Update of Indicator PCB Levels in Food in Southern Italy: Assessment of the Dietary Exposure for Adult and Elderly Population

(e levels of non-dioxin-like PCB indicators (iPCBs 28, 52, 101, 138, 153, and 180) were determined in food samples (seafood, meat and processed meat, milk and dairy products, hen eggs, olive oil, and other fats) to evaluate the exposure of adult and elderly population. iPCB levels in samples were in the following order: fishery products>meat and processed meat>milk and dairy products> olive oil and other fats. None of the samples had concentrations above the maximum permissible limits for human consumption established by the European Union legislation, except for salami samples.(e dietary intake for the total population was 12.33 ng·kg bw·d, while depending on the sex/age groups, exposure was estimated between 9.60 and 12.11 ng·kg bw·d, with seafood being the major contributor. (e exposure scenario indicates that further efforts must still be carried out to protect the consumer from these harmful chemicals.


Introduction
Polychlorinated biphenyls (PCBs) are a group of man-made substances comprising 209 congeners classified as persistent organic pollutants (POPs). is complex mixture of isomers differs in chemical structure according to the position of chlorine substituents around the biphenyl backbone. It is well recognized that different chlorination patterns deeply affect the biological activity and toxicological profile of these chemicals. PCBs with non-or mono-ortho chloro substitution, known as dioxin-like PCBs (dl-PCBs), exert their toxicity primarily through binding with the aryl hydrocarbon receptor, while the remaining congeners classified as non-dioxin-like PCBs (ndl-PCBs) appear to act via different modes determining a complex spectrum of adverse effects, such as neurological, neuroendocrinological, and immunological disorders [1]. To this latter group belong six congeners, the so-called indicator PCBs (iPCBs), considered a suitable tool to monitor the contamination degree because their sum constitutes about half of the total ndl-PCBs present in feed and foods [2]. In addition to their higher abundance compared to other congeners, recent studies indicate that ndl-PCBs are not eliminated as easily as dl-PCBs and that low ndl-PCBs doses, such as those similar to background contamination in food, can cause negative effects when exposure is prolonged over time [3]. Consequently, understanding the level, trend and impact of these contaminants on human health remain a salient point for the researchers [2]. Many studies have been carried out to evaluate the dietary exposure to these POPs worldwide lending, in general, greater attention to more toxic dl-PCBs than ndl-PCBs. Also, in Italy, the volume of existing data for iPCBs is limited and predominantly referred to contamination of fishery products [4][5][6][7][8] and milk [3,[9][10][11][12], while the other foods have been the subject of very few investigations, although some of them showed levels of contamination often exceeding the established limits for human consumption [13][14][15][16]. For example, the results obtained from the analyses of meat and hen egg samples from two regions of Southern Italy, Apulia [13] and Campania [16], revealed iPCB concentrations above maximum permissible limits set by the European Union. Consequently, this study is placed in the context of the need to acquire further data on the presence of such contaminants also in other food products in Italy. In making the analysis, we ascertain whether the concentrations of six iPCBs are compliant with the European Union food law requirements. Finally, we estimate the exposure level in the adult and elderly population and compare it with the acceptable daily intake of 10 ng·kg −1 bw·d −1 set for the sum of six ndl-PCB indicators at the national level by various countries.

Sample Collection.
In May-July 2019, food samples were randomly acquired from five supermarkets in cities (Bari, Lecce, Taranto, Foggia, Brindisi, and Matera) in Southern Italy. A total of 35 types of foods are classified into the following groups: (1) fish and seafood, (2) meat and meatbased products, (3) milk (whole cow milk) and dairy products, (4) eggs from free foraging chickens, and (5) extravirgin olive oil and fats were acquired from five supermarkets of each city. For each food, 120 individual items were taken and combined in two composite samples. For shellfish and crustaceans, 300 individual units were used to prepare the composite sample. For bluefin tuna, slices (n � 30) of about 0.1-0.2 kg of muscle tissue were taken. e composite samples (only edible parts) were homogenized and stored below −20°C.

Analytical Method and Instrumental Analysis.
e concentrations of six indicator PCBs (iPCBs: 28, 52, 101, 138, 153, and 180) were determined. e analytical method has been reported in detail in the previous paper [17]. Homogenized samples (0.5-3.0 g) were mixed with Na 2 SO 4 and spiked with PCB 143 used as the internal standard. e mixture was extracted with hexane, and the extracts were evaporated to dryness, permitting the gravimetric determination of the fat content [18]. e milk fat was isolated by using liquid-liquid extraction with diethyl ether and petroleum ether (1 : 1) [19]. An aliquot (about 100 mg) of the extracts was dissolved in hexane and cleaned by passing through 8 g of acid silica (H 2 SO 4 , 44% w.w.), using 50 mL of a mixture of hexane/dichloromethane (1/1, v/v) for elution of the analytes. e eluates were dried completely by using a stream of nitrogen and redissolved in 100 μL of iso-octane. To control the whole sample preparation process, appropriate C 13 -labelled standards (Wellington Laboratories Inc., Guelph, Ontario, Canada) were added to the samples before extraction (PCB-LCS-H) and at the clean-up stage (PCB-SCS-H). e identification and quantification of iPCBs were performed on an HRGH/HRMS system composed of a gas chromatograph trace series coupled with a mass spectrometer 2000 MAT 95 XL ( ermo Finnigan, Bremen, Germany). An Rtx-PCB capillary column (60 m × 0.25 mm i.d. × 0.25 μn film thickness; Restek, Cernusco Sul Naviglio, MI, Italy) was used for chromatographic separation of iPCBs. e initial oven temperature was set at 100°C for 2 min, then raised to 180°C, at a rate of 20°C/min, hence increased to 260°C at a rate of 2°C/min, and finally brought up to 300°C with a rate of 5°C/min held for 4 min. e carrier gas used was helium, 6.0 purity degree (99.9999%), in the constant flow mode of 1.0 mL/min. e injection temperature was set at 280°C, and the injections were performed in the splitless mode with a 1 ml vol. (for 1 min. and a split rate of 140 ml/min). e temperatures of the ion source and transfer line were set at 270 and 290°C, respectively. e electron ionization mode (E.I.) was used for mass method operations. Electron energy and detector voltages were 35 eV and 350 V, respectively. e detector resolving power was >10.000 (10% valley definition), and the two most intense ions were monitored for the determination of the single congeners. Six points (0.5, 2.0, 5.0, 10.0, and 20.0 ng·g −1 ) and calibration curves (r 2 > 0.999) in the linear response interval of the detector were created for quantification. To test the accuracy of the method, certified reference material and blank samples were analysed with each batch of samples. Levels of each iPCBs in the procedural blanks were always under the limit of detection. To correctly identify the tested compounds and to check matrix interferences, some food samples were fortified with a known amount of each PCB congener obtaining satisfactory recoveries (PCB 28 � 91%, PCB 52 � 94%, PCB 101 � 96%, PCB 138 � 100%, PCB 153 � 104%, and PCB 180 � 98%). All samples were analysed in triplicate. Precision was tested every day by running 3 replicates of the calibration curves at the beginning and at the end of the analyses (RDS � 15%). Mackerel oil (No 350-Community Bureau of Reference, BCR) from Promochem GmbH was used as reference material. For the replicate, standard reference materials, and recovery of labelled compounds, the relative standard deviations (RSD) were <10% for all the detected compounds. e recovery rates of labelled standards were between 85 and 120%. e limits of detection (LODs) were calculated on the blanks (n.°30) analysed along with unknown samples and were set as three times the signal-to-noise ratio (0.04-1.40 pg·g −1 for PCBs). e limits of quantification (LOQs) were between 0.0004 and 0.04 ng·g −1 for PCBs and corresponded to ten times the signal-to-noise ratio. Concentrations of ndl-PCBs are

Exposure Assessment.
e estimated dietary intakes (deterministic approach) were calculated by multiplying the food consumption data by the mean concentrations of iPCBs in each food followed by dividing by the body weight. Consumption (Table 1) and biometric data of the total population and of two sex/age classes (male adults: 18-64.9 years, body weight 78.4 kg; female adults: 18-64.9 years, body weight 62.2 kg; male elders: ≥65 years, body weight 78.1 kg; female elders: ≥65 years, body weight 65.0 kg; total population: body weight 70.0 kg) were obtained from the Italian national food consumption survey INRAN-SCAI [21,22]. To estimate the total intake of the sum of the six iPCBs (Σ6 iPCBs), individual daily intakes of each iPCBs from different food items were added up. For exposure calculations, the contamination level of each sample expressed in lipid weight was converted into wet weight by using the lipid content of the samples. Exposure data were estimated according to both the lower bound and upper bound hypotheses. e Kruskal-Wallis test was undertaken to compare the estimated weekly intake according to the gender for each age group. All p values below 0.05 were considered statistically significant.

Levels in Food and Compliance with the EU Regulation.
e analytical results of six iPCBs and their sum (Σ6 iPCBs) for each food category, expressed as ng·g −1 on a wet or lipid weight basis to facilitate the comparison with the maximum permissible limits (MPLs) set by the European Union (EU) [20], are illustrated in Table 2. e differences between upper and lower bound values are negligible, so the results are expressed in upper bound hypothesis.
Within the fishery products, the highest contamination mean level was in fish, followed by shellfish, crustaceans, and cephalopods that showed moderate concentrations. Because the species are from different trophic guilds, there is the potential for the observed concentration differences. In fact, when analysing the contaminant load, one must take into account the trophic niche in which the species are inserted. Consequently, the increase in concentrations from lowtrophic-level invertebrates to high-trophic-level organisms is an expected finding, demonstrating the effect of PCB biomagnification through the food chains. In the similar way, within fish, the top-level carnivore as the Atlantic bluefin tuna showed the highest levels, followed by benthic species rosefish and red mullet, while the lowest value was in the European pelagic hake. e contribution of individual six congeners to total iPCB was generally in the order of PCB 153 > PCB 138 > PCB 180, while the combined contributions of the lower chlorinated congeners to iPCB total strongly varied being 9.4%, 12.2%, 33.3%, and 2.3% in fish, shellfish, cephalopods, and crustaceans, respectively (see Figure 1). e distribution pattern of congeners is in very good match with results reported in the literature, where PCBs 138, 153, and 180 are described as the predominant congeners in marine biota [23]. e prevalence of these highly chlorinated congeners is the result of their greater resistance to metabolism and elimination than the lower congeners and is also linked to their major abundance in technical mixtures, such as Aroclor 1254 and 1260 [24], the most commonly used in Europe. For fishery products, the EU has set an MPL for the Σ 6 iPCBs of 75 ng·g −1 wet weight, and our results in all marine species did not exceed this limit. Within the meat group, the sum of the six congeners on a median basis was 39.71 ng·g −1 lipid weight, and the highest concentration was found in veal fillet, followed by poultry composite samples having values in a similar range, whereas the pork meat showed the lowest concentrations. A higher variability level was registered within processed meats with concentrations ranging from 6.34 ng·g −1 lipid weight in raw ham to 46.64 ng·g −1 lipid weight in salami. Differences in PCB contamination between meat samples could be explained by a multitude of factors such as the environmental quality in which the animal lives, the contamination level in the feed, and the life span of the animals [25]. For processed meats, the different manufacturing processes altering the fat content can be responsible for the observed variability in the levels of these lipophilic chemicals [26]. For example, Kuzukiran and Filazi [27] found in fermented air-dried meat products (salami, soudjouk, and sausage) from a variety of animals a wide concentration range (0.407-3.936 ng·g −1 lipid weight) corroborating the assumption of an influence of meat processing in PCB levels. e study of the distribution profiles of iPCBs highlighted a difference between meats and processed meats (see Figure 1). For meat samples, the profile was dominated by PCBs 153 (23.3%), 138 (21.6%), and 101 (18.6%), followed by PCBs 52 (16.4%) and 28 (14.2%) having a similar percentage contribution and by PCB 180 whose concentrations accounted for 5.8% of all six congeners. Instead, in processed meat, the percentage contribution of PCB 52 to the sum of six indicators was the highest (24.9%), followed by PCB 101 (23.5%) and PCBs 138 (21.8%) and 153 (17.9%), while the lowest percentages were for the congeners 28 (5.5%) and 180 (6.4%). However, regardless of the observed profile differences, the values in  Figure 1). However, concentrations exceeding the EU maximum permissible level were found solely in cow milk cheese samples. Concerning hen eggs, the concentrations of iPCBs did not exceed the MPL of 40 ng·g −1 lipid weight established by the EU. PCBs 138 (35.8%) and 153 (18.2%) were the major contributors to the Σ 6 iPCBs, followed by PCB 28 which also showed a consistent percentage (17.3%), while the smallest percentage quotas were for PCBs 101 (7.1%), 52 (9.2%), and 180 (12.4%) (see Figure 1). e olive oil samples with concentrations of 2.82 ng·g −1 lipid weight due solely to PCBs 28 (51.4%) and 52 (48.6%) met the requirements specified in the EU legislation. In the remaining fats, PCBs levels ranging from 11.64 ng·g −1 lipid weight for mayonnaise to 14.73 ng·g −1 lipid weight for butter samples were below the MPL.

Dietary Exposure to iPCBs.
Consumer health protection starts with the development of measures to reduce exposure to toxic substances, so reference toxicological doses have been set to ensure that people do not exceed a certain body load. ndl-PCBs account for the majority of total PCBs in the environment and food [2] and thus are responsible for a significant quota of human exposure. Nonetheless, establishing a guide value for ndl-PCB remains complicated due to the simultaneous presence of the much more potent dl-PCBs which makes it difficult to determine the precise contribution of ndl-PCB to consumer exposure. However, an acceptable daily intake (ADI) of 10 ng·kg −1 bw·d −1 for their sum has been set at the national level in some countries [28,29]. Taking into account this reference value, dietary intakes in the total population and the adult and elderly groups exceeded the ADI, with a single exception represented by elderly females for whom a value lower but close to the fixed limit was calculated (see Table 3).
However, the intake values did not reach levels of statistical significance within sex/age groups (p > 0.05) suggesting similarity in rates and patterns of consumption for population classes examined. In fact, in adults and elders, fishery products dominated in the exposure with percentages ranging from 34.1% to 38.2%. e food group milk and dairy products was the second most important source of iPCBs intake contributing between 27.1% and 32.2% in male adults and female elders, respectively. Meat and processed meat showed a contribution lower than dairy products, with percentages varying from 21.9% to 28.3%. Hen eggs, olive oil, and other fats played a minor role with contributions ranging between 6.6% and 9.0% and 3.4% and 4.5%, respectively. Similarly, in total population, fish and other seafood categories were the major contributors to total intake (36.0%), followed by milk and dairy products (27.8%), meat and processed meat (25.5%), hen eggs (6.5%), and olive oil and other fats (4.2%) (see Table 4).
An analogous contribution pattern has been observed in earlier dietary intake studies carried out in Italy [14] and in other European countries such as France [30], Spain [31], and Germany [32]. Exceptions have been encountered for the Belgian adult population where dairy products and Table 3: Estimated daily intakes of iPCBs (ngkg −1 bwd −1 ) from the consumption of various categories of food for the total population and for adults (18-  Journal of Food Quality 5 fishery products contribute almost equally to the total dietary exposure to the iPCBs [33] and in the Austrian adult population with milk and dairy products being the major contributing food group to the total dietary intake of ndl-PCBs (50-55%) followed by the seafood and meat group [34]. Table 5 illustrates the comparison among iPCBs estimates of dietary exposure in Europe, as described in the literature. In Italy, for the general population in 2008, the mean estimated intake of iPCBs was 10.9 ng·kg −1 bw·d −1 for adults [14]. In the case of France, the mean exposure to the six indicators of PCBs was 2.71 ng·kg −1 bw·d −1 in the adult population [30], while another French study for the same population group reported a higher exposure level equal to 7.7 ng·kg −1 bw· d−1 [35]. In a German survey carried out with 20 participants employing the duplicate diet method, Fromme et al. [36] showed an average intake for the sum of iPCB indicators in adults of 5·6 ng·kg −1 bw·d −1 . e dietary exposure of Belgians to iPCBs ranged from 5.33 ng·kg −1 bw·d −1 to 6.05 ng·kg −1 bw·d −1 for the lower and upper bound approaches, respectively [33]. In Austria, the mean dietary intake of iPCBs was estimated between 2.64 and 3.19 ng·kg −1 bw·d −1 for adult women and men [34]. A dietary intake of 12.6 ng·kg −1 bw·d −1 was reported from Denmark [37], while a Norwegian study found a median intake of 4.4 ng·kg −1 bw·d −1 and 5.4 ng·kg −1 bw·d −1 for adult and elderly consumers, respectively [38]. e most recent studies available in the literature (Table S. 2) present the intakes of these chemicals from specific food products. For example, for fish and fish products from Greece, different authors have reported exposure levels near our results [39,40]. Higher levels have been reported in Poland for chicken meat and hen eggs [41], whereas in Italy, a lower intake from free-range hen egg consumption was registered [42]. It is obvious that the comparison of the exposure results, already complicated by different methodological choices (sampling strategies, analytical methods, targeted congeners, and the level used to assess the dietary intake), must also consider the population's dietary preferences and the rates of consumption. Consequently, it should be clearly pointed out that the comparison of the different exposure results has to be interpreted with care. However, the estimated total intake of the exposure levels calculated in the present study were in line with data extracted from EFSA [43], reporting estimates of ndl-PCB exposure for the European population from 4.3 to 25.7 ng·kg −1 bw·day −1 . Nevertheless, the intakes estimated near the higher part of the ADI of 10 ng·kg −1 bw·day −1 could

Conclusions
e present results could serve as a set of reference data for exposure to the adult and elderly population in Southern Italy, for which no data are currently available, except for one of our previous surveys [44]. Concentrations found in foodstuffs examined appear to be below the currently recommended European non commercialization values, with the exception of salami samples. e estimation of the dietary intake due to consumption of the entire group of these foods producing values above or near to the amount declared by various countries as tolerable daily intake seems to indicate that efforts must still be carried out to observe a major reduction of human exposure. is is also in consideration of the fact that estimation of the intake does not consider other food items consumed which can be a source of these harmful chemicals. It is hence worth emphasizing the need to monitor the residues of ndl-PCBs in all foods and especially in fishery products which are considered the major contributor to the dietary ndl-PCB intake.

Data Availability
Raw data were generated at the University of Bari, Veterinary Laboratory. Derived data supporting the findings of this study are available from the corresponding author M.M.S. on request.

Conflicts of Interest
e authors declare that they do not have any conflicts of interest.

Authors' Contributions
Conceptualization was conducted by Grazia Barone