Decreases in Concentrations and Human Dietary Intakes of Polychlorinated Biphenyls (PCBs) and Polybrominated Diphenyl Ethers (PBDEs) in Korean Seafood Between 2005 and 2017

Concentrations of polychlorinated biphenyls (PCBs) and polybrominated diphenyl ethers (PBDEs) were measured in 23 seafood species widely consumed by the Korean population in the periods of 2005–2007, 2010–2011, and 2015–2017. The Σ82PCB (sum of 82 PCB congeners) and Σ19PBDE (sum of 19 PBDE congeners) concentrations in the seafood samples of 2015–2017 were 0.06–6.69 ng/g wet weight and 0.01–1.60 ng/g wet weight, respectively. The Σ82PCB and Σ19PBDE concentrations in the samples were significantly correlated. Elevated PCB and PBDE concentrations were found in fatty fish, such as herring, mackerel, and tuna. The current human intakes of PCBs and PBDEs were much lower than the tolerable daily intake or lowest observed adverse effect level. The levels and human dietary intakes of PCBs and PBDEs in the 2015–2017 survey showed decreases of 17–73% and 57–86%, respectively, compared with those in 2005–2007 and 2010–2011 surveys. This indicates that global bans on PCBs and PBDEs have been effective, and their levels and human exposure to them have been gradually declining.

Polychlorinated biphenyls (PCBs) and polybrominated diphenyl ethers (PBDEs) are persistent organic pollutants (POPs) that accumulate in organisms and are biomagnified through food chains with elevated levels in top predators (Moon et al. 2010). These contaminants can cause adverse health effects, such as dermal toxicity, immunotoxicity, reproductive deficits, teratogenicity, and endocrine interferences (Ankarberg et al. 2007). A global ban was imposed on the use of POPs at the Stockholm Convention on POPs. PCBs and PBDEs were classified as POPs in 2001, respectively (Stockholm Convention 2008. In Korea, approximately 4300 tons of PCBs was used until the 1990s, after which the use of PCBs was banned ). In Korea, PBDEs have been widely used as brominated flame retardants. The usage of decabrominated diphenyl ether (deca-BDE) in Korea in 2002 was 12,324 tons, while that of pentabrominated diphenyl ethers (penta-BDEs) and octabrominated diphenyl ethers (octa-BDEs) together was 84 tons. Penta-and octa-BDEs were banned in 2008, but deca-BDE was unrestricted. However, owing to a voluntary phase out, the usage of deca-BDE in Korea was decreased to 1405 ton in 2010 (MOE 2015). Because PBDEs are not bonded into materials but simply blended with the polymers, they are likely to be released into the environment and enter biological tissues through food chains. These contaminants are still found in seafood and coastal environments, and pose a substantial risk to human health (Leng et al. 2014). The evaluation of the effectiveness of the aforementioned global bans and efforts in terms of how fast the POPs, such as PCBs and PBDEs, in environmental levels and their corresponding risks decline in the world has become a recent concern.
A temporal-trend study is an essential tool for assessing the effectiveness of legislative action on target contaminants in the environment (Jeong et al. 2016;Toms et al. 2018). However, reporting temporal trends of dietary exposure to POPs, such as PCBs and PBDEs associated with seafood consumption is insufficient (Domingo et al. 2008;Perelló et al. 2015;Sun et al. 2015;Toms et al. 2018). We conducted a nationwide baseline study for the periods of [2005][2006][2007] and 2010-2011, wherein we investigated seafood contamination with a focus on determining the concentration of POPs, including PCBs and PBDEs, in the seafood species most consumed by the Korean population. The human exposures to PCBs and dioxin-like PCBs (DL-PCBs) via seafood consumption were estimated to be 4.4 ng/kg bw/day and 0.9 pg-TEQ/kg bw/day, respectively ).
In 2010-2011, the exposures to PBDEs via seafood intake were estimated to be 0.4 ng/kg bw/day (MLTM 2012). To investigate the temporal trend of the levels and the total dietary exposures to PCBs and PBDEs, the same seafood groups (fish, crustaceans, cephalopods, and bivalves) were again collected and analyzed during 2015-2017. We investigated the current levels of 82 PCBs and 19 PBDEs in seafood and their dietary intake via seafood and compared the present situation with the previous baseline studies.

Sample Collection
Thirty-one marine species (fish, shellfish, crustaceans, cephalopods, bivalves, and gastropods) that were caught in the Yellow Sea, South Sea, and East Sea areas of Korea were selected from the most commonly consumed and commercially important species in Korea, as in our previous nationwide survey (MLTM 2012; ). Thirty-one marine species were annually collected twice or thrice during the period of 2015-2017 (n = 230) from the Busan cooperative fish market, which is the largest and representative fish market of South Korea, along with information regarding fishing areas and vessels. The collected samples were stored in a cooler box with ice and immediately transported to the laboratory. After removing the skin of the fish and cephalopods, their muscles and tissues were homogenized using an ultra-disperser (T25; IKA, Staufen, Germany). The shells of the bivalves, gastropods, and crustaceans were removed, and all the soft tissues were pooled and homogenized for analysis. For the PCB and PBDE analyses, two or three composite samples (n = 76-77) were prepared for each seafood item for each year from 2015 to 2017. Each composite sample consisted of more than 20 individual units for fish, crustaceans, and cephalopods, and 100 individual units for bivalves. Only the edible parts of each seafood were included in the composites.

Chemical and Instrumental Analyses
Eighty-two PCB congeners including 12 DL-PCB congeners  and 6 non dioxin-like (NDL)-PCB congeners 52,101,138,153,180), and 19 PBDE congeners  were analyzed in the seafood samples. A total of 82 PCB stock standard solutions, 35 13 C 12 -labelled PCBs for surrogate standards, and 7 13 C 12 -labelled PCBs for internal standards were purchased from Wellington (Guelph, ON, Canada). Nineteen PBDE stock standard solutions, 11 13 C 12labelled PBDEs for surrogate standards, and a 13 C 12 -labelled PBDE (BDE-138) for an internal standard were purchased from Wellington (Guelph, ON, Canada).
The analyses of the PCBs and PBDEs in the seafood were performed according to the methods described in previous papers ). The edible tissues (approximately 50 g) were digested in 200 mL of 1 N ethanolic KOH solution [ethanol (min. 99.5%) and KOH (min. 85%) from Wako Pure Chemicals, Tokyo, Japan] for 2 h by mechanical shaking. The surrogate standards for the PCBs and PBDEs were spiked into the samples before the digestion process. The alkaline solutions were extracted twice with 150 mL of hexane (Ultra residue analysis, J.T. Baker, Phillipsburg, NJ). The extracts were washed using purified water and then dried over 50 g of anhydrous Na 2 SO 4 and subsequently reduced to 10 mL via rotary evaporation. Purification of PCBs and PBDEs was performed by using a multilayer silica gel column containing Na 2 SO 4 anhydrous (4 g), 10% AgNO 3 silica gel (3 g), silica gel (0.6 g), 22% (w/w) H 2 SO 4 silica gel (3 g), 44% (w/w) H 2 SO 4 silica gel (4 g), silica gel (0.6 g), and 2% (w/w) KOH silica gel (2 g) with 150 mL of 15% DCM in hexane. The eluents were concentrated to approximately 200 µL for the instrumental analysis.
The analyses of PCBs and PBDEs were performed using a gas chromatograph (Agilent 6890, Agilent Technologies, USA) coupled with a high-resolution mass spectrometer (HRMS; JEOL 700D, Tokyo, Japan). For the PCBs, the capillary column used was a DB-5MS column (length 60 m, inner diameter 0.25 mm, film thickness 0.25 µm; J&W Scientific, Palo Alto, CA). For the PBDEs, a DB-5HT capillary column was used (length 15 m, inner diameter 0.25 mm, film thickness 0.10 µm; J&W Scientific). A detailed description of instrumental analyses for the PCBs and PBDEs is provided in the Supplementary materials.

Quality Assurance and Quality Control
All glassware was washed with detergent, rinsed with hot water and distilled water, and dried in a fume hood. Before use, the dried glassware was rinsed three times with acetone and hexane to minimize the method blank levels. Procedural blanks for every ten seafood samples were extracted, cleaned, and analyzed in the same manner as the real samples. The accuracy of the method was determined by analyzing the certified reference materials (CRMs) of a fish tissue (SRM 1946), from the US National Institute for Standards and Technology (NIST; Gaithersburg, MD).
The mean concentrations of the target compounds found in the seven CRM samples were 71-105% and 68-98% of the certified PCB and PBDE concentrations, respectively. The mean recoveries of the spiked surrogate internal standards of the PCBs and PBDEs were 81 ± 15% and 84 ± 12%, respectively. The calculated limits of detection (LOD, i.e., the concentrations providing a signal-to-noise ratio of 3) were 0.001-0.023 ng/g for individual PCB congeners and 0.002-0.033 ng/g for individual PBDE congeners. In the procedural blanks, some PCB and PBDE were detected at negligible levels (< 10% of LOD).

Calculation of Dietary Intakes
The intake of an "average" Korean of a group of PCBs or PBDEs (ng/kg bw/day) in a seafood species was calculated by multiplying the mean concentration of the PCB or PBDE groups (ng/g wet weight) in the seafood species of interest by the mean daily per capita intake of the seafood species in Korea (g/day) and dividing the result by the mean body weight for Koreans (kg body weight). The concentrations of undetected compounds were considered to be half of the respective LOD. Daily consumption rates of each seafood for the general population of Korea were obtained from the National Health and Nutrition Survey (MOHW 2011(MOHW , 2015. Table 1 lists the concentrations of Σ 82 PCB (sum of 82 PCB congeners), Σ 6 NDL-PCB (sum of six NDL-PCB congeners), Σ 12 DL-PCB (sum of 12 DL-PCB congeners), Σ 19 PBDE (sum of 19 PBDE congeners), and Σ 7 PBDE (sum of  in 31 seafood species widely consumed in Korea during 2015-2017. The concentrations of Σ 82 PCB were 0.06-6.69 ng/g wet weight, and CB-153 was present in the highest proportion (14 ± 2.9%) among the PCB congeners, followed by CB-138 (7.2 ± 1.8%), CB-118 (5.6 ± 1.6%), and CB-101 (5.0 ± 1.7%). Herring, eel, tuna, and Spanish mackerel presented relatively high concentrations (6.69, 4.41, 3.73, and 3.68 ng/g wet weight, respectively), whereas shrimp and abalone presented low concentrations (0.06 and 0.07 ng/g wet weight, respectively).

Concentrations and Profiles of PCBs
The levels of Σ 6 NDL-PCB were 0.02-2.23 ng/g wet weight, and CB-153 contributed more than 41% to the Σ 6 NDL-PCB (Fig. 1). The levels of Σ 6 NDL-PCB contributed 35% to the Σ 82 PCB, and the levels of Σ 6 NDL-PCB were strongly correlated with those of Σ 82 PCB (r = 0.997, and p < 0.01). Herring, eel, tuna, and Spanish mackerel also presented the highest concentrations of Σ 6 NDL-PCB. None of the samples showed concentrations of Σ 6 NDL-PCB above the European Union's threshold concentrations, which are set at 75 ng/g wet weight for fish and 300 ng/g wet weight for eels (Official Journal of the European Union 2011). The levels of Σ 6 NDL-PCB observed in this study were lower than those in the Polish Baltic fishing area (Piskorska-Pliszczynska et al. 2012), Catalonia of Spain (Perelló et al. 2015), four areas (South China Sea, Bohai Sea, East China Sea, and Yellow Sea) of China (Liu et al. 2011;Qian et al. 2017), three areas (Adriatic, Ionian, and Tyrrhenian seas) of Italy (Miniero et al. 2014), and a nationwide survey of France (Arnich et al. 2009), as shown in Table 2.
The levels of Σ 12 DL-PCB were 0.01-0.71 ng/g wet weight and CB-118 contributed 55% to Σ 12 DL-PCB. The levels of Σ 12 DL-PCB contributed 10% to the Σ 82 PCB, and the levels of Σ 12 DL-PCB were significantly correlated with those of Σ 82 PCB (r = 0.980, and p < 0.01). Based on the World Health Organization toxic equivalent factor (WHO-TEF) values (2005), Σ 12 DL-PCB concentrations were 0.001-0.721 pg-TEQ/g wet weight, and the dominant congeners were CB-126 (89 ± 6.2%), which has the highest toxic equivalent factor (TEF 0.1) among the DL-PCBs, followed by CB-169 (8.4 ± 6.3%). Spanish mackerel (0.721 pg-TEQ/g wet weight), eel (0.679 pg-TEQ/g wet), herring (0.614 pg-TEQ/g wet weight), and tuna (0.475 pg-TEQ/g wet weight) presented high concentrations of DL-PCBs, which was similar to the case of the NDL-PCBs. The European Union's threshold concentrations of the sum of dioxins (PCDD/ Fs + DL-PCBs) have been set as 6.5 pg-TEQ/g wet weight for fish and 10 pg-TEQ/g wet weight for eels (Official Journal of the European Union 2011). The contributions of four seafood species (Spanish mackerel, eel, herring, and tuna) to the threshold concentrations were 6.8-11%, which indicates that the DL-PCB levels in seafood in Korea were safe. The levels of DL-PCBs in the present study were comparable to or lower than those in four major cities (Malmoe, Gothenburg, Uppsala, and Sundsvall) of Sweden (Törnkvist et al. 2011), and Catalonia of Spain (Perelló et al. 2015), the Ariake Sea of Japan (Naito et al. 2003), three areas (Adriatic, Ionian, and Tyrrhenian seas) of Italy (Miniero et al. 2014), and the Polish Baltic fishing area (Piskorska-Pliszczynska et al. 2012), as shown in Table 2.
The concentrations of Σ 7 PBDE (0.01-1.04 ng/g wet weight) in seafood in our study were similar to those reported in Catalonia of Spain (0.012-1.62 ng/g wet weight, Domingo et al. 2008) and Italy (0.06-2.98 ng/g wet weight; Miniero et al. 2014) but were lower than those reported in China (0.2-476 ng/g wet weight, Liu et al., 2011), the United States (1-300 ng/g wet weight; Gandhi et al. 2017), and Australia (0.61-21.39 ng/g wet weight; Shanmuganathan et al. 2011). Table 3 lists the exposures of the Korean population to total PCBs and PBDEs via the consumption of seafood. The body weight of 59.8 kg was used to calculate the daily intakes for an average Korean. For the Korean population, the dietary intakes of total PCB and six NDL-PCB were 45.2 ng/day (0.76 ng/kg bw/day) and 15.7 ng/day (0.26 ng/kg bw/day), respectively. The high contributions of the PCB and six NDL-PCB intakes corresponded to eel and mackerel, with approximately 40% of the total. The dietary DL-PCB intake was 7.27 pg-TEQ/day (0.12 pg-TEQ/kg bw/day), and the high contributions of DL-PCB intake were eel and mackerel, which accounted collectively for ~ 45% of the total. The squid, anchovy, tuna, Spanish mackerel, and saury demonstrated moderate contributions and collectively accounted for 30% of the total. This is associated with the relatively high consumption of this species and the higher concentrations of PCBs in it. In contrast, herring presented the highest concentration but was a minor contributor owing to its low consumption of 0.1 g/day, in Korea.

Human Exposure via Seafood Consumption
For the assessment of risk assessment of PCBs through seafood consumption, we compared the TDI (tolerable daily intakes) and TWI (tolerable weekly intakes) established by Korea, Denmark, and the WHO. The total PCB intake (0.76 ng/kg bw/day) was ~ 3.8% of the TDI (20 ng/kg bw/ day) of the WHO in 2003 and 0.8% of the TDI (100 ng/kg bw/day) of the Danish Veterinary and Food Administration in 2012 (Carlsson et al. 2014). The DL-PCB intake (0.12 pg-TEQ/kg bw/day) was 3.0% of the TDI (4 pg-TEQ/kg bw/day) of dioxins (PCDD/Fs + DL-PCBs) of the WHO and 6.0% of the TWI (Tolerable Weekly Intakes, 14 pg-TEQ/kg bw/ week) of the Korea Food and Drug Administration (KFDA) . These results suggest that the human exposure to the PCBs and DL-PCBs through seafood intake in Korea is low. If the PCDD/F concentrations are added to the DL-PCB concentrations in view of their dioxin-like toxicity, the combined concentrations do not exceed the TDI of WHO and TWI of KFDA, because the dietary intake contribution of the PCDD/Fs (average 30% of DL-PCBs) was lower than that of the DL-PCBs (Lee et al. 2007;).
The total PBDE and sum of seven PBDE dietary intakes were 9.86 ng/day (0.17 ng/kg bw/day) and 5.96 ng/day (0.10 ng/kg bw/day), respectively. The high contributions to the total PBDE and sum of seven PBDE intakes were from eel and mackerel, which comprised approximately 40-45% of the total. The squid, tuna, Spanish mackerel, saury, and anchovy also presented moderate contributions and collectively accounted for 30% of the total. To date, the TDI levels have not been set for PBDEs by the European Union due to limited available data. A lowest observed adverse effect level (LOAEL) of 1 mg/kg/day has been reported for the most sensitive toxic effects of PBDEs (Martí-Cid et al. 2007). Benchmark doses per day (computed and estimated "safe levels" of the PBDEs) are currently 309 µg/kg bw/day for BDE-47, 12 µg/kg bw/day for BDE-99, and 83 µg/kg bw/ day for BDE-153 (EFSA 2011). The European Food Safety Authority concluded that only BDE-99 would be a potential health concern for the European population (EFSA 2011).
In this study, the dietary intake of seven PBDEs, including BDE-47, BDE-99, and BDE-153 (0.10 ng/kg bw/day), was less than 1% of the safe levels. According to the LOAEL and safe levels, the PBDE intake through seafood results in a safety factor of various orders of magnitude.

Temporal Trends of Concentrations of PCBs and PBDEs
To investigate the temporal trends of PCBs and PBDEs in seafood (for the same 23 species), the concentrations, Table 3 Estimated human exposure to PCBs and PBDEs by the general population of Korea, 2015-2017 a Sum of 6 non-dioxin-like (NDL)-PCB congeners  b Sum of 12 DL-PCB congeners  c Sum of 7 PBDE congeners  d The body weight, 59.8 kg, was used to calculate the daily intakes for an average Korean  The rate of decrease of the average concentrations of DL-PCBs during the 12-year period was 73%. Decreases in DL-PCB levels were observed in 22 of 23 species, and a decline in DL-PCB levels of more than 80% was observed in six species (hairtail, mackerel, tuna, yellow croaker, crab, and shrimp). Kim and Yoon (2014) reported that average concentrations of dioxins/furans in Korean air samples decreased 15 times from 0.79 pg-TEQ/m 3 in 1999 to 0.052 pg-TEQ/ m 3 in 2009. Those of DL-PCBs also decreased by ten times during the periods of 1999. Jeong et al. (2016 also reported that levels of DL-PCBs in finless porpoises inhabiting Korean coastal waters presented a significant decrease of 60-68% between 2003 and 2010 and regulations on POPs have thus been effective for marine mammals in Korea. Hence, these results indicate that a reduction in PCB pollution might have occurred in Korea. Similar decreasing trends in PCB levels have recently been reported in Indo-Pacific humpback dolphins (Sousa chinensis) collected in South China during surveys in 2004-2009 and 1995-2001, which is a representative biomonitor for contaminants in aquatic ecosystems (Wu et al. 2013). Sun et al. (2015) also reported that the PCB concentrations in two fish species from the Pearl River Estuary (South China) in 2013 were significantly lower than those in 2005 (p < 0.05), and declines of 61-80% were observed in two fish species during the 8-year period. This result indicates that legislative actions on POPs such as PCBs have been effective in marine environments.
The levels of PBDEs in seafood were 0.02-1.55 ng/g wet weight (mean: 0.42) in 2010-2011 and 0.01-1.60 ng/g wet weight (mean: 0.35) in 2015-2017. The rate of decrease during the 7-year period was 17% for the average concentrations of PBDEs. A decrease in PBDE levels was observed in 14 of 23 species, and a decline in PBDE levels of more than 50% was observed in six species (anchovy, anger fish, mackerel, rockfish, crab, and shrimp). These results may reflect the ban on PBDEs, and their decreased use due to the effectiveness of regulations and controls. In Korea, penta-and octa-BDE were banned in 2008, and usage of deca-BDE decreased by nine times during the period of 2002(MOE 2015. Gauthier et al. (2008) evaluated the PBDEs in herring gull eggs obtained from the Laurentian Great Lakes . PBDE congeners derived mainly from  Table 1 penta-BDE and octa-BDE mixtures, i.e., BDE-47, -99, and -100, presented rapid increases up until 2000; however, there was no increasing trend after 2000, due to the regulation of penta-and octa-BDE mixtures. Since their phase-outs in the 2000s, PBDE levels in herring gull eggs in 2012-2013 were 30% lower than those in herring gull eggs in 2006(Su et al. 2015. The declining PBDE trend for edible portions of fish was observed in the Great Lakes between 2006/07 and 2012 (Gandhi et al. 2017). The levels of PBDEs in environmental media are expected to decline further due to regulatory actions.

Temporal Trends of Human Exposure
The total dietary intake of DL-PCB was 45.6 pg-TEQ/day in the 2005-2007 survey, 11.3 pg-TEQ/day in 2010-2011, and 6.18 pg-TEQ/day in 2015-2017 (Fig. 3) Fig. 3. A (e) (f) Fig. 3 Temporal trend of (a) consumption for each seafood species and (b) total seafood consumption of Korean population, and dietary exposure of Korean population to DL-PCBs (c) by each seafood con-sumption and (d) by total seafood consumption, to PBDEs (e) by each seafood species and (f) by total seafood consumption. Species names on the x-axes were presented in the same order as in Table 1 higher intake of PBDEs of 20.1 ng/day, was found in the 2010-2011 survey (MLTM 2012), while in the subsequent survey investigation, the estimated intake of PBDEs declined to 8.46 ng/day in the 2015-2017 survey. The rate of decrease in the dietary intake of PBDEs during the 7-year period was 58%. Significant decreases of 69-92% were found in the case of anchovy (6.16 ng/day vs. 0.51 ng/day) and mackerel (6.05 ng/day vs. 1.87 ng/day). The decreasing rates of the two species were 51-66% in PBDE concentrations and 37-75% in their consumption. A decrease in seafood intake was found in 14 of 23 species, and the rate of decrease in seafood consumption was 27%. Domingo et al. (2008) reported that the dietary intake of PBDEs through fish and shellfish intakes in Spain decreased by 14% between 2000 and 2006 due to a 26% decrease in fish and shellfish consumption. Toms et al. (2018) analyzed serum pool concentrations of BDE-47, -99, -100, and 153 from 2002/03 to 2012/13, and reported that temporal trends were ageand congener-specific. Only the two youngest age groups (0-4 years old and 5-15 years old) demonstrated statistically significant decreases over the time, probably because of a decline in infant and toddler exposures to PBDEs in the indoor environment as the use of PBDEs in consumer products has been phased out; however, older age groups showed no significant trend with time. These results indicate that legislative actions on POPs have been effective in decreasing human exposure to POPs through seafood.

Conclusions
The exposure of the Korean population to PCBs and PBDEs via seafood consumption has decreased significantly in 2015-2017 compared with those in 2005-2007 and 2010-2011. Significant decreases in PCB intakes were found in fatty fish, such as mackerel, tuna, hairtail, yellow croaker, and anchovy. This is associated with the decrease in the concentrations of PCBs and PBDEs in seafood between the surveys, due to regulatory actions. This indicates that the levels and human exposures to PCBs and PBDEs from seafood consumption are expected to continue to decrease.