Cadmium and arsenic levels in edible fi shes, Oreochromis niloticus (Nile tilapia) and Ompok bimaculatus (butter catfi sh) from Padaviya Reservoir, Sri Lanka and human health risk assessment associated with their dietary exposure

* Corresponding author (shyamalikaww@nara.ac.lk; https://orcid.org/0000-0003-0085-9221) This article is published under the Creative Commons CC-BY-ND License (http://creativecommons.org/licenses/by-nd/4.0/). This license permits use, distribution and reproduction, commercial and non-commercial, provided that the original work is properly cited and is not changed in anyway. Abstract: Consumption of fi sh contaminated with toxic heavy metals is a threat to human health. Cadmium and arsenic are suspected as potential risk factors for chronic kidney disease of unknown aetiology (CKDu), which is highly prevalent in agricultural settlements in dry zone districts of Sri Lanka causing a severe public health crisis. Objectives of this study were (i) to determine cadmium and arsenic levels in muscle, liver and kidney of two edible fi shes (Nile tilapia and butter catfi sh) from the Padaviya reservoir located in the North Central Province, Sri Lanka where CKDu is highly prevalent and (ii) to assess the potential human health risks associated with the dietary exposure through edible muscle of these fi shes. Cadmium and arsenic levels of the tissues in the two fi sh species were determined using inductively coupled plasma mass spectrometry. In both fi sh species, highest cadmium levels were found in kidney (p < 0.05) whereas no signifi cant tissue specifi c diff erences (p > 0.05) were evident for arsenic. In the edible muscle, the maximum detected cadmium level in Nile tilapia was 0.1 mg/kg in wet weight. However, cadmium levels in the muscle of all Butter catfi sh were < 0.05 mg/kg. Arsenic levels in the muscle of all Nile tilapia and Butter catfi sh were < 0.05 mg/kg. Cadmium contents in muscle of most fi shes and arsenic contents in all fi shes (n = 60) were within the maximum permissible limits set by the international food standards regulatory authorities. Based on the estimated daily intake and target hazard quotients (< 1) for lifetime exposure, it is highly unlikely that cadmium and arsenic contents of the muscle meat of these fi shes would pose human health risks to moderate level consumers.

Abstract: Consumption of fi sh contaminated with toxic heavy metals is a threat to human health. Cadmium and arsenic are suspected as potential risk factors for chronic kidney disease of unknown aetiology (CKDu), which is highly prevalent in agricultural settlements in dry zone districts of Sri Lanka causing a severe public health crisis. Objectives of this study were (i) to determine cadmium and arsenic levels in muscle, liver and kidney of two edible fi shes (Nile tilapia and butter catfi sh) from the Padaviya reservoir located in the North Central Province, Sri Lanka where CKDu is highly prevalent and (ii) to assess the potential human health risks associated with the dietary exposure through edible muscle of these fi shes. Cadmium and arsenic levels of the tissues in the two fi sh species were determined using inductively coupled plasma mass spectrometry. In both fi sh species, highest cadmium levels were found in kidney (p < 0.05) whereas no signifi cant tissue specifi c diff erences (p > 0.05) were evident for arsenic. In the edible muscle, the maximum detected cadmium level in Nile tilapia was 0.1 mg/kg in wet weight. However, cadmium levels in the muscle of all Butter catfi sh were < 0.05 mg/kg. Arsenic levels in the muscle of all Nile tilapia and Butter catfi sh were < 0.05 mg/kg. Cadmium contents in muscle of most fi shes and arsenic contents in all fi shes (n = 60) were within the maximum permissible limits set by the international food standards regulatory authorities. Based on the estimated daily intake and target hazard quotients (< 1) for lifetime exposure, it is highly unlikely that cadmium and arsenic contents of the muscle meat of these fi shes would pose human health risks to moderate level consumers.

INTRODUCTION
Consumption of fi sh provides an important source of protein, polyunsaturated fatty acids and essential minerals which are associated with health benefi ts and normal growth. However, contamination of fi sh with toxic heavy metals can aff ect the nutritional and other benefi cial eff ects of fi sh on human health (Castro-González & Méndez-Armenta, 2008). Due to the high degree of toxic eff ects, some heavy metals and metalloids such as cadmium and arsenic have been ranked as priority metals that are of public health signifi cance. Cadmium and arsenic are systemic toxicants that could induce multiple organ damage, even at lower levels of exposure (Tchounwou et al., 2012;Karri et al., 2016;Wise et al., 2017). Inorganic forms of arsenic in food appears to be more toxic than the organic arsenic forms (USEPA, 2000;Castro-González & Méndez-Armenta, 2008;Varol & Sünbül, 2018).
Chronic kidney disease of unknown aetiology (CKDu) is an emerging health problem in some low-

December 2020
Journal of the National Science Foundation of Sri Lanka 48 (4) and middle-income countries, including Sri Lanka, El Salvador and Nicaragua (Weaver et al., 2015). Heavy metals, especially cadmium and arsenic are suspected as potential risk factors for CKDu (Jayatilake et al., 2013). Yet, involvement of these elements in the pathogenesis of CKDu in Sri Lanka is still questionable (Nanayakkara et al., 2019). CKDu is highly prevalent in agricultural settlements in the Dry Zone districts of Sri Lanka causing a severe public health crisis among rural communities (Bandara et al., 2008;Jayatilake et al., 2013). Heavy metal exposure through the dietary intake of water and food may play signifi cant contribution. Reservoir fi sh is the main source of animal protein of the rural people living in the Dry Zone. These reservoirs could be contaminated with heavy metals due to excessive use of agrochemicals in the Dry Zone (Bandara et al., 2008). Heavy metals can accumulate in fi sh directly through contaminated water and indirectly via the food chain (Castro-González & Méndez-Armenta, 2008). Several studies have reported heavy metal contaminations in reservoir fi shes in the Dry Zone of Sri Lanka (Bandara et al., 2008;Allinson et al., 2009;Jayatilake et al., 2013;Jinadasa & Edirisinghe., 2013;Wijesinghe et al., 2018). Some studies showed that the measured heavy metal contents in the edible parts of the fi sh samples did not exceed the maximum limit for human consumption according to the international guidelines (Allinson et al., 2009;Jinadasa & Edirisinghe., 2013). More recently, Wijesinghe et al. (2018) reported that Pb, Cd, Cr, Cu and Zn contents detected in the head part of Oreochromis mossambicus, Oreochromis niloticus and Etroplus suratensis from Minneriya, Parakrama Samudraya and Kaudulla reservoirs were higher than the provisional tolerable weekly intake of these metals specifi ed by the FAO/WHO. The objectives of the present study were to determine the cadmium and arsenic levels in muscle, liver and kidney tissues of two commonly consumed fi sh species, Oreochromis niloticus (Nile tilapia) and Ompok bimaculatus (Butter catfi sh) from Padaviya reservoir located in an area in Sri Lanka where CKDu is highly prevalent. The potential human health risks associated with the consumption of cadmium and arsenic contaminated fi shes was also assessed using three approaches: comparison of metal levels with maximum permissible levels (MPL); comparison of estimated daily intake with reference doses; and estimation of target hazardous quotients for non-carcinogenic eff ects.

Site description
Padaviya Reservoir (8°48′32.30″ N; 80°45′26.04″ E; catchment area: 532.48 km 2 ) is a perennial reservoir located in the Anuradhapura District in the North Central Province of Sri Lanka (Figure 1). Land uses in the area include forestry, paddy cultivation and other agricultural lands. The reservoir is fed by two seasonal streams, namely, Mukunu Oya and Mora Oya. It also receives rainwater during northeast monsoon period (Ministry of Irrigation and Water Resources Management, 2015). Fishery is an important economic activity that takes place in the reservoir. Nile tilapia and Butter catfi sh are two important edible freshwater fi sh species that mainly contribute to the inland fi sheries sector.

Sampling fi sh
Freshly harvested samples of Nile tilapia (body weight: 180 -300g; total length: 19 -24 cm) and Butter catfi sh (body weight: 110 -225g; total length: 24 -34 cm) were randomly obtained from the fi sh landing sites of Padaviya reservoir in February, March and May 2018 covering a dry season where metals can be concentrated in the reservoir due to evapotranspiration. The fi shes (n = 10 fi sh per each visit for each fi sh species) were transported to the laboratory in sealed freezer bags under ice cold storage conditions (4 °C -5 °C). Upon arrival in the laboratory, body weight and total length of the fi shes were measured. From each fi sh, muscle, liver and kidney tissues were obtained using stainless steel dissecting kits and frozen separately at -20 °C until analysis.

Analysis of metals
Analysis of cadmium and arsenic in the fi sh tissues was carried out in the Industrial Technology Institute Colombo using Inductively Coupled Plasma Mass Spectrometry (ICP-MS). Briefl y, homogenised wet tissue samples (0.5 -1.0g) were placed in digestion tubes containing 7.00 mL of concentrated nitric acid (≥ 69 %, AR grade) and 3.00 mL of 30 % hydrogen peroxide. The samples were digested following the microwave digestion procedure for food in the CEM MARS 5 Microwave Digester (AOAC, 2012). The temperature was raised to 190 °C in 20 min and maintained for 20 min until the digestion was completed. The cooled digested solution was diluted to 25 mL with deionised water. Cadmium and arsenic levels in the samples were analysed using ICP-MS (Agilent 7900). Quality control and quality assurance aspects of the metal analyses included the use of analytical grade reagents for all analyses, use of calibration standards and certifi ed reference materials. The correlation coeffi cients and regression coeffi cients of the calibration standards for the linear regression were above 0.995 and 99.7, respectively. Freeze-dried certifi ed reference material (Dog fi sh liver; DOLT 04) from the National Research Council of Canada was used to validate the metal analysis following the same analytical procedure. Recovery of cadmium and arsenic levels in the certifi ed reference material was around 78 % and 91 %, respectively. Limit of Detection (LOD) was 0.05 mg/kg for both cadmium and arsenic.

Human health risk assessment
Potential human health risk associated with cadmium and arsenic through fi sh consumption was assessed using several approaches as described by Varol and Sünbül (2018). Firstly, heavy metal(loid) contents in fi sh samples were compared with the available maximum permissible levels (MPLs) of the respective metal(loid)s set by international food standards regulatory authorities (EC, 2006;FSANZ, 2013;MHPRC, 2013). Secondly, the estimated daily intake (EDI) of each heavy metal(loid), which depends on its concentration in edible fi sh parts (muscle meat) and the amount of fi sh consumption per day, was calculated and compared with reference dose (RfD) values reported by the USEPA (2000). The daily intake of metals from fi sh meals for adults was calculated using the following equation: where, EDI is the estimated daily intake (mg/kg bw/day), MC is the metal concentration in fi sh muscle (mg/kg wet weight), IRd is the daily average fi sh ingestion rate (kg/day) and BW is the average body weight for an adult. In Sri Lankan context, the IRd for fi sh was taken as 0.060 kg/day (Allinson et al., 2009;2010). Average body weight of an adult was considered as 70 kg (the USEPA default value). Considering the worst-case scenario, the maximum metal(loid) level detected in the edible muscle tissue was used as a more cautious approach for human health risk estimations. In cases where specifi c metal(loid) levels were below the LOD, LOD/√2 was used (Verbovšek, 2011) as its concentration in the fi sh muscle for the risk estimations.
Thirdly, the target hazard quotient (THQ) was used to assess the non-carcinogenic human health risks associated with fi sh consumption as described by USEPA (2000) using equation (2). THQ is the ratio between the estimated dose of a contaminant and the respective RfD value. RfD is an estimate of a daily exposure to which the human population that may be exposed without an appreciable risk of deleterious eff ects during a lifetime. If the THQ ˂ 1, the exposed population is unlikely to experience obvious non-carcinogenic adverse eff ects. If the THQ ≥ 1, there is a potential health risk (USEPA, 2000).

December 2020
Journal of the National Science Foundation of Sri Lanka 48(4) THQ = [(EF*ED*IRd*MC) (RfD*BW*AT)] ... (2) where, THQ is the target hazard quotient; EF is the exposure frequency (365 days/year); ED is the exposure duration (70 years); RfD is the oral reference dose (mg/kg body weight /day); AT is the averaging time for non-carcinogens (365 days/year × number of exposure years, assuming 70 years in this study). RfDs reported by USEPA (2000) for cadmium (0.001 mg/kg body weight /day) and inorganic arsenic (0.0003 mg/kg body weight /day) were used in the calculation. It was assumed that inorganic arsenic in the fi sh was 10 % of the total arsenic content (Zhong et al., 2018).
For non-carcinogenic eff ects of metals, the maximum allowable daily fi sh consumption limits for adults (CR lim in kg/day) were also estimated to determine the amount of fi sh that can be safely consumed daily using the following equation (USEPA, 2000): Maximum allowable daily consumption limits were converted to the number of allowable fi sh meals of a specifi ed meal size per week using the following equation (USEPA, 2000).
where, CR mw is the maximum allowable fi sh consumption rate per week (meals/week), T ap is the time averaging period (7 days per week) and MS is the meal size (kg fi sh/meal). In this estimation, USEPA default value of 0.227 kg fi sh/meal for adults was used as the meal size instead of 0.060 kg fi sh/meal, considering the high-end exposure groups due to heavy fi sh meal consumption.

Statistical analysis
All metal(loid) levels in fi sh are reported as wet tissue weight bases. For statistical interpretations, concentrations of the particular metal (loid), which were less than the LOD were estimated as LOD/√2, as recommended by Verbovsek (2011). The same criterion was applied for risk assessment estimations when a particular metal(loid) level in all samples were below the LOD. Overall metal(loid) levels are presented as mean ± SEM, median and the range (minimum and maximum) using descriptive statistics. For each fi sh, the measured metal(loid) levels in diff erent tissues in diff erent sampling events were compared separately by nonparametric Mann Whitney U test. The same statistical test was used for comparison of respective metal levels in the two fi sh species. Pearson's product moment correlation coeffi cient test was used to assess potential associations of metal(loid) levels in fi sh tissues with body weight and total length of the fi shes. p < 0.05 was considered as statistically signifi cant. Statistical analysis was carried out using Minitab 17 statistical software package.

RESULTS AND DISCUSSION
Cadmium and arsenic levels in fi sh tissues Cadmium levels in muscle, liver and kidney tissues of the two fi sh species are presented in Table 1. Cadmium contents in the muscle of majority of Nile tilapia were below the limit of detection (< 0.05 mg/kg wet weight) whereas only fi ve fi sh sampled in the third sampling event contained detectable amounts of cadmium in the muscle tissues (range 0.05 -0.10 mg/kg). Comparatively high levels of cadmium were found in the kidneys of Nile tilapia in the third sampling event (range 0.2 -0.9 mg/kg).
Of the 30 Nile tilapia analysed in this study, livers of 10 fi sh and kidneys of 21 fi sh contained detectable levels of cadmium. In all Butter catfi sh (n = 30), cadmium content in the muscle was below the detection limit. However, the livers of 10 fi sh and kidneys of 20 fi sh contained detectable levels of cadmium. Comparatively high levels of cadmium were found in the kidney tissues of Butter catfi sh in the fi rst and second sampling events (Table 1). In both fi sh species, the highest cadmium level was found in kidney tissues (Nile tilapia: 0.9 mg/kg; Butter catfi sh 0.40 mg/kg) in comparison to the muscle and liver tissues.
In both fi sh species, overall cadmium levels in kidney tissues were signifi cantly higher than that of the other tissues (p < 0.05). However, no signifi cant diff erences (p > 0.05) were found between the two fi sh species with respect to overall cadmium levels.
Arsenic levels in the muscle tissues of all Nile tilapia and Butter catfi sh (Table 2) were below the limit of detection (< 0.05 mg/kg wet weight). Of the total number of fi sh analysed (n = 30 for each species), arsenic was detected only in liver tissues of fi ve Nile tilapia and one butter catfi sh, and kidney tissues in three Nile tilapia ( Table 2). The highest arsenic level was found in kidney tissues (0.46 mg/kg) of Nile tilapia. In Butter catfi sh, the detected highest arsenic level was in liver tissue (0.06 mg/kg). For arsenic, no signifi cant diff erences (p > 0.05) in tissue specifi c or species specifi c bioaccumulative patterns were found.
Bioaccumulation of heavy metals in fi shes may be related to various biological factors such as species, Journal of the National Science Foundation of Sri Lanka 48 (4) December 2020   Limit of detection of arsenic is 0.05 mg/kg wet weight. Sampling event-specifi c results are presented only as minimum and maximum ranges: the numbers in parantheses indicate the ratio of fi sh with analysed metalloid above detection limits to the total number of fi sh analysed for each sampling event (n = 10 fi sh). No signifi cant diff erences in the arsenic levels were found among diff erent tissues or the two species (p > 0.05). Both fi sh species used in the study were within the edible size range (Table 3). Of the cadmium and arsenic levels in the three tissues of Nile tilapia and Butter catfi sh tested, the cadmium contents in the liver of Nile tilapia showed statistically signifi cant positive correlations with the total length of the fi sh (p = 0.004). No other statistically signifi cant relationships were found between fi sh body size and cadmium or arsenic levels in the analysed fi shes (Table 3). The results may indicate that cadmium accumulation capacity in the liver of Nile tilapia may increase as the fi sh gets longer with age. As cadmium can induce increased metallothionein production in Nile tilapia liver (Chandrasekera et al., 2008) bioaccumulation of cadmium in liver tissues may be increased with prolong exposure. Indrajith et al. (2008) reported strong positive correlations between body size (body weight and body length) and cadmium levels in the muscles of two estuarine food fi shes, Etroplus suratensis and Ambassis comersoni from the Negombo estuary, Sri Lanka indicating high health risks associated with the consumption of bigger size fi shes.
Limited scientifi c information is available on the levels of toxic metals in reservoir fi shes in Sri Lanka. Most of the available studies have focused on metal levels in tilapia species. Allinson et al. (2009)   In general, concentrations of cadmium in the three tissues of both fi sh species followed the order: muscle ≈ liver < kidney showing greater affi nity of the kidney for cadmium bioaccumulation. In both fi sh species, the mean concentration of cadmium in the kidney was nearly fi ve times greater than that in the muscle. Compared to other tissues, high levels of cadmium were also found in kidneys of some freshwater fi sh species caught from the Lot River, France (Andres et al., 2000). The results suggest that kidney is the main site of cadmium accumulation in fi sh. Arsenic levels in both species of fi sh were below the detection limits in most cases and did not show signifi cant tissue-specifi c diff erences (p >0.05).
A consistent tissue-specifi c arsenic accumulation pattern was also not evident with respect to arsenic levels in muscle, liver, intestine and gills of ten freshwater fi sh species caught from the Manchar Lake, Pakistan (Shah et al., 2009).
Journal of the National Science Foundation of Sri Lanka 48(4) December 2020 that cadmium levels in the muscle of tilapia captured in 2002 from three reservoirs of Sri Lanka (Kaudulla, Rajanganaya, and Udawalawe reservoirs) ranged from < 0.15 to 0.2 mg/kg. Bandara et al. (2008) reported that mean levels of cadmium in the edible muscle of Nile tilapia captured from two reservoirs in the North Central Province, Sri Lanka ranged from 0.057 to 0.202 mg/kg. The highest cadmium levels detected in Nile tilapia from these two reservoirs were 0.115 mg/kg (Karapikkada reservoir) and 0.425 mg/kg (Thuruwila reservoir). Previously reported cadmium levels in tilapia from other reservoirs were higher than the levels found in muscles of Nile tilapia captured in 2018 from the Padaviya reservoir. Mean cadmium level in muscles of tilapia (Oreochromis sp.) collected in 2010 from selected reservoirs from the North Central, Southern and Eastern Provinces in Sri Lanka was 0.03 mg/kg on a dry weight basis (Jinadasa & Edirisinghe, 2013), which would be much lower when converted to wet weight basis. Nevertheless, the reported mean level of cadmium in muscle of tilapia (Jinadasa & Edirisinghe, 2013) is somewhat comparable to the levels of cadmium found in the two fi sh species analysed in the present study.
In this study, arsenic content in the edible muscle tissues of the two fi sh species was less than 0.05 mg/kg (estimated mean 0.035 mg/kg). Reliable scientifi c reports on arsenic levels in the reservoir fi sh species in Sri Lanka are not available for comparison with the data in the present study. When compared with the recently reported arsenic levels in muscle tissues (wet weight basis) of freshwater fi shes in other regions, the highest arsenic concentrations detected in Nile tilapia (0.46 mg/kg in kidney) and Butter catfi sh (0.06 mg/kg in liver) in this study were much lower than the mean arsenic levels reported in the muscle of Cyprinus carpio from Kabul River(0.86 mg/kg) and Chenab River (2.98 mg/kg) and Labeo rohita from upper Indus and lower Indus River (0.67 and 2.5 mg/kg respectively) in Pakistan (Nawab et al., 2018) and of Cyprinus carpio (0.0759 mg/kg) and Euphrates barbell (0.132 mg/kg) from Keban Dam Reservoir in Turkey (Varol & Sünbül, 2018).

Human health risk assessment
Consumption of fi sh contaminated with toxic heavy metals is a threat to human health (Castro-González & Méndez-Armenta, 2008). Heavy metals, especially cadmium and arsenic are suspected to be potential risk factors for CKDu, which is highly prevalent in agricultural settlements in the Dry Zone districts of Sri Lanka and has posed a severe public health crisis among rural communities in recent times (Jayatilake et al., 2013). The most likely avenue of metal exposure for these rural communities would be through the dietary intake of food and water (Bandara et al., 2008;Jayatilake et al., 2013). Reservoir fi shes including Nile tilapia and Butter catfi sh are the main sources of animal protein of the rural people living in the Dry Zone districts of Sri Lanka. In this study, potential human health risks associated with dietary intake of cadmium and arsenic levels through Nile tilapia and Butter catfi sh from the Padaviya reservoir were assessed using three approaches: comparison of metal levels with maximum permissible levels (MPL); comparison of estimated daily intake with reference doses; and estimation of target hazardous quotients for non-carcinogenic eff ects as described by Varol and Sünbül (2018). Maximum permissible levels for heavy metal(loids) in fi sh for consumption of the general public have not yet been established by the food regulatory authorities in Sri Lanka. However, there is a gazette notifi cation with respect to maximum levels of cadmium, lead and mercury in fi sh for export, which states that the maximum level of cadmium in export fi sh should be 0.05 mg/kg wet weight (GOSL, 2007) which is similar to the MPL of cadmium (0.05 mg/kg wet weight) for muscle meat of fi shes specifi ed by the European Commission for human consumption (EC, 2006). Previous studies indicate that cadmium levels in the edible tissues of tilapia captured from some other reservoirs of Sri Lanka (Allinson et al., 2009;Jinadasa & Edirisinghe, 2013) were within the acceptable international food regulatory limits for human consumption. In the edible muscles of Nile tilapia, the detected highest concentration of cadmium was 0.1 mg/kg whereas in all Butter catfi sh, cadmium concentrations in the muscles were below 0.05 mg/kg (Table 1). When compared with the specifi ed MPL for cadmium (Table 4), cadmium content in the edible muscle has exceeded in 3% of Nile tilapia, while cadmium contents in the muscle of all Butter catfi sh were below the respective MPL. Yet, cadmium contents in the kidney (53 % of Nile tilapia and 67 % of Butter catfi sh) and liver (23 % Nile tilapia and 20 % Butter catfi sh) tissues of a considerable portion of the sampled fi sh have exceeded the MPL for cadmium (Table 4). The MPL for total arsenic in fi sh has not been provided by most food standards of international regulatory authorities. The MPL for inorganic arsenic in fi sh for human consumption has been reported as 2 mg/kg wet weight by food regulatory authorities in Australia and New Zealand (FSANZ, 2013) while the Ministry of Health of the People's Republic of China has lowered the MPL value for inorganic arsenic in fi sh to 0.1 mg/kg wet weight (MHPRC, 2013). In the present study, even the total arsenic levels in three tissues of Nile tilapia December 2020 Journal of the National Science Foundation of Sri Lanka 48 (4) and Butter catfi sh were below these inorganic arsenic MPL for human consumption (Table 4). When using the assumption that a 0.06 kg portion is consumed each day (Allinson et al., 2009), estimated daily intake values of cadmium and arsenic (Table 4) from Nile tilapia and Butter catfi sh muscle meat are several folds lower than the corresponding reference doses reported by the USEPA (2000) for edible portions of fi sh, indicating low health risks associated with the fi sh consumption. Target hazardous quotient values estimated for the respective metal(loid)s at this daily fi sh consumption rate (Table 4) are distinctively much lower than the hazard quotient threshold value of one (01) indicating low noncarcinogenic health risks associated with the cadmium and arsenic levels in muscle meat of these fi shes.
Based on estimated consumption limits (Table 4), large quantities of Nile tilapia muscle tissue would need to be consumed daily on a regular basis to exceed the intake limits of cadmium (> 0.7 kg) and inorganic arsenic (> 6 kg) to pose non-carcinogenic health risks. Similarly, daily consumption of muscle tissue of Butter catfi sh should be greater than 2 kg/day and 6 kg/day, respectively to exceed the intake limits of cadmium and inorganic arsenic. Estimated maximum allowable fi sh consumption rates per week (meals/week) also indicate extremely heavy consumption conditions. These fi sh intake levels are considered highly unlikely for general population who consume fi sh at moderate levels.
In the human health risk assessment, it was assumed that cooking has no eff ect on cadmium and inorganic arsenic content in the muscle tissue portions and ingested dose is equal to the absorbed heavy metal dose. There is a possibility to ingest more cadmium if the kidney and liver tissues of the fi shes are consumed as they showed greater cadmium accumulative capacities. The kidney of fi shes may not be removed before cooking as it is fi rmly attached to the ventral part of the vertebral column of the fi sh. It is advisable to remove all liver and kidney tissues when preparing these fi sh for food. Cadmium and inorganic arsenic are classifi ed as human carcinogens (Tchounwou et al., 2012;Wise et al., 2017). Hence, carcinogenic health risks associated with cadmium and arsenic exposure through these fi shes cannot be excluded from the study. There can be synergistic or additive eff ects caused by exposure to low levels of multiple carcinogenic heavy metals from these fi shes on human. Periodic monitoring of toxic metal and metalloid levels in reservoir fi shes in areas where CKDu is highly   (2000) Journal of the National Science Foundation of Sri Lanka 48(4) December 2020 prevalent in Sri Lanka is recommended considering the need to ensure safety of fi sh food for the vulnerable rural communities.

CONCLUSION
This study generated new information on cadmium and arsenic levels in two of the mostly consumed reservoir fi sh species in the Padaviya reservoir, Sri Lanka and their suitability for human consumption. In both fi sh species, the highest cadmium levels were found in the kidney tissue showing greater cadmium bioaccumulative capacity in the kidney tissues. However, for arsenic, no signifi cant tissue specifi c bioaccumulative pattern was found in these fi shes. In the edible muscle, the highest cadmium level detected was 0.1 mg/kg and arsenic levels were < 0.05 mg/kg on wet weight basis. Human health risk assessment revealed that cadmium contents in edible muscle of most fi shes and arsenic contents in all fi shes were within the maximum permissible limits set by international food regulatory authorities. According to the estimated daily intake and target hazard quotients, it is highly unlikely that cadmium and arsenic contents of the muscle meat of these fi shes from Padaviya Reservoir would pose non-carcinogenic human health risks to moderate level consumers. However, carcinogenic health risks associated with heavy metal(ioid) exposure via fi sh food cannot be excluded.