Bioaccumulation of heavy metals in some commercially important fishes from a tropical river estuary suggests higher potential health risk in children than adults

The Karnaphuli River estuary, located in southeast coast of Bangladesh, is largely exposed to heavy metal contamination as it receives a huge amount of untreated industrial effluents from the Chottagram City. This study aimed to assess the concentrations of five heavy metals (As, Pb, Cd, Cr and Cu) and their bioaccumulation status in six commercially important fishes, and also to evaluate the potential human health risk for local consumers. The hierarchy of the measured concentration level (mg/kg) of the metals was as follows: Pb (13.88) > Cu (12.10) > As (4.89) > Cr (3.36) > Cd (0.39). The Fulton’s condition factor denoted that fishes were in better ‘condition’ and most of the species were in positive allometric growth. The bioaccumulation factors (BAFs) of the contaminants observed in the species were in the following orders: Cu (1971.42) > As (1042.93) > Pb (913.66) > Cr (864.99) > Cd (252.03), and among the specimens, demersal fish, Apocryptes bato appeared to be the most bioaccumulative organism. Estimated daily intake (EDI), target hazard quotient (THQ), hazard index (HI) and carcinogenic risk (CR) assessed for potential human health risk implications suggest that the values were within the acceptable threshold for both adults and children. However, calculated CR values indicated that both age groups were not far from the risk, and HI values demonstrated that children were nearly 6 times more susceptible to non-carcinogenic and carcinogenic health effects than adults.


Introduction
Heavy metals pollution has become a major concern worldwide due to their toxicity, intrinsic persistence, non-biodegradable nature, and accumulative behaviors [1]. These metals differ from other toxic materials in a way that they are neither created nor destroyed by human. They are inert in the environment and are often considered to be conservative pollutants if left a1111111111 a1111111111 a1111111111 a1111111111 a1111111111

Sampling
The study area (Karnaphuli River estuary) is located from 22.234008 N and 91.821105 E to 22.289695 N and 91.794403 E (Fig 1). A total of six commercial species (i.e. Apocryptes bato (Chewa), Pampus chinensis (Rup chanda), Liza parsia (Bata), Mugil cephalus (Flathead bata), Hyporhamphus limbatus (Ek Thuitta), and Tenualosa toil (Chandana ilish) were collected from fishermen for a period of seven months (February 2018 to August 2018) using seine net. The collected samples were carried in plastic ice container and immediately stored at -20˚C. Afterwards in the lab, total length (cm) and weights (gm) were measured carefully to the nearest 0.1 cm using a vernier caliper; total weight was determined with an electronic balance to 0.01 g accuracy. Muscles of each specimen were dissected with stainless steel scissors for further chemical analysis using inductively coupled plasma mass spectrometry (ICP-MS, Model: ELAN9000, Perkin-Elmer, Germany) for metal detection. Data were analysed statistically by fitting a straight line adopting the least square method.

Chemical analysis procedures and accuracy
A 1.5 g of dissected muscle was used for subsequent analysis. The samples were dried in a Zirbus freeze drying machine (Model: VaCo 2, Germany, CFC free, condenser dimensions-Ø200×200 mm) for 24 h [30]. After cooling, 0.5 g of muscle samples were placed in 10ml of of 100% HNO 3. The solution was heated on an oil bath and added with 4 drops of H 2 O 2 repeatedly until the mixture turned clear. The solution was mineralized using microwave digester (model: WX-6000, China). After filtering, solution was diluted for 2h and kept in a 50 ml tezaron tube [3]. A standard reagent (Merck VI; Germany) was used to analyze the metal concentration in the targeted fish species, which was prepared from the same acid matrix [1]. Moreover, certified reference material (CRM 320, Merck KGaA, Germany) was employed for validation and accuracy check [3].

Metal pollution index (MPI)
To assess the metal pollution, the metal pollution index (MPI) was adopted following [31] and [32]. The equation is as follows: where, CM 1 is the concentration value of first concerned metal, CM 2 is the concentration value of second concerned metal, CM 3 is the concentration of third concerned metal, CM n is the concentration of n th metal (mg/kg dry wt) in the tissue sample of a certain species.

Statistical analysis
The mean and standard deviation of metal concentrations were calculated. The Kolmogorov-Smirnov, Shapiro-Wilk and Kruskal-Wallis tests were performed using SPSS 23. The Kolmogorov-Smirnov and Shapiro-Wilk tests were performed to check the normality of the data [33]. The Kruskal-Wallis test was carried out to identify significant variance of the targeted elements in the specimens of the studied area where p � 0.05 was used as the cut-off for significance (confidence level in 95%). In addition, the Levene's test with the significant level at p � 0.05 was adopted to determine the homogeneity of variances in terms of ANOVA tests. Correlation matrix (CM) and principal component analysis (PCA) were performed to determine the correlations and association between heavy metals in the studied fish species [34]. The correlation matrix used to analyze heavy metals could result in both positive and negative outcomes [35][36][37][38]. Hierarchical cluster analysis is one of the most widely used hierarchical algorithms which results in clusters in which variables or individuals are added in sequence considering hierarchy of the cluster [39]. Additionally, the similar groups of the studied elements within the sampling sites measured as a distance between the two closest members, and is conducted by Origin 9 software [3,40]. Metals with similar properties were pooled in one or associated cluster while the dissimilar groups of elements were plotted in a separate cluster, and thus differentiated the contamination status of the samples [41-43].

Bioaccumulation factor
Bioaccumulation factors (BAFs) were calculated as a ratio between the concentration level of biota (those in water) and the living environment of the specimens, and was expressed as follows [44-46]: where Cn Biota is the concentration of metal in the tissues (mg/kg) and Cn Water is the metal concentration in the aquatic environment (mg/l). BAF is categorized as follows: BAF < 1000: no probability of accumulation; 1000 < BAF < 5000: bioaccumulative; BAF > 5000: extremely bioaccumulative [47].

Length-weight relationship and condition factor
The length-weight relationship of the fish samples were calculated using Fulton condition factor following the equation [48-51]: where W is the total body weight of fish (gm), L is the total length of fish (cm). Fulton's Q is categorized as follows: Q = 1: Condition is poor, Q = 1.2: condition is moderate, Q � 1.40: condition is proportionally good [52]. The equation can be expressed by the following formula [12,53]: The equation can be estimated using the least-square formula adopted with the logarithm form of the equation is shown as [54]: where 'a' is the calculated intercept of the regression line, and 'b' is the coefficient of that regression. The 'b' values signify the growth pattern of an organism which can be classified as follows: b < 3: negative allometric, b = 3: Isometric and b > 3: positive allometric [48, 55]. As Fulton's Q is substantially correlated with the length-weight relationship, exponent 'b' acts an identical role of determining the well-being of the organisms [48]. The deviation of the condition, further, depends on the food availability and the divergence of reproductive organ development [56].

Estimated daily intake (EDI)
Estimated daily intake (EDI) was calculated by the following equation [22,57,58]: where Cn is the concentration level of metal in the selected fish tissues (mg/kg dry-wt); IGr is the acceptable ingestion rate, which is 55.5 g/day for adults and 52.5 g/day for children [59,60]; Bwt is the body weight: 70 kg for adults and 15 kg for children [59].

Target hazard quotient (THQ) for non-carcinogenic risk assessment
THQ was estimated by the ratio of EDI and oral reference dose (RfD). RfDs of the different metals for example As, Pb, Cd, Cr, and Cu are 0.0003, 0.002, 0.001, 0.003 and 0.3, respectively [59,61]. The value of ratio < 1 implies a non-significant risk effects [62]. The THQ formula is expressed as follows [60,[63][64][65]: Where E d is exposure duration (

Hazard index (HI)
Hazard index (HI) was calculated for the multiple elements (Hg, As, Mn, and Cr) found in the fish samples and the equation is as follows [3,19,66,67]: where, THQs is the estimated risk value for individual metal [68]. When HI value is higher than 10, the non-carcinogenic risk effect is considered high for exposed consumers [69][70][71].

The concentration of heavy metals and source identification
The average concentration of Pb, Cu, As, Cr, and Cd from the fish tissues were 13 Table 1). From the results, Pb was 1.14, 2.8, 4.13, and 35.44 folds higher than Cu, As, Cr, and Cd respectively, and contributed 40% of all the elements in the study area. Along with that, Pb attributed a higher concentration of 15.73 mg/kg at S1, whereas Cd was accounted for lower concentration of 0.29 mg/kg on average observed at S1. The evaluated MPIs ranged from 3.65 mg/kg to 4.70 mg/kg with the mean of 4.02 mg/kg (Table 1). Due to higher concentration level, the maximum MPI value (4.70 mg/kg) was corresponded to A. bato, followed by P. chinensis From the statistical results, Kolmogorov-Smirnov and Shapiro-Walk test revealed that the metals in the targeted fish species were non-normally distributed along the study area. The adopted Levene's tests showed that metals were non-homogenously distributed. Kruskal-Wallis test identified that the distribution of metal was significantly different (p � 0.05) in the fish species along the sampling stations. From Table 2, among the species, L. parsia and H. limbatus exhibited significant relationship (regression line) with As only, where T. toil showed significant relationship with As and Pb (p � 0.05). None of the other metals showed a significant linear relationship with the organisms. Among the species, T. toil showed the maximum response for Pb (R 2 = 99.5%), followed by L. parsia for As (R 2 = 83.7%). Table 3  To execute hierarchical cluster dendrogram, the Ward-Linkage method was employed with Euclidean distance, which resulted in three distinct clusters, presented in Fig 3. Cluster 1 included As and Cr that could have been originated from anthropogenic activities like chemical industries. Pb, and Cu confined in cluster 2 could have been attributed from the textile organization, fertilization and oil droppings from the boats/ships in the study area. Lastly, Cd was included in cluster 3, and source of the Cd could have come from battery industry.

Length-weight Relationship and condition factor evaluation
Higher 'b' value reflected the appetite state and reproductive organ development of the species [55]. The identified b value of L. parsia from the length weight relationship was close to 3, hence, it represented isometric growth pattern that was considered as ideal shape (Table 4). Meanwhile, among all, species, P. chinensis exhibited the highest positive allometric growth, which was 37 times higher, than average value and 14 folds, on average, higher than other species.

Bioaccumulation (BAF) status of targeted species
The estimated BAFs are depicted in Fig 4. The BAFs were ranged from 110.53 for Cd observed in S4 to 3353.7 for Cu as well. The minimum value was found for T. toil, on the

Estimated daily intake (EDI)
The explored EDI of two concerned age groups, adults and children, is presented and summarized in Table 5. Adults and children showed comparably higher EDIs for demersal species than pelagic ones. In the consequence, high doses of demersal species were exposed to the consumers through consuming metal affected fish species as the food items. The EDIs for both groups were organized in the following order: Pb > Cu > As > Cr > Cd.

Target hazard quotient (THQ) and Hazard Index (HI) for noncarcinogenic risk
The assessed target hazard quotient (THQ) for the studied fish species are displayed in Table 6. THQs in the adult group induced to As, Pb, Cd, Cr, and Cu were 0.016, 0.003, 3.0E-04, 0.001 and 2.38E-04, respectively, whereas for children the values were 0.097, 0.014, 0.001, 0.004 and 0.001, respectively. Moreover, the rank of the THQs of the elements was as follows: As > Pb > Cr > Cd > Cu. While, for the cumulative scenario of HI, children were 5.83 times more susceptible than adults. However, the investigated HI was not surpass the recommended limit (Table 6).

Carcinogenic risk (CR)
Exposure of CR was estimated for a particular element and summarized in The results showed that children were exposed to higher CRs than adults. But, calculated CR values for both age groups were noted far from the risk as acceptable range is 10 −6 to 10 −4 .

Concentration of heavy metals and source identification
In the present study, T. toil had the highest concentration of As, whereas A. bato exhibited the maximum concentration for Pb, Cd and Cu, and P. chinensis had the highest concentration of Cr. The metal concentrations in the edible tissues were ranked in the following sequence: A. bato > P. chinensis > H. limbatus > T. toli > M. cephalus > L. parsia. Data from the previous literatures showed that metal concentrations in fish muscles varied widely depending on the location and species. The As and Pb concentrations in this study were higher than all other findings and recommended guidelines. Although Cd concentration was at the lower level comparing Northeast coast, Ganga River, and Pearl River, it surpassed all the guideline values   [86], which were generally lower than the present findings. In this study, the demersal species had comparatively higher concentration level of metals than pelagic ones as they inhabit close to the bottom or sediment. The MPI values of 3.65 to 4.70 in the present study were much lower than those of Blackchin tilapia, Sarotherodon melanotheron and Silver catfish, Chrysichthys nigrodigitatus (8.1 to 17.76) from Okrika Estuary, Nigeria. This is most likely due to the oil bunkering and transportation activities along the study sites [94]. The findings of MPI in the present study were almost similar to that of Rutilus rutilus in Pluszne Lake [12]. However, the metal accumulation in fishes could be highly influenced by sampling locations and habitats [95,96]. Previous studies have shown that heavy metals in the aquatic environment could come from different natural and anthropogenic sources [1,3,39,97], where many factors influence their concentrations e.g., the original levels of rocks and parent materials, processes of soil formation, contamination by human activities, and other anthropogenic factors [98]. Generally, high correlations between specific heavy metals in the environment may reflect similar levels of contamination and/or release from the same sources of pollution [40]. In this study, strong positive correlations were found between Cd and Pb, Cr and Pb, Cr and Cd, and Cr and Cu indicating that they had the same source either natural or anthropogenic. The employed PCA revealed that the source of origin of the metal was anthropogenic. Pb, Cd, and Cu were the dominant compounds in PCA analysis due to their high loading scores. PCA also revealed the changes of geochemical composition and the justification of cluster analysis [99]. In the cluster analysis, Pb, Cd and Cu are grouped together in cluster 2, indicating that the anthropogenic sources of these heavy metals are closely related to the sediments of the study area. There are several manufacturing industries near the study area which largely use alloy, paints and poisonous chemicals containing As. The commercial uses of heavy metals in modern microelectronic and optical industries are notable sources of As intrusion in the aquatic environment [100]. Non-essential element, Pb originated from extreme agriculture, poultry farms, industries and textile mills might end up in the aquatic ecosystem [101], thus contributing to the metal in the study area. Thus the benthic feeders are to be greatly affected by the deposited Pb in the ecosystem [91]. Cd is typically found at a low concentration in the aquatic environment, however, indiscriminate use of phosphate fertilizer and industries are two primary sources of Cd introduction [102]. In the study area, nickel-cadmium battery manufacturing along with industries engaged with Cd metal incineration and production might increase the  Cd concentration level in the aquatic environment [103,104]. Besides Cd and Pb, Cr is also widely introduced in the textile industries [105]. Near the bank of the Karnaphuli estuary, such commercial textile industries produce colour pigment and thus become a common contaminant for the aquatic ecosystem [106]. Notably, a considerable level of Cu become swelled up in the study area due to oil droppings from ships and boats, recurrent usage of antifouling paints and other boating interferes [107].

Bioaccumulation (BAF) status of target species
The bioaccumulation potential of metals was assessed in the muscles of various fish species, and was found to vary from species to species. The hierarchy suggested that most of the species were tend to be bioaccumulative (BAF value close to 1000). A. bato exhibited the highest bioaccumulation in the studied area. The accumulation of the metal elements in an aquatic organism depends upon the classification of species, invasion pathways, metabolic characters of the sampled tissues and finally, the surrounding environmental condition in which the species live [108]. In our results, the BAFs of As, Cd, Cr, Pb and Cu were relatively higher than those of Pearl River estuary [90], where the BAFs in Tilapia were reported to be Cd > Cu > Pb > Cr. Such reports were mostly in line with our results. The fact is that, Cu is actively persistent in muscles due to being an essential element of living tissue [60,109]. Notably, the bioaccumulation capability of Cd takes a long time to spare, thus making it relatively infirm [20].

Human health risk evaluation
EDI, based on the oral reference dose (RfD) for an individual element [110], reflects the daily exposure to the toxic element, and is executed to avoid any harmful effect on human health [63]. The records of EDI of the people were compared with recommended daily allowance (RDA), provided by WHO, and introduced that, mean EDI values of the metals were still lower than RDAs. The values lower than RDA guidelines suggested a lower possible health effect of the elements to the consumers. However, it would not be wise to take it as a permanent measurement to reach a final conclusion [61,63]. The <1 THQ values for both adult and children suggested that the adverse effect on human health might not occur. Similarly, the HI results also followed the THQ trend. Hence, there is no potential non-carcinogenic effect for the consumers due to intake of the fish species. Studies carried out by several authors in similar conditions were in line with our results [32, 69,[111][112][113]. In general, the assessment of THQ for human health risk evaluation has no doseresponse relation of the examined elements [114]. However, human can dramatically suffer in the long run due to multiple simultaneous pollutants [77].
The CR values lower than 10 −6 indicates a negligible health risk whereas values in the range of 10−6-10 −4 are in the acceptable belt [63]. The CR values found in this study suggested an acceptable limit, and consumers are therefore less prone to carcinogenic risk. In fact, 90% of the carcinogenic risk is observed in the As contaminated aquatic food items. The inorganic state of As is more lethal than organic one [20,115], and only 10% of the total As can be assessed as inorganic form [63]. The findings of the present study were in the acceptable range (10 −6 to 10 −4 ) compared with Okogwu et al. [116], except for Cr which surpassed the CR limits. Again, in the Persian Gulf, consumers at the threshold limit for As were in concern of carcinogenic risk [117]. For this reason, carcinogenic risk should be given more attention due to intake of aquatic products, especially for the study area.

Conclusions
Five heavy metals in the muscles of six fish species from the Karnaphuli River estuary were measured to investigate their potential sources, bioaccumulation rate and human health risk. Relatively high concentration of Cu was observed in A. bato. In most cases, the metal concentrations exceeded the recommended limits. The maximum metal accumulation was recorded in A. bato. A. bato, P. chinensis and H. limbatus were observed to be extreme bio-accumulative species. The mean bioaccumulation factors of the metals were observed in the studied species as follows: Cu > As > Pb > C > Cd.). Among the fishes studied, demersal fishes had higher levels of heavy metals compared with those fishes in the upper layers, suggesting higher level of heavy metals in the bottom water or sediment compared with the surface water. Accumulation of heavy metals in fish could be resulted from surface contact with the water, by breathing, and via the food chain. Heavy metals in the sediment enter the food chain through the feeding of benthic animals. Pearson correlation analysis showed strong positive correlations between Cd and Pb, Cr and Pb, Cr and Cd, and Cr and Cu indicating that they had the same source, either natural or anthropogenic. As per cluster analysis, Pb, Cd, and Cu, were grouped together, indicating the local anthropogenic sources. Existing textile mills, fertilizer factory, leather industry and agricultural activities in the catchment area of the river might be the sources of high concentration of Pb and Cd. Carcinogenic risk assessment suggested local consumers were free from the risk of cancer for the time being but they might be affected in future upon consumption of fish from studied region. Finally, it is found that children were six times more vulnerable to non-carcinogenic and carcinogenic risks than adults. Nonetheless, further study required ensuring the same conclusions are reached.