Caramel colour and process by-products in foods and beverages: Part I - Development of a UPLC-MS/MS isotope dilution method for determination of 2-acetyl-4-(1,2,3,4-tetrahydroxybutyl)imidazole (THI), 4-methylimidazole (4-MEI) and 2-methylimidazol (2-MEI).

Caramel colours are used by the food industry in a wide range of foods and beverages. During their manufacturing, low molecular weight compounds such as 4-methylimidazole (4-MEI), the structural isomer of 4-MEI, 2-methylimidazole (2-MEI) and 2-acetyl-4-tetrahydroxy-butylimidazole (THI) are generated. The presence of these inevitable by-products of caramel manufacturing can be hazardous to human health. This publication describes an isotope dilution Ultra-High-performance Liquid Chromatography tandem mass spectrometric method (UHPLC-MS/MS) that was developed and validated for the simultaneous quantification of these impurities in both beverages/liquids and foods. A limit of quantification of 5 µg/kg was obtained for 4-MEI and THI. The expanded measurement uncertainty (U; k = 2) for these compounds was below 51% in beverages/liquids and below 56% in foods. As higher measurement uncertainties were obtained for 2-MEI, the developed analytical procedure can only be used in a semi-quantitative way for this compound.

Caramel colour and process contaminants in foods and beverages: Part II - Occurrence data and exposure assessment of 2-acetyl-4-(1,2,3,4-tetrahydroxybutyl)imidazole (THI) and 4-methylimidazole (4-MEI) in Belgium.

In Europe, 2-acetyl-4-(1,2,3,4-tetrahydroxybutyl)imidazole (THI) and 4-methylimidazole (4-MEI) are - to a certain level - allowed to be present in the food colours ammonia caramel (E 150c) and sulphite ammonia caramel (E 150d). Besides their presence in food colours, exposure to these contaminants may also include other dietary sources. This study describes the occurrence of THI and 4-MEI in a wide variety of food products (n = 522) purchased from the Belgian market and their dietary intake in Belgian consumers from 15 years old onwards. THI was found to be present in 22.4% of the investigated foods at a level up to 551 µg/kg. For 4-MEI (57.7% quantifiable), concentrations up to 2,835 µg/kg were observed. The average dietary intake amounted to 0.02-0.36 µg kg-1 bw-1 day for THI and 0.4-3.7 µg kg-1 bw-1 day for 4-MEI. Coffee, cola and beer were contributing most to the dietary THI and 4-MEI intake in Belgium.

Modelling the exposure to chemicals for risk assessment: a comprehensive library of multimedia and PBPK models for integration, prediction, uncertainty and sensitivity analysis - the MERLIN-Expo tool.

MERLIN-Expo is a library of models that was developed in the frame of the FP7 EU project 4FUN in order to provide an integrated assessment tool for state-of-the-art exposure assessment for environment, biota and humans, allowing the detection of scientific uncertainties at each step of the exposure process. This paper describes the main features of the MERLIN-Expo tool. The main challenges in exposure modelling that MERLIN-Expo has tackled are: (i) the integration of multimedia (MM) models simulating the fate of chemicals in environmental media, and of physiologically based pharmacokinetic (PBPK) models simulating the fate of chemicals in human body. MERLIN-Expo thus allows the determination of internal effective chemical concentrations; (ii) the incorporation of a set of functionalities for uncertainty/sensitivity analysis, from screening to variance-based approaches. The availability of such tools for uncertainty and sensitivity analysis aimed to facilitate the incorporation of such issues in future decision making; (iii) the integration of human and wildlife biota targets with common fate modelling in the environment. MERLIN-Expo is composed of a library of fate models dedicated to non biological receptor media (surface waters, soils, outdoor air), biological media of concern for humans (several cultivated crops, mammals, milk, fish), as well as wildlife biota (primary producers in rivers, invertebrates, fish) and humans. These models can be linked together to create flexible scenarios relevant for both human and wildlife biota exposure. Standardized documentation for each model and training material were prepared to support an accurate use of the tool by end-users. One of the objectives of the 4FUN project was also to increase the confidence in the applicability of the MERLIN-Expo tool through targeted realistic case studies. In particular, we aimed at demonstrating the feasibility of building complex realistic exposure scenarios and the accuracy of the modelling predictions through a comparison with actual measurements.

A semi-probabilistic modelling approach for the estimation of dietary exposure to phthalates in the Belgian adult population.

In this study, a semi-probabilistic modelling approach was applied for the estimation of the long-term human dietary exposure to phthalates--one of world's most used families of plasticisers. Four phthalate compounds were considered: diethyl phthalate (DEP), di-n-butyl phthalate (DnBP), benzylbutyl phthalate (BBP) and di(2-ethylhexyl) phthalate (DEHP). Intake estimates were calculated for the Belgian adult population and several subgroups of this population for two considered scenarios using an extended version of the EN-forc model. The highest intake rates were found for DEHP, followed by DnBP, BBP and DEP. In the Belgian adult population, men and young adults generally had the highest dietary phthalate intake estimates. Nevertheless, predicted dietary intake rates for all four investigated phthalates were far below the corresponding tolerable daily intake (TDI) values (i.e. P99 intake values were 6.4% of the TDI at most), which is reassuring because adults are also exposed to phthalates via other contamination pathways (e.g. dust ingestion and inhalation). The food groups contributing most to the dietary exposure were grains and grain-based products for DEP, milk and dairy products for DnBP, meat and meat products or grains and grain-based products (depending on the scenario) for BBP and meat and meat products for DEHP. Comparison of the predicted intake results based on modelled phthalate concentrations in food products with intake estimates from other surveys (mostly based on measured concentrations) showed that the extended version of the EN-forc model is a suitable semi-probabilistic tool for the estimation and evaluation of the long-term dietary intake of phthalates in humans.

Modelling the environmental transfer of phthalates and polychlorinated dibenzo-p-dioxins and dibenzofurans into agricultural products: the EN-forc model.

This study aimed to predict the occurrence of four phthalates, two polychlorinated dibenzo-p-dioxins and two polychlorinated dibenzofurans in environmental and agricultural media from observed concentrations in air, sludge, manure and concentrate. For the environmental and agricultural fate modelling, the newly developed multimedia model "EN-forc" (ENvironmental Food transfer model for ORganic Contaminants) was used. To validate EN-forc calculations, the predicted concentrations of the considered chemicals in soil, groundwater, drinking water, plants and animal products were compared with both observed and modelled concentrations available in the literature. For the majority of the considered matrices, predicted phthalate and dioxin levels differed one order of magnitude at most with observed concentrations. Unfortunately, the transfer models implemented in EN-forc lacked power to predict levels of some phthalates and dioxins in pasture, root crops and/or tubers. Concentrations of phthalates and dioxins in offal could not be predicted due to the absence of suitable models that have an acceptable level of complexity to implement in EN-forc. For this type of food products, further research is highly encouraged. In a next step, the modelling framework of EN-forc will be extended in order to be able to predict human dietary exposure to organic chemicals like phthalates and dioxins.

Transfer of eight phthalates through the milk chain--a case study.

This survey determined the levels of eight phthalates - i.e. dimethyl phthalate (DMP), diethyl phthalate (DEP), diisobutyl phthalate (DiBP), di-n-butyl phthalate (DnBP), benzylbutyl phthalate (BzBP), di(2-ethylhexyl) phthalate (DEHP), dicyclohexyl phthalate (DCHP) and di-n-octyl phthalate (DnOP) - in several Belgian milk and dairy products. Samples were obtained from various farms, a dairy factory and from different shops in order to investigate phthalate contamination "from farm to fork". At several stages in the milk chain, product contamination with phthalates - mostly DiBP, DnBP, BzBP and DEHP - was observed. At farm level, the mechanical milking process and the intake of phthalate containing feed by the cattle were found to be possible contamination sources. At industry and retail level, contact materials including packaging materials were additional contamination sources for phthalates in milk and dairy products.

Effect of cooking at home on the levels of eight phthalates in foods.

Food products can be contaminated with toxic compounds via the environment. Another possibility of food contamination is that toxicants are generated in foods or that chemicals migrate from food contact materials into foods during processing. In this study, the effect of cooking at home on the levels of phthalates - world's most used group of plasticisers - in various food types (starchy products, vegetables and meat and fish) was examined. Eight compounds were considered, namely dimethyl phthalate (DMP), diethyl phthalate (DEP), diisobutyl phthalate (DiBP), di-n-butyl phthalate (DnBP), benzylbutyl phthalate (BBP), di(2-ethylhexyl) phthalate (DEHP), dicyclohexyl phthalate (DCHP) and di-n-octyl phthalate (DnOP). Food products were analysed before as well as after cooking (boiling, steaming, (deep-)frying or grilling). In general, phthalate concentrations in foods declined after cooking, except in vegetables, where almost no effect was seen. Several factors influenced the degree of this decline (e.g. weight difference, fat uptake, etc.). Of all phthalates, DEHP, DiBP and BBP were affected the most. In conclusion, cooking at home definitely affected phthalate concentrations in foods and thus needs to be considered in order to correctly assess humans' dietary exposure to these contaminants.

Phthalates in Belgian cow's milk and the role of feed and other contamination pathways at farm level.

This study investigated the occurrence of dimethyl phthalate (DMP), diethyl phthalate (DEP), diisobutyl phthalate (DiBP), di-n-butyl phthalate (DnBP), benzylbutyl phthalate (BBP), di(2-ethylhexyl) phthalate (DEHP), dicyclohexyl phthalate (DCHP) and di-n-octyl phthalate (DnOP) in raw cow's milk and feed from Belgian farms in order to determine their most relevant contamination pathways in milk. Measurable levels of DMP, DEP, DnBP, DCHP and DnOP were found in various feed samples, although they were not observed in milk. A plausible explanation for this is that they are rapidly metabolised in cows. DEHP and in a smaller degree also DiBP and BBP levels in milk seemed to vary across seasons and farms. DiBP and BBP levels were lower in summer than in winter milk, which was in contrast with what was observed for DEHP. This is possibly due to another feed composition during summer and winter. Comparing BBP and DEHP concentrations in manually with those in mechanically obtained milk revealed that, besides environmental contamination via feed ingestion, contact materials used during the mechanical milking process is another important contamination pathway. Concentrations observed in this study confirm the decreasing trend of DEHP in European cow's milk owing to the substitution of DEHP by other plasticisers.

Analysis of phthalates in food products and packaging materials sold on the Belgian market.

Phthalates are organic lipophilic compounds that are principally used as plasticiser to increase the flexibility of plastic polymers. Other applications are a.o. the use of phthalates in printing inks and lacquers. Human exposure to phthalates mainly occurs via food ingestion and can induce adverse health effects. In this study, the presence of eight phthalate compounds--dimethyl phthalate (DMP), diethyl phthalate (DEP), diisobutyl phthalate (DiBP), di-n-butyl phthalate (DnBP), benzylbutyl phthalate (BBP), di(2-ethylhexyl) phthalate (DEHP), dicyclohexyl phthalate (DCHP) and di-n-octyl phthalate (DnOP)--was investigated in 400 food products, divided over eleven groups, and packages sold on the Belgian market. For this purpose, suitable extraction techniques were developed and validated for four different matrices, namely high-fat foods, low-fat food products, aqueous-based beverages and packaging materials. The instrumental analysis was performed by means of gas chromatography-low resolution-mass spectrometry with electron impact ionisation (GC-EI-MS). A wide variety of phthalate concentrations was observed in the different groups. DEHP was found in the highest concentration in almost every group. Moreover, DEHP was the most abundant phthalate compound, followed by DiBP, DnBP and BBP. This survey is part of the PHTAL project, which is the first project that discusses phthalate contamination on the Belgian food market.

Health impact assessment and monetary valuation of IQ loss in pre-school children due to lead exposure through locally produced food.

A case study has been performed which involved the full chain assessment from policy drivers to health effect quantification of lead exposure through locally produced food on loss of IQ in pre-school children at the population level across the EU-27, including monetary valuation of the estimated health impact. Main policy scenarios cover the period from 2000 to 2020 and include the most important Community policy developments expected to affect the environmental release of lead (Pb) and corresponding human exposure patterns. Three distinct scenarios were explored: the emission situation based on 2000 data, a business-as-usual scenario (BAU) up to 2010 and 2020 and a scenario incorporating the most likely technological change expected (Most Feasible Technical Reductions, MFTR) in response to current and future legislation. Consecutive model calculations (MSCE-HM, WATSON, XtraFOOD, IEUBK) were performed by different partners on the project as part of the full chain approach to derive estimates of blood lead (B-Pb) levels in children as a consequence of the consumption of local produce. The estimated B-Pb levels were translated into an average loss of IQ points/child using an empirical relationship based on a meta-analysis performed by Schwartz (1994). The calculated losses in IQ points were subsequently further translated into the average cost/child using a cost estimate of €10.000 per loss of IQ point based on data from a literature review. The estimated average reduction of cost/child (%) for all countries considered in 2010 under BAU and MFTR are 12.16 and 18.08% as compared to base line conditions, respectively. In 2020 the percentages amount to 20.19 and 23.39%. The case study provides an example of the full-chain impact pathway approach taking into account all foreseeable pathways both for assessing the environmental fate and the associated human exposure and the mode of toxic action to arrive at quantitative estimates of health impacts at the individual and the population risk levels alike at EU scale. As the estimated B-Pb levels fall below the range of observed biomonitoring data collected for pre-school children in 6 different EU countries, results presented in this paper are only a first approximation of the costs entailed in the health effects of exposure to lead and the potential benefits that may arise from MFTR measures inscribed in Commission policies.

Predicting blood lead levels from current and past environmental data in Europe.

The present case study on lead in Europe illustrates the use of the Integrated Monitoring Framework Strategy to assess the health outcome of environmental pollution by evaluating the associations between lead in various environmental compartments (air, soil, dust, drinking water and diet) and lead concentrations in blood (B-Pb) for various age-related sub-populations. The case study was aimed to investigate whether environmental, exposure and biomonitoring data at general population level, covering all EU member states, could be integrated. Although blood lead has been monitored extensively in Europe, consistent datasets are not yet available. Data diverge with regard to objectives, regional scale, sampling years, gender, age groups and sample size. Significant correlations were found between B-Pb and the concentrations of Pb in air and diet. The significant decrease of the Pb in air over time from 0.31 µg/m(3) (P95: 0.94; n=98) prior to 1990 to 0.045 µg/m(3) (P95: 0.11; n=256) in 2007 (latest observations included) (Δ=-85%) corresponds to a decline in B-Pb by 48% and 57% in adult women and adult men, respectively. For pre-school children a more shallow decline in B-Pb of 16% was calculated over the same period. Similarly, the reduction in Pb-dietary intake from on average 68.7 µg/d (P95: 161.6; n=19) in 1978 to 35.7 µg/d (P95: 82.3; n=33) in the years post 2000 (Δ=-48%) is paralleled by a decline in B-Pb of 32, 33 and 19% in adult women, primary- and pre-school children, respectively. Insufficient data exist for other age groups to calculate statistically significant correlations. Although regression models have been derived to predict B-Pb for different sub-populations in Europe based on Pb concentrations in air and soil as well as dietary intake, it is concluded that the available data are insufficient to accurately predict actual and future simultaneous exposure to Pb from various environmental compartments, and as a consequence the health impact of Pb for various target populations at EU scale. At least due to data availability, air Pb remains the best predictor of B-Pb in the population. However, lead emission sources have largely been reduced and inhalation of lead in air is not causal to B-Pb levels. Therefore, there is a need of adequate data for Pb in soil and house dust, and in diet and drinking water as these are causal exposure sources with a longer Pb half-life than air. An extended and more harmonized surveillance system monitoring B-Pb, especially in children, is urgently required in order to identify, quantify and reduce still remaining sources of Pb exposure.

Dietary cadmium intake by the Belgian adult population.

The aim of this study was to estimate the dietary cadmium (Cd) intake of the Belgian adult population, to compare this dietary Cd exposure to the tolerable weekly intake (TWI) recently established by the European Food Safety Authority (EFSA) and to determine the major food groups that contribute to dietary Cd exposure in Belgium. Food consumption data were derived from the 2004 Belgian food consumption survey (two 24 h recalls, 3083 participants). Cadmium concentrations in food items (n = 4000) were gathered from the control program of the Belgian Federal Agency for the Safety of the Food Chain for the period 2006-2008. Dietary intake per individual was calculated from consumption data and median Cd concentrations. The population mean, median and 95th percentile of the dietary intake values were 0.98, 0.85 and 2.02 µg kg⁻¹ body weight per week respectively. Two percent of the Belgian adult population has a dietary Cd intake above the recent TWI of 2.5 µg kg⁻¹ body weight established by EFSA in 2009. Cereal products and potatoes contribute for more than 60% to Cd intake.