Occurrence, quantities and probable human health risks of indicatory polychlorinated biphenyls in processed Lates niloticus (L.) products from Lake Victoria in Tanzania

A study was conducted in Lake Victoria to assess the occurrence, levels and risks of indicator polychlorinated biphenyls (PCBs) in four processed Lates niloticus (L.) products (salted-sundried, trims, smoked and deep-fried). Samples extractions were done using QuEChERS method while detection and quantification of congeners was done using a GC- ECD and GC- MS. Six PCBs (CB 28, CB 52, CB 118, CB 138, CB 153 and CB 180) were detected at measurable quantities in fish products. The PCBs; CB 138, CB 153 and CB 180 dominated the loading due to their structures and high degree of chlorination. However, the mean concentration of ΣPCBs in this study were below MRL of 75 µg/kg set for fish by European Commission, implying that the fish products were safe for human consumption in regard to indicator PCBs. Similarly, indicator PCBs, CB 138, CB 153 and CB 180 were more prevalent (20 to 80%) in all fish products than other congeners. For both adults and children the cancer risks were low-to-moderate (ranging from 2.0E-04 to 3.0E-04 for adults and 2.0E-04 to 1.0E-03 for children) while the non-cancer risks were insignificant as the Hazard Indices were less than one.


INTRODUCTION
Polychlorinated biphenyls (PCBs) are synthetic organic compounds that are characterized by their lipophilicity (Cok et al., 2007;Bordajandi et al., 2008;Bjermo et al., 2013), persistence in the environment (due to longer halflives), toxicity, long range atmospheric transport (LRAT) (Liu et al., 2007), accumulation in biota and increase in concentration with time at higher trophic levels (Polder et al., 2014;Ssebugere et al., 2014;Oluoch-Otiego et al., 2016).The compounds are also regarded as endocrine disruptors (EDs) (Bell, 2014) as they alter the normal functioning of the endocrine system and are difficult to degrade in the environment (Field and Sierra-Alvarez, 2008;Frouin et al., 2013).Most are reported as potential carcinogens being responsible for breast, liver and testicular cancers.They also have negative reproductive effects such as low birth weights, small head circumferences, miscarriages, poor sperm quality and low sperm counts (Bell, 2014).
In recent years Lake Victoria fisheries sector has had an abusive history as in 1998, fish exports from Lake Victoria to the European Union were temporarily banned following observations of tainted fish, which were later proved to have been harvested using endosulfan (Henry and Kishimba, 2006).There are also unsubstantiated claims that in order to extent the shelf life of products in the markets some unfaithful fish sellers store their fish products using chemicals such as pesticides and other unknown repellents which might be potential sources of PCBs in fish products along the fish value chain from fishing to consumption.It is also suspected that some processors use transformer oils for fish frying and consumption of smoked fish products is common in the area.Both transformer oil and smoke are good sources of PCBs in the environment and environmental compartments (Wenaty et al., 2019a;Witczak, 2012).
A study carried out in Poland revealed that smoking fish increases the levels of PCBs specifically; CB 101, CB 118, CB 138, CB 153 and CB 180.The reason for this was found to be due to a decrease in co-distillation with steam and the consequence of their penetration with smoke to the fish meat tissue (Witczak andCiereszko, 2006, 2012).There is therefore a need to investigate the impacts of smoking on fish products from Lake Victoria.Another study which was done on fried fish products revealed that deep frying reduces the levels of PCBs in fish due to the fact that; (i) deep frying process creates unique cooking conditions that accelerates drying of the fillets (ii) evaporation of water and PCBs from the fillets as a result of high temperature of the cooking oil and by transfer of PCBs to the cooking oil which itself could be acting as an extraction solvent (Witczak and Ciereszko, 2012), though the same compounds are likely to be reintroduced into fish muscles as the cooking oil is reused in subsequent fish processing.Such studies have not been undertaken in Africa and in Lake Victoria in particular thus it is necessary that a study be designed to reveal the prevalence of PCBs and how safe the fish products are.Being lipophilic in nature, high levels of PCBs could be anticipated in fish products such as fish trims that are mainly the fatty tissues of the fish mass.We could also expect high prevalence and levels of PCBs in smoked products due to the reason that the process takes place in closed system and therefore there is no room for escape of PCBs together with water vapour.PCBs are likely to stick in the walls of smoking chambers and go back to surfaces of the fish tissues being processed (Witczak, 2012).
This study therefore was designed to assess the prevalence, levels and risks of indicator PCBs in four processed L. niloticus products sold such as saltedsundried, fish trims, smoked fish and deep-fried.The products are mainly consumed by the low and middle income communities in the domestic and regional markets and that to our knowledge no studies regarding the prevalence, levels and risks of PCBs in such products (salted-sundried and trims) have been reported anywhere around the globe and limited studies on smoked and deep fried fish have been reported in developed nations.Studies of such kind have mainly been undertaken for fish products intended for export markets such as fish fillets.Similar studies need to be done for products that are processed for domestic and regional markets as they are consumed by the majority of the low income population and that the safety of such products in terms of chemical hazards particularly PCBs is still unknown.This study focused on only indicatory PCBs because they are known to be more persistent and bio accumulative in food chain compared to other congeners.They are therefore assumed to be a suitable representative for all PCBs.

Description of the study area
The study was conducted at the Kirumba International Fish Market in Mwanza between April and August 2018.The Kirumba International Fish Market was purposively selected out of other markets in Lake Victoria because it is the largest fish market in the zone that collects fish products from all other regions making up the Tanzanian side of Lake Victoria.Fish folks and other processors bring their already processed fish products at the market and sell them to buyers who then deliver the products to different markets located within and outside the country, such as Uganda, DRC, Burundi, Rwanda, Zambia, Kenya and Malawi where they are used as source of protein as well as income (LVFO, 2013).
Fish samples extraction and clean-up for determination of indicator PCBs was effected using a quick, easy, cheap, effective, rugged and safe (QuEChERS) procedure with some modifications at the National Fish Quality Control Laboratory in Mwanza, Tanzania.Thirty grams of each sample was measured in triplicates and blended to homogenize.Thirty grams of the composite samples were transferred into 200 ml centrifuge tubes.Thereafter, 2.5 g of sodium bicarbonate (NaHCO3), 60 ml of ethyl acetate and 15 g of anhydrous Na2SO4 were added and placed in a shaking machine to homogenize for 20 min.The supernatants were transferred into 15 ml centrifuge tubes containing 0.125 g of Primary Secondary Amine (PSA) and 0.75 g of anhydrous MgSO4 (Anastassiades et al., 2003;Wenaty et al., 2019a, b).The mixture was centrifuged at 2500 rpm for 10 min and left to separate for further 5 min.The supernatants were transferred into vials ready for GC analysis.

Recoveries and analytical quality control
Recovery tests were done for six indicator PCBs of interest.Blank samples were spiked with standards and were subsequently extracted and analysed in the same way as other samples.To maintain the quality of analytical results blanks and standards were run every after five samples.

Chemical analysis
Chemical analysis was performed at the National Fish Quality Control Laboratory in Mwanza.The samples of fish collected were analysed for Polychlorinated biphenyls (Σ-7PCBs); with IUPAC numbers: 52,101,118,138,153 and 180.The seven indicator PCB congeners were chosen based on their persistence in food web and their tendency to increase in concentration at higher trophic levels.The studied compounds are listed in the Stockholm Convention on POPs for initial elimination and reduction in use because of their effects on environment as well as living organisms.

Gas chromatographic analysis of samples
The determination of the PCBs in the fish samples was carried out using gas chromatography (GC).A gas chromatograph (GC-2010, Shimadzu) equipped with 63 Ni Electron Capture Detector (ECD) and a non-polar (HP-5MS) capillary column of 30 m length × 0.25 mm i.d.× 0.25 μm film thickness was used.Nitrogen was used as both a carrier and make-up gas at a flow rate of 23.7 ml min -1 .The temperature programme was: initial temperature of 120°C held for 2 min, then increased at a rate of 10°C min -1 to 270°C held for 1 min, and at a rate of 2°C min -1 to the final temperature of 290°C held for 3 min.The injector and detector temperatures were 220 and 290°C, respectively.The GC was operated in a splitless mode with an injection volume of 1 μL.The standard mixture was injected in the beginning and after every five samples.Samples were injected in duplicate.The confirmation of the findings was done using gas chromatography-mass spectrometry (Shimadzu GC-MS QP 2010 Ultra equipped with a mass selective detector-MSD, fused silica capillary column Rtx-5MS of 30 m length×0.25 mm i.d×0.25 μm film and an autosampler) applying the procedures described by Mahugija et al. (2018).The GC-MS was performed in splitless injection mode and the mass spectrometer was operated in electron impact (EI) ionization and full scan mode.The calibration/working standard solutions were prepared by dissolving portions of the stock solutions in the same solvents as used for the samples.Calibration curves were prepared by running series of mixtures of standard solutions and plotting the peak areas against Wenaty et al. 419 concentrations.Identification of the compounds involved checking the matching of the retention times and the mass spectra of the PCBs in samples to those of external reference standards that were prepared and run at the same conditions as for the samples.Quantification was carried out by linear integration of the standards and sample data based on peak areas.

Data analysis
Statistical analysis used for data analysis includes subjecting the measured PCBs data to descriptive statistics for the deduction of minimum, maximum, mean concentrations and standard deviations of the detected PCBs.Data was further subjected to SPSS, Version 16.0.Data on PCB concentration were presented as mean ± SD.One-way ANOVA was used to compare concentrations between products.In data processing, the concentrations of PCBs in samples below the limit of detection (<LOD) were treated as zero.
Separation of means was done using Duncan's Multiple Range Test.Significance was declared different at p<0.05 for all analyses.

Risk assessment model
The estimated dose (CDI) received through consumption of fish products was calculated using Equation 1 and the cancer risk (CR) using Equation 2, which were adopted from the Environmental Protection Agency of the United States (USEPA, 1997(USEPA, , 2009;;Man et al., 2013). ( (2) For non-carcinogenic risks, the hazard quotients (HQ) of each congener PCBs measured were calculated by using Equation 3 and overall non-cancer risk using Equation 4. (3) Where; CDI (mg/kg-day) is the estimated chronic daily intake, CR is the cancer risk via consumption of fish products contaminated with PCBs, C (mg/kg) is the measured concentration of indicator PCBs in fish products, IR (kg/day) is the fish consumption rate; for this study 0.37 kg/day was used (Wenaty et al., 2018) (Man et al., 2013;Ge et al., 2013;ATSDR, 1995).For non-carcinogenic risks, hazard index (HI) greater than one was considered risky while HI less than one was considered no risk associated with consumption of fish products (Wenaty et al., 2019a, b).

Recovery experiment for PCBs extraction procedure
The results for the recovery experiment are shown in Table 1.The mean percentage recoveries for PCBs extraction procedure ranged from 71.42±0.21% to 92.48±0.31%based on triplicate determinations.Studies have indicated that recoveries ranging between 70 and 120%, the extraction procedure is considered perfect (Afful et al., 2013a, b).Results herein suggest a perfect extraction method that is recommended for use in further PCBs studies, hence needing no corrections for the recoveries.The pattern of the total PCBs loading was Deep-fried products<Salted-sundried products<Smoked products< Trims.The loading for deep fried products was significantly different (DF = 3.00, F = 1.53 and P = 0.038) from the rest of the products.The total PCBs loading for other investigated products were quite similar.Low levels of PCBs in deep fried products could be attributed to the fact that at high temperatures the cooking oil acts as an extracting solvent, thus high levels of PCBs are expected to be left with the oil (Witczak, 2009a, b).

Concentrations of PCBs in L. niloticus fish products from Lake Victoria
The trims are the fatty parts of fish, thus PCBs being highly lipophilic are mainly concentrated in fatty tissues (Bjermo et al., 2013;Polder et al., 2014).For smoked products, high concentrations of PCBs could be due to the reason that there is reduced co-distillation of the components with water vapour (Witczak and Ciereszko, 2006) and removal of water from the product as the PCBs are soluble in fat and lipids.Similarly, studies have shown that smoke consist certain amounts of PCBs (Witczak, 2012) and therefore acting as sources of these persistent organochlorine compounds in foods.
For all fish products considered in this study, the PCBs loading was dominated by CB 138, CB 153 followed by CB 180.Comparable results are also reported in previous studies (Polder et al., 2014;Ssebugere et al., 2014;Oluoch-Otiego et al., 2016).The domination tendency of CB 138, CB 153 and CB 180 are also reported in other previous studies (Polder et al., 2014;Oluoch-Otiego et al., 2016).This is due to the fact that CB 138, CB 153 and CB 180 are not metabolized by certain organisms compared to the rest of congeners (Ssebugere et al., 2014).Boon et al. (1997) also reported that the rate of metabolisms of PCBs depends mainly on structure and the degree of chlorination of the molecule.Being highly chlorinated, CB 138, CB 153 and CB 180 tend to have longer half-lives, persistent to biodegradation and therefore easily detected in environmental samples.The contribution of the three congener PCBs to total loading was 53.1% for salted and sundried products, 66.7% for trims, 53.3% for smoked products and 60.1% for deep fried products.
However, the mean concentration of ΣPCBs in this study were within the limit of 75 µg/kg set for fish by European Commission (EC, 2011), implying that the fish products were safe for human consumption in regard to indicator PCBs.Furthermore the total loading as per this study are far higher than that found by Polder et al. (2014) (0.57 µg/kg) from Lake Tanganyika for fresh Oreochromis niloticus samples, higher than those detected by Ssebugere et al. (2014) (0.229 to 0.716 µg/kg) in fresh L. niloticus from the Ugandan side of Lake Victoria.This indicates that processed products have higher levels than fresh fish.It is therefore suggested that some fish processing technologies such as smoking are main sources of PCBs in food products.

Prevalence of the detected indicator PCBs in different processed L. niloticus fishery products
The prevalence (%) of the indicator PCBs detected in different L. niloticus processed products are shown in Figure 1.For the six indicator PCBs that were detected in four fish products, all were less prevalent in deep fried products being detected in only 20% of all samples considered in this study.Generally, for the rest of the products the prevalence was in the following ranges: 40 to 60% for salted-sundried fish products, 50 to 70% for trims and 40 to 80% for smoked products.The mean percentage prevalence followed this trend: Deep fried products<Salted-sundried products<Trims<Smoked products.Indicator PCBs, CB 138, CB 153 and CB 180 were more prevalent (20 to 80%) in all fish products than other congeners.This is attributed to their structures and high degree of chlorination.

Human health risk assessment
Human health risk assessment for the indicator PCBs measured in four processed fish products considered in this study was evaluated using Equations 1 to 4 and ATSDR standard for adults and children.As shown in Table 4, the cancer risks based on indicator PCBs loading for adults were in a range between 2.0E-04 and 3.0E-04 while for children were between 2.0E-04 and 1.0E-03.These values are within the range 1.0E-04≤value<1.0E-03and classified as low to moderate risk (ATSDR, 2007;Man et al., 2013;Wenaty et al., 2019a, b).This observation suggests that there are only few cancer risks of indicator PCBs associated with consumption of L. niloticus products from Lake Victoria.Based on ATSDR standard, the cancer risks for PCBs in this study are between low to moderate.
Table 5 shows the hazard quotients (HQs) and hazard indices (HI) defining the non-cancer risks of indicator PCBs associated with consumption of L. niloticus products from Lake Victoria.The Hazard Indices (HI)(sum of Hazard Quotients (HQs)) ranged between 8.5E-03 and 1.8E-02 for adults and between 1.0E-02 and 2.0E-02 for children.In both cases, the HI values were very low (less than one).The United States Environmental Protection Agency (USEPA, 2009), recommends that HI values less than one indicates no risk.Therefore results from this study suggest that the risks associated with consumption of the analysed fish products from Lake Victoria are insignificant for both adults and children in regards to indicator PCBs (Table 4).

CONCLUSION AND RECOMMENDATIONS
This study analysed the occurrence, levels and risks of indicator PCBs in processed L. niloticus products from Lake Victoria in Tanzania.The levels of the detected PCBs were below the maximum recommended limits for fish and fishery products.The investigated fish products are therefore safe for human consumptions in regards to indicator PCB residues.Human health risk assessment indicated low cancer risks and insignificant non-cancer risks suggesting that the fish products do not present a health risk.However, follow up studies to assess the cooking oil and the influence of fish processing such as deep-frying and smoking on levels of indicator PCBs in fish products are hereby recommended.

Table 1 .
Results of the percentage recoveries for PCBs extraction procedure.

Table 2 .
Mean concentrations (µg/kg) of individual PCBs and ∑PCBs in processed fish products from Lake Victoria in Tanzania.

Table 3 .
Analysis of Variance for the detected PCBs in processed fish products from Lake Victoria.* Means are significantly different at 0.05 level.DF: Degree of freedom, F: F -Value and P: P -Value. *

Table 4 .
Lifetime cancer risks for the indicator PCBs detected in processed L. niloticus products from Lake Victoria.stands for adults, CHI for children, CR for lifetime cancer risk and ND for Not Determined as the concentration was <LOD. AD

Table 5 .
Non-carcinogenic risks of the indicator PCBs measured in processed L. niloticus products from Lake Victoria.stands for adults, CHI for children, HQ for hazard quotient, HI for hazard index and ND for not determined. AD