Food and Chemical Toxicology 139 (2020) 111292

Contents lists available at ScienceDirect

Food and Chemical Toxicology

journal homepage: www.elsevier.com/locate/foodchemtox

Health risk assessment to dioxins, furans and PCBs in young children: The T
first French evaluation
Marion Hulina, Véronique Sirota, Paule Vasseurb, Aurelie Mahea, Jean-Charles Leblanca,
Julien Jeanb, Philippe Marchandc, Anaïs Venisseauc, Bruno Le Bizecc, Gilles Rivièrea,∗
a
ANSES, Risk Assessment Department (DER), 14 rue Pierre et Marie Curie, F-94701, Maisons-Alfort, France
b
CNRS UMR 7360, University of Lorraine, F-57070, Metz, France
c
LABERCA, Oniris, INRA, F-44300, Nantes, France

A R T I C LE I N FO A B S T R A C T

Keywords: A total diet study (TDS) was conducted between 2010 and 2016 to characterize the health risk related to che-
Total diet study mical residues in food of French not breastfed children under three years of age (infant TDS). Among the targeted
PCDD/F and PCBs substances, polychlorinated dibenzo-p-dioxins (PCDDs), polychlorinated dibenzofurans (PCDFs) and poly-
Exposure assessment chlorinated biphenyls (PCBs) have been characterized as they accumulate through the food chain, especially in
Population exposure
lipid-rich food items, and because they have been associated with a number of adverse effects in humans.
Contaminants
Food samples (n = 180) were collected to be representative of the dioxins and PCB exposure through the
Risk characterization
whole diet of non-breastfed children from 1 to 36 months old and prepared as consumed (including cooking)
prior to analysis.
Dietary exposure was then assessed for 705 representative children under 3 years of age based on their food
consumptions recorded through a 3-consecutive-days record. Levels of PCDD/Fs and PCBs in infant food were
lower than those observed in common food, leading to significant differences in exposure according to age
groups. Mean exposures to PCDD/Fs ranged from 0.22 to 0.44 pg TEQWHO05.kg bw−1.d−1 (0.40–0.65 at the 90th
percentile), depending on the age group and the hypothesis considered to manage left-censored data. Mean
exposure to non-dioxin-like PCBs ranged from 0.87 ng kg bw−1.d−1 (1.55 at the 90th percentile) in the 1–4
months old children to 3.53 ng kg bw−1.d−1 (5.44 at the 90th percentile) in the 13–36 months old children. For
dioxins and NDL-PCBs, the tolerable daily intake (TDI) was exceeded for some age groups, in particular for older
ones.
Therefore, appropriate management measures must continue for reducing exposure; it concerns mainly
common milk in youngest children, ultra-fresh dairy products and fish. For PCBs, recommendations on fish
consumption should be reminded. Moreover, toxicity studies focusing on mixtures of dioxin-like compounds
should be encouraged in order to take into account effect of mixtures.

1. Introduction substances for the general population with food representing more than
90% of the total exposure (EFSA, 2005; EFSA, 2010; EFSA, 2018).
Polychlorinated dibenzo-p-dioxins (PCDDs), polychlorinated di- Certain congeners of the dioxin, furan, and polychlorinated biphenyl
benzofurans (PCDFs) are generated during thermal (fires, incinerations, family are known as -dioxin-like- as a result of chemical structures,
etc.) or chemical processes, whereas polychlorobiphenyls (PCBs) have physico-chemical properties, and toxic responses similar to 2,3,7,8-
been produced and used until 1987 for their excellent electric insulators tetrachlorodibenzo-p-dioxin (TCDD), the index chemical for dioxin-like
explaining their extensive use as coolant fluids in power transformers compounds (also known as Seveso dioxin).
and capacitors. PCDDs, PCDFs and PCBs regroup 75, 135 and 209 The toxicity of PCBs varies significantly depending on the species
congeners respectively and are classified as persistent organic pollu- and on the different congeners within the same species, the number and
tants. They accumulate along the food chain and are mainly stored in position of the chlorine atoms on the phenyl nuclei determining their
fatty tissues. Food appears to be the main route of exposure of these toxicity. Based on structural and toxicological properties, PCBs are

∗
Corresponding author. Risk Assessment Department (DER), French Agency for Food, Environmental and Occupational Health & Safety (ANSES), 14 rue Pierre et
Marie Curie, 94701, Maisons-Alfort, France.
E-mail address: [email protected] (G. Rivière).

https://doi.org/10.1016/j.fct.2020.111292
Received 30 November 2019; Received in revised form 16 March 2020; Accepted 17 March 2020
Available online 21 March 2020
0278-6915/ © 2020 Elsevier Ltd. All rights reserved.
M. Hulin, et al. Food and Chemical Toxicology 139 (2020) 111292

grouped into 2 categories, the “dioxin-like PCBs” (dl-PCBs) and the 2. Material and methods
“non dioxin-like PCBs” (ndl-PCBs). The sum of 6 “non-dioxin-like” PCBs
(28, 52, 101, 138, 153, 180) is generally considered as representing 2.1. Consumption data and sample collection
50% of the total concentration of PCBs (Afssa, 2007).
Besides the activation of the Ah (aryl hydrocarbon) receptor by dl- Consumption data are those from the cross-sectional survey on in-
PCBs, several other receptors are involved in the PCB response for both dividual dietary consumption in children under 3 years conducted by
categories: receptors of sexual steroids (estrogens, androgens), thyroid the Syndicat Français des Aliments de l’Enfance et de la Nutrition
hormones, neurotransmitters or calcic receptors (RyR). The number of Clinique, © « Etude SOFRES 2005/Université de Bourgogne – Pr M.
targets affected by a same congener and the multiplicity of congeners in Fantino pour le Syndicat Français des Aliments de l’Enfance » (Fantino,
each PCB mixtures resulting in pollution explain the diversity of ad- 2005; Fantino and Gourmet, 2008). In this survey, dietary consumption
verse effects observed in contaminated individuals: cutaneous, hepatic, data of 705 French children aged from 1 month to 3 years were re-
metabolic, immunologic, neurologic, endocrine disruption. Both human corded between January and March 2005. Children were selected
and animal studies identified the developing brain as a vulnerable through proportionate quota sampling based on the age, the occupation
target of PCBs (Zoeller et al., 2000; Sagiv et al., 2010). Recent studies of the mother and the family socioeconomic category. People who took
suggested that the non-dioxin-like PCBs are primarily responsible for care of the children have completed 3-consecutive-days records to de-
the developmental neurotoxicity associated with PCBs (Klocke and scribe all food consumptions including brands, quantities and portion
Lein, 2020). sizes. Breast-fed children, even partially, were excluded from this
The international agency for research on cancer (IARC) classified survey. Information on the body weight was also recorded on the child's
PCBs as carcinogenic to humans in 2013, so as TCDD (2,3,7,8-tetra- health record.
chlorodibenzo-p-dioxin) already classified in 1997. Adverse non-cancer The sampling plan has already been described (Hulin et al., 2014).
effects associated with oral TCDD exposure include hepatic, neurolo- Briefly, on the basis of the consumption survey, the most consumed
gical, immunological, reproductive, endocrine, and developmental ef- foods by the children in terms of quantity and/or consumer rates were
fects. In the offspring of rat, effects on reproduction and development as selected. In addition, some foods known to contribute significantly to
well as immunological effects were observed after administration of the exposure to one of the chemicals of interest were added to the list.
TCDD (Seveso dioxin). The main mechanism of action of TCDD and Foods were sampled between July 2011 and July 2012. For each food
dioxin-like compounds such as PCDDs, PCDFs, and dl-PCBs, is their item identified, 12 subsamples of equal weight were analyzed: every
binding to the Ah receptor (EFSA, 2018). Exposure of infants to PCBs month, one subsample of each food item was bought and prepared “as
and PCDD/Fs which are neurodevelopmental toxicants is therefore a consumed”, according to the results of a specific on-line survey on food
concern. preparation (Hulin et al., 2014). After collection, the 12 subsamples
Due to a number of regulatory measures since the 1980s, the have been grouped, homogenized and frozen (−18 °C) prior to analysis.
emission of dioxins into the environment and, consequently, human In total, 5 484 food products were purchased to produce 457 composite
exposure to dioxins, have decreased significantly (Aylward and Hays, samples. Selected food items covered more than 97% of the children
2002; EFSA, 2018). However, due to their high persistence and bioac- diet.
cumulation and their toxic potential, dietary dioxins and PCB exposure
can be of concern. Different studies and evaluations showed a general 2.2. Sample analyses
decrease in dietary exposure of the European population to dioxins and
PCBs since the 2000's However, exceedance of the reference values was Among the 457 food samples, PCBs and PCDD/Fs were only ana-
still observed, toddlers and other children being the most exposed lyzed in food items known or supposed to contain PCDD/Fs and PCBs
groups. In 2012, Sirot et al. concluded that less than 4% of the French and for which data were not available in our previous study (Sirot et al.,
population exceeded the health-based guidance value (HBGV) for 2012). In total, PCBs and PCDD/Fs concentrations were measured in
PCDD/F + dl-PCBs set by JECFA in 2001. For total PCB, it was esti- 178 samples: 169 infant foods, i.e. products commercialized especially
mated that 2.6 and 6.5% of, respectively, French adults and French for infant and young children, and 9 common foods. LABERCA, that is
children exceeded the health based guidance value set by Afssa in 2007 the French Reference Laboratory for dioxin and PCB analysis in food,
(Afssa, 2007; Sirot et al., 2012). However this study did not take into performed the analyses of the samples using method already described
account children under 3 years old. This population is known for being elsewhere (Antignac et al., 2006; Costera et al., 2006).
more susceptible to contaminants (Landrigan et al., 2003), especially Picograde® quality solvents were purchased from LGC Promochem
during the prenatal and postnatal period (Makri et al., 2004; Sly and (Wesel, Germany), and dichloromethane was from Biosolve (Dieuze,
Flack, 2008; Diamanti Kandarakis et al., 2009). Moreover, they con- France). Sulfuric acid was purchased from Panreac Quimica (Barcelona,
sume more food than adults in proportion to their body weight. As a Spain). Silica gel and sodium sulfate were from Merck (Darmstadt,
result, this population is considered more vulnerable than other groups Germany), Florisil® and Carbopack™ C from LGC Promochem (Wesel,
to chemicals' exposure. There are few studies in Europe that have in- Germany). 13C-PCDD/Fs, 13C-dl-PCB-77/81/105/114/118/123/126/
vestigated non breasted infant and toddler exposure to PCDD/Fs and 156/157/167/169/189, 13C-ndl-PCB-28/52/101/138/153/180, were
PCBs (Pandelova et al., 2011; EFSA, 2012; De Filippis et al., 2014). purchased from Wellington Laboratories (Guelph, Ontario, Canada). All
More recently, in 2018, EFSA concluded that the exposure to PCDD/Fs calibration standards and spiking solutions were prepared by serial
and dl-PCBs was of concern for all age groups including infants and dilutions in toluene.
toddlers (EFSA, 2018). Food samples were first freeze-dried before grinding. The 13C-la-
In our study, we aimed at evaluating the exposure of the French belled internal standards were added to the samples before the ex-
infants to PCBs and PCDD/Fs present in food as contaminants using the traction step. Lipids were extracted from lyophilised samples by
Total Diet Study (TDS) approach. TDSs aim at providing contamination Pressurised Liquid Extraction (Büchi, Rungis, France) using a toluene/
data for food prepared as consumed by the population, and exposure acetone mixture (70:30, v/v), pressure set at 100 bar and temperature
data, in order to help risk managers in their decisions (EFSA et al., at 120 °C. The solvent was then evaporated and the lipid content of the
2011). In 2010, the French Agency for Food, Environmental and Oc- sample was determined gravimetrically. Finally, before purification, the
cupational Health & Safety (Anses) launched the first French infant TDS dried fat extract was redissolved in n-hexane.
focusing on children under 3 years old. This paper aims at presenting After removal of fat on a silica gel column loaded with sulfuric acid,
the contamination and the exposure data generated during the course of PCBs were separated from PCDD/Fs by means of a Florisil® column. The
this study for PCBs and PCDD/Fs. PCDD/F fraction was further cleaned up onto a column consisting of a

2
M. Hulin, et al. Food and Chemical Toxicology 139 (2020) 111292

mixture of Carbopack™ C/Celite® 545. Separation of coplanar (non- common food as the present study focused on infant foods (Sirot et al.,
ortho) PCBs from non-planar PCBs was achieved on an activated mix- 2012). These data were based on a sampling of 1 319 food samples
ture of Florisil®/Carbopack™ C/Celite® 545 (overnight at 130 °C). After collected in the whole metropolitan French territory divided into eight
addition of external standards for the recovery calculation (13C12- regions between 2007 and 2009. For each food samples, 15 subsamples
1,2,3,4-TCDD for the PCDD/Fs, 13C12-PCB111 for the PCBs), the final representative of the population food consumption have been bought
extract was reconstituted by addition of toluene in the 3 fractions. and prepared as consumed by the population prior to analyses, ac-
Chromatographic separation of PCDD/F congeners was carried out cording to the general TDS methodology and as it was done in the
on a DB-5MS column (60 m × 0.25 mm i.d., 0.25 μm film thickness; present study.
Agilent Technologies, USA), while for PCBs a HT8-PCB column For each food items, the analytic result obtained on the pooled
(60 m × 0.25 mm i.d., 0.25 μm film thickness; SGE Analytical Science, sample was associated with the individual consumption values, based
UK) was used. PCDD/Fs and PCBs were quantified by gas chromato- on a deterministic approach as suggested by EFSA, FAO and WHO for
graphy (7890A; Agilent Technologies, USA) coupled to high-resolution TDS approach (EFSA et al., 2011). When available, the allocation of
mass spectrometry (double sector, JMS-700D and 800D; Jeol, Japan). those concentrations to a food consumed by an individual took into
All the procedures integrated the necessary quality assurance account its home region. If no regional data was available, national data
parameters to fulfil the requirements of the European Commission were used. For infant formulae that are diluted prior to use, the type
Regulation (EU) N° 252/2012 of March 2012 laying down methods of and brand of water used to reconstitute the products were also con-
sampling and analysis for the official control of levels of dioxins, dioxin- sidered (Hulin et al., 2014).
like PCBs and non-dioxin-like PCBs in certain foodstuffs. Analysis were Exposure was therefore assessed individually, for each child of the
performed upon an accredited system ISO-17025. The method used consumption survey, according to the following formula:
have been validated and is accredited ISO-17025.
n
To ensure the quality of the analysis, besides the use of appropriate ∑k = 1 Ci,k × Lk, j
Ei, j =
internal standards in each sample, labelled standards were added at the BWi
end of each process in order to calculate recoveries. Moreover, cleaned
laboratory glassware was rinsed with dichloromethane prior to use and where Ei,j is the mean daily exposure to contaminant j of individual i, n
the analyses were carried out in an over-pressurised room to minimise is the number of foods in the diet, Ci,k is the daily consumption of food k
environmental contamination. by individual i, Lk,j is the concentration of contaminant j in food k, and
Further, a continuous monitoring of the analytical procedure was BWi is the body weight of individual i.
implemented through procedural blanks. As the analytical contamina- Mean, standard deviation, and 90th percentile (P90) of exposure
tion is fully under control, blank control was not deducted. were then calculated for the population divided into four age groups:
Reproducibility was assessed using quality control samples (QC) reg- 1–4 months (N = 124), 5–6 months (N = 127), 7–12 months
ularly characterised over years. The accuracy of the analytical method (N = 195) and 13–36 months (N = 259). Total exposure levels (LB-UB)
is further ensured by regular participation of the laboratory to profi- to PCDD/Fs, dl-PCBs + PCDD/Fs and to the 6 ndl-PCBs of children
ciency tests organized by the European Reference Laboratory (EURL) under 1 year according to the type of milk consumed have also been
for POPs. calculated. In order to study the impact of milk consumption on ex-
posure, student tests have been performed between the type of con-
2.3. Contamination data, exposure calculation and risk assessment sumers.
The food contribution to the mean exposure has been assessed as the
In order to evaluate the risk associated with exposure to PCDD/Fs, percentage of the total exposure due to the consumption of each food
World Health Organization toxic equivalency factors (WHO TEF) were group.
used to express exposure in the toxic equivalents (TEQs) (Afssa The health risk associated with the dietary exposure to each che-
2005)Therefore, for PCDD/Fs, the sum of the 17 congeners (2.3.7.8 - mical has been assessed by calculating, for each age group, the per-
TCDD, 1.2.3.7.8 - PeCDD, 1.2.3.4.7.8 - HxCDD, 1.2.3.6.7.8 - HxCDD, centage of children over the health-based guidance value and its 95%
1.2.3.7.8.9 - HxCDD, 1.2.3.4.6.7.8- HpCDD, OCDD, 2.3.7.8 - TCDF, confidence interval (CI95%). PCDD/Fs and PCBs have been evaluated
1.2.3.7.8 - PeCDF, 2.3.4.7.8 - PeCDF, 1.2.3.4.7.8 - HxCDF, 1.2.3.6.7.8 - separately in this study in order to facilitate the implementation of
HxCDF, 1.2.3.7.8.9 - HxCDF, 2.3.4.6.7.8 - HxCDF, 1.2.3.4.6.7.8 management measures as sources of each compound's family are dif-
-HpCDF, 1.2.3.4.7.8.9 -HpCDF, OCDF) was calculated by multiplying ferent. Additionally, exposure to the sum of PCDDs, PCDFs and dl-PCBs
the analytical concentration of each congener by their individual TEFs was estimated to make it possible to compare this exposure with the
(Van den Berg et al., 2006; EFSA, 2012) named TEFWHO05. Exposure health-based guidance value set by Efsa in 2018 (EFSA, 2018). For
to PCBs was measured by the mean of 6 ndl-PCBs (28, 52, 101, 138, PCDD/Fs, the toxicological reference value of US-EPA (US-EPA, 2012)
153, 180), as their sum represents 50% of the total concentration of of 0.7 pg TEQ.kg bw−1.d−1 has been considered; for PCB a tolerable
PCBs (Afssa, 2007). daily intake (TDI) of 10 ng kg bw−1.d−1 has been considered to eval-
Censored data (results below the limits of detection (LOD) and uate the risk associated with the 6 NDL-PCB, mostly found in foodstuffs
quantification (LOQ)) were processed according to a substitution and which represent 50% of all PCB congeners in food (Afssa, 2007).
method based on the WHO recommendations (GEMS/Food-EURO, For the sum of PCDD/Fs and dl-PCBs, the health-based guidance value
2013). It involved framing the actual level using the lowest (lower- set at 2 pg TEQ.kg bw−1.week−1 by Efsa (EFSA, 2018) on the basis of a
bound (LB)) and highest (upper-bound (UB)) values possible. The LB decrease of the sperm quality in 9 years old infants exposed during
was calculated by assuming that all values below the LOD were equal to childhood was considered to evaluate the risk.
zero and those between the LOD and the LOQ were equal to the LOD.
The UB was calculated by assuming that all values below the LOD were
equal to the LOD and those between the LOD and the LOQ were equal to 2.4. Collective appraisal
the LOQ. Exposure data were then estimated according to both the LB
and UB hypotheses. For the 6 ndl-PCB, as all samples were detected, The collective assessment of the risk linked to PCDD/F and PCB
concentration was the same for LB and UB hypothesis. Therefore only exposures has been conducted with the expert panel dealing with
one result was presented. chemical contaminants in food.
In order to complete the diet coverage, some additional regional and
national data from French TDS2 have been used for contamination of

3
M. Hulin, et al. Food and Chemical Toxicology 139 (2020) 111292

Table 1
Levels of PCDD/Fs, dl-PCBs (a) and 6 ndl-PCBs (b) in food consumed by French infants and toddlers.
a. PCDD/F and dl-PCB

N % detection Fresh weight Fat basis

LB UB LB UB

Mean SD Min Max Mean SD Min Max Mean SD Min Max Mean SD Min Max

PCDD/Fs (pg TEQWHO05.g−1)

Milk-based beverage 8 77.9 0.001 0.000 0.001 0.002 0.002 0.000 0.002 0.003 0.062 0.013 0.041 0.080 0.102 0.018 0.065 0.119
Cereals-based food 5 64.7 0.002 0.003 0.000 0.006 0.005 0.005 0.001 0.010 0.068 0.059 0.021 0.162 0.219 0.207 0.059 0.480
Milk-based dessert 6 83.3 0.004 0.001 0.002 0.005 0.004 0.001 0.003 0.005 0.146 0.042 0.076 0.199 0.174 0.021 0.150 0.206
Growing-up milk 9 65.4 0.003 0.002 0.001 0.006 0.004 0.002 0.002 0.007 0.150 0.121 0.036 0.358 0.213 0.110 0.095 0.405
Soup, puree 11 88.2 0.002 0.001 0.001 0.003 0.002 0.001 0.001 0.004 0.105 0.031 0.066 0.164 0.139 0.041 0.083 0.204
Fruit purée 4 55.9 0.001 0.000 0.000 0.001 0.002 0.001 0.002 0.003 0.166 0.090 0.062 0.278 0.441 0.200 0.252 0.723
Vegetable-based ready-to-eat meal 20 63.5 0.001 0.001 0.000 0.002 0.002 0.001 0.001 0.004 0.093 0.068 0.034 0.347 0.205 0.096 0.109 0.528
Meat/fish based ready-to-eat meal 45 82.5 0.002 0.004 0.001 0.025 0.003 0.004 0.001 0.025 0.145 0.181 0.041 1.127 0.197 0.195 0.068 1.127
Infant formula 28 82.1 0.001 0.001 0.000 0.004 0.002 0.001 0.001 0.004 0.059 0.046 0.017 0.239 0.088 0.050 0.044 0.239
Follow-on formula 33 89.3 0.002 0.001 0.001 0.005 0.002 0.001 0.001 0.005 0.102 0.058 0.044 0.327 0.122 0.057 0.067 0.328

dl-PCBs (pg TEQWHO05.g−1)

Milk-based beverage 8 59.4 0.001 0.000 0.000 0.001 0.001 0.000 0.001 0.001 0.028 0.015 0.002 0.049 0.046 0.010 0.037 0.063
Cereals-based food 5 28.3 0.000 0.000 0.000 0.000 0.002 0.002 0.000 0.005 0.007 0.014 0.000 0.033 0.080 0.059 0.029 0.151
Milk-based dessert 6 87.5 0.008 0.004 0.004 0.013 0.008 0.003 0.005 0.013 0.316 0.109 0.247 0.531 0.321 0.106 0.249 0.531
Growing-up milk 9 92.6 0.004 0.003 0.000 0.008 0.004 0.003 0.001 0.008 0.217 0.207 0.022 0.564 0.219 0.206 0.025 0.564
Soup, puree 11 62.1 0.001 0.001 0.000 0.005 0.002 0.001 0.000 0.005 0.089 0.060 0.029 0.235 0.100 0.058 0.040 0.236
Fruit purée 4 64.6 0.001 0.001 0.001 0.002 0.001 0.001 0.001 0.002 0.181 0.068 0.126 0.274 0.203 0.051 0.156 0.275
Vegetable-based ready-to-eat meal 20 87.5 0.002 0.001 0.000 0.005 0.002 0.001 0.001 0.005 0.149 0.080 0.046 0.413 0.162 0.076 0.077 0.413
Meat/fish based ready-to-eat meal 45 80.4 0.004 0.012 0.000 0.081 0.005 0.012 0.000 0.081 0.235 0.553 0.002 3.641 0.243 0.551 0.020 3.641
Infant formula 28 76.8 0.001 0.001 0.000 0.005 0.001 0.001 0.000 0.005 0.033 0.061 0.001 0.307 0.039 0.061 0.010 0.307
Follow-on formula 33 79.0 0.001 0.002 0.000 0.008 0.001 0.002 0.000 0.008 0.070 0.120 0.012 0.625 0.075 0.118 0.016 0.625

b. 6 ndl-PCBs

N % detection Fresh weight Fat basis

Mean SD Min Max Mean SD Min Max

6 ndl-PCBs (ng.kg−1)

Milk-based beverage 8 100 5.4 1.8 3.6 8.1 260.7 97.6 160.7 437.6
Cereals-based food 5 100 6.6 8.3 0.8 19.8 277.8 214.0 117.5 602.9
Milk-based dessert 6 100 40.3 25.7 23.1 91.2 1679.3 1001.0 968.1 3687.1
Growing-up milk 9 100 15.2 10.9 3.5 30.1 929.3 827.4 213.9 2450.7
Soup, puree 11 100 12.9 16.8 2.5 60.9 701.8 599.4 309.4 2363.6
Fruit purée 4 100 7.6 3.2 4.6 10.8 1230.2 132.2 1060.9 1384.0
Vegetable-based ready-to-eat meal 20 100 11.6 4.6 3.4 21.4 885.6 409.6 578.2 2373.1
Meat/fish based ready-to-eat meal 45 100 41.5 128.4 2.1 852.7 2149.6 5821.5 197.8 38383.3
Infant formula 28 100 4.5 3.8 1.8 16.6 200.7 193.8 68.5 951.7
Follow-on formula 33 100 8.9 6.8 2.6 31.7 464.1 443.5 104.5 2578.9

N: number of composite samples analyzed.

LB: Lower bound hypothesis; UB: Upper bound hypothesis. For the 6 ndl-PCBs, concentrations were equal under the LB and UB hypothesis.

3. Results and discussion (corresponding to 2150 ng kg−1 and 1679 ng kg−1 on fat basis,
Table 1) were measured. For dl-PCB, concentrations were respectively
3.1. Food contamination and exposure of 0.004–0.005 pg TEQWHO05.g−1 fresh weight according to LB-UB
hypothesis for “meat and fish based ready-to-eat meals” and 0.008 pg
Most analyses were performed on infant foods, as data on common TEQWHO05.g−1 fresh weight for “milk-based dessert”. For these sub-
foods were available elsewhere (Sirot et al., 2012). The percentages of stances, concentrations in infant foods were generally low compared
detection and concentration data are shown in Table 1. The highest with common foods. Details on concentrations by congeners for each
mean concentrations of PCDD/Fs were measured in the food category food items analyzed in this study are available online (https://www.
“infant milk-based desserts” (0.4 pg TEQWHO05g−1 fresh weight under data.gouv.fr/fr/datasets/donnees-etude-de-lalimentation-totale-
LB-UB). However, the most contaminated infant food item was a pool of infantile/).
jars of salmon cooked with sorrel with a concentration of 0.025 pg Based on these data and the ones measured in the previous French
TEQWHO05.g−1 fresh weight, followed by two samples of infant biscuits TDS (Sirot et al., 2012), exposure of children was estimated for different
(0.062–0.010 and 0.041–0.010 pg TEQWHO05.g−1 fresh weight ac- age groups according to food diversification (Table 2). For PCDD/Fs as
cording to LB-UB hypothesis, Table 1). For both dl and ndl-PCBs, the well as PCBs, exposure levels were higher for 13–36 months than for
highest levels in infant foods were found in the food category “meat or 1–4 months. Mean PCDD/F exposures ranged between 0.22 pg
fish based ready-to-eat meals” and “milk-based desserts”. For ndl-PCB TEQWHO05.kg bw−1.d−1in 1–4 months and 0.38 pg TEQWHO05.kg
concentrations of 41.5 and 40.3 ng kg−1 fresh weight, respectively, bw−1.d−1 in 13–36 months under LB hypothesis. Under UB, mean

4
M. Hulin, et al. Food and Chemical Toxicology 139 (2020) 111292

Table 2
Exposure levels (LB-UBa) to PCDD/Fs, dl-PCBs + PCDD/Fs and 6 ndl-PCBs of children under 3 years of age and percentage of exceedance of the health-based
guidance value.
Mean ± SD P90 exposure % children exceeding the health-based guidance value

LB UB LB UB LB UB

PCDD/Fs (pg TEQwho05.kg bw−1.d−1) Reference dose = 0.7 pg TEQWHO05.kg bw−1.d−1 (US-EPA, 2012)

1–4 months 0.218 ± 0.117 0.318 ± 0.131 0.400 0.488 NC* NC*
5–6 months 0.239 ± 0.066 0.292 ± 0.071 0.410 0.454 0 NC*
7–12 months 0.288 ± 0.121 0.347 ± 0.136 0.491 0.574 4.5 [0.7; 8.2] 4.5 [0.7; 8.2]
13–36 months 0.383 ± 0.239 0.444 ± 0.254 0.579 0.651 5.1 [3.1; 7.0] 7.4 [5.0; 9.7]
dl-PCBs + PCDD/Fs (pg TEQ WHOwho05.kg bw−1.d−1) Reference dose = 0.3 pg WHO05.kg bw−1.d−1 (EFSA, 2018)
1–4 months 0.366 ± 0.356 0.484 ± 0.378 0.600 0.69 48.9 [37.9; 60.0] 82.9 [74.6; 91.3]
5–6 months 0.458 ± 0.325 0.524 ± 0.336 1.00 1.04 64.9 [50.0; 79.7] 82.7 [70.9; 94.4]
7–12 months 0.675 ± 0.477 0.747 ± 0.499 1.18 1.32 84.0 [77.3; 90.6] 92.6 [87.8; 97.3]
13–36 months 0.989 ± 0.529 1.067 ± 0.540 1.52 1.61 99.2 [98.4; 100.0] 99.6 [99.1; 100.0]
−1 −1 −1 −1
6 ndl-PCBs (ng.kg bw .d ) TDI = 10 ng kg bw .d (Afssa, 2007)
1–4 months 0.874 ± 0.861 1.55 –
5–6 months 1.40 ± 0.550 2.77 –
7–12 months 2.28 ± 1.99 3.74 NC*
13–36 months 3.53 ± 4.13 5.45 2.7 [1.2; 4.2]

Mean ± SD correspond to the mean of exposure and associated standard deviation. [] correspond to the 95% confidence interval.
LB: Lower bound hypothesis; UB: Upper bound hypothesis * NC: not calculated as less than 5 children exceeded the HBGV.
a
LB = Lower bound – UB = Upper bound.

Fig. 1. Contributions under LB hypothesis (expressed as %)2 of the main food groups to the mean exposure (total population) and 90th percentile (highly exposed
population) exposure for PCDD/Fs (3a), dl-PCBs + PCDD/Fs (3b) and 6 ndl-PCBs (3c)
2
In this table a “.” signify that the contribution of the food groups for the related age groups has not been considered as any product consumed has been analyzed
(low consumption rate and not a major contributors).
4a. Contribution to PCDD/F exposure
4b. Contribution to dl-PCB + PCDD/F exposure
4c. Contribution to 6 ndl-PCB exposure
Only food groups contributing to more than 10% are represented individually.

5
M. Hulin, et al. Food and Chemical Toxicology 139 (2020) 111292

Table 3
Total exposure levels (LB) to PCDD/Fs and 6 ndl-PCBs of children under 1 year according to the type of milk consumed, common cow milk (C), infant formulae (IA)
and the regimen.
1–4 months 5–6 months 7–12 months

Type of milk consumed C IA C C + IA IA C C + IA IA

N (%) 3 (2.4%) 121 (97.6%) 11 (8.7%) 9 (7.1%) 107 (84.2%) 107 (84.2%) 17 (8.7%) 156 (80%)
Exposure to PCDD/Fs (pgTEQ WHO05.kg bw−1.d−1) 0.74 0.20 0.43a 0.29b 0.21c 0.46a 0.31b 0.25c
Exposure to PCDD/Fs + dl-PCBs (pgTEQ WHO05.kg bw−1.d−1) 1.76 0.33 1.09a 0.62b 0.37c 1.19a 0.73a 0.57a
Exposure to 6 ndl-PCBs (ng.kg bw−1.d−1) 6.16 0.72 3.45a 1.77b 1.14c, 3.87a 2.37a 1.98b

a,b,c: on a same line between the different types of consumers, and separately for the age groups, the values with a different index letter are significantly different
p < 0.05 (Student test). Due to small sample size, statistical comparison has not been performed for 1–4 months.

exposure ranged from 0.29 to 0.44 pg TEQWHO05.kg bw−1.d−1. When means that 96% of foods consumed by children and known to con-
considering PCDD/Fs and dl-PCBs, mean exposure of 13–36 months tribute to PCDD/Fs and PCBs exposure have been taken into account in
reached 1.07 pg TEQWHO05.kg bw−1.d−1 in UB hypothesis and 1.61 pg the exposure estimation. This estimation is therefore a good re-
TEQWHO05.kg bw−1.d−1 for P90. Mean exposure to ndl-PCBs (UB = LB) presentation of young children dietary exposure. However, this study
ranged from 0.87 ng kg bw−1.d−1 in the 1–4 months old children to did not take into account breastfed children and breastmilk is known to
3.53 ng kg bw−1.d−1 in the 13–36 months old children. P90 ranged be more contaminated that infant formulae. Data on European coun-
from 1.55 ng kg bw−1.d−1 to 3.53 ng kg bw−1.d−1. Starting at the age tries as analyzed in the frame of WHO coordinated studies showed
of 7 months, milk, ultra-fresh dairy products and cheese contributed to concentration of PCDD/F between 2.40 and 15.9 pg TEQWHO05.g−1 on
the exposure of PCCD/Fs and PCBs between 30 and 60% (Fig. 1). High fat basis (EFSA, 2018) whereas we found concentration of 0.059 and
differences were indeed observed in exposure according to the type of 0.0102 pg TEQWHO05.g−1 for respectively infant and follow-up for-
milk consumed by children. Among our study population, three chil- mulae. Therefore, it would be important to have a global view on ex-
dren (2 among the 1–4 months and 1 among the 13–36 months) ap- posure of young children to take into account PCDD/Fs and PCBs
peared to be highly exposed to the PCDD/Fs (between 1.03 and 1.40 pg burden in breastmilk and to consider the risk/benefit of breastfeeding.
TEQWHO05.kg bw−1.d−1 under the UB hypothesis). These children have However, data on consumption and contamination of human milk is
consumed during the three days of the dietary survey a high quantity of indeed very difficult to obtain. In France, the breastfeeding rate is quite
common milk (cow milk, 700 to 1 068 g d−1) as well as cream for the low compared to other OECD countries (OECD, 2012): only two out of
oldest child. Children consuming exclusively common cow milk were three newborns are breastfed immediately after birth and only 42% of
indeed exposed at least 2 fold more than the ones consuming infant these infants are still breastfed at the end of the 4th month (Bonet et al.,
formulae (p < 0.03 to p < 0.001, Table 3). For older children, fish 2013). However, it is now recognised that some contaminants, espe-
appeared to be a main contributor of PCB in highest exposed children cially persistent organic pollutants, are found in high concentrations in
between 13 and 36 months. Details on food groups’ contribution for human milk and lead to a potential risk for newborns (EFSA, 2011;
each age groups are available in supplementary materials (Table S1). Ulaszewska et al., 2011). Therefore, collecting data on breast milk
Our results on PCDD/Fs and dl-PCBs were in the same range as contamination in France would be valuable for estimating exposure in
those observed in previous studies (Pandelova et al., 2011; EFSA, 2012; this particular population.
De Filippis et al., 2014). In an European study (Pandelova et al., 2011),
the authors estimated that non breast-fed children under 9 months were 3.2. Risk assessment
exposed at levels between 0.14 and 2.79 pg TEQWHO05.kg bw−1.d−1
depending on the age and the type of infant formula they consumed Based on exposure assessment, the percentages of children ex-
(milk, soy or hypoallergenic). However, for the 6 ndl-PCBs, con- ceeding the health-based guidance value (HBGV) were estimated
centrations of infant and baby foods with fish or meat were higher than (Table 2). In children over seven months of age, the TDI proposed by
ours (10 900 ng.kg−1on fat basis for the mean and 17 300 ng.kg−1for the US EPA in 2012 for PCDD/Fs (i.e. 0.7 pg kg bw−1.d−1) was ob-
the p99) as well as exposure (between 8.5 and 25.7 ng kg bw−1.d−1 served to be significantly exceeded by 4.5% of children from 7 to 12
under UB and up to 53.5 ng kg bw−1.d−1for the most exposed children) months, and by 5.1–7.4% in the 13–36 months old children, depending
(EFSA, 2012). This can be explained either by a difference of con- on the hypothesis of management of the censored data (LB-UB). In
tamination of fish and meat or of the quantity of these ingredients in children under six months of age, it was not possible to precisely esti-
food composition. Concerning infant formulae, our results were in the mate the percentage of exceedance, as less than 5 children exceeded the
same order of magnitude as some found in previous studies for both HBGV, and they were considered as non-representative cases. The cases
contaminant families (Loran et al., 2009; Pandelova et al., 2011). Due of exceedance of HBGV were mainly explained by higher consumption
to a decreasing food consumption in proportion to body weight with of common milk (or cow milk, the single main contributor in children
ages, we expected to have higher exposure levels in young children under six months), fish and ultra-fresh dairy products, than in con-
compared to older ones. However, exposure in 1–4 months children was sumers as a whole. When comparing our results on PCDD/F + dl-PCBs
lower than 13–16 months ones. Moreover, 13–36 months children were with the recent TWI offset by EFSA (EFSA, 2018), between 48.9% and
exposed in the same range of magnitude than 3–6 years old children 99.6% of children exceed the HBGV according to age groups and the
from the second French TDS for PCDD/Fs, but lower for the 6 ndl-PCBs censorship hypothesis.
(Sirot et al., 2012). This can be explained by the gradual introduction of For the 6 ndl-PCBs, the cases in which the TDI selected (10 ng kg
foods, in conjunction with diversification, that are more contaminated bw−1.d−1) was exceeded were significant from the age of one year
than infant formulae. Cow milk consumption instead of infant formulae (2.7% CI95% [1.2; 4.1]). The limits were not exceeded for children
lead also to higher exposure in young children. These observations re- under six months of age. Between seven months and one year, once
inforce the importance of following recommendations on food diversi- again it was not possible to estimate precisely the extent to which the
fication and not consuming common cow milk before the age of limits were exceeded, given the limited number of children concerned
1 year as already recommended based on nutritional needs. . (n < 5). In children exceeding the heath-based guidance value, 82% of
The data used for exposure assessment covered 68% of the total diet the exposure was from fish consumption. They had all consumed high
and 96% of the diet theoretically contributing to the exposure. That quantity of salmon in the 3 day of the consumption survey: 21 g d−1 for

6
M. Hulin, et al. Food and Chemical Toxicology 139 (2020) 111292

those children vs. 12 g d−1 for all children. life exposure by combining food consumption and data contamination
Results on risk assessment for PCDD/Fs or PCBs in young children at different life steps as shown in the case of cadmium by Pruvost-
differ according to studies (Loran et al., 2009; Pandelova et al., 2011; Couvreur et al., (2020). This evaluation has been conducted separately
Karjalainen et al., 2012; De Filippis et al., 2014; Morales-Suarez-Varela for PCB and PCDD/F exposure levels, in view of risk management since
et al., 2018). The differences can be explained by methodological sources of pollution for these two groups of substances are different.
choices. In our study for example, we were not able to cover all specific Even if such approach appears appropriate from a health point of view,
consumptions or populations. There may therefore be various un- the method for risk characterization of cumulated exposure can be
certainties regarding the representativeness of the subjects and their discussed: (1) other components having “dioxin-like” mechanism could
consumption regarding the population under consideration, i.e. infants also be considered such as certain brominated flame retardants and (2)
and young children living in metropolitan France. Moreover, the use of some substances present concomitantly with dioxins and furans can
a food diary has already been associated with reporting biases that may show antagonistic effects that are not taken into account (De Waard
lead to (Lioret et al., 2011; Berta Vanrullen et al., 2014). Thus, certain et al., 2008). To better assess the risk associated to “dioxin-like” effects
food intakes are likely to be unrecorded, particularly food ingested as a whole, it would be necessary to conduct toxicity studies on mixture
between the main meals within a day. This uncertainty may lead to an effects.
underestimation of the risk. Finally, the consumption study used to
estimate chronic intakes and exposures is based on a collection of in- 4. Conclusion and recommendations
takes over 3 days. This short period leads to a high variability compared
to a longer collection period and may lead to uncertainty in the esti- Dietary exposure to PCDD/Fs and PCBs was identified as a concern
mates in relation to the exposure period considered (from a few months for children above 6 months. Exposure, mainly via common food pro-
to 2 years depending on the age groups considered). Nevertheless, EFSA ducts making a major contribution to exposure of the most exposed
recommends 2 non-consecutive 24th-recalls for food consumption data children to these compounds should therefore be reduced, i.e. milk,
collection (EFSA, 2014). Even if it was consecutive day, in our study we ultra-fresh dairy products and fish. These data allowed to complete the
had recalls on 3 days. It has to be known that there are statistical exposure assessment of French population focusing on a sensitive po-
methods to reduce this uncertainty, known as intra-individual variance pulation and strengthen the recommendations on diversification and
reduction methods, but these have not been used in the present work. fish consumption.
This choice is protective because it will tend to overestimate high ex- Through their diet, individuals are exposed to a multitude of con-
posures (the mean will be correctly estimated), and thus overestimate taminants from diverse chemical classes. Phenomena of competition,
the possible risk (Mancini et al., 2015). additivity or synergy can occur between several substances, which may
As already explained, the TDS approach in the present study en- lead to the organism responding in an unexpected way with regard to
sured to have a good representation of the whole diet. Nonetheless, the known toxicological effects for each substance. This question of
milk and milk products as well as fish were the main contributors in all mixtures cannot be limited solely to a mixture of compounds belonging
studies. Concerning fish, the contamination data in our study only fo- to the same chemical class. It is therefore necessary to identify mixtures
cused on a limited number of species, but the literature showed that (or "cocktails") of substances that are relevant in health terms and
other species can contain PCB concentrations equal to or higher than realistic from the point of view of population exposure. Data collected
salmon (Marchand et al., 2006). In their study, Morales-Suarez-Varela in this study are a real opportunity for identifying the cocktails to which
and al. showed that even if fish is under the regulatory levels in most children are actually exposed, and could therefore provide input for
samples, daily intake in consumers of large quantities of fish were ex- studies of the potential associations between these cocktails of sub-
ceeding toxicological reference values for ndl-PCBs (Morales-Suarez- stances and the health effects.
Varela et al., 2018). Therefore, even though PCB levels in fish have
been falling over the past few years (Anses, 2015), recommendations on CRediT authorship contribution statement
fish consumption have to be reminded. Moreover, according to our
results and previous ones, a debate should be conducted on the max- Marion Hulin: Conceptualization, Data curation, Formal analysis,
imum values laid down by the regulations and the choice of matrices to Investigation, Writing - original draft. Véronique Sirot:
be regulated for both substance families. Special attention should be Conceptualization, Data curation, Formal analysis, Investigation,
given to the youngest age groups due to their special vulnerability and
Writing - original draft. Paule Vasseur: Formal analysis, Writing - re-
higher exposure. view & editing. Aurelie Mahe: Formal analysis. Jean-Charles Leblanc:
In terms of health risks, it has also to be reminded that exposures
Conceptualization, Funding acquisition, Writing - review & editing.
estimated in this study correspond to exposures over a given period of Julien Jean: Formal analysis, Writing - review & editing. Philippe
time. On one hand, persistent organic pollutants (POPs), as PCBs and
Marchand: Formal analysis. Anaïs Venisseau: Formal analysis. Bruno
PCDD/Fs can accumulate in the body: thus the dose to which an in- Le Bizec: Formal analysis, Writing - review & editing. Gilles Rivière:
dividual is actually exposed at a given time corresponds not only to the
Formal analysis, Writing - original draft.
ingested dose but also to the dose present in the body as a result of
previous exposure. However, for infants, the body burden due to ex-
Declaration of competing interest
posure during the first months of life, in particular through breast-
feeding, is limited in time. However, exposure during fetal life, via
maternal feeding, can lead to a significant body burden. On the other The authors declare that they have no known competing financial
hand in the case of substances for which infants are not more sensitive interests or personal relationships that could have appeared to influ-
than adults and where effects are associated with a long period of ex- ence the work reported in this paper.
posure, exceedances observed over a limited time period do not ne-
cessarily lead to a health risk. In the second French TDS, exceedance Acknowledgments
was also observed in children and young adults (Sirot et al., 2012).
Moreover, based on an integrative risk assessment approach on French The authors are grateful to the Anses expert committee panel in
contamination data, Béchaux et al. predicted the body burden of the charge of assessing the chemical risk in food and the experts from the
French population in 2030, assuming that the dietary exposure will infant TDS scientific committee and the analytic methods in food,
remain stable at the 2009 level (Bechaux et al., 2014). Therefore, in namely: Jean-Pierre Cravédi, Claude Atgié, Pierre-Marie Badot, Jacques
order to conclude on health risk, it will be of interest to estimate long- Bélégaud, Catherine Bennetau-Pelissero, Emmanuelle Bichon, Valérie

7
M. Hulin, et al. Food and Chemical Toxicology 139 (2020) 111292

Camel, Martine Claw, Christophe Cordella, Guillaume Duflos, Camille EFSA, FAO, WHO, 2011. Joint Guidance of EFSA, FAO, WHO. Towards a Harmonised
Dumat, Cyril Feidt, Jean-Marc Fremy, Jérôme Gay-Queheillard, Total Diet Study Approach: a Guidance Document. The EFSA Journal, vol. 9 EFSA,
FAO, WHO, Parma, Rome, Geneva.
Philippe Glorennec, Konrad Grob, Thierry Guérin, Laurence Guldner, Fantino, M., 2005. [Etude SFAE sur la consommation alimentaire des nourrissons et en-
Nicole Hagen–Picard, Dary Inthavong, Florence Lacoste, Laïla Lakhal, fants en bas âge français de 1 mois à 36 mois - Analyse des données nutritionnelles
Béatrice Lalère, Claude Lambré, Michel, Laurentie, Raphaëlle Le Garrec, (non publié)].
Fantino, M., Gourmet, E., 2008. Apports nutritionnels en France en 2005 chez les enfants
Catherine Leclercq, Eric Marchioni, César Mattéi, André Mazur, Sakina non allaités âgés de moins de 36 mois. Arch. Pediatr. 15 (4), 446–455.
Mhaouty-Kodja, Fabrice Nesslany, Laurent Noel, Alain-Claude Roudot, GEMS/Food-EURO, 2013. ADDENDUM 2013 - Second Workshop on Reliable Evaluation
Patrick Sauvegrain, Rémy Slama, Karine Tack, Eric Verdon and Jean- of Low-Level Contamination of Food. Workshop in the Frame of GEMS/Food-EURO.
Technical Report. WHO, Kulmbach, Germany.
Paul Vernoux. Hulin, M., Bemrah, N., Nougadère, A., Volatier, J.L., Sirot, V., Leblanc, J.C., 2014.
The infant TDS was supported by the Ministry for food, agriculture Assessment of infant exposure to food chemicals: the French Total Diet Study design.
and fisheries, the Ministry for health the Ministry for ecology and Food Addit. Contam. Part A Chem Anal Control Expo Risk Assess 31 (7), 1226–1239.
Karjalainen, A.K., Hirvonen, T., Kiviranta, H., Sinkko, H., Kronberg-Kippila, C., Virtanen,
sustainable development and the French Agency for Food,
S.M., Hallikainen, A., Leino, O., Knip, M., Veijola, R., Simell, O., Tuomisto, J.T., 2012.
Environmental and Occupational Health & Safety (ANSES). Long-term daily intake estimates of polychlorinated dibenzo-p-dioxins and furans,
polychlorinated biphenyls and polybrominated diphenylethers from food in Finnish
Appendix A. Supplementary data children: risk assessment implications. Food Addit. Contam. Part A Chem Anal
Control Expo Risk Assess 29 (9), 1475–1488.
Klocke, C., Lein, P.J., 2020. Evidence implicating non-dioxin-like congeners as the key
Supplementary data to this article can be found online at https:// mediators of polychlorinated biphenyl (PCB) developmental neurotoxicity. Int. J.
doi.org/10.1016/j.fct.2020.111292. Mol. Sci. 21 (3).
Landrigan, P.J., Kimmel, C.A., Correa, A., Eskenazi, B., 2003. Children's health and the
envrionment: public health issues and challenges for risk assessment. Environ. Health
References Perspect. 12 (2), 257–265.
Lioret, S., Touvier, M., Balin, M., Huybrechts, I., Dubuisson, C., Dufour, A., Bertin, M.,
Maire, B., Lafay, L., 2011. Characteristics of energy under-reporting in children and
Afssa, 2005. Rapport de l'Afssa relatif aux dioxines, furanes et PCB de type dioxine :
adolescents. Br. J. Nutr. 105 (11), 1671–1680.
Evaluation de l'exposition de la population française. Maisons-Alfort, France, Afssa,
Loran, S., Conchello, P., Bayarri, S., Herrera, A., 2009. Evaluation of daily intake of
pp. 57.
PCDD/Fs and indicator PCBs in formula-fed Spanish children. Food Addit. Contam.
Afssa, 2007. Avis de l'Afssa relatif à l'établissement de teneurs maximales pertinentes en
Part A Chem Anal Control Expo Risk Assess 26 (10), 1421–1431.
polychlorobiphényles qui ne sont pas de type dioxine (PCB "non dioxin-like", PCB-
Makri, A., Goveia, M., Balbus, J., Parkin, R., 2004. Children's susceptibility to chemicals: a
NDL) dans divers aliments. (saisine n°2006-SA-0305). Maisons-Alfort, Afssa, pp. 28.
review by developmental stage. J. Toxicol. Environ. Health B Crit. Rev. 7 (6),
Anses, 2015. Avis de l'Anses et rapport d'expertise collective relatifs à la consommation de
417–435.
poissons d'eau douce et PCB : aspects réglementaires, méthodologiques et sanitaires.
Mancini, F.R., Sirot, V., Busani, L., Volatier, J.L., Hulin, M., 2015. Use and impact of usual
(saisines n°2014-SA-0122 et 2011-SA-0039). Maisons-Alfort, Anses, pp. 107.
intake models on dietary exposure estimate and risk assessment of chemical sub-
Antignac, J.P., Marchand, P., Gade, C., Matayron, G., Qannari el, M., Le Bizec, B., Andre,
stances: a practical example for cadmium, acrylamide and sulphites. Food Addit.
F., 2006. Studying variations in the PCDD/PCDF profile across various food products
Contam. Part A Chem Anal Control Expo Risk Assess 32 (7), 1065–1074.
using multivariate statistical analysis. Anal. Bioanal. Chem. 384 (1), 271–279.
Marchand, P., Antignac, J.P., Brosseau, A., Gade, C., Venisseau, A., Sabatie, M.-R., Sirot,
Aylward, L.L., Hays, S.M., 2002. Temporal trends in human TCDD body burden: decreases
V., Tard, A., Volatier, J.L., Leblanc, J.C., André, F., Le Bizec, B., 2006. Factors
over three decades and implications for exposure levels. J. Expo. Anal. Environ.
(trophic levels, fish specie, habitat, fat content…) influencing PCDD/F, PCB and
Epidemiol. 12 (5), 319–328.
PBDE concentration in fish retailed in France. Organohalogen Compd. 68, 608–611.
Bechaux, C., Zeilmaker, M., Merlo, M., Bokkers, B., Crepet, A., 2014. An integrative risk
Morales-Suarez-Varela, M., Lopez Santana, N., Marti Requena, P., Beser Santos, M.I.,
assessment approach for persistent chemicals: a case study on dioxins, furans and
Peraita-Costa, I., Llopis-Gonzalez, A., 2018. Estimation of daily intake of poly-
dioxin-like PCBs in France. Regul. Toxicol. Pharmacol. 70 (1), 261–269.
chlorinated biphenyls not similar to dioxins (NDL-PCB) from fish consumption in
Berta Vanrullen, I., Volatier, J.L., Bertaut, A., Dufour, A., Dallongeville, J., 2014.
Spain in different population groups. Publ. Health Nutr. 21 (16), 2959–2968.
Characteristics of energy intake under-reporting in French adults. Br. J. Nutr. 111 (7),
OECD, 2012. OECD Family Database. C01.5 : Breastfeeding Rates. OECD, Paris.
1292–1302.
Pandelova, M., Piccinelli, R., Lopez, W.L., Henkelmann, B., Molina-Molina, J.M.,
Bonet, M., Marchand, L., Kaminski, M., Fohran, A., Betoko, A., Charles, M.A., Blondel, B.,
Arrebola, J.P., Olea, N., Leclercq, C., Schramm, K.W., 2011. Assessment of PCDD/F,
2013. Breastfeeding duration, social and occupational characteristics of mothers in
PCB, OCP and BPA dietary exposure of non-breast-fed European infants. Food Addit.
the French 'EDEN mother-child' cohort. Matern. Child Health J. 17 (4), 714–722.
Contam. Part A Chem Anal Control Expo Risk Assess 28 (8), 1110–1122.
Costera, A., Feidt, C., Marchand, P., Le Bizec, B., Rychen, G., 2006. PCDD/F and PCB
Pruvost-Couvreur, M., Le Bizec, B., Bechaux, C., Riviere, G., 2020. A method to assess
transfer to milk in goats exposed to a long-term intake of contaminated hay.
lifetime dietary risk: example of cadmium exposure. Food Chem. Toxicol. 137,
Chemosphere 64 (4), 650–657.
111130.
De Filippis, S.P., Brambilla, G., Dellatte, E., Corrado, F., Esposito, M., 2014. Exposure to
Sagiv, S.K., Thurston, S.W., Bellinger, D.C., Tolbert, P.E., Altshul, L.M., Korrick, S.A.,
polychlorinated dibenzo-p-dioxins (PCDDs), polychlorinated dibenzofurans (PCDFs),
2010. Prenatal organochlorine exposure and behaviors associated with attention
dioxin-like polychlorinated biphenyls (DL-PCBs) and polybrominated diphenyl ethers
deficit hyperactivity disorder in school-aged children. Am. J. Epidemiol. 171 (5),
(PBDEs) through the consumption of prepared meals in Italy. Food Addit. Contam.
593–601.
Part A Chem Anal Control Expo Risk Assess 31 (6), 1114–1126.
Sirot, V., Tard, A., Venisseau, A., Brosseaud, A., Marchand, P., Le Bizec, B., Leblanc, J.C.,
De Waard, W.J., Aarts, J., Peijnenburg, A., De Kok, T.M., Van Schooten, F.-J.,
2012. Dietary exposure to polychlorinated dibenzo-p-dioxins, polychlorinated di-
Hoogenboom, L., 2008. Ah receptor agonist activity in frequently consumed food
benzofurans and polychlorinated biphenyls of the French population: results of the
items. Food Addit. Contam. 25 (6), 779–787.
second French Total Diet Study. Chemosphere 88 (4), 492–500.
Diamanti Kandarakis, E., Bourguignon, J.P., Guidice, L.C., Huauser, R., Prins, G.S., Soto,
Sly, P.D., Flack, F., 2008. Susceptibility of children to environmental pollutants. Ann. N.
A.M., Zoeller, T., Gore, A.C., 2009. Endocrine-disrupting chemicals: an endocrine
Y. Acad. Sci. 1140, 163–183.
society scientific statement. Endocr. Rev. 30 (4), 293–342.
Ulaszewska, M.M., Zuccato, E., Davoli, E., 2011. PCDD/Fs and dioxin-like PCBs in human
EFSA, 2005. Opinion of the Scientific Panel on Contaminants in the Food Chain on a
milk and estimation of infants' daily intake: a review. Chemosphere 83 (6), 774–782.
Request from the Commission Related to the Presence of Non Dioxin-like
US-EPA, 2012. Rome. Italy. EPA's Reanalysis of Key Issues Related to Dioxin Toxicity and
Polychlorinated Bisphenyls (PCB) in Feed and Food. The EFSA Journal, vol. 284
Response to NAS Comments, vol. 1 EPA/600/R-10/038F.
EFSA, Parma.
Van den Berg, M., Birnbaum, L.S., Denison, M., De Vito, M., Farland, W., Feeley, M.,
EFSA, 2010. Results of the Monitoring of Dioxin Levels in Food and Feed. EFSA Scientific
Fiedler, H., Hakansson, H., Hanberg, A., Haws, L., Rose, M., Safe, S., Schrenk, D.,
Report. vol. 8 EFSA, Parma.
Tohyama, C., Tritscher, A., Tuomisto, J., Tysklind, M., Walker, N., Peterson, R.E.,
EFSA, 2011. Scientific Opinion on Polybrominated Diphenyl Ethers (PBDEs) in Food.
2006. The 2005 World Health Organization reevaluation of human and Mammalian
EFSA Panel on Contaminants in the Food Chain (CONTAM). Parma, Italy, EFSA.
toxic equivalency factors for dioxins and dioxin-like compounds. Toxicol. Sci. 93 (2),
EFSA, 2012. Scientific Opinion of the EFSA panel on Contaminants in the Food Chain on
223–241.
the presence of dioxins (PCDD/Fs) and dioxin-like PCBs (DL-PCBs) in commercially
Zoeller, R.T., Dowling, A.L., Vas, A.A., 2000. Developmental exposure to polychlorinated
available foods for infants and young children. EFSA J. 10 Parma, EFSA.
biphenyls exerts thyroid hormone-like effects on the expression of RC3/neurogranin
EFSA, 2014. Guidance on the EU menu methodology. EFSA J. 12, 3944 Parma, EFSA.
and myelin basic protein messenger ribonucleic acids in the developing rat brain.
EFSA, 2018. Scientific Opinion of the EFSA Panel on Risk for Animal and Human Health
Endocrinology 141 (1), 181–189.
Related to the Presence of Dioxins and Dioxin-like PCBs in Feed and Food. The EFSA
Journal, vol. 16 EFSA, Parma.