Introduction

Arsenic (As) is a naturally occurring metalloid element that is commonly found in the environment (Chen et al. 2000a; Tanner et al. 2000). It is widely used in medicine, electronics, agriculture and metallurgy (Suzuki and Mandal 2002). Arsenic in the aquatic environment in Pearl River Delta (PRD) is closely related to anthropogenic sources (Wong et al. 1995). Aquatic products for human consumption, such as fish and shellfish, can accumulate high concentrations of arsenic from water and sediment. It can be biomagnified in food chains, and this has been reported in Lake Macquarie (Australia) and the Sulu Sea (Philippines) (Asante et al. 2010; Barwick and Maher 2003).

The PRD is one of the most important industrial and agricultural districts in China. Freshwater fish production is a major aquaculture economic resource in the PRD and the aquatic products produced are catered for the residents of Hong Kong and surrounding areas. However, freshwater fish produced from the PRD contained rather high concentrations of As (Cheung et al. 2008). These included cultured fish and fish available from local markets which contained 0.42–3.44 mg kg−1 wet weight (w. wt) of As (Cheung et al. 2008).

Monoculture and polyculture are two major freshwater fish culture models commonly practiced in PRD. Monoculture is based mainly on rearing carnivorous fish, and polyculture is based mainly on vegetarian and omnivorous fish (Cheung et al. 2008; Zhou and Wong 2000). In polyculture pond ecosystem, fish species of higher trophic levels have narrow foraging ranges than other aquatic ecosystems and therefore a simpler food chain relationship (Cheng et al. 2011). In our previous studies conducted in the PRD region, stable isotopes of nitrogen were used to investigate the relationship between the trophic levels and mercury concentration in biota tissues (Cheng et al. 2011), and showed that mercury biomagnifaction in the aquaculture pond ecosystem was lower than other marine and freshwater ecosystems.

With increasing public awareness concerning the safety of foods originating from mainland China, the major objectives of the present study were to investigate the As contents contained in the freshwater aquatic products, As bioaccumulation and biomagnification in aquatic food chains of freshwater fish ponds in the PRD (the main source for consumption in Hong Kong), and evaluate the potential health risks of exposure to the Hong Kong residents via dietary intake of these products.

Materials and methods

Sampling

Fish and corresponding sediment samples were collected from 18 fish ponds (11 monoculture pond and seven polyculture ponds) (Table 1) around the PRD during October 2008 and April 2009. This is a rather long term study with samples collected since the end of 2008. There are several sub-projects under this main project and the present paper is mainly focused on bioaccumulation of arsenic and its potentials health risk. The stations were located in Huadu, Nanhai, Shunde, Zhongshan, Jiangmen, Nansha, Dongguan, and Huizhou (Fig. 1). Triplicate samples were collected from each pond. Surface sediment (0–10 cm) was collected using a stainless steel shovel. Fish samples collected included northern snakehead (Channa argus), mandrarin fish (Siniperca chuatsi), largemouth bass (Lepomis macrochirous), bighead carp (Aristichthys nobilis) and grass carp (Ctenopharyngodon idellus), were collected from the fishponds using a nylon net. Fish lengths and weights were recorded. Commercial fish feed pellets and live forage fish (mud carp [Cirrhina molitorella]) commonly used as fish feeds for feeding commercial fish were purchased from fish farmers at each site. Sediment and fish samples were wrapped in aluminum foil and placed in polyethylene bags during transportation and stored at −20 °C in the laboratory for subsequent analyses. Zooplankton organisms were sampled at approximately 0.5–10 m depth from fish ponds and were collected using a non-metallic plankton net (202 μm) for multiple vertical tows, stored in 100 ml acid-treated Teflon vials (Chen et al. 2000b), transported to the laboratory and stored at −20 °C for subsequent analyses.

Table 1 Mean arsenic concentrations and chemistry parameters in sediment
Full size table
Fig. 1
figure 1

Map of the sampling locations in PRD, China. Huadu = HD1 and HD2; Nanhai = NH1, NH2 and NH3; Shunde = SD1, SD2, SD3 and SD4; Jiangmen = JM1 and JM2; Zhongshan = ZS1 and ZS2; Nansha = NS1 and NS2; Dongguan = DG1 and DG2; Huizhou = HZ

Full size image

Analytical methods

All the samples were freeze-dried and ground into powder. The sediment samples were homogenized by passing through a stainless steel 0.154-μm sieve. Next, 2 g of sediment samples was placed in high-pressure Teflon containers, added with 12 ml aqua regia (hydrochloric acid/nitric acid, 3:1 ratio [v/v]), while 0.5 g of fish muscle samples or fish feed pellets, added with 10 ml nitric acid, was placed in high-pressure Teflon containers and stored at room temperature (25 °C) for predigestion in 6–8 h. Subsequently, all samples were further digested using the MARS Microwave Reactions System (CEM, USA).” At the end of microwave digestion, samples were transferred to disposable tubes and diluted to a volume of 25 ml with Milli-Q water. The extracts were analyzed by inductively coupled plasma-mass spectrometry (Perkin-Elmer, Elan 9000, Norwalk, CT, USA). Subsample of sediment was used for the analysis of total organic carbon (TOC), using thermal partitioning at 550 °C (USEPA 1997). For determining the geochemical fraction of arsenic in sediment samples, the European Community Bureau of Reference (BCR) sequential extraction procedure modified by Rauret et al. (1999) was adopted in this study.

Quality control

The accuracy of the As measurements for each analytical batch was determined using four certified reference materials (CRM): NIST 8704 (Buffalo River sediment, National Institute Standards and Technology, USA); NIST 1944 (New York/New Jersey Waterway Sediment, National Institute Standards and Technology, USA); NIST 1566b (Oyster tissue, National Institute Standards and Technology, USA) and TORT-2 (lobster hepatopancreas, National Research Council of Canada). The recoveries of the CRM for arsenic ranged from 90 % to 103 % (NIST 8704, n = 6, 90 ± 10 %; NIST 1944, n = 6, 94 ± 7 %; NIST 1566b, n = 6, 92 ± 3 % and TORT-2, n = 6, 103 ± 5 %). CRM NIST 8704 and NIST 1944 were used to check the BCR sequential extraction, the total concentration (Acid soluble faction + Reducible faction + Oxidizable faction + Residual faction) of BCR sequential extraction yielded good accuracy of analysis (80–95 %, n = 6).

Assessment of sediment quality

The enrichment factor (EF) estimates the anthropogenic impact on sediment by using the following formula (Loska et al. 2003)

$$ \mathrm{EF}=\frac{{{{{{{\mathrm{A}}_x}}} \left/ {{\mathrm{F}{{\mathrm{e}}_x}}} \right.}}}{{{{{{{\mathrm{A}}_b}}} \left/ {{\mathrm{F}{{\mathrm{e}}_b}}} \right.}}} $$
(1)

where A x , A b , Fe x and Fe b are the concentrations of Arsenic (A) and Fe in the sample (x) and in the background reference (b) (Presley et al. 1992; Tanner et al. 2000), respectively. In this study, iron is used as the reference element, because its content in the studied sediment is predominantly (97 %) associated with the parent material matrix of the sediment (Diaz-de Alba et al. 2011).

The geoaccumulation index (I geo) determines pollution levels by comparing current metal contents with preindustrial levels (Müller 1981). The geoaccumulation index value is calculated using the following equation:

$$ {I_{geo }}={\log_2}\left( {\frac{{{C_n}}}{{1.5{B_n}}}} \right) $$
(2)

where C n is arsenic content in test sediment and B n is the background content in the Earth’s crust (Hamilton 2000).

The risk assessment code (RAC) is a method to assess the ability of metals to be released and to enter in the food chain (Rath et al. 2009). The RAC refers to the percentage of a certain metal associated with this weak acid exchangeable and carbonate fraction. Criteria for these pollution indictors are given in Table 2 (Loska et al. 2003; Müller 1981; Sutherland 2000).

Table 2 Contamination categories base on the enrichment factor, geoaccumulation index and risk assessment code (Loska et al. 2003; Müller 1981; Sutherland 2000)
Full size table

Stable isotope analysis

Stable isotope analysis has been used as a tool to investigate the relationship between the trophic level and contaminate concentration in biota tissues (Cheng et al. 2011; Tu et al. 2011). The biota samples were analyzed for stable isotopes at Institute of Soil Science (Nanjing, China), Chinese Academy of Sciences.

Stable isotope values were expressed as

$$ {\delta^{15 }}\mathrm{N}=\left( {\frac{{{R_{sample }}}}{{{R_{{s\tan dard}}}}}-1} \right)\times 1,000 $$
(3)

where R sample is the corresponding ratios of 15N/14N (Hobson et al. 1995), while the R standard values are based on atmospheric N2 (air). Replicate measurements of internal laboratory standards (albumen) showed measurement errors of ±0.3‰ for stable nitrogen isotope measurement.

Biota-sediment accumulation factor (BSAF)

BSAF is calculated by the following equation (Szefer et al. 1999):

$$ \mathrm{BSAF}=\frac{{{C_t}}}{{{C_s}}} $$
(4)

where C t is the arsenic concentration in the tissues and C s is the arsenic concentration in sediment.

Risk assessment

Risks of arsenic exposure to Hong Kong residents via dietary freshwater fish intakes were assessed according to the guidelines recommended by the USEPA (2000, 2009). For non-carcinogenic effects, the estimated daily intake was compared with the recommended reference doses (RfD, 0.3 μg kg−1 day−1) (USEPA 2010) as stated in Eq. 5:

$$ \mathrm{Hazard}\,\mathrm{ratio}\,\left( {\mathrm{HR}} \right)=\frac{\mathrm{EDI}}{\mathrm{RfD}}=\frac{{{C_{As }}\times \mathrm{DR}\times \mathrm{EF}\times \mathrm{ED}}}{{\mathrm{BW}\times \mathrm{AT}\times \mathrm{RfD}}} $$
(5)

where EDI is the estimated daily intake, DR is the daily consumption rate (kg person−1 day−1), C As is the arsenic concentration (mg kg−1, w. wt), BW is the average body weight (kg), EF is the exposure frequency (365 day year−1), ED is the average duration of exposure (year), BW is the average body weight (kg), and AT is the average exposure time (365× ED day year). Average adult and child body weights were 58.6 kg (Wang et al. 2005) and 21.8 kg (Leung et al. 2000), respectively, and the DR was 93 and 50 g day−1 for adults and children, respectively, in Hong Kong (Leung et al. 2000). The HR exceeding 1 indicates that there is potential risk to human health. For the carcinogenic effects, the cancer risk (CR) was obtained by using the oral slope factor of arsenic (OSFAs) (1.5 mg kg−1 day−1) (USEPA 2010) in the following equation:

$$ \mathrm{CR}=\mathrm{EDI}\times \mathrm{OS}{{\mathrm{F}}_{\mathrm{As}}} $$
(6)

Data analyses

The data analyses were performed using SPSS 19.0 for Windows. Normality was confirmed by the Kolmogorov–Smirnov test. Data of As concentrations were analyzed using two independent t-tests, Wilcoxon rank sum test, one-way ANOVA and Kruskal–Wallis test as the requirement.

Results and discussions

As concentration in sediment

Arsenic concentrations in sediment samples are shown in Table 1. The mean concentration in PRD of As was 13.93 ± 3.68 mg kg–1 d. wt, ranging from 8.41 to 22.76 mg kg–1 d. wt. The HD1 site had the highest As concentration (22.76 ± 2.02) in sediment among all sampling sites. In different sampling regions, the highest As concentration was detected in sediment samples collected from Zhongshan (18.50 ± 2.39 mg kg–1 d. wt, min 15.49 mg kg–1 d. wt, max 21.33 mg kg–1 d. wt) and Nansha (17.93 ± 0.77 mg kg–1 d. wt, min 16.69 mg kg–1 d. wt, max 18.61 mg kg–1 d. wt). These levels were significantly higher (p < 0.05) than those from Nanhai (10.69 ± 1.27 mg kg–1 d. wt, min 8.93 mg kg–1 d. wt, max 11.43 mg kg–1 d. wt) and Huizhou (8.41 ± 2.98 mg kg–1 d. wt, min 5.21 mg kg–1 d. wt, max 11.20 mg kg-1 d. wt). These levels were significantly higher (p < 0.05) than those from Nanhai (10.69 ± 1.27 mg kg–1 d. wt) and Huizhou (8.41 ± 2.98 mg kg–1 d. wt). The mean value was similar to those reported in the surface sediment of Guangzhou section of the Pearl River which ranged from 16.7 to 33.4 mg kg–1 (Wang et al. 2010). The As concentrations of sediment samples collected from Zhongshan, Dongguan and Huadu were lower than those reported in our previous study with sediment samples collected from PRD (62.7 mg kg–1 d. wt, 60.3 mg kg–1 d. wt and 62.4 mg kg–1 d. wt) (Cheung et al. 2008). This may be due to the traditional aquaculture farming model in the PRD region, where the pond water is drained once every few years to maintain water quality and reduce the incidence of fish diseases. Removal of the surface sediment would also reduce As concentration. The pond mud is also removed and used as fertilizer for corps growing on the dykes of fish ponds (Wong et al. 2004). In addition, the significant reduction in organic matter content of the surface sediment will significantly lower the ability for arsenic accumulation in sediment (Pikaray et al. 2005).

The ecotoxicity and mobility of metal and metalloid elements in the environment are strongly dependent on their specific chemical forms or method of binding (Quevauviller 1998). The As distribution in sediment (in percentage) following the BCR procedure is shown in Fig. 2. The residual faction of As accounted for 42.7–76.2 % of the total content with an average of 62.5 %, has indicated that the freshwater fish ponds in PRD may be subjected to strong influence devised from extensive human activities. Arsenic metal is commonly used as an alloying additive and widely used in the manufacture of semiconductor and solar cells. Arsenic and its compounds are also used in the manufacture of pesticides, preservatives, dyes, and pharmaceuticals (Grund et al. 2008). Higher As concentrations observed in fish ponds in Nansha, Huadu, Zhongshan and Dongguan may be explained by the highly developed manufacturing industry, and Huizhou had the lowest concentration of As, which may be due to its relatively less developed manufacturing industry compared to the other cities (Chen et al. 1997).

Fig. 2
figure 2

The distribution of As in different geochemical phases of sediments collected from freshwater fish ponds in Pearl River Delta

Full size image

The Fe/Mn oxides fraction and oxidizable fraction of As accounted for 13–32 % and 0.6–18.9 % of its total content with an average of 22.4 % and 5.7 %, respectively. The As oxidizable fraction in Zhongshan (18.9 %) was higher than other sampling sites. For the As in non-residual fractions, a major portion is bound to the Fe/Mn oxides with comparable amounts associated with the acid soluble and oxidizable fractions. The reducible As predominating in non-residual fractions has been reported by many researchers (Diaz-de Alba et al. 2011; Wang et al. 2010).

The TOC content of sediment varied from 1.32 % to 7.37 % with an average of 3.75 %. Huangdu and Dongguan had the highest TOC contents of sediments. In the present study, positive correlations between organic carbon and As concentration of sediments were observed (R 2 = 0.418, p < 0.05). Sediment organic carbon could adsorb As and the concentration could increase according to the increase of organic carbon content by the formation of organo-As complexes (Pikaray et al. 2005). The concentration of total As and the ratios of residual As indicated that organic carbon may play a significant role in controlling As transport in the sediment of the freshwater fish ponds. Arsenic speciation in sediment and water samples was not monitored, and these should be investigated for future studies on As transport in sediment. The high-density culture model used large amounts of forage and high protein feed pellets, the fish feeds residues and fish metabolites could increase the organic carbon content of sediments and adsorb more As into the sediments. High levels As concentrations were detected in trash fish and pellet feeds, those which were used as fish feeds (Table 3). This could be a reason in increased concentrations of As in the pond sediments.

Table 3 Arsenic concentrations in fish, trash fish, and feed pellets collected from different sampling sites in PRD, China
Full size table

Figure 3 show the pollution indicators of EF, I geo and RAC for As concentrations in sediment collected from the freshwater fish ponds in PRD, and the criteria for these pollution indictors are shown in Table 2 (Loska et al. 2003; Müller 1981; Sutherland 2000). The mean of EF values in the fresh water culture pond in PRD was 1.52 ± 0.44, suggesting that a significant portion of metal was delivered from non-crustal materials, or non-natural weather process, with anthropogenic source (urbanization and economic development) an important contributor (Diaz-de Alba et al. 2011; Loska et al. 2003). In accordance with the criteria of EF (Loska et al. 2003) (Table 2), the highest EF values were recorded in Dongguan (2.25) and Huadu (2.06), meaning that those two sites were moderately contaminated by As, and the other sites (EF value <2) were less polluted by As. The mean I geo of As was 2.30 ± 0.39, ranging from 1.63 to 2.76, which was considerably polluted (Table 2), and all RAC of As in sediment samples posed no potential mobility (RAC value <1) (Rath et al. 2009).

Fig. 3
figure 3

The geoaccumulation index (I geo), enrichment factors (EF) and risk assessment code (RAC) in sediments for arsenic analysis of each sampling area (n = 54) (a) and total area (b) in freshwater fish Ponds in the Pearl River Delta. Each box represents interquartile range (25th and 75th percentile) of I geo, EF and RAC

Full size image

As concentration in fish

The concentrations of As detected in different fish species collected from different ponds ranged from 0.05 to 3.01 mg kg−1 w. wt. The largemouth bass had the highest As concentration (2.23 ± 0.57 mg kg−1 w. wt), which was 1 order of magnitude higher than the other species (Table 3). In addition, As concentrations in muscle of bighead cap (0.20 ± 0.09 mg kg−1 w. wt), grass carp (0.12 ± 0.08 mg kg−1 w. wt) and mud carp (0.47 ± 0.08 mg kg−1 w. wt) were similar to difference species of freshwater fish available in the local markets in PRD (Cheung et al. 2008). The feeding modes and food items of the fish are listed in Table 3. The mud carp (0.47 ± 0.08 mg kg−1 w. wt) was used to feed mandarin fish (0.13 ± 0.04 mg kg−1 w. wt), and trash fish (6.76 ± 0.59 mg kg−1 w. wt) was used to feed largemouth bass. Arsenic uptake and adsorption over the gills and other body surfaces may be an important route of exposure for invertebrates and fish (Barwick and Maher 2003). Food ingestion may be another pathway to take up As from fish food (Asante et al. 2010; Barwick and Maher 2003).

The freshwater fish are mainly fed with trash fish or pellet feeds, with fishmeal as a major protein adds into the pellet feeds, and most fishmeal is derived from trash fish (Cheung et al. 2008; Wong et al. 2006). Trash fish is a fish resource having low market value (Weatherley and Cogger 1977). The fish resource may be of low value due to low demand for that particular species/size, in addition to it being too costly to handle and process compared to market value and is often spoiled during transportation or handling (Weatherley and Cogger 1977). In this study, the trash fish (feed for mandarin fish) (6.76 ± 0.59 mg kg−1 w. wt) and the feed pellets (feed for northern snakehead) (2.75 ± 0.93 mg kg−1 w. wt) contained high As concentrations, the contaminated fish feeds may be the main source of As intake by freshwater fish.

The BSAFs for arsenic in individual fish species and zooplankton, BSAF in zooplankton (0.286), mud carp (0.195) and largemouth bass (0.191) were much higher than those in carnivorous mandarin fish (0.012) and northern snakehead (0.022), filter feeder bighead carp (0.041) and herbivorous grass carp (0.016). It has been observed that BSAF of As are higher in most omnivorous species (zooplankton, mud carp and bighead carp). This is due to the fact that As dissolved in water may be adsorbed onto particulates in the sediment, which may accumulate in aquatic organisms via absorption of water and ingestion of particulates from the water column (Casado-Martinez et al. 2010; Chen et al. 2000b). The As source of biota sample is mainly derived from the trash fish and pellet feeds in the aquaculture pond ecosystem as well as As concentrations occurring in higher trophic species by predation. It could explain why carnivorous fish species possess a relatively lower BSAF.

The ability of a contaminant to biomagnify can be expressed in terms of biomagnification factors (BMF), where BMF >1 indicate biomagnification. The BMF calculations are based on predator–prey relationship from the literature, and corrected to unity for trophic level differences (Chen et al. 2000b). In the present study, As appeared to diminish (BMF < 1) from the fish food (pellet feeds, zooplankton, mud carp and trash fish) to the predatory and non-predatory fish in the aquaculture pond ecosystem. The stable nitrogen (δ 15N) ranged from 6.61‰ to 17.6‰ for the investigated species and average δ 15N values are listed in Table 3. To determine if the biomagnifications of As in the freshwater fish ponds in PRD, the trophic level (TL) of the selected aquatic organisms was calculated based on δ 15N, as described by Fisk et al. (2001):

$$ \mathrm{T}{{\mathrm{L}}_{\mathrm{consumer}}}={{{2+\left( {{\delta^{15 }}{{\mathrm{N}}_{\mathrm{consumer}}}-{\delta^{15 }}{{\mathrm{N}}_{\mathrm{zooplankton}}}} \right)}} \left/ {3.4 } \right.} $$
(7)

where TL consumer is the trophic level and trophic level (TL) of zooplankton is assumed to be 3.0, and 3.4 is the isotopic EF. The δ 15N for grass carp and mud carp (average TLs of 2.48 and 3.00, respectively) indicated that they were the secondary consumers. Bighead carp (TL = 3.88), largemouth bass (TL = 4.32), Northern snakehead (TL = 4.60) and Mandarin fish (TL = 4.01) were in between tertiary and quartus consumers in omnivorous and predatory food chains, respectively. In the present study, two typical food chains of freshwater fish ponds in the PRD were selected for investigation: (1) omnivorous food chain consisted of zooplankton, grass carp, and bighead carp; (2) predatory food chain consisted of zooplankton, mud carp and mandarin fish. The concentrations of As in zooplankton and fish were negatively correlated (p < 0.05) with δ 15N, indicating that As in biotissues was not biomagnified (Fig. 4). Two significant linear relationships were obtained between log As and δ 15N (Fig. 4). The slope of the regression (−0.066 and −0.078) of the log transformed As concentrations and δ 15N values, as biomagnifications power, suggested there was no biomagnification or biodiminution of As from lower trophic levels (zooplankton and trash fish) to fish in the ponds. Aquatic organisms are able to bioaccumulate As from the surrounding water and contaminated prey. However, As compounds could be efficiently and rapidly excreted from fish, reducing the possibility of accumulation and trophic transfer in food chains (Campbell et al. 2005; Chen et al. 2000a; Nfon et al. 2009). This phenomenon could be explained by the accumulation efficiency of As and the necessary regulatory and detoxification mechanisms of aquatic organisms (Dallinger 1994). In addition, many physical and biological factors such as geography, sex, age and body conditions will influence As deposition in aquatic organisms (Barwick and Maher 2003; Campbell et al. 2005; McGeer et al. 2003). In fact, As diminution has been observed in a number of freshwater ecosystems, e.g., Upper Mystic Lake, USA (Chen et al. 2000a); tropical mangrove ecosystem, Vietnam (Tu et al. 2011); Yellow River Delta, China (Cui et al. 2011).

Fig. 4
figure 4

Regression relationship between, respectively, logarithm of As concentration and δ 15N for omnivorous and predatory food chains in zooplankton and fish species from the freshwater fish ponds in Pearl River Delta, China. See text for details regarding the types of sample analyzed and species codes. Filled square zooplankton, dashed line mud carp, filled triangle grass carp, filled circle bighead carp, filled diamond Mandarin fish

Full size image

No biomagnification or biodiminution of As from lower trophic levels to the cultured fish in these aquaculture ponds would be attributed to the aquaculture pond ecosystems, which are characterized by short food chains, lower species diversity, and abundance of different food sources (Table 3) available to different fish species. There are more aquatic species, longer food chains and more complex food webs existed in marine water systems than in freshwater systems. In marine water systems, an intriguing pattern has been observed where As biomagnification at the food chains consist of high trophic levels species (e.g., carnivores), and diminution at the food chains consist of lower trophic levels species (e.g., planktons and herbivores) (Asante et al. 2010; Nfon et al. 2009). Kuroiwa et al. (1994) showed that methylated forms of As, which are less toxic, could increase with trophic levels, but total As diminishes with increasing trophic levels in freshwater food chain. Arsenic speciation (organic and inorganic arsenic compounds) in the fish muscle samples of was not monitored, and this should be investigated for future toxicological studies. Based on the previous reports in fresh water fish muscle samples (Goessler et al. 1997; Kuroiwa et al. 1994; Schaeffer et al. 2006; Slejkovec et al. 2004), about of 85 % to >90 % of arsenic found in edible portions of fish and shellfish is organic arsenic (arsenobetaine [AsB] and dimethylarsinic acid [DMA]) and that approximately 10 % is inorganic arsenic. In general, inorganic arsenic species are more toxic than organic forms to living organisms (Bissen and Frimmel 2003; Ng 2005).

Human risk from consuming fish

Arsenic is toxic to both plants and animals and inorganic arsenicals are proven carcinogens in humans (Ng 2005). Toxicity of arsenic to human health ranges from skin lesions to cancer of the brain, liver, kidney and stomach (Smith et al. 1992). To estimate the human health risk from consuming As contaminated fish, the estimated expose doses were calculated for inorganic As.

According to previous studies reported the proportion of inorganic As in fish muscle (Table 4), assuming that 10 %, 20 % and 30 % of the total average As in muscle tissues is the inorganic toxic form.

Table 4 Arsenic speciation proportion of total arsenic in fish muscle
Full size table

Figure 5 shows the HR and CR of inorganic through consumption of freshwater fish by both adults and children in Hong Kong. An HR higher than 1 implies that the EDI exceeded the RfD for the contaminant of interest and that systemic effect may occur.

Fig. 5
figure 5

Hazard ratios and cancer risk of inorganic As through freshwater fish by adults and children in Hong Kong. The consumption rates are 93 g day−1 for adults and 50 g day−1 for children, respectively. Each box represents interquartile range (25th and 75th percentiles) of hazard ratios and cancer risk of each fish. MF mandarin fish, NSH northern snakehead, LB largemouth bass, BC bighead carp, GC grass carp

Full size image

In the present study, there was only largemouth bass HR value (1.18 for adult, 1.71 for children) was over 1 in the 10 % inorganic As of total As condition, all the HR values of other fish in three assuming conditions were lower than 1, suggesting that the residents may have a significant potential health risk by consuming largemouth bass and children have higher risk than adult. A risk above the 10−6 value considered by the USEPA (1989) as an acceptable risk for cancer when estimating the lifetime excess CR of As. The CR values of fish were all above 10−6, largemouth bass (5.3 × 10−4) in the 10 % inorganic As of total As condition, northern snakehead (1.75 × 10−4) in the 20 % inorganic As of total As condition and bighead carp (1.42 × 10−4) in the 30 % inorganic As of total As condition CR value were was higher than the upper limit of the acceptable risk levels (10−4) (USEPA 2010).

It should be noted that the assessment results may be influenced by uncertainty factors such as fish ingestion rates and further investigation would be necessary to investigate the bioavailability of inorganic As in fish to humans (Moreda-Pineiro et al. 2011). However, the present results provide useful information showing that the consumption of largemouth bass, northern snakehead and bighead carp might impose health risks to the local population, especially pregnant and nursing mothers who are more likely to be susceptible to inorganic As.

Conclusions

EF, I geo and RAC showed that the sediment samples of freshwater fish ponds in PRD are polluted moderately by As. BASF showed that omnivorous fishes and zooplankton accumulated more As from sediment than carnivorous fish. There was no biomagnification or biodiminution of As from lower trophic levels to fish in freshwater fish ponds. Risk assessment indicated the consumption of largemouth bass, northern snakehead and bighead carp might impose health risks to the local population, due to the fact that the CR value exceeded the upper limit of the acceptable risk levels (10−4) (USEPA 2010).