Spatial distribution, source identification, and risk assessment of organochlorines in wild tilapia from Guangxi, South China

Occurrence of the target OCs

The following individual OCs were found in fish samples with a detection frequency lower than 50%: i.e. p,p’-methoxychlor, heptachlor, heptachlor exo-epoxide, CB-66, 77, 81, 105, 114, 118, 123, 126, 128, 138, 153, 156, 157,167, 169, 170, and 180. These OCs are not discussed later. The concentrations of seven OC compounds in the muscle samples of 41 Nile tilapia and 34 Redbelly tilapia are shown in Supplementary Table S4. Since there were no significant interspecies differences between Nile tilapia and Redbelly tilapia (t-test, p = 0.16), the results of OCs analysis will be reported by tilapia genus in this study.

Median concentrations of OCPs and PCBs in tilapia samples from the main rivers system in the southern Guangxi are summarized in Table 1. PCBs and OCPs were detected in the muscle of all tilapia samples. DDTs were the predominant contaminant with a median concentration of 15.2 ng/g lw, and endosulfan was the second most common contaminant with a median concentration of 12.2 ng/g lw. PCBs, Drins, HCB, and HCHs concentrations in the fish were relatively low with median concentrations between 1.37 and 9.11 ng/g lw. The concentrations of the various OC compounds measured in the tilapia samples in this study were lower than those in tilapia collected from Guangdong province, China⁹, Africa^10,11,12, Europe, and America^13,14,15 (see Supplementary Table S5 online). This study showed that the main rivers in the southern Guangxi have low levels of OCs pollution, and the fish muscle contamination might be related to the low levels of pollution in the water and sediment. According to data from the National Bureau of statistics of China, the gross output value of industry and agriculture in Guangxi has been lower than that of other provinces or regions in China in the past few decades¹⁶. Therefore, the low levels of OCs pollution found in this study area are mainly the result of lower pollution input. In addition, most of the study area is located in the tropics, which have a relatively high perennial temperature. A warm climate is very conducive to enhance the metabolism rate of OCs by organisms¹⁷. The metabolism of organic pollutants by organisms occurs under the catalysis of a series of enzymes^18,19. Factors affecting the enzymatic reaction, such as enzyme concentration and temperature, will affect the metabolism of OCs in organisms. Temperature also affects the air–water partitioning, which influences the volatilization of chemical pollutants from water²⁰. Thus, dissolved chemical concentrations tend to be higher in cooler than in warmer waters²¹. In alignment with this supposition, Sobek et al. (2010) reported a largely reduced difference in bioaccumulation factor of PCBs between the Arctic and the temperate food webs, after adjustment for temperature effect²².

Distribution characteristics

Spatial distribution of OCs

The spatial distributions of seven OC compounds are presented in Fig. 1. The spatial distribution did not show a gradient in selected OCs concentrations. The spatial distribution of OCPs was under the double influence of a global distillation effect and the usage of OCPs²³. Human activities can affect the distribution of OCPs in hilly areas²⁴. However, there was no significant correlation between elevation and the residues of OCPs in this study (non-parametric test, p > 0.05) (Fig. S2). Therefore, the distribution pattern of OCs in this study was hardly affected by global distillation. High levels of OCPs were found in TD and GG, where endosulfan or DDTs were the predominant contributors. Endosulfan is a cyclodiene pesticide extensively used throughout the world to control a wide variety of insects and mites²³. Endosulfan levels were remarkably higher (10–411 times) in tilapia samples from TD than in samples from other sites. This observation was consistent with the fact that the local fruit and vegetable farming industry is the primary income source in the TD²⁵. Therefore, we believe that the high levels of endosulfan in this study could be attributed to local pesticide practices specific to pest control needs over a short period²⁶. Similarly, the higher levels of DDTs observed in GG also might be related to local short-term agricultural activities.

PCBs are ubiquitous in tilapia samples from the study area, with a detection rate of 100%. In contrast to the OCP compounds, the overall trend of the PCBs was fairly homogenous. A relatively high median PCB concentration was detected in tilapia samples from TX, while slightly lower concentrations were detected from PN. There were no significant differences among different sampling locations (t-test, p > 0.05). The minor differences could be explained by the migration and spread of PCBs in the environment. The limited historical use of PCBs in the present study area is another important factor contributing to this phenomenon²⁵.

Spatial differences in pollutant metabolites

The ratio of parent compounds to their metabolites can provide useful information for the diagnosis of their sources^23,24,27. The scatter plots for isomeric ratios of selected OCPs are shown in Fig. 2.

The ratios between p,p’-DDT, p,p’-DDE and p,p’-DDD have been regarded as an indication of increasing or decreasing inputs to the environment. A ratio of (p,p’-DDE + p,p’-DDD)/p,p’-DDT < 1.0 indicates a fresh application and a ratio of more than 1.0 indicates historical DDTs^23,28. In the present study, the ratios of (p,p’-DDE + p,p’-DDD)/p,p’-DDT were all higher than 1.0, except in TD. Furthermore, p,p’-DDT was not detected in fish samples from NN, TX and WZ. These two facts imply that there is no recent introduction of technical DDT to the river in these regions (Fig. 2a and Table S6). The ratio of (p,p’-DDE + p,p’-DDD)/p,p’-DDT for the TD site indicated the fresh application of technical DDTs in this area. The major source of DDT pollution in China is through the application of technical DDTs and Dicofol in agriculture^3,24. Technical DDT contains higher than 85% of p,p-DDT, and less than 15% of o,p-DDT³. Therefore, the ratio of o,p’-DDT/p,p’-DDT also can be used to determine whether DDT pollution is caused by technical DDT or Dicofol²⁹. Generally, values of o,p’-DDT/p,p’-DDT in the range of 1.3 to 9.3 or higher is indicative of a Dicofol source, while a range of 0.2–0.3 is indicative of a technical DDT source²³. Our study shows o,p’-DDT/p,p’-DDT ratios predominantly ranging between 0 and 1.80. Higher o,p’-DDT/p,p’-DDT ratios (> 1.3) were found in two fish samples from FS (1.80) and CZ (1.71) districts, which indicates that Dicofol may be the main contributor to DDTs in these areas. In summary, the DDT residues in wild tilapia from rivers of the southern Guangxi originated mainly from the historical application of Dicofol and technical DDTs, whereas recent application of technical DDTs are indicated in TD.

Technical endosulfan includes two active ingredients: α-endosulfan (70%) and β-endosulfan (30%)²³. Because α-endosulfan decomposes more easily than β-endosulfan, a α-/β-endosulfan ratio of < 2.33 can be used to distinguish between historical use and recent use. In fish samples in which α- and β-isomer were co-detected, the isomer ratios ranged from 0 to 2.1 (Fig. 2b). The higher values of the α-/β-endosulfan ratio (> 2.33) present in the tilapia samples from TD and FS, indicate continual use of endosulfan in these areas. In the other sites, those ratios are all below 2.33, indicating there was no recent application of technical endosulfan in that area. It is noteworthy that one sample (from TX site) contained β-endosulfan at a level below the limit of detection, but had appreciable levels of α-endosulfan, which may have been transported in from other areas. Because the Henry’s law constant for α-endosulfan is higher than the constant for β-endosulfan, there is a greater tendency for α-endosulfan to evaporate from the surface medium to air^23,30.

The concentrations of ten PCB congeners in the present study area are illustrated in Fig. 3. Using degree of chlorination, these congeners can be divided into light PCBs (2–3 chlorines), medium PCBs (4–6 chlorines), and heavy PCBs (7–10 chlorines). The PCB sources of the 75 fish samples can be classified into the same categories since the PCBs in all sampling sites generally exhibited the following order: heavy PCBs (63.3–86.1% of ∑₁₀PCBs) > medium PCBs (9.72–18.2% of ∑₁₀PCBs) > light PCBs (4.66–18.3% of ∑₁₀PCBs). The higher residual content of heavy PCBs may be related to historical production and use, or relate to their stably chemical structure³¹. Tri-CBs and penta-CBs were the major PCB products manufactured in China from 1965 until they were banned in 1974^31,32. The proportion of these compounds in PCBs was only 2.86–24.7% in this study. On the other hand, light PCBs have higher volatility and a lower octanol–water distribution coefficient than heavy PCBs³³. Once absorbed into the organisms, light PCBs are usually more rapidly metabolized than the more highly chlorinated congeners³⁴. Our results also indicated that the proportion of deca-CBs in heavy PCBs and PCBs was 78.3–98.1% and 51.8–88.6%, respectively. And the sampling sites with high deca-CBs ratio were distributed in the main agricultural farming areas (middle and upper reaches of rivers). And China banned the production of deca-CBs as early as 1974³⁵. Therefore, we believe that historical heritage was the main source of deca-CBs in the study area.

Correlation among biological parameters and OC compounds

We studied the effects of biological parameters (including total length, body mass, age, and lipid content) on the bioaccumulation of individual contaminants in tissue samples of wild tilapia (based on dry weight). The loading and scores plot of PCA based on the concentrations of OCs in the tilapia muscle samples are displayed in Fig. 4. The PC1 explained 58.6% of the total variance and PC2 accounted for 23.5% of the variance. Table S7 lists the correlation coefficients between OC compounds and biological parameters, while the correlation coefficients between OC congeners and biological parameters are listed in Table S8. A significant relationship between growth parameters (i.e. total length, age, and body mass) was found in the tilapia samples, but only age and lipid content were significantly correlated (p < 0.01) (Fig. 4 a). The level of OCs in TD was higher than in other sampling sites and was significantly correlated with endosulfan concentration (p < 0.05) (Fig. 4 b). PCA analysis showed that DDTs, CHLs, OCPs and PCBs were significantly correlated with the biological parameters, suggesting that these compounds were continuously accumulated during the growth of the tilapia. Surprisingly, there was no significant correlation between OCPs and lipid content, which is different from previous studies^35,36,37. Drins and endosulfan were negatively correlated with growth parameters and lipid content. There may be two main reasons for this result. First, endosulfan is likely to be degraded in tilapia because the bioaccumulation capacity of endosulfan is poor (log K_ow = 3.66) when compared to other compounds³⁸. This also may be the main reason for the poor correlation between OCPs and biological parameters. Secondly, Drins might undergo biodilution during the growth of tilapia if the bioaccumulation rate of the pollutant is less than the growth rate³⁴. A significant correlation was observed among HCH isomers, DDT isomers, and its metabolites. HCB was significantly related to DDTs, which suggested that they probably originated from similar contamination sources²³. Similarly, CHLs and endosulfan probably came from similar sources.

In this study, higher correlations were found between contaminants and growth parameters than between contaminants and lipids, which may indicate that growth parameters are the predominant factor in determining OCs bioaccumulation^39,40. This may be due to the fact that the intake of OCs in fish is closely related to their feeding intensity, and the feeding intensity increases with the length and weight of the fish⁴⁰. Furthermore, exposure time (age of fish) is also a governing parameter in the uptake of contaminants. Considering the relatively higher correlation of OCPs and PCBs with total length and body mass rather than age (Table S7), we thought the food intake rate of fish better reflects the degree of pollution than age. After all, feeding is an important way for fish to be exposed to pollutants.

Risk assessment of OCs via fish consumption

Estimated daily intakes (EDIs)

Based on the assessment method are shown in Text S4, the EDIs of OCPs through wild tilapia consumption from ten sites are summarized in Table S9. The assessment assumes a worst-case scenario for residents in the study area by using the highest measured concentrations in the calculation. The results of the study showed that the EDIs for males were almost equal to those for females, although previous studies reported significant differences⁹. For OCPs, dietary intakes from TD were 1.6 to 40 times higher than exposures from the other sites. The EDIs of all OCs through consumption of wild tilapia were far below the recommended acceptable daily intake (ADI)⁴¹, indicating minimum risk caused by these pollutants from tilapia consumption. However, fish and fish products account for higher than 10% of total food consumption in southern China⁹, so higher dietary intake of OCPs or PCBs can be expected when all food sources are taken into account.

Risk assessment

Relevant oral reference dose (RfD) values and cancer slope factors were used following the US Environmental Protection Agency Integrated Risk Information System (IRIS) (Table 2 and Text S5)⁴². Comparing the cancer benchmark concentrations derived from the main rivers in the southern Guangxi (this study) with those from the Chenab River, Pakistan⁴³, almost all cancer benchmark concentrations for the residents in this study were lower than the Chenab River’s values. The primary reasons for this are the lower daily fish consumption rates and pollutant content in the southern Guangxi as compared to in Pakistan^2,9.

Two HRs at the 50th and 95th percentiles were evaluated and summarized to assess both non-cancer and cancer risks associated with consumption of fish containing OC contaminants (Fig. 5). The HRs calculated for non-cancer risk assessment using the 50th and 95th percentiles of PCBs were greater than one, a result of the relatively high concentrations of PCBs in the fish samples collected from the main rivers in Guangxi, South China. The non-cancer risks based on 95th percentile concentrations of endosulfan, DDTs, and chlordane were also greater than one (Fig. 5 a). These results indicated that several compounds ingested via consumption of tilapia have a specific non-carcinogenic risk. On the other hand, the cancer risks associated with fish consumption based on 50th and 95th percentile concentrations of OC compounds were all less than one (Fig. 5 b), suggesting that daily exposure to these contaminants due to fish consumption has a lifetime cancer risk of less than one in one million. The cancer risk for DDTs (0.88) was found to be the highest among all organochlorines, followed by dieldrin (0.43), and PCB (0.29). Therefore, OCs do not present a lifetime cancer risk in this study, there is an increased risk for those who consume large quantities of wild tilapia.

There are several limitations in the present study. This investigation did not consider (1) potential risk based on different consumer groups and co-intake of many foods; (2) possible interactions among various toxic chemicals; (3) over-conservative risk assessment. Research on the influence of these limitations on the assessment results is in progress.

Source: Ecology - nature.com