Science of the Total Environment 484 (2014) 121–128

Contents lists available at ScienceDirect

Science of the Total Environment

j ourna l homepage: www.e lsev ie r .com/ locate /sc i totenv

Maternal diet, prenatal exposure to dioxin-like compounds and birthoutcomes in a European prospective mother–child study (NewGeneris)

Eleni Papadopoulou a,b,c,⁎, Manolis Kogevinas b,c,d,e,f, Maria Botsivali g, Marie Pedersen b,c,e,h, Harrie Besselink i,Michelle A.Mendez c, Sarah Fleming j, Laura J. Hardie j, Lisbeth E. Knudsen k, JohnWright l, Silvia Agramunt c,d,m,Jordi Sunyer b,c,e, Berit Granum a, Kristine B. Gutzkow a, Gunnar Brunborg a, Jan Alexander a,Helle Margrete Meltzer a, Anne Lise Brantsæter a, Katerina Sarri n, Leda Chatzi n, Domenico F. Merlo o,Jos C. Kleinjans p, Margaretha Haugen a

a Division of Environmental Medicine, Norwegian Institute of Public Health, Lovisenberggata 8, 0456 Oslo, Norwayb Pompeu Fabra University, Plaça de la Mercè 10-12, 08002 Barcelona, Spainc Centre for Research in Environmental Epidemiology (CREAL), Doctor Aiguader 88, 08003 Barcelona, Spaind IMIM (Hospital del Mar Research Institute), Doctor Aiguader 88, 08003 Barcelona, Spaine CIBER Epidemiología y Salud Pública (CIBERESP), Instituto de Salud Carlos III, Melchor Fernández Almagro 3-5, 28029 Madrid, Spainf National School of Public Health, Alexandras Avenue 196, 11521 Athens, Greeceg National Hellenic Research Foundation, Institute of Biological Research and Biotechnology, Vassileos Constantinou Avenue 48, 11635 Athens, Greeceh INSERM (National Institute of Health Medical Research), Team of Environmental Epidemiology Applied to Reproduction and Respiratory Health, Institute Albert Bonniot, BP 170, La Tronche,F-38042 Grenoble Cedex 9, Francei BioDetection Systems B.V., Science Park 406, 1098 XH Amsterdam, The Netherlandsj Centre for Epidemiology and Biostatistics, Leeds Institute of Genetics, Health and Therapeutics, University of Leeds, Woodhouse Lane, Leeds LS2 9JT, UKk Section of Environmental Health, Department of Public Health, University of Copenhagen, Øster Farimagsgade 5, 1014 Copenhagen, Denmarkl Bradford Institute for Health Research, Temple Bank House, Bradford Royal Infirmary, Duckworth Lane, Bradford BD9 6RJ, UKm Parc de Salut Mar, Obstetrics and Gynaecology Department, Passeig Marítim 25-29, 08003 Barcelona, Spainn Department of Social Medicine, Medical School, University of Crete, Voutes Campus, Heraklion, Crete GR-71003, Greeceo Epidemiology, Biostatistics, and Clinical Trials, National Cancer Research Institute, Largo Rosanna Benzi 10, 16132 Genoa, Italyp Department of Toxicogenomics, Maastricht University, Universiteitssingel 40, 6229ER Maastricht, The Netherlands

H I G H L I G H T S

• Dioxin-diet score was defined using RRR in a five country population.• Dioxin-diet score was positively correlated to dioxin-like activity in maternal blood.• Women following a diet high in meat and fish had a higher dioxin-diet score.• High dioxin-diet score was associated with a 121 g reduction in birth weight.• Maternal diet can contribute to foetal exposure to dioxins and dioxin-like compounds.

⁎ Corresponding author at: Department of Exposure anOslo, Norway. Tel.: +47 21076511.

E-mail address: Eleni.Papadopoulou@fhi.no (E. Papado

http://dx.doi.org/10.1016/j.scitotenv.2014.03.0470048-9697/© 2014 Elsevier B.V. All rights reserved.

a b s t r a c t

a r t i c l e i n f o

Article history:Received 20 January 2014Received in revised form 7 March 2014Accepted 7 March 2014Available online 30 March 2014

Editor: Damia Barcelo

Maternal diet can result in exposure to environmental contaminants including dioxins whichmay influence foe-tal growth. We investigated the association betweenmaternal diet and birth outcomes by defining a dioxin-richdiet. We used validated food frequency questionnaires to assess the diet of pregnant women from Greece, Spain,United Kingdom, Denmark and Norway and estimated plasma dioxin-like activity by the Dioxin-ResponsiveChemically Activated LUciferase eXpression (DR-CALUX®) bioassay in 604 maternal blood samples collectedat delivery. We applied reduced rank regression to identify a dioxin-rich dietary pattern based on dioxin-likeactivity (DR-CALUX®) levels in maternal plasma, and calculated a dioxin-diet score as an estimate of adherence

d Risk Assessment, Division of Environmental Medicine, Norwegian Institute of Public Health, PO Box 4404 Nydalen, NO-0403

poulou).

122 E. Papadopoulou et al. / Science of the Total Environment 484 (2014) 121–128

Keywords:Dietary patternsPregnancyDioxinsBirth weightCohort studyDR-CALUX

to this dietary pattern. In thefive country population, dioxin-diet scorewas characterised byhigh consumption ofred and white meat, lean and fatty fish, low-fat dairy and low consumption of salty snacks and high-fat cheese,during pregnancy. The upper tertile of the dioxin-diet score was associated with a change in birth weight of−121 g (95% confidence intervals: −232, −10 g) compared to the lower tertile after adjustment for con-founders. A small non-significant reduction in gestational age was also observed (−1.4 days, 95% CI: −3.8,1.0 days). Our results suggest that maternal diet might contribute to the exposure of the foetus to dioxins anddioxin-like compounds andmay be related to reduced birthweight. More studies are needed to develop updateddietary guidelines for women of reproductive age, aiming to the reduction of dietary exposure to persistentorganic pollutants as dioxins and dioxin-like compounds.

© 2014 Elsevier B.V. All rights reserved.

1. Introduction

Polychlorinated dibenzo-p-dioxins (PCDDs), dibenzo-furans (PCDFs)and polychlorinated biphenyls (PCBs) are environmental contaminants,unintentionally or industrially produced. Seven PCDD congeners,10 PCDF congeners and 12 PCB congeners, the dioxin-like PCBs, sharesimilar chemical and toxicological properties and form the group of “di-oxins and dioxin-like compounds”. Dioxins and dioxin-like compoundsare lipophilic, stable against degradation and able to bio-accumulate,thus they are highly persistent and ubiquitous in the environment andthe food chain (EFSA, 2004; Kulkarni et al., 2008). Human exposure oc-curs mainly through food consumption (EFSA, 2012; Liem et al., 2000).Dioxins and dioxin-like compounds are introduced into food products ofanimal origin when terrestrial animals consume contaminated soil, feedand pasture (Rose et al., 2012). Additionally, aquatic sediments are con-sidered the biggest environmental sink of dioxins and dioxin-like com-pounds (Kulkarni et al., 2008). Persistent compounds bioaccumulatemore efficiently in the biota of the aquatic rather than the terrestrial envi-ronment, so seafood is the main source of dietary exposure (EFSA, 2012).

The dominating role of seafood as a contributor to the dietary intakeof dioxins and dioxin-like compounds has been confirmed by severalepidemiological studies (Fattore et al., 2006; Kiviranta et al., 2004;Sirot et al., 2012; Tornkvist et al., 2011), while others have reported asubstantial contribution of meat and dairy products (De Mul et al.,2008; Fernandez et al., 2004; Schecter et al., 2001). Variations in expo-sure between andwithin populations are often explained by differencesin the levels of food contamination or in dietary habits (Fernandez et al.,2004; Perello et al., 2012;Windal et al., 2010). Dietary habits of popula-tions can be described by dietary patterns. Dietary patterns can providean overall exposure estimate, considering multiple dietary sources ofexposure to environmental contaminants, but have been scarcely usedwhen investigating the association between food intake and bloodlevels of dioxins and dioxin-like compounds (Arisawa et al., 2011;Kvalem et al., 2009).

The diet of pregnantwomen has been related to levels of dioxins anddioxin-like compounds in maternal blood, cord blood and human milk(Halldorsson et al., 2008; Huisman et al., 1995; McGraw and Waller,2009; Wang et al., 2009). During pregnancy dioxins and dioxin-likecompounds can be transferred from the mother to the placenta andthe foetus and may lead to a variety of toxic effects (Tsukimori et al.,2013). The toxicity of dioxins and dioxin-like compounds is mainly me-diated through their binding to the aryl hydrocarbon receptor (AhR)(Arisawa et al., 2005; Denison et al., 2011). The Dioxin-ResponsiveChemically Activated LUciferase eXpression (DR-CALUX®) bioassaywas developed to provide an estimate of the binding affinity of dioxinsand dioxin-like PCBs to the AhR and is considered a valid and rapidscreening tool for human exposure (Brouwer et al., 2004; Laier et al.,2003; Van Wouwe et al., 2004). A previous study in the NewGenerisproject did not found associations between maternal DR-CALUX levelsand birth outcomes (Vafeiadi et al., 2014). There are few studiesassessing blood levels of dioxins and dioxin-like compounds, by theDR-CALUX bioassay, and diet in pregnant women (Halldorsson et al.,2009; Koppen et al., 2009; Pedersen et al., 2012).

The aim of our study was to identify a dietary pattern relatedto blood levels of dioxins and dioxin-like compounds of pregnantwomen from five European countries (NewGeneris project), andto estimate the association between the dietary pattern and birthoutcomes.

2. Materials and methods

2.1. Study population

This study is part of the European NewGeneris project (Newbornsand Genotoxic exposure risks) and includes five European mother–child studies (Merlo et al., 2009): The Rhea study is a population-based mother–child study at the prefecture of Heraklion, Crete, Greece(Chatzi et al., 2009). Pregnant women were recruited within a year,from February 2007. The INMA-NewGeneris sub-cohort includedwomen with singleton pregnancies, enrolled from May 2007 to March2010, in Sabadell and Barcelona, Spain (Guxens et al., 2012). TheMoBa-NewGeneris sub-cohort included pregnant women with full-term singleton pregnancies from Oslo and Akershus in Norway, partici-pants of the BraMat and BraMiljo studies, recruited from 2007 to 2008.These women were already enrolled in the Norwegian Mother andChild Cohort Study (MoBa) (Magnus et al., 2006). The Danish study in-cluded 2 enrolment periods, from December 2006 to December 2007and from September to December 2009, in Copenhagen, Denmark.Women with singleton pregnancies and planned Caesarean sectionswere recruited in the 1st period, while singleton pregnancies wererecruited during the 2nd period (Pedersen et al., 2009). The BiB-NewGeneris sub-cohort includes pregnant women participants of theBorn in Bradford (BiB) study with elective Caesarean section, recruitedfrom January 2008 to May 2009 in Bradford, UK (Hepworth et al.,2012; Wright et al., 2012).

Dietary information and plasma dioxin levels were available for 665women with singleton pregnancies. After excluding 61 women dueto implausible values of total energy intake (outside the range of1000–4000 kcal/day), 604 women were included in this analysis(Davey et al., 2003). Questionnaires and medical records were usedto collect information on several maternal lifestyle and socio-demographic characteristics, as well as birth outcomes. Gestationalage for participants from Denmark, Greece, and Spain was estimatedfrom the interval between lastmenstrual period anddate of the deliveryand was corrected by ultrasound measurement if there was a discor-dance of ≥7 days between estimates. Ultrasound-based estimationwas provided for most participants from England and Norway.Finally, 537 mothers had available information on gestational age andbirth weight, as well as on other socio-demographic and lifestylecharacteristics.

This study was conducted according to the guidelines laid down inthe Declaration of Helsinki and all procedures involving humans wereapproved by the research ethics committee in each country. Writteninformed consent for participation of the women and their childrenwas obtained from all participating women.

123E. Papadopoulou et al. / Science of the Total Environment 484 (2014) 121–128

2.2. Dioxins and dioxin-like compounds in maternal blood (DR-CALUX®bioassay)

Peripheral blood samples from the mothers were collected arounddelivery. Dioxin-like activity was determined through the DR-CALUX®bioassay at BioDetection Systems B.V., Amsterdam, The Netherlands.The protocol for sample processing has been presented elsewhere(Vafeiadi et al., 2014). In brief, the DR-CALUX® assay is based on agenetically modified H4IIE rat hepatoma cell line which contains thefirefly luciferase reporter gene under the transcriptional control ofAhR. Upon exposure of the cells to dioxins or dioxin-like chemicalslight, that is proportional to the strength of the AhR binding, is emitted.The luminance is calibrated with respect to 2,3,7,8-TCDD toxic equiva-lency quantities. Dioxin-like activity was estimated in 604 maternalblood samples and was expressed as pg CALUX®-TEQ per g of lipid.Samples below the level of detection (LOD) were assigned a value of0.5 × LOD.

2.3. Maternal dietary patterns

Information on maternal diet during pregnancy was collected usingvalidated cohort-specific food frequency questionnaires (FFQs)(Brantsæter et al., 2008; Chatzi et al., 2011; Hepworth et al., 2012;Pedersen et al., 2012; Vioque et al., 2007). All FFQs were semi-quantitative and included similar questions on food consumption, useof dietary supplements and dietary changes due to pregnancy. The re-ported frequency of consumption for each food item was multipliedwith the portion size and converted into daily food intake (g/day).Further, intake of nutrients was calculated using food compositiontables.

Food items which, according to the literature, are established orpotential sources of dietary dioxins and dioxin-like compounds wereintegrated in 13 food groups: red meat, white meat, processed meat,fatty-fish, lean fish, mixed fish dishes, high-fat milk/yoghurt, high-fatcheese, low-fat dairy, eggs, butter, salty snacks and fast-food (Table 1).Food items that might contribute to the total exposure to dioxins anddioxin-like compounds but were not common for all studies were addi-tionally included in the country-specific analysis. Intake of fish-oildietary supplements is common for Norwegian pregnant women andwas included in the analysis for the Norwegian study (Haugen et al.,2008). Each FFQ was designed to assess the overall maternal dietaryhabits in the period of pregnancy and the consumption of more food

Table 1Food items included in the food groups and available information for pregnant women in the N

Food groups Food items included in the food group

Red meat Beef, veal, pork, lamb, mutton, minced meat, burger, meatballsWhite meat Chicken, turkey, rabbitProcessed meat Cold turkey/chicken/ham cuts smoked or boiled, bacon, sausages (p

mortadella, salami, chicken/turkey nuggets, ham/chicken croquetteFatty fish Anchovy, sardines, mackerel, salmon, trout, herring, canned sardineLean fish Cod, tuna, saithe, haddock, halibut, plaice, flounder, perch, pike, hakMix fish dishes Fish dishes with no defined fish type (fish croquette, fried fish, fish

mayonnaise, sushi, fish curry)High-fat milk/yoghurt Milk and yoghurt with total fat content higher than 2%High-fat cheese Cheesewith total fat content higher than 25% (feta, Roquefort, edam

goat cheese, blue cheese, brie, smoked cheese)Low-fat dairy Milk and yoghurt with total fat content lower than 2% and cheese wEggsButter Added butterSalty snacks Potato chips, crackers, pop-corn, salty snacksFast-food French-fries, pizza, wrap/tortilla/taco/pita, burger/kebab (fast-food)Meat offals Liver, liver paste, kidney, brain, tripe, pate (beef, pork, chicken, lam

Mix meat dishes Meat dishes with not defined meat type (pasta/lasagne/noodles wiGame Birds, hareShellfish Shrimps, lobster, crab, mussels, squid, cuttlefish, octopus, oysters, c

Fish offals Cod roe, cod roe paste, fish liver, fish liver paste

items was reported. However, in our analysis we have included fooditems, in aggregated food groups, that are relevant to the intake of di-oxins and dioxin-like compounds from food consumption. Further, weaim to derive dietary patterns of these specific food groups which areoften identified as sources of the dioxins and dioxin-like compounds.

Reduced rank regression (RRR) was applied to identify dietarypatterns. RRR is a dimension reduction technique that determines linearfunctions of predictors (food groups), called dietary patterns, bymaximising the explained variation of the response (Hoffmann et al.,2004). RRR is similar to other data exploratory methods for dietary pat-terns, such as principal component or cluster analysis, while it addition-ally requires the definition of a response variable based on the scientificknowledge of the disease physiology (Ambrosini et al., 2012).

In our study daily intakes of 13 food groups (g/day), common for allcountries, were included as predictors and the response variable wasdioxin-like activity in maternal plasma (pg CALUX-TEQ/g lipid), asestimated by the DR-CALUX bioassay. Hence, we extracted a dietarypattern, based on maternal dioxin-like activity, to study the associationbetween maternal diet, maternal levels of dioxin-like activity and birthoutcomes. The extracted dietary pattern was termed “dioxin-dietarypattern”. The adherence of each woman to the “dioxin-dietary pattern”was described by the “dioxin-diet score”. This is a standardised score(with a mean of zero) calculated as the sum of the products of food in-takemultiplied by the corresponding factor loading. The factor loadingsrepresent the correlation of each food group with the dietary patternscore andhigh absolute values of factor loadings depict higher contribu-tion of the food group to the dietary pattern. The percentage of the var-iation of dioxin-like activity in maternal blood (pg CALUX-TEQ/g lipid)explained by the dietary pattern, was further estimated.

2.4. Statistical analysis

Maternal dioxin-diet score was our exposure variable and was usedas continuous or in tertiles (low, middle, upper tertiles). We describedmaternal characteristics, gestational age and birth weight by tertiles ofmaternal dioxin-diet score.

Simple andmultiple linear regressionmodelswere fitted to estimatethe association between maternal dioxin-diet score (continuous or intertiles) with either gestational age (days) or birth weight (g). The po-tential non-linear relationship was evaluated by including the exposurevariable in tertiles and trend tests were performed to evaluate dose–response relations. Several potential confounders were assessed from

ewGeneris project.

Available information

All studies (n = 604)All studies (n = 604)

ork, chicken, beef), hot-dogs,s, doner kebab, cracklings

All studies (n = 604)

s/mackerel, smoke salmon/trout All studies (n = 604)e, sole, canned tuna, dried cod All studies (n = 604)soup, fish cake/balls, fish paste with All studies (n = 604)

All studies (n = 604), gouda, white cheese, cured cheese, All studies (n = 604)

ith total fat content lower than 25% All studies (n = 604)All studies (n = 604)All studies (n = 604)All studies (n = 604)All studies (n = 604)

b, venison), foie gras Greek, Spanish, Norwegian and Danish studies(n = 530)

th meat, meat pie) Norwegian, Danish and British studies (n = 405)Greek and Norwegian studies (n = 247)

lams, cockles Greek, Spanish, Norwegian and Danish studies(n = 530)Norwegian and Danish studies (n = 331)

124 E. Papadopoulou et al. / Science of the Total Environment 484 (2014) 121–128

a variety of maternal and pregnancy-related characteristics withestablished association with maternal dietary habits and birth out-comes. Characteristics, identified as predictors of maternal plasmadioxin-like activity (pg CALUX-TEQ/g lipid) by backward eliminationin stepwise regression models, were included in the adjusted models.Child gender and gestational age were included as a priori confoundersin the models of birth weight. The potential effect modification ofgender was assessed by stratified analysis.

In theNorwegian study only full-termbirthswere included,while allBritish women and Danishwomen from the 1st recruitment period hadplanned Caesarean sections. Thus, due to the differences in the recruit-ment plan of pregnantwomen in each of the 5mother–child studies,weconducted a sensitivity analysis for the association between dioxin-dietscore and gestational age, after excluding 319 Norwegian, British andDanish women. Further, we repeated the RRR procedure to derivecountry-specific dietary patterns with more food groups.

Analyses were performed using STATA 10.1 (Stata Corporation,College Station, Texas) and the PROC PLS procedure, in the SAS version9.2 (SAS Institute, Inc.) was used for the RRR method (Hoffmann et al.,2004).

3. Results

3.1. Maternal dioxin-dietary pattern

The dioxin-dietary pattern was characterised by high intakes of redand white meat, fatty and lean fish, mixed fish dishes, low-fat dairyproducts (factor loadings ≥0.20) and low intakes of salty snacks andhigh-fat cheese (factor loadings ≤−0.20). This pattern explained 7.9%of the dioxin-like activity variation in maternal blood (pg CALUX-TEQ/glipid) (Table 2). Women from Spain and Denmark, non-Caucasians andnon-smokers tended to have a higher dioxin-diet score (Table 3). Ahigher dioxin-diet score was linked to higher maternal age, daily energyand fat intake. Dioxin-like activity in maternal blood was also positivelyrelated to the dioxin-diet score (Spearman's rho = 0.29, p-valueb 0.001).

3.2. Association between maternal dioxin-dietary pattern and birthoutcomes

We found an inverse dose–response association between maternaldioxin-diet score and birth weight after adjustment for confounders(p for trend = 0.034) (Table 4). The reduction in birth weight forinfants born by mothers in the upper tertile of the dioxin-diet scorewas −121 g (95% CI: −232, −10) compared to mothers in the lowertertile. Similar negative associations were obtained in a stratified

Table 2Factor loadings for 13 food groups included in dioxin-dietary pattern and the explainedvariation of dioxin-like activity in maternal blood (in pg CALUX-TEQ/g lipid) by thedioxin-dietary pattern in the NewGeneris project.

Dioxin-dietary pattern factor loadings

Food groupsRed meat 0.27White meat 0.27Processed meat −0.14Fatty-fish 0.27Lean fish 0.35Mix fish dishes 0.24High-fat milk/yoghurt −0.15High-fat cheese −0.54Low-fat dairy 0.30Eggs 0.14Butter −0.10Salty snacks −0.35Fast-food 0.16

Explained variation of dioxin-likeactivity in maternal blood (%)

7.9

analysis by child gender, though statistically significant results were ob-served only in boys (boys: p for trend = 0.039 and beta for high tertileof dioxin-diet score = −170 g, 95% CI = −332, −8 and girls: p fortrend = 0.159 and beta for high tertile of dioxin-diet score = −117 g,95% CI =−277, 43) (data not shown).

We observed a negative non-significant association betweenmaternal dioxin-diet score in the upper tertile and gestational age(β = −1.4 days, 95% CI −3.8, 1.0 days) (Table 4). In a sensitivityanalysis where 319 Norwegian, British and Danish mother–child pairswere excluded, the association between dioxin-diet score and gesta-tional agewas also negative, though not significant andwithwider con-fidence intervals (middle tertile of dioxin-diet score: β = −1.4 days,95% CI: −4.9, 2.1; upper tertile of dioxin-diet score: β = −3.0 days,95% CI:−6.9, 0.8). Pregnancy-related conditions such as preeclampsia,gestational diabetes or gestational hypertension might influence thebirth outcomes. In our study, 5.7% out of 488 pregnantwomen reportedsuch conditions, and further adjustment of our models for such condi-tions did not alter our results (results not shown).

Finally, we identified country-specific dietary patterns related tomaternal dioxin-like activity. The derived dioxin-dietary patterns variedsubstantially between countries and the explained dioxin-like activityvariation ranged from 9.8 to 27.6%. Additionally, the country-specificdioxin-diet scores were positively correlated with dioxin-like activitylevels (pg CALUX TEQ/g lipid) in maternal blood (range of correlation:0.26–0.50) (Supplementary Table 1).

4. Discussion

We investigated the association between maternal diet, dioxin-likeactivity in maternal blood and birth outcomes in a large Europeanstudy. Our main finding was that high adherence to a dietary patternthat was positively linked to dioxin-like activity blood levels was associ-ated with a reduction in birth weight. Less consistent findings wereobserved for gestational age.

We found that a diet high in red and white meat and lean and fattyfish, was linked to high levels of dioxins and dioxin-like compounds inmaternal blood of European pregnant women. Fish and shellfish con-sumption have been previously identified as predictors of dioxin levelsin maternal blood (Glynn et al., 2007; Halldorsson et al., 2008; Wanget al., 2009). Halldorsson et al. found that high fat intake, explained byhigh consumption of red meat and low consumption of fatty-fish waspositively associated with plasma dioxin-like activity of 100 Danishpregnant women (Halldorsson et al., 2009). Additionally, “meat”, “sea-food” and a “dairy” dietary patterns were independently and positivelylinked to blood levels of PCDD/Fs and dioxin-like PCBs of 1656 Japaneseadults (Arisawa et al., 2011). However, there are no reports for pregnantwomen and the effect of combined food consumption, as dietarypatterns, on dioxin and dioxin-like levels in maternal blood.

In our study, a pattern with positive meat and fish loadings waslinked to high dioxin levels in maternal blood and high adherence tothis pattern was associated with reduced birth weight. A previousstudy in the same study population did not find associations betweenmaternal DR-CALUX levels and birth outcomes, while in the presentstudy maternal dietary patterns linked to maternal DR-CALUX levelswere associated with a reduction in birth weight. This previous studyshowed, however, a negative association between DR-CALUX levelsmeasured in cord blood and gestational length in male neonates(Vafeiadi et al., 2014). High meat intake combined with low fish intakeduring pregnancy has been associated with lower birth weight andincreased risk for being small-for-gestational age (Ricci et al., 2010;Timmermans et al., 2012) and vice versa, high fish and low meatconsumption has been linked to protective effects on foetal growth(Chatzi et al., 2012). Findings on the relationship between consumptionof seafood during pregnancy and foetal growth are discrepant, with re-ports of either negative effects (Halldorsson et al., 2007, 2008; Mendezet al., 2010; Rylander et al., 2000;Weisskopf et al., 2005), positive or null

Table 3Maternal characteristics, birth weight and gestational age by tertiles of dioxin-diet score for pregnant women and their children in the NewGeneris project.

Tertiles of dioxin-diet score

All Low Middle Upper

n % n % n % n % p-Valuea

CountryGreece 85 14.1 62 30.8 21 10.4 2 1.0 b0.001Spain 114 18.9 18 9.0 31 15.4 65 32.2UK 74 12.2 36 17.9 24 11.9 14 6.9Denmark 169 28.0 5 2.5 55 27.4 109 54.0Norway 162 26.8 80 39.8 70 34.8 12 5.9

Educational levelLow 120 19.9 42 23.1 33 18.1 45 25.3 0.132Middle 202 33.4 77 42.3 67 36.8 58 32.6High 220 36.4 63 34.6 82 45.1 75 42.1Missing 62 10.3

Maternal ethnicityCaucasian 523 86.6 188 93.5 181 90.0 154 77.0 b0.001Non-Caucasian 79 13.1 13 6.5 20 10.0 46 23.0Missing 2 0.3

ParityPrimiparous 216 35.8 66 33.3 79 40.1 71 35.5 0.361Multiparous 379 62.7 132 66.7 118 59.9 129 64.5Missing 9 1.5

Smoking during pregnancyNo 518 85.8 158 79.4 177 88.9 183 92.0 0.001Yes 79 13.1 41 20.6 22 11.1 16 8.0Missing 7 1.1

Median IQR Median IQR Median IQR Median IQR p-Valuec

Age (years)b 31.3 4.8 30.1 5.5 31.5 4.2 32.2 4.5 b0.001Pre-pregnancy BMI (kg/m2) 22.9 5.4 23.0 5.6 23.1 5.1 22.6 5.3 0.620Energy intake (kcal/day) 2280 979 2141 822 2204 940 2557 1078 b0.001Fat intake (g/day) 96.3 55.7 92.0 44.3 87.3 51.5 113.8 73.5 b0.001Dioxin-like activity in maternal blood (pg CALUX-TEQ/g lipid) 39.3 28.2 35.5 31.1 38.5 30.5 45.5 30.0 b0.001Gestational age (days)b 273.8 9.8 273.7 9.6 274.8 9.8 272.9 9.9 0.140Birth weight (g) 3452 620 3425 614 3520 620 3440 650 0.363

Abbreviations: IQR = interquartile range is equal to the difference between Q1 and Q3.a p-Value for χ2 test for differences between proportions.b Presented values are means and SD.c p-Value for one-way ANOVA or non-parametric Kruskal–Wallis tests for differences between tertiles of the dioxin-diet score.

125E. Papadopoulou et al. / Science of the Total Environment 484 (2014) 121–128

effects (Brantsæter et al., 2012; Drouillet-Pinard et al., 2010; Guldneret al., 2007; Heppe et al., 2011; Lucas et al., 2004; Thorsdottir et al.,2004). In a recent analysis of 152,000 mother–child pairs from 19European birth cohorts researchers concluded that moderate fish con-sumption during pregnancy can be linked with positive birth outcomes(Leventakou et al., 2013). Hence, dietary patterns related to blood bio-markers of contaminants might be a promising methodology to studythe complex link between overall maternal diet, prenatal exposure toenvironmental dietary contaminants and health effects.

Table 4Association between dioxin-diet score with gestational age and birth weight for pregnant wom

Gestational age (days)a

n β (95% CI)

Continuous dioxin-diet scorec 520 0.0 (−0.9, 0.8Dioxin-diet score in tertilesd

Low (b−0.4 points) 175 ReferenceMiddle (−0.4 to 0.4 points) 173 −0.4 (−2.4, 1.6Upper (N0.4 points) 172 −1.4 (−3.8,1.0)

p for trend e 0.260

a Model is adjusted for maternal educational level (low, middle, high), energy (kcals/day)smoking during pregnancy (yes/no) and country.

b Model is adjusted for the same variables and additionally for gestational age (days) and gec Regression coefficient of dioxin-diet score used as a continuous variable corresponds to the

diet score.d Regression coefficient of dioxin-diet score used as tertiles corresponds to the adjusted chang

dioxin-diet score.e p for trend is estimated for the adjusted change across increasing tertiles of the dioxin-die

In a further analysis we identified dietary patterns based on dioxin-like activity in maternal blood by country. Moving from northern tosouthern Europe, the substantial contributors to the dioxin-dietary pat-tern were white meat and game for Norwegian women, fatty fish forDanish women, dairy products for British women and meat and fattyfish for Greek women. These food groups were positively related withblood dioxin levels. For Spanish women a more complicated patternwas derived which was characterised by high consumption of mix-fish dishes (mainly fried small fish and fish croquettes), high-fat milk,

en and their children in the NewGeneris project.

Birth weight (g)b

p β (95% CI) p

) 0.998 −24 (−63, 15) 0.222

Reference) 0.710 −39 (−130, 52) 0.397

0.254 −121 (−232, −10) 0.0320.034

, maternal age (years), pre-pregnancy BMI (kg/m2), parity (primiparous/multiparous),

nder.adjusted change in gestational age or birth weight for an increase of 1 point of the dioxin-

e in gestational age or birthweight by increasing tertiles compared to the low tertile of the

t score.

126 E. Papadopoulou et al. / Science of the Total Environment 484 (2014) 121–128

fast-food and meat offals. Differences in the POP profile and the back-ground levels of POPs along with different dietary habits and the smallsample size by country might explain such discrepancies betweenpopulations.

Overall, maternal diet explained a small percentage of the DR-CALUX variation (8%). However, similar small percentages of explainedvariation in blood biomarkers have been reported by previous studiesthat have used the RRR procedure to derive dietary patterns (Centrittoet al., 2009; Fung et al., 2012; Heroux et al., 2010; Liese et al., 2010;McNaughton et al., 2009). Similarly low proportions of variance inblood levels of dioxins and dioxin-like compounds explained by food in-take has been reported in previous studies of pregnantwomen, suggest-ing that other lifestyle factors or sources of exposure might influenceblood levels (Glynn et al., 2007; Huisman et al., 1995). It has beenpointed out that the explained variation in an outcome variable (inour case dioxin levels) may not be a good metric of the associationand the risk attributed to a predictor variable (in our case food groups)(Pearce, 2011). This may happen when there is not enough variationin the predictor variables, which are women's dietary habits inour study. If, for example, intake of meat and seafood does not varymuch among pregnant women then the explained plasma dioxinvariation by diet would be small, even though diet still would be themain contributor to dioxin exposure. Moreover, given that there is abackground long term exposure of the population to persistent environ-mental pollutants, the identification of risk factors, like diet, and high-risk groups is challenging. Finally a direct comparison between studiesis difficult due to different methodologies in dietary assessment orblood analysis.

The present study has strengths and limitations. The large sample ofpregnant women from 5 European countries, the use of a biomarker toassess maternal and foetal body burden and the use of RRR to derivedietary patterns, are among the strengths of our study. RRR deriveddietary patterns have been linked to obesity, coronary heart disease,type 2 diabetes, breast cancer and other chronic diseases, by usingblood estimates, anthropometric measurements or nutrient intakewith an established association with the disease (Ambrosini et al.,2012; McNaughton et al., 2008; Meyer et al., 2011; Schulz et al.,2008). Since biomarkers reflect the body burden of environmental tox-icants that might lead to adverse health effects, dietary patterns basedon exposure biomarkers might be more relevant for the effect of dieton disease.

Nevertheless, uncontrolled measurement errors introduced by theuse of FFQs and unmeasured long-term diet may have influenced ourresults. The use of country-specific FFQs could make the harmonisationof the dietary data for the study of dietary patterns in the overall popu-lation of 604 women from five countries harder. For this reason westudied dietary patterns by including general aggregated food groups,consisting of a variety of food items, instead of single specific fooditems. Further, we performed country-specific analysis in order toderive specific dietary patterns after includingmore food groups. More-over, individual differences in absorption, distribution and metabolismof persistent organic pollutantsmight add uncertainty to the estimationof dietary exposure. Additional uncertainties related to the assessmentmethod for blood levels of dioxins and dioxin-like compounds mightalso introduce uncertainty to our estimates. However, the DR-CALUXbioassay has been suggested as a valid ranking method to categorise apopulation according to their exposure to dioxins and dioxin-likecompounds and has been used by other epidemiological studies(Halldorsson et al., 2009; Koppen et al., 2009; Pedersen et al., 2012;Vafeiadi et al., 2014). Maternal diet was assessed during pregnancy;hence pregnancy-related dietary changes might have influenced theestimates of long-term exposure to dioxin-like compounds throughdiet. However, changes in diet and intake of dioxins and PCBs due topregnancy are considered small (Chan-Hon-Tong et al., 2013; Crozieret al., 2009).Wehave noted that themain dietary contributors ofmater-nal dioxin-like activity levels differ somehow between countries which

might be attributed to relatively small country-specific sample sizes,while the main analyses are based on the larger pooled study popula-tion. An additional limitation of our study is the lack of information onmaternal residence (urban/rural) as it might confound the associationbetween maternal diet and birth outcomes, also regarding the dietaryexposure to dioxin-like compounds. Finally, the lack of information onpreconceptional maternal lifestyle factors, such as smoking, is alsoacknowledged as a limitation of our study, due to the relationshipwith birth outcomes (Haas et al., 2005).

We acknowledge that the body burden and toxicity of dioxin-likecompounds depend on their bioaccumulation resulting from long-term exposure and is not directly linked with the amount consumedat a given time. However, there are several reports on the positiverelationship between dietary intake and blood levels of dioxins andPCBs, either assessed as calculated intakes of dioxins and PCBs or asfood consumption concluding on the importance of dietary intake as asignificant predictor of the body burden (Gasull et al., 2011; Huismanet al., 1995; Knutsen et al., 2011; Kvalem et al., 2009). Given the mainexposure pathway to dioxins and PCBs through diet and their longhalf-lives, long-term dietary intake can reflect the body burden andare used as a proxy of human exposure to dioxins and PCBs and torank individuals in the population (Bilau et al., 2008; Huisman et al.,1995; Kvalem et al., 2012). Finally, we would like to comment that theobserved reduction in birth weight of 120 g on the current study corre-sponds to a 3–4% reduction of themean birth (3452 g, SD: 620 g), whichis not trivial. However, the interpretation and clinical relevance of smallshifts in birth weight, associated with prenatal exposures to environ-mental contaminants are a topic of vigorous discussion in epidemiolog-ical studies (Savitz, 2007; Wilcox, 2001).

5. Conclusions

In conclusion, we observed that maternal diet was associated withdioxin-like activity in maternal blood during pregnancy and this prena-tal exposuremight lead to reduced birthweight. More studies are need-ed to develop updated dietary guidelines for women of reproductiveage, aiming for the reduction of dietary exposure to persistent organicpollutants such as dioxins and dioxin-like compounds.

Supplementary data to this article can be found online at http://dx.doi.org/10.1016/j.scitotenv.2014.03.047.

Acknowledgements

The NewGeneris (Newborns and Genotoxic exposure risks) studywas funded by the European Union (EU Contract Food-CT-2005-016320). The study was also supported by grants obtained locally in-cluding: the National Institute for Health Research, UK (ProgrammeGrant RP-PG-0407-10044), the Norwegian Ministry of Health, theNorwegian Ministry of Education and Research, the Norwegian Re-search Council/FUGE (Grant No. 151918/S10), the EU funded HiWATE(Contract No. Food-CT-2006-036224), the US National Institutes ofHealth/National Institute of Environmental Health Sciences (ContractNo. NO-ES-75558) and the US National Institutes of Health/NINDS(Grant No. 1 UO1 NS 047537-01). HB is employed by BioDetectionSystems B.V., Amsterdam, The Netherlands. EP holds the YggdrasilGrand (Yggdrasil mobility programme 2012–2013) awarded by theResearch Council of Norway (219671/F11). MP holds a ‘Juan de laCierva’ post-doctoral fellowship awarded from the Spanish Ministry ofScience and Innovation (JCI-2011-09479).

References

Ambrosini GL, Emmett PM, Northstone K, Howe LD, Tilling K, Jebb SA. Identification of adietary pattern prospectively associated with increased adiposity during childhoodand, adolescence. Int J Obes (Lond) 2012;36:1299–305.

Arisawa K, Takeda H, Mikasa H. Background exposure to PCDDs/PCDFs/PCBs and its po-tential health effects: a review of epidemiologic studies. J Med Invest 2005;52:10–21.

127E. Papadopoulou et al. / Science of the Total Environment 484 (2014) 121–128

Arisawa K, Uemura H, Hiyoshi M, Kitayama A, Takami H, Sawachika F, et al. Dietarypatterns and blood levels of PCDDs, PCDFs, and dioxin-like PCBs in 1656 Japaneseindividuals. Chemosphere 2011;82:656–62.

Bilau M, Matthys C, Baeyens W, Bruckers L, De BG, Den HE, et al. Dietary exposure todioxin-like compounds in three age groups: results from the Flemish environmentand health study. Chemosphere 2008;70:584–92.

Brantsæter AL, Haugen M, Alexander J, Meltzer HM. Validity of a new food frequencyquestionnaire for pregnant women in the Norwegian Mother and Child CohortStudy (MoBa). Matern Child Nutr 2008;4:28–43.

Brantsæter AL, Birgisdottir BE, Meltzer HM, Kvalem HE, Alexander J, Magnus P, et al. Ma-ternal seafood consumption and infant birth weight, length and head circumferencein the Norwegian Mother and Child Cohort Study. Br J Nutr 2012;107:436–44.

Brouwer A, Sonneveld E, Botschuijver S, Besselink H, van den Burg B. A low volumemethod for the analysis of dioxins and dioxin-like compounds in serum and wholeblood using BDS' DR CALUX® bioassay. Ogranohalogen Compd 2004;66:687–90.

Centritto F, Iacoviello L, Di GR, De CA, Costanzo S, Zito F, et al. Dietary patterns, cardiovas-cular risk factors and C-reactive protein in a healthy Italian population. Nutr MetabCardiovasc Dis 2009;19:697–706.

Chan-Hon-Tong A, Charles MA, Forhan A, Heude B, Sirot V. Exposure to food contami-nants during pregnancy. Sci Total Environ 2013;458–460:27–35.

Chatzi L, Plana E, Daraki V, Karakosta P, Alegkakis D, Tsatsanis C, et al. Metabolic syndrome inearly pregnancy and risk of preterm birth. Am J Epidemiol 2009;170:829–36.

Chatzi L, Melaki V, Sarri K, Apostolaki I, Roumeliotaki T, Georgiou V, et al. Dietary patternsduring pregnancy and the risk of postpartum depression: the mother–child ‘Rhea’cohort in Crete, Greece. Public Health Nutr 2011;14:1663–70.

Chatzi L, Mendez M, Garcia R, Roumeliotaki T, Ibarluzea J, Tardon A, et al. Mediterraneandiet adherence during pregnancy and fetal growth: INMA (Spain) and Rhea (Greece)mother–child cohort studies. Br J Nutr 2012;107:135–45.

Crozier SR, Robinson SM, Godfrey KM, Cooper C, Inskip HM. Women's dietary patternschange little from before to during pregnancy. J. Nutr. 2009;139:1956–63.

Davey GK, Spencer EA, Appleby PN, Allen NE, Knox KH, Key TJ. EPIC-Oxford: lifestylecharacteristics and nutrient intakes in a cohort of 33 883 meat-eaters and 31 546non meat-eaters in the UK. Public Health Nutr 2003;6:259–69.

De Mul A, Bakker MI, Zeilmaker MJ, Traag WA, Leeuwen SP, Hoogenboom RL, et al.Dietary exposure to dioxins and dioxin-like PCBs in The Netherlands anno 2004.Regul Toxicol Pharmacol 2008;51:278–87.

Denison MS, Soshilov AA, He G, DeGroot DE, Zhao B. Exactly the same but different: pro-miscuity and diversity in themolecular mechanisms of action of the aryl hydrocarbon(dioxin) receptor. Toxicol Sci 2011;124:1–22.

Drouillet-Pinard P, Huel G, Slama R, Forhan A, Sahuquillo J, Goua V, et al. Prenatal mercurycontamination: relationship with maternal seafood consumption during pregnancyand fetal growth in the ‘EDEN mother–child’ cohort. Br J Nutr 2010;104:1096–100.

EFSA. Scientific colloquium summary report: dioxins. Methodologies and principles forsetting tolerable intake levels for dioxins, furans and dioxin-like PCBs; 2004.

EFSA. Update of the monitoring of levels of dioxins and PCBs in food and feed. EFSA J2012;10:2832.

Fattore E, Fanelli R, Turrini A, Di DA. Current dietary exposure to polychlorodibenzo-p-dioxins, polychlorodibenzofurans, and dioxin-like polychlorobiphenyls in Italy. MolNutr Food Res 2006;50:915–21.

Fernandez MA, Gomara B, Bordajandi LR, Herrero L, Abad E, Abalos M, et al. Dietaryintakes of polychlorinated dibenzo-p-dioxins, dibenzofurans and dioxin-likepolychlorinated biphenyls in Spain. Food Addit Contam 2004;21:983–91.

Fung TT, Schulze MB, Hu FB, Hankinson SE, Holmes MD. A dietary pattern derived tocorrelate with estrogens and risk of postmenopausal breast cancer. Breast CancerRes Treat 2012;132:1157–62.

Gasull M, Bosch de BM, Puigdomenech E, Pumarega J, Porta M. Empirical analyses of theinfluence of diet on human concentrations of persistent organic pollutants: a system-atic review of all studies conducted in Spain. Environ Int 2011;37:1226–35.

Glynn A, Aune M, Darnerud PO, Cnattingius S, Bjerselius R, Becker W, et al. Determinantsof serum concentrations of organochlorine compounds in Swedish pregnant women:a cross-sectional study. Environ Health 2007;6:2.

Guldner L, Monfort C, Rouget F, Garlantezec R, Cordier S. Maternal fish and shellfish intakeand pregnancy outcomes: a prospective cohort study in Brittany, France. EnvironHealth 2007;6:33.

Guxens M, Ballester F, Espada M, Fernandez MF, Grimalt JO, Ibarluzea J, et al. Cohortprofile: the INMA–INfancia y Medio Ambiente–(Environment and Childhood)Project. Int J Epidemiol 2012;41:930–40.

Haas JS, Fuentes-Afflick E, Stewart AL, Jackson RA, Dean ML, Brawarsky P, et al.Prepregnancy health status and the risk of preterm delivery. Arch Pediatr AdolescMed 2005;159:58–63.

Halldorsson TI, Meltzer HM, Thorsdottir I, Knudsen V, Olsen SF. Is high consumption offatty fish during pregnancy a risk factor for fetal growth retardation? A study of44,824 Danish pregnant women. Am J Epidemiol 2007;166:687–96.

Halldorsson TI, Thorsdottir I, Meltzer HM, Nielsen F, Olsen SF. Linking exposure topolychlorinated biphenyls with fatty fish consumption and reduced fetal growthamong Danish pregnant women: a cause for concern? Am J Epidemiol 2008;168:958–65.

Halldorsson TI, Thorsdottir I, Meltzer HM, Strom M, Olsen SF. Dioxin-like activity inplasma among Danish pregnant women: dietary predictors, birth weight and infantdevelopment. Environ Res 2009;109:22–8.

Haugen M, Brantsaeter AL, Alexander J, Meltzer HM. Dietary supplements contributesubstantially to the total nutrient intake in pregnant Norwegian women. Ann NutrMetab 2008;52:272–80.

Heppe DH, Steegers EA, Timmermans S, Breeijen H, Tiemeier H, Hofman A, et al. Maternalfish consumption, fetal growth and the risks of neonatal complications: the Genera-tion R Study. Br J Nutr 2011;105:938–49.

Hepworth SJ, Hardie LJ, Fraser LK, Burley VJ, Mijal RS, Wild CP, et al. Deoxynivalenolexposure assessment in a cohort of pregnant women from Bradford, UK. FoodAddit Contam Part A Chem Anal Control Expo Risk Assess 2012;29:269–76.

Heroux M, Janssen I, Lam M, Lee DC, Hebert JR, Sui X, et al. Dietary patterns and therisk of mortality: impact of cardiorespiratory fitness. Int J Epidemiol 2010;39:197–209.

Hoffmann K, Schulze MB, Schienkiewitz A, Nothlings U, Boeing H. Application of a newstatistical method to derive dietary patterns in nutritional epidemiology. Am JEpidemiol 2004;159:935–44.

HuismanM, Eerenstein SE, Koopman-Esseboom C, Brouwer M, Fidler V, Muskiet FA, et al.Perinatal exposure to polychlorinated biphenyls and dioxins through dietary intake.Chemosphere 1995;31:4273–87.

Kiviranta H, Ovaskainen ML, Vartiainen T. Market basket study on dietary intake of PCDD/Fs, PCBs, and PBDEs in Finland. Environ Int 2004;30:923–32.

Knutsen HK, Kvalem HE, Haugen M, Meltzer HM, Brantsaeter AL, Alexander J, et al. Sex,BMI and age in addition to dietary intakes influence blood concentrations andcongener profiles of dioxins and PCBs. Mol Nutr Food Res 2011;55:772–82.

Koppen G, Den HE, Nelen V, Van De Mieroop E, Bruckers L, Bilau M, et al. Organochlorineand heavy metals in newborns: results from the Flemish Environment and HealthSurvey (FLEHS 2002–2006). Environ Int 2009;35:1015–22.

Kulkarni PS, Crespo JG, Afonso CA. Dioxins sources and current remediation technologies—a review. Environ Int 2008;34:139–53.

Kvalem HE, Knutsen HK, Thomsen C, Haugen M, Stigum H, Brantsaeter AL, et al. Roleof dietary patterns for dioxin and PCB exposure. Mol Nutr Food Res 2009;53:1438–51.

Kvalem HE, Brantsaeter AL, Meltzer HM, Stigum H, Thomsen C, Haugen M, et al. Develop-ment and validation of prediction models for blood concentrations of dioxins andPCBs using dietary intakes. Environ Int 2012;50:15–21.

Laier P, Cederberg T, Larsen JC, Vinggaard AM. Applicability of the CALUX bioassay forscreening of dioxin levels in human milk samples. Food Addit Contam 2003;20:583–95.

Leventakou V, Roumeliotaki T, Martinez D, Barros H, Brantsaeter AL, Casas M, et al. Fishintake during pregnancy, fetal growth, and gestational length in 19 European birthcohort studies. Am J Clin Nutr 2013;99:506–16.

Liem AK, Furst P, Rappe C. Exposure of populations to dioxins and related compounds.Food Addit Contam 2000;17:241–59.

Liese AD, Nichols M, Hodo D, Mellen PB, Schulz M, Goff DC, et al. Food intake patterns as-sociated with carotid artery atherosclerosis in the Insulin Resistance AtherosclerosisStudy. Br J Nutr 2010;103:1471–9.

Lucas M, Dewailly E, Muckle G, Ayotte P, Bruneau S, Gingras S, et al. Gestational age andbirth weight in relation to n-3 fatty acids among Inuit (Canada). Lipids 2004;39:617–26.

Magnus P, Irgens LM, Haug K, Nystad W, Skjaerven R, Stoltenberg C. Cohort profile: theNorwegian Mother and Child Cohort Study (MoBa). Int J Epidemiol 2006;35:1146–50.

McGraw JE, Waller DP. Fish ingestion and congener specific polychlorinated biphenyl andp, p′-dichlorodiphenyldichloroethylene serum concentrations in a great lakes cohortof pregnant African American women. Environ Int 2009;35:557–65.

McNaughton SA, Mishra GD, Brunner EJ. Dietary patterns, insulin resistance, andincidence of type 2 diabetes in the Whitehall II Study. Diabetes Care 2008;31:1343–8.

McNaughton SA, Mishra GD, Brunner EJ. Food patterns associated with blood lipids arepredictive of coronary heart disease: the Whitehall II Study. Br J Nutr 2009;102:619–24.

Mendez MA, Plana E, Guxens M, Foradada Morillo CM, Albareda RM, Garcia-Esteban R,et al. Seafood consumption in pregnancy and infant size at birth: results from aprospective Spanish cohort. J Epidemiol Community Health 2010;64:216–22.

Merlo DF, Wild CP, Kogevinas M, Kyrtopoulos S, Kleinjans J. NewGeneris: a Europeanstudy on maternal diet during pregnancy and child health. Cancer Epidemiol Bio-markers Prev 2009;18:5–10.

Meyer J, Doring A, Herder C, Roden M, Koenig W, Thorand B. Dietary patterns, subclinicalinflammation, incident coronary heart disease and mortality in middle-aged menfrom the MONICA/KORA Augsburg cohort study. Eur J Clin Nutr 2011;65:800–7.

Pearce N. Epidemiology in a changing world: variation, causation and ubiquitous riskfactors. Int J Epidemiol 2011;40:503–12.

Pedersen M, Wichmann J, Autrup H, Dang DA, Decordier I, Hvidberg M, et al. Increasedmicronuclei and bulky DNA adducts in cord blood after maternal exposures totraffic-related air pollution. Environ Res 2009;109:1012–20.

Pedersen M, Halldorsson TI, Autrup H, Brouwer A, Besselink H, Loft S, et al. Maternal dietand dioxin-like activity, bulky DNA adducts and micronuclei in mother–newborns.Mutat Res 2012;734:12–9.

Perello G, Gomez-Catalan J, Castell V, Llobet JM, Domingo JL. Assessment of the temporaltrend of the dietary exposure to PCDD/Fs and PCBs in Catalonia, over Spain: healthrisks. Food Chem Toxicol 2012;50:399–408.

Ricci E, Chiaffarino F, Cipriani S, Malvezzi M, Parazzini F. Diet in pregnancy and risk ofsmall for gestational age birth: results from a retrospective case–control study inItaly. Matern Child Nutr 2010;6:297–305.

Rose M, Fernandes A, Foxall C, Dowding A. Transfer and uptake of polychlorinateddibenzo-p-dioxins and furans (PCDD/Fs) and polychlorinated biphenyls (PCBs) intomeat and organs of indoor and outdoor reared pigs. Food Addit Contam Part AChem Anal Control Expo Risk Assess 2012;29:431–48.

Rylander L, Stromberg U, Hagmar L. Lowered birth weight among infants born to womenwith a high intake of fish contaminated with persistent organochlorine compounds.Chemosphere 2000;40:1255–62.

Savitz DA. Guest editorial: biomarkers of perfluorinated chemicals and birth weight.Environ Health Perspect 2007;115:A528–9.

128 E. Papadopoulou et al. / Science of the Total Environment 484 (2014) 121–128

Schecter A, Cramer P, Boggess K, Stanley J, Papke O, Olson J, et al. Intake of dioxins and re-lated compounds from food in the U.S. population. J Toxicol Environ Health A 2001;63:1–18.

Schulz M, Hoffmann K, Weikert C, Nothlings U, Schulze MB, Boeing H. Identification of adietary pattern characterized by high-fat food choices associated with increased riskof breast cancer: the European Prospective Investigation into Cancer and Nutrition(EPIC)—Potsdam Study. Br J Nutr 2008;100:942–6.

Sirot V, Tard A, Venisseau A, Brosseaud A, Marchand P, Le BB, et al. Dietary exposure topolychlorinated dibenzo-p-dioxins, polychlorinated dibenzofurans and polychlorinatedbiphenyls of the French population: results of the second French Total Diet Study.Chemosphere 2012;88:492–500.

Thorsdottir I, Birgisdottir BE, Halldorsdottir S, Geirsson RT. Association of fish andfish liver oil intake in pregnancy with infant size at birth among women of nor-mal weight before pregnancy in a fishing community. Am J Epidemiol 2004;160:460–5.

Timmermans S, Steegers-Theunissen RP, Vujkovic M, Den BH, Russcher H, Lindemans J,et al. The Mediterranean diet and fetal size parameters: the Generation R Study. BrJ Nutr 2012;108:1399–409.

Tornkvist A, Glynn A, AuneM, Darnerud PO, Ankarberg EH. PCDD/F, PCB, PBDE, HBCD andchlorinated pesticides in a Swedish market basket from 2005—levels and dietaryintake estimations. Chemosphere 2011;83:193–9.

Tsukimori K, Morokuma S, Hori T, Takahashi K, Hirata T, Otera Y, et al. Characterization ofplacental transfer of polychlorinated dibenzo-p-dioxins, dibenzofurans andpolychlorinated biphenyls in normal pregnancy. J Obstet Gynaecol Res 2013;39:83–90.

Vafeiadi M, Agramunt S, Pedersen M, Besselink H, Chatzi L, Fthenou E, et al. In uteroexposure to compounds with dioxin-like activity and birth outcomes. Epidemiology2014;25:215–24.

VanWouwe N, Windal I, Vanderperren H, Eppe G, Xhrouet C, Massart AC, et al. Validationof the CALUX bioassay for PCDD/F analyses in human blood plasma and comparisonwith GC–HRMS. Talanta 2004;63:1157–67.

Vioque J, Weinbrenner T, Asensio L, Castello A, Young IS, Fletcher A. Plasma concentra-tions of carotenoids and vitamin C are better correlated with dietary intake in normalweight than overweight and obese elderly subjects. Br J Nutr 2007;97:977–86.

Wang RY, Jain RB, Wolkin AF, Rubin CH, Needham LL. Serum concentrations of selectedpersistent organic pollutants in a sample of pregnant females and changes in theirconcentrations during gestation. Environ Health Perspect 2009;117:1244–9.

Weisskopf MG, Anderson HA, Hanrahan LP, Kanarek MS, Falk CM, Steenport DM, et al.Maternal exposure to Great Lakes sport-caught fish and dichlorodiphenyldichloroethylene, but not polychlorinated biphenyls, is associated with reducedbirth weight. Environ Res 2005;97:149–62.

Wilcox AJ. On the importance – and the unimportance – of birthweight. Int J Epidemiol2001;30:1233–41.

Windal I, Vandevijvere S, Maleki M, Goscinny S, Vinkx C, Focant JF, et al. Dietary intake ofPCDD/Fs and dioxin-like PCBs of the Belgian population. Chemosphere 2010;79:334–40.

Wright J, Small N, Raynor P, Tuffnell D, Bhopal R, Cameron N, et al. Cohort profile: theBorn in Bradford multi-ethnic family cohort study. Int J Epidemiol 2012;42:978–91.