children (1–6 years) Food groups

Cereals Potatoes Sugars Beans Seeds Vegetables Mushrooms Fruits Meats Eggs Fish and shellfish Seaweeds Milk, dairy products Fat and oils Beverages Spices and Condiments

Contribution (%) to total dietary exposure Lead

Cadmium

Mercury

11.1 3.1 0.6 1.6 0.0 15.4 0.3 23.4 3.9 1.3 11.3 2.3 21.3 0.7 1.8 2.0

37.9 3.5 0.0 2.9 0.1 8.6 1.8 2.0 0.3 0.1 22.1 17.6 1.5 0.0 0.1 1.5

25.4 0.7 0.3 1.0 0.1 7.9 0.6 4.7 5.3 1.4 31.3 0.5 18.3 0.9 0.6 0.8

In 2006, the dietary exposures of the Korean population to lead, cadmium, and mercury were reported using 1998 KNHANES dietary data.22 Average dietary exposure to lead, cadmium, and mercury were 24.4, 14.3, and 1.61 μg/person/day. Since individual body weight was not available and target population was not children, it was not comparable with our study. The authors compared the PTWIs with average exposure of population mean, therefore, distribution of exposure were not presented. Recently, another study was conducted using one-day 24-hour recall from 2007 KNHANES dietary data.23 In that study, 114 representative foods were collected and analyzed in 2009. The mean individual exposures to lead, cadmium, and mercury were 0.84, 1.30, and 0.28 μg/kg bw/day, respectively, which were higher than the results in our study. Both studies compared mean heavy metal exposure divided by standard body weight and concluded that dietary exposure to heavy metals posed no hazard to Koreans’ health. It was not directly comparable to our study because our study assessed long-term individual dietary exposure using a distribution analysis, whereas the previous Korean study focused on short-term exposure using point estimates of concentration data. In the context of these consumption surveys, it is important to know whether the day selected represents a typical eating day.24 We conducted 2-day 24HRs to diminish the impact of withinindividual variation using statistical modeling because 2 is the minimum number of days required to control within-individual variation. In 2012, the long-term dietary exposures to lead in young children were reported by combining food consumption data of 11 European countries with concentration data on lead from different Member States.6 In this study, food consumption was recorded on 2 non-consecutive days of 1279 children aged 2 to 6 years using a Dietary Record method from the Dutch National Food Consumption Survey, and the beta binomial-normal © 2014 Macmillan Publishers Limited


1325 model was used to estimate the long-term dietary exposure to lead. In the medium bound concentration scenario, the P95 of exposure ranged from 0.73 μg/kg bw/day in 6-year-olds to 1.0 ug/kg bw/day in 2-year-olds. Overall, the dietary exposure to lead in our study was somewhat lower than their medium bound scenario. Although linking specific foodstuff to their specific consumption levels may be more accurate estimation of the exposure than when a food grouping approach of linkage is used, missing linkage of foodstuffs in our study may cause some underestimation of the exposure estimate. When assessing long-term exposure, there are sources of uncertainty in the concentration data that may affect the results, including the sampling procedure, the analytical precision of nonsampled foods, and the concentrations assigned to the samples with no detected metals.6 EFSA has published a guideline on how to handle these uncertainties.25 In the present study, the uncertainty created by the limited number of measurements was characterized quantitatively with confidence intervals by using a bootstrap procedure.26 Another important factor affecting the exposure results is the extent to which all foods are covered. Based on the 118 core foods, our concentration data covered 70.4% of all food items (613 of 871). However, 613 food items covered 93.5% of the total consumed food, given that the core foods were selected based mainly on the food consumed by Koreans. Only 6.6% of foods were not included in our analysis, and among the 16 food groups, the top two food groups with low coverage were sugar and seeds (50.0 and 9.5% coverage, respectively). However, the uncovered amount of those two food groups represented only 0.6% of the total amount consumed. Our exposure results might be underestimated because Koreans consume mostly foods from the cereal group (coverage = 88.4%). A long-term exposure estimation study that is similar to ours was carried out in 2003, the Dutch National Food Consumption Survey (DNFCS) consumption data were re-analyzed to assess the long-term exposure of the Dutch population to lead, cadmium, and mercury.27 The DNFCS consumption data included information about daily consumption over two days, and the authors selected the ISU method to eliminate the within-individual variation. In their study, the median values for the long-term usual exposure of children aged 1 to 6 years to lead, cadmium, and mercury were 0.10, 0.32, and 0.014 μg/kg bw/day, respectively. Overall, the dietary exposure to the three heavy metals in this study was higher than in their study, although their estimates were lower bound estimates using a zero value when the limit of detection was unknown. On the other hand, we used a value of 1/2 LOD, since the levels of contaminants in food usually follows a log-normal distribution; assigning a value of 1/2 LOD to all nondetected data is considered as a conservative approach. In addition, GEMS/FOOD-Euro (1995) recommended the value of 1/2 LOD, if there is more than 60% of censored values among the data.28 In this study, we compared the estimated distributions with the reference cut-off values. The guideline value for cadmium intake was set based on the critical renal cadmium concentration for individuals ⩾ 50 years of age because cadmium accumulates in the kidneys over the years. Cadmium exposure in young children tends to be relatively high for their body weight. Thus, in young children, cadmium exposure that is slightly higher than the guideline value will not directly impact health. However, continuous monitoring of cadmium intake is needed for children with high cadmium intake. The amount of lead that leads to developmental neurotoxicity is directly related to children’s health. Children with high lead intake in the present study may be exposed to potential health risks, especially neurodevelopmental disorders. To our knowledge, this study is the first to estimate the longterm heavy metal exposure of young Korean children considering within-individual variation and the uncertainty of concentration © 2014 Macmillan Publishers Limited

data. However, some limitations have to be discussed. First, the selection of the core foods was based on food consumption data for the entire population of Koreans. Because of this limitation, we may have missed foods that are episodically consumed but have high concentrations of metals and some essential foods consumed by children. Therefore, our results may be underestimated. Another limitation of this study is related to the linkages between 118 core foods and 613 assigned foods. For some foods, no concentration data were available. In those cases, the most appropriate and realistic alternative from the 118 analyzed foods was linked. This assumption was based on the author’s opinion and is certainly better than assuming a zero concentration for those foods not to underestimate of the exposure, but it still could result in an over- or underestimation of the exposure. Next, the individual food consumption surveys were conducted in the summer; hence, we could not consider seasonal effects of food intake. Because the seasonal variation is relatively greater for some food groups, this issue may have affected the heavy metal exposure estimates. Future studies should cover the four seasons of food intake to estimate heavy metal exposure. We have tried to estimate the distribution of heavy metal exposure. Since all are estimates, we can never be sure about the direction and degree of uncertainties in the analysis. However, we tried to produce robust result by using the robust estimate, median, of 10 000 bootstrap re-sampled values along with the range. Nevertheless, we could not explain the un-linked food consumption data with concentration data, resulting in possible underestimation of exposure in our study. CONCLUSIONS In summary, the long-term exposure to lead and the long-term exposure to cadmium were somewhat greater than the reference values, whereas exposure to mercury was well below the TWI in this population. Fruits were an important source of lead exposure; cereal and fish and shellfish contributed the most to the total cadmium and mercury exposure. Further studies may be necessary to evaluate the food items that contribute to heavy metal exposure, and continuous monitoring is needed to ensure the safety of food intake and dietary patterns among vulnerable groups in Korea. CONFLICT OF INTEREST The authors declare no conflict of interest.

ACKNOWLEDGEMENTS This research was supported by a grant from the Korean Food and Drug Administration (10162KFDA994).

REFERENCES 1 MacIntosh DL, Spengler JD, Ozkaynak H, Tsai L, Ryan PB. Dietary exposures to selected metals and pesticides. Environ Health Perspect 1996; 104: 2. 2 European Food Safety Authority. Metals as contaminants in food. 2011. 3 de Burbure C, Buchet JP, Leroyer A, Nisse C, Haguenoer JM, Mutti A et al. Renal and neurologic effects of cadmium, lead, mercury, and arsenic in children: evidence of early effects and multiple interactions at environmental exposure levels. Environ Health Perspect 2006; 114: 584–590. 4 Llobet JM, Falco G, Casas C, Teixido A, Domingo JL. Concentrations of arsenic, cadmium, mercury, and lead in common foods and estimated daily intake by children, adolescents, adults, and seniors of Catalonia, Spain. J Agric Food Chem 2003; 51: 838–842. 5 Mykkanen H, Rasanen L, Ahola M, Kimppa S. Dietary intakes of mercury, lead, cadmium and arsenic by Finnish children. Hum Nutr Appl Nutr 1986; 40: 32–39. 6 Boon PE, TeBiesebeek JD, Sioen I, Huybrechts I, Moschandreas J, Ruprich J et al. Long-term dietary exposure to lead in young European children: comparing a pan-European approach with a national exposure assessment. Food Addit Contam Part A Chem Anal Control Expo Risk Assess 2012; 29: 1701–1715.



1326 7 Zukowska J, Biziuk M. Methodological evaluation of method for dietary heavy metal intake. J Food Sci 2008; 73: R21–R29. 8 Blanusa M, Lead Juresa D. cadmium, and mercury dietary intake in Croatia. Arh Hig Rada Toksikol 2001; 52: 229–237. 9 Erzen I, Ursic S, Bosnjak K. Assessment of dietary intake of cadmium, lead and mercury via foods of the plant and animal origin in Slovenia. Med Arh 2002; 56: 105–109. 10 Galal-Gorchev H. Dietary intake, levels in food and estimated intake of lead, cadmium, and mercury. Food Addit Contam 1993; 10: 115–128. 11 Garcia-Rico L, Leyva-Perez J, Jara-Marini ME. Content and daily intake of copper, zinc, lead, cadmium, and mercury from dietary supplements in Mexico. Food Chem Toxicol 2007; 45: 1599–1605. 12 Marti-Cid R, Llobet JM, Castell V, Domingo JL. Dietary intake of arsenic, cadmium, mercury, and lead by the population of Catalonia, Spain. Biol Trace Elem Res 2008; 125: 120–132. 13 Munoz O, Bastias JM, Araya M, Morales A, Orellana C, Rebolledo R et al. Estimation of the dietary intake of cadmium, lead, mercury, and arsenic by the population of Santiago (Chile) using a Total Diet Study. Food Chem Toxicol 2005; 43: 1647–1655. 14 Wilhelm M, Wittsiepe J, Schrey P, Hilbig A, Kersting M. Consumption of homegrown products does not increase dietary intake of arsenic, cadmium, lead, and mercury by young children living in an industrialized area of Germany. Sci Total Environ 2005; 343: 61–70. 15 Hart A, Smith GC, Macarthur R, Rose M. Application of uncertainty analysis in assessing dietary exposure. Toxicol Lett 2003; 140-141: 437–442. 16 Finley B, Paustenbach D. The benefits of probabilistic exposure assessment: three case studies involving contaminated air, water, and soil. Risk Anal 1994; 14: 53–73. 17 Andersen LF, Bere E, Kolbjornsen N, Klepp KI. Validity and reproducibility of self-reported intake of fruit and vegetable among 6th graders. Eur J Clin Nutr 2004; 58: 771–777. 18 Nusser SM, Carriquiry AL, Dodd KW, Fuller WA. A semiparametric transformation approach to estimating usual daily intake distributions. J Am Stat Assoc 1996; 91: 1440–1449.


19 Prevention KCfDCa The Fourth Korea National Health and Nutrition Examination Survey (KNHANES IV), Korea Centers for Disease Control and Prevention: Seoul, 2009. 20 Korean Food and Drug Administration. Korean Food Standards Codex vol. 2: Korean Food and Drug Administration: Cheongwon, Korea, 2010. 21 Hirvonen T, Sinkko H, Hallikainen A, Kiviranta H, Pietinen P, Valsta L et al. Modelling the intake of polychlorinated dibenzo-p-dioxins and dibenzofurans: impact of energy under-reporting and number of reporting days in dietary surveys. Food Addit Contam Part A Chem Anal Control Expo Risk Assess 2010; 27: 1170–1176. 22 Lee H-S, Cho Y-H, Park S-O, Kye S-H, Kim B-H, Hahm T-S et al. Dietary exposure of the Korean population to arsenic, cadmium, lead and mercury. J Food Comp Anal 2006; 19: S31–S37. 23 Choi D. Dietary Intake and Risk Assessment of Heavy Metals (TDS). National Institution of Food and Drug Safety Evaluation: Cheongwon-gun, Korea, 2009. 24 Freedman LS, Midthune D, Carroll RJ, Krebs-Smith S, Subar AF, Troiano RP et al. Adjustments to improve the estimation of usual dietary intake distributions in the population. J Nutr 2004; 134: 1836–1843. 25 EFSA. Guidance of the scientific committee on a requrst from EFSA related to uncertainties in dietary exposure assessment. EFSA J 2006; 438: 1–54. 26 de Boer WJ, van der Voet H, Bokkers BG, Bakker MI, Boon PE. Comparison of two models for the estimation of usual intake addressing zero consumption and non-normality. Food Addit Contam Part A Chem Anal Control Expo Risk Assess 2009; 26: 1433–1449. 27 de Winter-Sorkina R, Bakker MI, van Donkersgoed G, van Klaveren JDDietary intake of heavy metals (cadmium, lead and mercury) by the Dutch population, RIKILT. Institute of Food Safety [etc.]: Wageningen [etc.], 2003. 28 Tressou J, Leblanc J, Feinberg M, Bertail P. Statistical methodology to evaluate food exposure to a contaminant and influence of sanitary limits: application to Ochratoxin A. Regul Toxicol Pharmacol 2004; 40: 252–263.

© 2014 Macmillan Publishers Limited