Oral bioaccessibility and human risk assessment of organochlorine pesticides (OCPs) via fish consumption, using an invitro gastrointestinal model
Hong-Sheng Wang,Yin-Ge Zhao,Yu-Bon Man,Chris K.C. Wong,Ming-Hung Wong
DOI: https://doi.org/10.1016/j.foodchem.2011.02.035
IF: 8.8
2011-01-01
Food Chemistry
Abstract:The oral bioaccessibility of dichlorodiphenyltrichloroethane (DDT) and hexachlorocyclohexanes (HCHs) in twenty fish species were evaluated using an in vitro gastrointestinal model. Relatively high concentrations of ∑HCH (0.33–9.88 ng g −1 , mean 1.57 ng g −1 ) and ∑DDT (0.74–131 ng g −1 , mean 12.2 ng g −1 ) were observed in all market fish. The average bioaccessibilities for ∑HCHs were 3.35% and 8.73% in gastric and intestinal conditions, respectively, and for ∑DDTs were 5.48% and 17.6%, respectively. Significant ( p < 0.05) correlations were observed between OCP congener digestible concentrations in fish muscle and their corresponding concentrations in human tissues of Hong Kong residents. The health risk assessments based on solvent and digestible concentrations were found to be contradictory due to the fact that only a limited proportion of OCP was bioaccessible. It was concluded that human health risk assessment based on solvent concentrations should be modified by taking bioaccessibility of the contaminant into account. Keywords DDTs HCHs Oral bioaccessibility Fish muscle Risk 1 Introduction Organochlorine pesticides (OCPs) including dichlorodiphenyltrichloroethane (DDT) and hexachlorocyclohexanes (HCHs) are among the most widely applied chemicals in the world. These chemicals are hard to degrade and hence are capable of remaining in the environment up to decades ( Gonzalaz-Rodriguez, Rial-Otero, Cancho-Grande, & Simal-Gandara, 2008 ). Presently, fairly high levels of HCHs and DDTs are still being detected in various environmental media of Hong Kong ( Cheung, Leung, Kong, & Wong, 2007; Leung, Kwok, Nie, Cheung, & Wong, 2009; Wong, Leung, Chan, & Choi, 2005 ). For example, DDTs concentrations in human milk collected from Hong Kong ranked number 5 among 19 examined countries or regions ( Wong et al., 2005 ). The main pathway for human exposure to OCPs is the consumption of contaminated seafood. It was reported that 75% of DDT intake by the Chinese was attributed to the consumption of marine food ( Nakata et al., 2002 ). Dougherty et al. (2000) also revealed that freshwater fish consumption accounted for 75% of the exposure to OCPs such as DDTs (whereas other food items contributed not more than 5%) across the United States. For Hong Kong, the concentrations of DDT and HCH in human milk were significantly correlated with fish consumption frequency of the donors ( Wong et al., 2005 ). Therefore, the consumption of OCP contaminated fish has led to growing public health concerns, particularly for high risk groups, such as pregnant women and children. Fish and shellfish are major components in the diet of Hong Kong residents with an average consumption rate of 164.4 g d −1 per person ( Dickman & Leung, 1998 ), which is higher than the fish consumption rate (142.2 g d −1 ) set by USEPA for subsistence consumers ( USEPA, 2000 ). In addition, OCP concentrations in fish collected from Hong Kong have been found to be very high: DDT ranged from 1.10 to 1018 ng g −1 (wet wt, ww) with 45% of muscle samples of freshwater fish species and 70% muscle samples of marine fish species exceeding the screening value for DDTs set by USEPA ( Cheung et al., 2007 ). It has been indicated that the main contribution to DDTs among different food items (including meat, fish, vegetable and others) was from seafood for Hong Kong students ( Chung, Kwong, & Yau, 2008 ), and fish contributed the largest portion of the dietary intake of DDTs (57%) among different seafood types in South China ( Guo, Wu, Shen, & Zeng, 2010 ). Therefore, serious attention should be paid on OCP contamination in fish in order to safeguard the public health of Hong Kong residents. It has been observed that solvent concentrations of pollutants may overestimate the absorbed amount through oral ingestion ( Ramesh et al., 2004 ). To accurately assess the oral exposure risk, the two most critical issues are bioaccessibility and bioavailability. At first, contaminants become accessible after gastrointestinal digestion, then after being absorbed and metabolized by the intestinal epithelium, contaminants will enter the blood circulation system and be transported to target organs ( Lopez-Perez et al., 2006; Yu et al., 2010 ). Therefore, in vitro methods have been used to estimate the bioaccessibility of contaminants in food ( Xing, Yang, Chan, Tao, & Wong, 2008; Yu et al., 2010 ). The bioaccessibility of a contaminant in food is defined as the fraction of the contaminant mobilized from the food matrices during gastrointestinal digestion ( Oomen, Sips, Groten, Sijm, & Tolls, 2000 ). In terms of OCPs, there have only been a few studies concerning their bioaccessibilities in soil samples. It was reported that recoveries of pesticides from certified reference materials (CRM) 805–050 (soil sample) following extraction with gastric fluid were below the limit of detection, while following intestinal extraction ranged from 5.3% to 12.8% ( Scott & Dean, 2005 ). Tao et al. (2009) suggested that the bioaccessibilities of OCP congeners varied from 4% to 97% depending on the properties of soils and compounds. However, there is no published data concerning the bioaccessibility of OCPs via fish consumption. In the present study, the bioaccessibility of OCPs in twenty species of common edible fish collected from Hong Kong markets were investigated. More specifically, the objectives were: (1) to develop an in vitro digestion method to analyse the bioaccessibility of OCPs in fish muscle and their contributing factors such as chemical-physical parameters; and (2) to assess potential health risks and related consumption advisories based on total and digestible OCP concentrations in fish muscle. To our knowledge, this is the first study to investigate the bioaccessibility of OCPs via fish consumption using the in vitro gastrointestinal model. 2 Materials and methods 2.1 Sample collection and preparation Twenty species of common edible fish which included ten freshwater and ten marine species were collected from Hong Kong local food markets during May to November 2009. Information including common and scientific name, length, weight, water and lipid content, feeding mode and origin of the fish are listed in Table S1 . There were altogether 279 fish samples. All samples were packed with aluminium foil, transported to the laboratory and kept at −20 °C. The muscle of fish (including axial and ventral muscle) was removed, freeze-dried, and homogenized by grounding into powder prior to analyses. 2.2 Digestible fraction The detailed procedures of digestible fraction followed the method described by Goni, Serrano, and Saura-Calixto (2006) and our previous study ( Wang et al., 2010b ) with slight modification. Briefly, freeze-dried samples (3 g) were first placed in 30 ml of synthetic gastric juice (2.0 g l −1 pepsin in 0.15 M NaCl, acidified with HCl to pH 1.8) and shaken at 100 rpm for 2 h at 37 °C. The mixture was then centrifuged (15 min, 37 °C, 1500 rpm) and all the gastric suspensions were transferred to centrifuge tubes for further clean up and analyses. Artificial intestinal juice (30 ml, 2.0 g l −1 pancreatin, 2.0 g l −1 amylase and 5 g l −1 bile salts, in 0.15 M NaCl, pH 6.8) was added. Then, the mixture was re-suspended and shaken at 30 rpm for 6 h at 37 °C. Finally, the tubes were centrifuged at 1500 rpm at 37 °C for 15 min to separate supernatant and solids. The entire digestion process was performed in capped Teflon centrifuge tubes (50 ml) in the dark to simulate the anaerobic condition of the stomach. Both gastric and intestinal supernatants were filtered though 0.45 Millipore filters. Then, the surrogate (4,4′-dichlorobiphenyl) and acetone (10 ml) were added. The mixture was extracted using a separatory funnel, with hexane (50 ml) twice and DCM (50 ml) once. Concentrated sulfuric acid (2–3 ml) was subsequently added to the organic phase and the mixture was shaken vigorously and left at room temperature (25 °C) for several hours to remove biomolecules such as protein, chlorophyll, and fat/lipid. The organic layer was washed with water three times, dried over Na 2 SO 4 , concentrated to 1 ml, and passed through a florisil column for purification of the extract. The eluant was concentrated to 200 μl and internal standard 2,4,5,6-tetrachloro-mxylene (TCmX) was added and kept at −20 °C before instrumental analyses. Details about the OCP analyses and QA/QC are described in Supplementary material. 2.3 Calculation The bioaccessibility (% BA) of OCPs was calculated as the ratio of the amount of OCPs in the liquid phase to that in both the liquid and solid according to formula (1) ( Oomen et al., 2000 ). The total OCPs for intestinal digestion were considered as equal to the total OCPs in fish samples subtracted by OCPs in gastric juice. (1) % BA = ( BA extracted OCPs / total OCPs ) × 100 The relative accumulation ratio ( R nn ) was used to evaluate the accumulation of OCP congeners during the digestion process, and calculated by formula (2) ( Carlson & Hites, 2005 ): (2) R nn = P n , l / P n , w where R nn is the relative accumulation ratio of congener n in sample n ; P n,l and P n,w are percentage contribution of HCH or DDT congener n in the digestion liquid and whole sample, respectively. The screening value (SV) is defined as the concentration of chemical in edible tissue that is of potential public health concern, and it is used as a threshold value against tissue residue level of contamination in similar tissue collected from the environment ( USEPA, 2000 ). The SV for OCPs was calculated using formula (3) ( Fairey et al., 1997 ): (3) SV = [ ( RL / OSF ) × BW ] / CR where SV is screening value (ng g −1 ), RL maximum acceptable risk level (dimensionless), OSF oral slope factor (mg kg −1 d −1 ), BW body weight (kg) and CR consumption rate (g d −1 ). In the present study, BW and CR were 60 kg and 142.2 g d −1 , respectively ( Cheung et al., 2007 ). RL is defined as if a person weighing 60 kg consumed 142.2 g of fish per day containing the same concentration of contaminant, for 70 years, the increased risk would be at most one additional cancer death per 100,000 persons ( USEPA, 1989 ). The oral slope factors for HCHs and DDTs were 1.8 and 0.34 mg kg −1 d −1 , respectively ( USEPA, 1993; USEPA, 2006 ). To derive advisory consumption recommendations for multiple carcinogens in fish, the methodology developed by USEPA ( USEPA, 2000 ) was calculated according to formula (4) , based on the cancer slope factor for each compound and calculated to prevent 1 excess cancer in 100,000 over a lifetime exposure: (4) CR max = RL × BW / Σ m = 1 X C m × OSF m where CR max is the maximum allowable fish consumption rate (g d −1 ), C m is the measured concentration of contaminant m in a given species of fish (μg g −1 ), and OSF m is the cancer oral slope factor for contaminant m (mg kg −1 d −1 ). 2.4 Statistical analyses The data analyses were performed using SPSS 17.0 for Windows. Normality was confirmed by the Kolmogorov–Smirnov test. Data of OCP concentrations and bioaccessibility in gastric and intestinal condition were analysed using two independent t -tests, Wilcoxon rank sum test, one-way ANOVA and Kruskal–Wallis test. 3 Results and discussion 3.1 OCP concentrations in Hong Kong market fish The ∑HCHs ranged from 0.33 to 3.09 ng g −1 (mean 1.22 ng g −1 ) in freshwater fish and from 0.40 to 9.88 ng g −1 (mean 1.92 ng g −1 ) in marine fish, respectively. The ∑DDTs ranged from 0.81 to 34.8 ng g −1 (mean 4.61 ng g −1 ) in freshwater fish and from 0.74 to 131 ng g −1 (mean 19.8 ng g −1 ) in marine fish, respectively. The detailed results of ∑HCH, ∑DDT and their congeners are shown in Fig. S1 . The marine fish had significantly ( p < 0.01) higher concentrations of both HCHs and DDTs than freshwater fish. The highest concentrations of HCHs and DDTs were observed in Snubnose pompano ( Trachinotus blochii ), and the result was in line with our previous study ( Cheung et al., 2007 ). The feeding mode of fish is an important factor affecting the contamination levels. Former studies revealed that carnivorous fish had significantly higher levels of OCPs in their muscle than fish with other feeding modes ( Hu et al., 2010 ). Furthermore, the lipid content (15.1%, ww) of Snubnose pompano was the highest among the twenty species, and will undoubtedly accumulate more OCPs due to their lipophilic nature. Mandarin fish had the highest ∑DDT concentrations (mean 23.2 ng g −1 ) among the ten freshwater species, and the result was also in line with our previous studies ( Cheung et al., 2007; Kong, Cheung, Wong, & Wong, 2005 ). As a carnivorous fish, mandarin fish is situated at the highest trophic level in freshwater fishponds, and is fed with trash fish or fingerlings as the major food sources ( Guo, Yu, & Zeng, 2009 ). In comparison to data generated from various studies worldwide, the ∑HCH levels (0.33–9.88 ng g −1 , mean 1.57 ng g −1 ) obtained in the present study were comparable with those from the coast of Korea (0.38–5.6 ng g −1 , ww) ( Yim, Hong, Shim, & Oh, 2005 ), but higher than that from Daya Bay and the Pearl River Estuary (PRE) (median 0.26 and 0.20 ng g −1 , ww, respectively) ( Guo et al., 2008a ). The ∑DDT concentrations (0.74–131 ng g −1 , mean 12.2 ng g −1 ) were comparable to fish collected from freshwater fishponds of the Pearl River Delta (PRD) (1.5–62 ng g −1 , ww) ( Kong et al., 2005 ), higher than fish collected from the Hong Kong market of our previous study (3.9–26 ng g −1 , ww) ( Chan, Chan, & Dickman, 1999 ), Korean coastal waters (0.84–27.0 ng g −1 , ww) ( Yim et al., 2005 ), East China Sea (0.94–8.62 ng g −1 , ww) ( Jiang et al., 2005 ) and Shanghai (mean 9.1 ng g −1 , ww) ( Nakata et al., 2002 ), but lower than fish collected from the PRE (1.8–287 ng g −1 , median 80 ng g −1 , ww) and Daya Bay (1.7–462 ng g −1 , median 40 ng g −1 , ww) ( Guo et al., 2008a ). In general, DDT contamination of Hong Kong market fish was relatively serious. 3.2 Composition profile and potential source apportionments The relative abundance of HCH congeners (to ∑HCHs) in freshwater fish followed the trend of: β -HCH (41.4%) > γ -HCH (30.0%) > α -HCH (20.1%) > δ -HCH (8.54%), while in marine fish followed the same trend of β -HCH (61.7%) > γ -HCH (16.2%) > α -HCH (16.1%) > δ -HCH (5.98%). β -HCH was the predominant HCH congener in both marine and freshwater fish, accounting for 51.6% of the ∑HCH among the twenty fish species. This was consistent with former studies whereby β -HCH was the most dominant HCH congener in biota samples collected from China ( Guo, Meng, Tang, & Zeng, 2008b ). It has been observed that high ratios of α -/ γ -HCH (3–7) indicate the input of technical-HCHs, while low ratios point to the dominant use of lindane ( Willett, Ulrich, & Hites, 1998 ). In the present study, the mean/median ratio of α -/ γ -HCH was 0.89/0.67 in freshwater fish and 1.44/0.93 in marine fish, respectively. The present relatively low ratios of α -/ γ -HCH in fish muscle indicated that the use of lindane was a potential main source ( Guo et al., 2009 ). The relative abundances of DDT congeners (to ∑DDT) were not significantly different ( p > 0.05) among different species. As the predominant compound, p , p ′-DDE accounted for 51.3% and 41.2% of ∑DDT in freshwater and marine fish, respectively. Ratios of DDD/DDE and DDT/(DDD + DDE) have been used to assess the living environment of aquatic organisms, estimate the extent of DDT decomposition, and identify any recent inputs of DDTs ( Lee, Tanabe, & Koh, 2001 ). As shown in Fig. S2 , the ratios of DDD/DDE in marine fish (0.41–2.28, mean 0.87) were significantly ( p < 0.05) higher than that in freshwater fish (0.22–0.83, mean 0.51), which may indicate that mariculture environments were more reductive than freshwater fish ponds ( Wang et al., 2010a ), because seawater fish are usually farmed in meshwork cages, and the high population density of fish, would create more anaerobic conditions. All DDT/(DDD + DDE) values (except Orange-spotted grouper) were below 0.6, suggesting there were limited fresh inputs of DDT in Hong Kong market fish. 3.3 Oral bioaccessibility of HCHs and DDTs in fish muscle Fig. 1 shows the bioaccessibility of ∑HCH and ∑DDT in gastric and intestinal conditions for the twenty species. The bioaccessibility of ∑HCH in the gastric condition (0.37–5.50%, mean 3.35%) was significantly ( p < 0.05) lower than that in intestinal condition (0.73–22.7%, mean 8.73%). Similarly, the bioaccessibility of ∑DDT in gastric condition (0.64–12.8%, mean 5.48%) was significantly ( p < 0.05) lower than that in intestinal condition (4.31–33.1%, mean 17.6%). The higher bioaccessibility of OCPs in intestinal condition was also noted in previous studies concerning the bioaccessibility of trace organic pollutants using in vitro gastrointestinal models (PAHs, ( Tang et al., 2006 ); OCPs, ( Tao et al., 2009 )), which might have been due to the micelles formed in the aqueous suspensions by bile constituents ( Van de Wiele, Verstraete, & Siciliano, 2004 ). No significant ( p > 0.05) difference in bioaccessibility between marine and freshwater fish was observed in gastric and intestinal conditions for both HCH and DDT. The low bioaccessbility of OCPs might suggest that OCPs intake through fish consumption by conventional procedures with organic solvent is considerably overestimated than actual human exposure. For OCP congeners, the average oral bioaccessibility ranged from 3.08 ( β -HCH) to 8.05% ( p , p ′-DDT) in gastric conditions and 8.09 ( γ -HCH) to 25.9% ( p , p ′-DDT) in intestinal conditions ( Fig. S3 ). For both gastric and intestinal conditions, the bioaccessibility of DDTs was significantly ( p < 0.01) higher than HCHs in fish muscle. This was in line with the results of bioaccessibility of OCPs in soil samples collected from Haihe Plain, North China ( Tao et al., 2009 ). The relative accumulation ratio ( R nn ) was used to evaluate the accumulation of OCP congeners during the digestion process ( Carlson & Hites, 2005 ). If the R nn value of the congener is greater than 1, this suggests relative accumulation, otherwise it indicates relative dilution during digestion ( Carlson & Hites, 2005 ; Xing et al., 2008 ). In fish muscle, the R nn value of γ -HCH, DDE ( o , p ′-DDE, p , p ′-DDE) and DDT ( o , p ′-DDT, p , p ′-DDT) was greater than 1 ( Fig. S4 ), suggesting that these congeners could be accumulated during gastrointestinal digestion. The R nn values of DDT congeners were significantly ( p < 0.05) correlated with their K OW (R 2 = 0.80), which was in line with our previous study that R nn values of PCB congeners were significantly correlated with their physicochemical parameters ( Xing et al., 2008 ). Our early study suggested that the OCP concentrations in Hong Kong human tissues were very high, which could be attributed to diet intake especially seafood consumption ( Wong et al., 2005 ). The available data of OCP congener concentrations in the subcutaneous fat ( Qin et al., 2010 ), breast milk ( Wong, Leung, Poon, Lan, & Wong, 2002 ) and adipose tissue ( Poon, Leung, Wong, & Wong, 2005 ) of Hong Kong residents were linearly regressed to their corresponding digestible concentrations via fish consumption ( Fig. 2 ). Significant ( p < 0.05) correlations observed between OCP digestible concentrations in fish muscle and their concentrations in subcutaneous fat (R 2 = 0.82) and adipose tissue (R 2 = 0.72) of Hong Kong people further confirmed the causal relations between OCP contamination in human tissues and fish consumption. 3.4 Difference among individual OCPs As mentioned above, the bioaccessibility of individual OCPs in the gastrointestinal model could vary from very low (<0.01% for γ -HCH in Snubnose pompano) to very high (50.9% for p , p ′-DDT in Grass carp). The variations in the bioaccessibility among PAHs, OCPs and PCBs have also been reported ( Tang et al., 2006; Tao et al., 2009; Xing et al., 2008 ), which suggested that different bioaccessibilities of individual OCPs tend to link with their physicochemical properties. In the present study, the influences of physicochemical parameters including molecular weight (MW), log transformed water solubility (μg l −1 ), octanol–water partition coefficient ( K OW ) and soil organic carbon–water partitioning coefficient ( K OC ) to the bioaccessibility of individual OCPs were investigated. As shown in Fig. S5 , the bioaccessibilities of HCH and DDT congeners in fish muscle were negatively ( p < 0.05) correlated with their water solubility ( Fig. S5B ) and positively ( p < 0.05) correlated with their MW, K OW and K OA ( Fig. S5A, C and D , respectively) in both gastric and intestinal conditions (R 2 shown in Fig. S5 ). Our results suggested that OCPs possessing higher hydrophobicity in fish muscle are more orally bioaccessible, which was consistent with a study regarding OCPs bioaccessibility in soil samples that showed that K OW was linearly correlated (R 2 = 0.909, p < 0.05) to the log transformed in vitro bioaccessibility ( Tao et al., 2009 ), but contrasted with another study whereby bioaccessibility of PCBs in vegetable was negatively correlated to their K OA and MW ( Xing et al., 2008 ). As suggested by Van de Wiele et al. (2004) , bile salts in the digestive fluid play an important role in contaminant mobilisation by complexing the contaminant in the fluid. It is reasonable to expect that the complexation is more favourable for more hydrophobic compounds. Further study is needed to address such effects. High lipid foods such as fish and milk could accumulate more persistent organic pollutants (POPs) such as OCPs and PAHs, and therefore act as the major contributors for human exposure to POPs ( Carlson & Hites, 2005 ). However, the in vitro gastrointestinal model revealed that high lipid contents in food items may in fact decrease the intake of benzo(a)pyrene ( Weber & Lanno, 2001 ) and PCBs ( Xing et al., 2008 ), because lipid is difficult to digest and thereby POPs will be more difficult to be released from food matrix. The different abundances of contaminants in the samples could also be another reason causing the difference in oral bioaccessibility ( Tao et al., 2009 ). However, no linear correlation was observed either between the lipid contents and bioaccessibility or between the abundances of OCPs and bioaccessibility in fish muscle. The role of lipid content and contaminant abundance in food matrix on the bioaccessibility of organic pollutants needs further investigation. The constituents in the digestion juice could also significantly affect the oral bioaccessibility of POPs. The bioaccessibility of HCHs and DDTs under intestinal conditions was significantly ( p < 0.01) higher than that under gastric conditions, which may be due to the fact that bile extract at levels higher than the critical micelle concentration (CMC) of 0.15 g l −1 could reduce the surface tension of digestion juice effectively ( Tang et al., 2006 ). The concentrations of bile extract diluted in the present study (5 g l −1 ) were more than 30 times higher than CMC, which could promote the formation of bile micelles, and therefore enhance bioaccessibility. 3.5 Risk assessment based on bioaccessibility results Intake via fish consumption is a major route of human exposure to organic contaminants such as OCPs ( Dougherty et al., 2000 ). The concentrations of DDTs in human breast milk collected from Hong Kong ranked number 5 among 19 countries or regions ( Wong et al., 2005 ) and had a significant correlation with the frequency of fish consumption ( Wong et al., 2002 ). Considering that the average fish consumption rate (164.4 g per day per person) ( Dickman & Leung, 1998 ) was higher than 142.2 g d −1 recommended by USEPA for subsistence consumers ( USEPA, 2000 ), the potential health hazards of consuming fish containing high concentrations of OCPs for Hong Kong residents should not be overlooked. Based on the fish consumption rate of 142.2 g d −1 and the average 60 kg body weight, the screening value for ∑DDT and ∑HCH was 12.4 and 2.34 ng g −1 (ww) for Hong Kong residents ( USEPA, 2000 ). In this present study, 10.0% and 13.3% of the freshwater fish contained elevated levels of ∑HCH and ∑DDT, respectively, which exceeded the screening value, while for marine fish, the percentage of exceedence for ∑HCH and ∑DDT were 16.7% and 40.0%, respectively. The ∑DDT concentrations in Snubnose pompano exceeded the screening value by 7.2 times ( Table 1 ). The results suggested that some freshwater fish and a large proportion of marine fish available in Hong Kong might not be suitable for long-term consumption (large amount), and that regulatory attention is needed to focus on this threat. It seems that the solvent concentrations of organic pollutants would tend to overestimate the absorbed amount of POPs in the body because not all pollutants are bioavailable ( Ramesh et al., 2004 ). Table 1 shows the total digestible concentrations including gastric and intestinal condition in this in vitro digestion model for each fish species. The digestible ∑HCH ranged from 0.01 to 0.28 ng g −1 (mean 0.13 ng g −1 ), accounting for 1.07–26.9% (mean 11.9%) for the solvent ∑HCH concentrations. Digestible ∑DDT ranged from 0.29 to 10.9 ng g −1 (mean 1.48 ng g −1 ), accounting for 4.52–36.1% (mean 20.2%) for the original ∑DDT concentrations. The digestible ∑HCH and ∑DDT in all samples were well below the screening value for fish consumption of Hong Kong people. In order to estimate the potential risk to human health, percent probability associated with a particular point of departure was estimated from a plot of percent cumulative probability as a function of compound concentrations in fish ( Olsen et al., 2003 ). Hazard ratios (HRs) were calculated by dividing the average exposure by the benchmark concentrations (the benchmark concentration is derived by setting the cancer risk to 1,000,000 for lifetime exposure) ( Jiang et al., 2005 ). HRs exceeding 1 indicated that there was potential cancer risk to human health. The 50th percentile and 95th percentile concentration were estimated to assess the potential risk. The HRs of 95th percentile concentration for α -, β -HCH, p , p ′-DDE, p , p ′-DDD and ∑DDT in marine fish and for ∑DDT in freshwater fish were greater than 1 ( Table S2 ), suggesting that daily exposure from fish consumption had a lifetime cancer risk of greater than 1 in 1,000,000. Similarly, HRs of all the contaminants were below 1 both for marine and freshwater fish when bioaccessibility was taken into consideration ( Table S2 ). The results related to risk assessment based on solvent and digestible concentrations were contradictory, which was mainly due to the fact that only a limited proportion of OCPs in fish muscle were digestible. Considering that a large proportion of OCPs were not bioavailable, former studies about human health risk assessments based on solvent concentrations of OCPs in fish should be adjusted by taking bioaccessibility into account ( Meng, Zeng, Yu, Guo, & Mai, 2007; Shaw et al., 2006 ). 3.6 Relevant fish consumption advisories When 10–50% of sampled fish are found to be contaminated with DDT or HCHs at or above screening values, consumers are advised to reduce consumption of these species ( Foran, Cox, & Croxton, 1989 ). Based on solvent concentrations, 13.3% and 26.7% of the fish investigated in this study contained elevated levels of ∑HCH and ∑DDT, respectively, which exceeded the screening values, suggesting that Hong Kong residents should reduce their fish consumption, especially marine fish. The maximum allowable fish consumption rate (CR max , g d −1 ) of each species was calculated, based on the total and digestible OCP concentrations to prevent 1 excess cancer in 100,000 over a lifetime exposure ( Fig. 3 ). Based on the total solvent concentrations, the CR max ranged from 47.6 (Mandarin fish) to 244 g d −1 (Grass carp) in freshwater fish and from 12.8 (Snubnose pompano) to 400 g d −1 (Orange-spotted grouper) in marine fish. As shown in Fig. 3 , Hong Kong residents should take caution when consuming various kinds of fish (such as freshwater fish – Mandarin fish and Snakehead, and most marine fish – especially Snubnose pompano and Yellow croaker) because the CR max for certain fishes are lower than the USEPA recommended limit of 142.2 g d −1 . When bioaccessibility was taken into consideration, the CR max increased sharply and all of the species exceeded the present average consumption rate except for Snubnose Pompano (CR max , 115 g d −1 ). The digestible CR max ranged from 423–2480 g d −1 (mean 1243 g d −1 ) for freshwater fish and 115–2196 g d −1 (mean 1022 g d −1 ) for marine fish ( Fig. 3 ). Based on this health risk assessment, almost all of the fish species were considered safe for consumption because the digestible CR max was much higher than that actually consumed in daily life. The present study focused only on OCP contamination. If other pollutants (such as PAHs, PCBs, PBDEs and heavy metals) were to be taken into consideration, it is assumed that the effects of these pollutants would be additive. Therefore, the advised CR max for both total and digestible pollutants would decrease, but further studies are needed before a more concrete conclusion can be drawn. Acknowledgements The authors thank Mr. K.W. Chan and Dr. H.M. Leung for technical assistance, Dr. A.O.W. Leung for improving the manuscript. This research was supported by the Research Grants Council of the University Grants Committee of Hong Kong (Collaborative Research Fund, HKBU1/CRF/08 and Special Equipment Grant, HKBU09), the Mini-AOE (Area of Excellence) Fund from the Hong Kong Baptist University and the Super FRG (FSRS/07/08/02)/HKBU. Appendix A Supplementary material Supplementary material (OCP analyses and QA/QC, five figures and two tables) associated with this article can be found, in the online version, at doi:10.1016/j.foodchem.2011.02.035 . Appendix A Supplementary material Supplementary material OCP analyses and QA/QC, five figures and two tables. References Carlson and Hites, 2005 D.L. Carlson R.A. Hites Polychlorinated biphenyls in salmon and salmon feed: Global differences and bioaccumulation Environmental Science and Technology 39 19 2005 7389 7395 Chan et al., 1999 H.M. Chan K.M. Chan M. Dickman Organochlorines in Hong Kong fish Marine Pollution Bulletin 39 1–12 1999 346 351 Cheung et al., 2007 K.C. Cheung H.M. Leung K.Y. Kong M.H. Wong Residual levels of DDTs and PAHs in freshwater and marine fish from Hong Kong markets and their health risk assessment Chemosphere 66 3 2007 460 468 Chung et al., 2008 S.W.C. Chung K.P. Kwong J.C.W. Yau Dietary exposure to DDT of secondary school students in Hong Kong Chemosphere 73 1 2008 65 69 Dickman and Leung, 1998 M.D. Dickman K.M.C. Leung Mercury and organochlorine exposure from fish consumption in Hong Kong Chemosphere 37 5 1998 991 1015 Dougherty et al., 2000 C.P. Dougherty S.H. Holtz J.C. Reinert L. Panyacosit D.A. Axelrad T.J. Woodruff Dietary exposures to food contaminants across the United States Environmental Research 84 2 2000 170 185 Fairey et al., 1997 R. Fairey K. Taberski S. Lamerdin E. Johnson R.P. Clark J.W. Downing Organochlorines and other environmental contaminants in muscle tissues of sportfish collected from San Francisco Bay Marine Pollution Bulletin 34 12 1997 1058 1071 Foran et al., 1989 J. Foran M. Cox D. Croxton Sport fish consumption advisories and projected cancer risks in the Great Lakes basin American Journal of Public Health 79 3 1989 322 325 Goni et al., 2006 I. Goni J. Serrano F. Saura-Calixto Bioaccessibility of beta-carotene, lutein, and lycopene from fruits and vegetables Journal of Agricultural and Food Chemistry 54 15 2006 5382 5387 Gonzalaz-Rodriguez et al., 2008 R.M. Gonzalaz-Rodriguez R. Rial-Otero B. Cancho-Grande J. Simal-Gandara Occurrence of fungicide and insecticide residues in trade samples of leafy vegetables Food Chemistry 107 3 2008 1342 1347 Guo et al., 2008b Y. Guo X.Z. Meng H.L. Tang E.Y. Zeng Tissue distribution of organochlorine pesticides in fish collected from the Pearl River Delta, China: Implications for fishery input source and bioaccumulation Environmental Pollution 155 1 2008 150 156 Guo et al., 2008a L.L. Guo Y.W. Qiu G. Zhang G.J. Zheng P.K.S. Lam X.D. Li Levels and bioaccumulation of organochlorine pesticides (OCPs) and polybrominated diphenyl ethers (PBDEs) in fishes from the Pearl River Estuary and Daya Bay, South China Environmental Pollution 152 3 2008 604 611 Guo et al., 2010 J.Y. Guo F.C. Wu R.L. Shen E.Y. Zeng Dietary intake and potential health risk of DDTs and PBDEs via seafood consumption in South China Ecotoxicology and Environmental Safety 73 7 2010 1812 1819 Guo et al., 2009 Y. Guo H.Y. Yu E.Y. Zeng Occurrence, source diagnosis, and biological effect assessment of DDT and its metabolites in various environmental compartments of the Pearl River Delta, South China: A review Environmental Pollution 157 6 2009 1753 1763 Hu et al., 2010 G.C. Hu J.Y. Dai B.X. Mai X.J. Luo H. Cao J.S. Wang Concentrations and accumulation features of organochlorine pesticides in the Baiyangdian lake freshwater food web of North China Archives of Environmental Contamination and Toxicology 58 3 2010 700 710 Jiang et al., 2005 Q.T. Jiang T.K.M. Lee K. Chen H.L. Wong J.S. Zheng J.P. Giesy Human health risk assessment of organochlorines associated with fish consumption in a coastal city in China Environmental Pollution 136 1 2005 155 165 Kong et al., 2005 K.Y. Kong K.C. Cheung C.K.C. Wong M.H. Wong The residual dynamic of polycyclic aromatic hydrocarbons and organochlorine pesticides in fishponds of the Pearl River Delta, South China Water Research 39 9 2005 1831 1843 Lee et al., 2001 K.T. Lee S. Tanabe C.H. Koh Distribution of organochlorine pesticides in sediments from Kyeonggi Bay and nearby areas, Korea Environmental Pollution 114 2 2001 207 213 Leung et al., 2009 S.Y. Leung C.K. Kwok X.P. Nie K.C. Cheung M.H. Wong Risk assessment of residual DDTs in freshwater and marine fish cultivated around the Pearl River Delta, China Archives of Environmental Contamination and Toxicology 58 2 2009 415 430 Lopez-Perez et al., 2006 G.C. Lopez-Perez M. Arias-Estevez E. Lopez-Periago B. Soto-Gonzalez B. Cancho-Grande J. Simal-Gandara Dynamics of pesticides in potato crops Journal of Agricultural and Food Chemistry 54 5 2006 1797 1803 Meng et al., 2007 X.Z. Meng E.Y. Zeng L.P. Yu Y. Guo B.X. Mai Assessment of human exposure to polybrominated diphenyl ethers in China via fish consumption and inhalation Environmental Science and Technology 41 14 2007 4882 4887 Nakata et al., 2002 H. Nakata M. Kawazoe K. Arizono S. Abe T. Kitano H. Shimada Organochlorine pesticides and polychlorinated biphenyl residues in foodstuffs and human tissues from China: Status of contamination, historical trend, and human dietary exposure Archives of Environmental Contamination and Toxicology 43 4 2002 473 480 Olsen et al., 2003 G.W. Olsen T.R. Church J.P. Miller J.M. Burris K.J. Hansen J.K. Lundberg Perfluorooctanesulfonate and other fluorochemicals in the serum of American Red Cross adult blood donors Environmental Health Perspectives 111 16 2003 1892 1901 Oomen et al., 2000 A.G. Oomen A. Sips J.P. Groten D. Sijm J. Tolls Mobilization of PCBs and lindane from soil during in vitro digestion and their distribution among bile salt micelles and proteins of human digestive fluid and the soil Environmental Science and Technology 34 2 2000 297 303 Poon et al., 2005 B.H.T. Poon C.K.M. Leung C.K.C. Wong M.H. Wong Polychlorinated biphenyls and organochlorine pesticides in human adipose tissue and breast milk collected in Hong Kong Archives of Environmental Contamination and Toxicology 49 2 2005 274 282 Qin et al., 2010 Y.Y. Qin C.K.M. Leung A.O.W. Leung S.C. Wu J.S. Zheng M.H. Wong Persistent organic pollutants and heavy metals in adipose tissues of patients with uterine leiomyomas and the association of these pollutants with seafood diet, BMI, and age Environmental Science and Pollution Research 17 1 2010 229 240 Ramesh et al., 2004 A. Ramesh S.A. Walker D.B. Hood M.D. Guillen K. Schneider E.H. Weyand Bioavailability and risk assessment of orally ingested polycyclic aromatic hydrocarbons International Journal of Toxicology 23 5 2004 301 333 Scott and Dean, 2005 W.C. Scott J.R. Dean An assessment of the bioavailability of persistent organic pollutants from contaminated soil Journal of Environmental Monitoring 7 7 2005 710 715 Shaw et al., 2006 S.D. Shaw D. Brenner M.L. Berger D.O. Carpenter C.S. Hong K. Kannan PCBs, PCDD/Fs, and organochlorine pesticides in farmed Atlantic salmon from Maine, eastern Canada, and Norway, and wild salmon from Alaska Environmental Science and Technology 40 17 2006 5347 5354 Tang et al., 2006 X.Y. Tang L. Tang Y.G. Zhu B.S. Xing J. Duan M.H. Zheng Assessment of the bioaccessibility of polycyclic aromatic hydrocarbons in soils from Beijing using an in vitro test Environmental Pollution 140 2 2006 279 285 Tao et al., 2009 S. Tao Y. Lu D.Y. Zhang Y.F. Yang Y. Yang X.X. Lu Assessment of oral bioaccessibility of organochlorine pesticides in soil using an in vitro gastrointestinal model Environmental Science and Technology 43 12 2009 4524 4529 USEPA, 1989 USEPA (1989). Risk assessment guidance for superfund Vol. 1. Human health evaluation manual (Part A). Washington, DC: U.S. Environmental Protection Agency. USEPA, 1993 USEPA (1993). Dieldrin (CASRN 60-57-1): US Environmental Protection Agency, Office of Research and Development, Integrated Risk Information System. http://www.epa.gov/iris/subst/0225.htm#carc . USEPA, 2000 USEPA (2000). Guidance for assessing chemical contaminant, data for use in fish advisories, Vol. 1: fish sampling and analysis, third ed. EPA 823-R-95-007. Washington, DC: Office of Water. USEPA, 2006 USEPA (2006). Assessment of lindane and other Hexachlorocyclohexane isomers. Prevention, Pesticides and Toxic Substances (7508C). Washington, DC: U.S. Environmental Protection Agency. Van de Wiele et al., 2004 T.R. Van de Wiele W. Verstraete S.D. Siciliano Polycyclic aromatic hydrocarbon release from a soil matrix in the in vitro gastrointestinal tract Journal of Environmental Quality 33 4 2004 1343 1353 Wang et al., 2010a H.S. Wang Z. Cheng P. Liang D.D. Shao Y. Kang S.C. Wu Characterization of PAHs in surface sediments of aquaculture farms around the Pearl River Delta Ecotoxicology and Environmental Safety 73 5 2010 900 906 Wang et al., 2010b H.S. Wang Y.B. Man F.Y. Wu Y.G. Zhao C.K. Wong M.H. Wong Oral bioaccessibility of polycyclic aromatic hydrocarbons (PAHs) through fish consumption, based on an in vitro digestion model Journal of Agricultural and Food Chemistry 58 2010 11517 11524 Weber and Lanno, 2001 L.P. Weber R.P. Lanno Effect of bile salts, lipid, and humic acids on absorption of benzo[ a ]pyrene by isolated channel catfish (Ictalurus punctatus) intestine segments Environmental Toxicology and Chemistry 20 5 2001 1117 1124 Willett et al., 1998 K.L. Willett E.M. Ulrich R.A. Hites Differential toxicity and environmental fates of hexachlorocyclohexane isomers Environmental Science and Technology 32 15 1998 2197 2207 Wong et al., 2002 C.K.C. Wong K.M. Leung B.H.T. Poon C.Y. Lan M.H. Wong Organochlorine hydrocarbons in human breast milk collected in Hong Kong and Guangzhou Archives of Environmental Contamination and Toxicology 43 3 2002 364 372 Wong et al., 2005 M.H. Wong A.O.W. Leung J.K.Y. Chan M.P.K. Choi A review on the usage of POP pesticides in China, with emphasis on DDT loadings in human milk Chemosphere 60 6 2005 740 752 Xing et al., 2008 G.H. Xing Y. Yang J.K.Y. Chan S. Tao M.H. Wong Bioaccessibility of polychlorinated biphenyls in different foods using an in vitro digestion method Environmental Pollution 156 3 2008 1218 1226 Yim et al., 2005 U.H. Yim S.H. Hong W.J. Shim J.R. Oh Levels of persistent organochlorine contaminants in fish from Korea and their potential health risk Archives of Environmental Contamination and Toxicology 48 3 2005 358 366 Yu et al., 2010 Y.X. Yu J.L. Li X.Y. Zhang Z.Q. Yu T. Van de Wiele S.Y. Han Assessment of the bioaccessibility of polybrominated diphenyl ethers in foods and the correlations of the bioaccessibility with nutrient contents Journal of Agricultural and Food Chemistry 58 1 2010 301 308