Polychlorinated biphenyls (PCB) cover a group of 209 different PCB congeners which can be divided into two groups according to their toxicological properties. One group, consisting of 12 congeners, show toxicological properties similar to dioxins, is therefore termed "dioxin-like PCB" (DL-PCB), and these have been included in the "Risk Assessment of Dioxins and Dioxin-Like PCBs in Food" performed by the EU Scientific Committee on Food (SCF). The other PCB, referred to as "non dioxin-like PCB" (NDL-PCB), have not been previously evaluated by the SCF or EFSA. Both groups of PCB, NDL-PCB as well DL-PCB, are usually found in feed and food. PCB were widely used in a number of industrial and commercial applications. It is estimated that more than 1 million tons of technical PCB mixtures were produced world-wide since their first commercial use in the late 1920s. Although produced by comparable production processes, technical PCB mixtures contain both DL and NDL-PCB and may vary considerably with respect to their congener composition due to differences in the amount of chlorine and the reaction conditions applied. Moreover, technical PCB mixtures contain other dioxin-like compounds as impurities, such as polychlorinated dibenzofurans (PCDF). The different compositions as well as the presence of toxicologically relevant impurities may have a significant impact on the results of toxicological studies with technical PCB mixtures. Although the manufacture, processing and distribution of PCB has been prohibited in almost all industrial countries since the late 1980s, their entry into the environment still occurs, especially due to improper disposal practices or leaks in electrical equipment and hydraulic systems still in use. PCB are highly persistent and are globally circulated by atmospheric transport and thus are present in all environmental media. Data on the occurrence of NDL-PCB in food and feed have been reported in different ways for example as the sum of three PCB congeners (PCB 138, 153 and 180), as the sum of six PCB congeners (PCB 28, 52, 101, 138, 153, 180) often referred to as indicator PCB or as the sum of seven (sum of six indicator PCB plus PCB 118). This lack of consistency often hampers a direct comparison of occurrence data. The Panel decided to use the sum of the six indicator PCB as the basis for the evaluation in this opinion, because these congeners are appropriate indicators for different PCB patterns in various sample matrices and are most suitable for a risk assessment of NDL-PCB on the basis of the available data. The Panel noted that the sum of the six indicator PCB represents about 50% of total NDL-PCB in food. Following exposure of farm animals, NDL-PCB will accumulate in meat, liver and particularly in fat tissues. In addition, NDL-PCB will be transferred into milk and eggs, and levels in these products will reach a steady state following exposure over a period of several weeks. PCB 138 and 153, both with six chlorine atoms, show the highest carry-over into milk and eggs, in the order of 50-60%. After cessation of exposure, levels in eggs and milk initially decrease rapidly to about 50%, followed by a slower elimination phase. In fattened animals like calves, piglets, and poultry, and also farmed fish, no steady state is obtained, due to the fact that these animals are slaughtered at a young age. For risk assessment of domestic animals, the Panel compared the effect concentrations in the experimental diet with the NDL-PCB concentration in animal feed. Following a conservative approach, the 90th percentile of the sum of six NDL-PCB in compound feed, 0.02 mg/kg feed was taken as the point of comparison. This figure corresponds to about 0.04 mg total NDL-PCB, which is more than two orders of magnitude below the concentrations causing effects in most domestic animals studied. Mink are usually given feed based on fish. The 90th percentile of the sum of the six NDL-PCB in fish and fish products is 0.067 mg/kg corresponding to 0.13 mg/kg total NDL-PCB. This is only about five times below the concentration of PCB in feed that produced pronounced effects on reproduction in mink. The Panel therefore concluded that current background levels of NDL-PCB in animal feed are of no health concern for most domestic animals, with the possible exception of mink. Congener patterns in feed, particularly that of plant origin and in edible tissue may differ considerably. Due to the different sources of contamination, different origins of the feed and of food commodities, there is generally no correlation between the concentrations of NDL-PCB and DL-PCB Toxic Equivalents (TEQ) or the total TEQ (polychlorinated dibenzo-p-dioxins (PCDD), polychlorinated dibenzofurans (PCDF) and DL-PCB) with the exception of samples where the circumstances of contamination are known. More than 90% of the NDL-PCB exposure in the general population is via food. Average daily dietary intakes of total NDL-PCB can be estimated to be in the range of 10-45 ng/kg body weight (b.w.) per day. Limited exposure data for young children, up to six years of age, indicates that the average intake (breastfeeding excluded) of total NDL-PCB is about 27-50 ng/kg b.w. per day. However, where data on both adults and children within a specific population were available, in general children had exposure levels 2.5 fold higher than adults. In specific subpopulations with high dietary PCB exposure such as Baltic Sea fishermen the daily intake from fish of the sum of the six NDL-PCB could be about 40 ng/kg b.w., corresponding to an intake of total NDL-PCB of 80 ng/kg b.w. per day before taking into account the rest of the diet. Breastfed infants are a group of high NDL-PCB intake which might be two orders of magnitude higher than adult exposure. Other routes of exposure such as ambient and indoor air, dust and soil, do not usually contribute significantly to the body burden of the general population. However, there are situations in which contribution from contaminated indoor air could be considerable. Technical PCB mixtures used in toxicity studies contain both NDL-PCB and dioxin-like compounds such as DL-PCB. These mixtures exert a variety of toxicological effects such as effects on liver, thyroid, immune function, reproduction and behaviour as well as carcinogenicity. The adverse effects reported in laboratory animals following exposure to individual NDL-PCB were effects on the thyroid, liver and brain biochemistry, as well as immunotoxicity, oestrogenicity, and reproductive and neurodevelopmental effects. The latter effects are particularly found in the offspring of rodents following in utero exposure. However, these effects are not all specific for NDL-PCB but are also to be seen following exposure to polychlorinated dibenzo-p-dioxins, polychlorinated dibenzofurans, and DL-PCB. Several NDL-PCB congeners are metabolised to hydroxy-PCB and/or methylsulfonyl-PCB. Some of these metabolites may contribute to hormone-like effects seen with PCB. Results of in vitro and in vivo genotoxicity studies indicate that PCB are not mutagenic at the gene or chromosome level. Some NDL-PCB, in particular the lower chlorinated congeners, caused DNA damage, probably resulting from the formation of reactive oxygen species. In two-stage initiation-promotion studies, technical PCB mixtures containing NDL-PCB as well as DL-PCB promote liver carcinogenesis in rats, following initiation with genotoxic carcinogens. Data from animal experiments with several technical mixtures (Aroclor 1016, 1242, 1254 and 1260) indicate that PCB can cause liver and thyroid neoplasms in rats. The International Agency for Research on Cancer (IARC) classified PCB in Group 2A (probably carcinogenic to humans), based on limited evidence in humans and sufficient evidence in animals. Evaluation of the cancer studies in rats with technical PCB mixtures, and comparison with data obtained with TCDD, indicate that the dioxin-like components in technical PCB mixtures are likely to be responsible for the carcinogenic response of these mixtures. No peer reviewed data are available on the carcinogenicity of individual NDL-PCB congeners. Occupational exposures to PCB have been reported to be associated with an increased risk of cancer of the digestive system and possibly other sites. Some studies suggest that environmental PCB exposure may be linked to the development of breast cancer, although perhaps only in certain vulnerable sub-groups. Among non-cancer effects reported to be associated with environmental PCB exposure, adverse reproductive outcomes, delayed neurodevelopment and impairment of the immune system during development are considered to be the most important. The epidemiological studies however do not allow an estimation of the toxicity that may specifically be attributed to the NDL-PCB. Benchmark dose calculations have been based on human studies on developmental neurotoxicity and immunotoxicity after perinatal exposure to total DL and NDL-PCB. The 95% lower confidence limit of benchmark dose (BMDL) of approximately 1 mu g PCB/g lipid is only about four times higher than the current median concentration in human milk. The Panel noted that the comprehensive toxicological database on health effects of technical PCB mixtures was not suitable for the separate assessment of NDL-PCB, and that the human data on exposure to environmental mixtures containing PCB could not differentiate between the effects of NDL-PCB and DL-PCB and polychlorinated dibenzo-p-dioxins/polychlorinated dibenzofurans. Therefore, in its assessment the Panel concentrated on the toxicological information available for individual NDL-PCB congeners. Although the absence of mutagenicity indicates that a threshold approach is appropriate for the hazard characterisation, the toxicological database however, was considered to be too limited to allow the establishment of a health based guidance value for NDL-PCB. The Panel therefore decided to perform its health risk characterisation on the basis of a margin of exposure approach. The most sensitive effects seen in studies with individual NDL-PCB congeners in experimental animals were liver and thyroid toxicity. The NOAELs for these effects in 90-day rat studies with the individual NDL-PCB congeners PCB 28, 128, and 153 were in the range of 30-40 mu g/kg b.w. per day. For compounds that accumulate in the body, such as NDL-PCB, evaluations based on body burden (BB) calculations are considered more appropriate than evaluations based on the external dose. The Panel therefore applied a body burden approach to the results of the 90-day rat studies, and estimated body burdens at the "no observed adverse effect level" (NOAEL) of 400, 800, and 1,200 mu g/kg b.w. for PCB 28, 128, and 153, respectively. The Panel compared estimated body burdens at the NOAEL for different effects in animals with the estimated median human body burden derived from the analyses of human milk. The "margin of body burdens" at the NOAEL (NOAEL MoBB) were calculated by dividing the estimated animal body burden with the estimated median human body burden. NOAEL MoBBs of 900, 6,300, and 85 were obtained for the effects on liver and/or thyroid of PCB 28, 128, and 153, respectively. Although PCB 28, 128, and 153 showed similar potencies in the 90-day toxicity studies, PCB 153 had the lowest NOAEL MoBB due to its abundance in human tissues. Application of the same approach to experimental animal studies on reproductive and developmental effects, oestrogenicity, thyroid effects and effects on the immune system and the developing nervous system, revealed NOAEL MoBBs that were higher than 1,600. In order to evaluate the impact of exposure to total NDL-PCB, the Panel noted that the available toxicological database on NDL-PCB covered a number of congeners present in food and human tissues. Considering that the "lowest observed adverse effect level" BB for the most sensitive effects (liver, thyroid) were 10 times higher than the NOAEL BB (400, 800, and 1,200 mu g /kg b.w. for PCB 28, 128, and 153, respectively), the Panel chose an overall body burden of 500 mu g /kg b.w. as a representative conservative body burden at the NOAEL (NOAEL BB) for all individual NDL-PCB and for the sum of NDL-PCB occurring in human tissues. Based on the median total concentration of all NDL-PCB measured in human milk sampled in European countries of about 240 ng/g fat, and assuming 20% fat content in the human body, a median human body burden of about 50 mu g/kg b.w. was estimated. Consequently the overall NOAEL MoBB is about 10. Although this margin appears rather small, it should be stressed that the endpoints considered in the evaluation of the individual NDL-PCB congeners, can also be observed after treatment with polychlorinated dibenzo-p-dioxins, polychlorinated dibenzofurans, or DL-PCB. Since a number of these latter compounds have relatively high potencies for these effects in rats, minor contamination (in the range of 0.1%) of the NDL-PCB congeners studied with potent dioxin-like compounds might be sufficient to explain the effects observed. Thus, any estimate of a NOAEL for NDL-PCB is hampered by the uncertainty in the extent to which NDL-PCB congeners might have been contaminated with polychlorinated dibenzofurans and/or DL-PCB. Therefore the "true" NOAEL MoBB for NDL-PCB might be larger. On the other hand, the MoBB was calculated on the basis of the median concentrations of NDL-PCB in human milk, and some populations in Europe may have considerably higher body burdens. During the nursing period, breastfed infants may have daily intakes, on a body weight basis, of NDL-PCB estimated to be about two orders of magnitude higher than the average adult intake. This elevated intake by the infants is related to the mother's long-term intake of NDL-PCB with food. However, the subtle neurodevelopmental effects that were reported in some studies of human infants were mainly associated with exposure to a mixture of NDL-PCB, DL-PCB, and polychlorinated dibenzo-p-dioxins/polychlorinated dibenzofurans, and any causal role of NDL-PCB is unclear. The Panel noted that in many other studies of infants, breastfeeding was associated with beneficial effects, in spite of the contaminants present in human milk. In conclusion, no health based guidance value for humans can be established for NDL-PCB because simultaneous exposure to NDL-PCB and dioxin-like compounds hampers the interpretation of the results of the toxicological and epidemiological studies, and the database on effects of individual NDL-PCB congeners is rather limited. There are however indications that subtle developmental effects, being caused by NDL-PCB, DL-PCB, or polychlorinated dibenzo-p-dioxins/polychlorinated dibenzofurans alone, or in combination, may occur at maternal body burdens that are only slightly higher than those expected from the average daily intake in European countries. Because some individuals and some European (sub)-populations may be exposed to considerably higher average intakes, a continued effort to lower the levels of NDL-PCB in food is warranted.
