Vea is a community under the Bongo District of the Upper East Region of Ghana. The Vea reservoir is located between latitude 10˚45'N and longitude 1˚W. The vegetation in the study area is dry guinea savannah, distinguished by short grasses and short trees. The climate has a mean minimum and maximum temperature of 14˚C and 40˚C, respectively. The mean annual rainfall ranges from 850 to 1000 mm which occurs in the months from May to October, followed by a prolonged dry season [15]. The first part of the dry season from November to mid-February is characteristically cold and dry with dusty Harmattan winds. The rest of the dry season is usually characterized by a wide temperature range from 14˚C at night and to over 35˚C during the day. Humidity is also very low, making the daytime temperature high and less comfortable. The Vea irrigation project is constructed on the tributary of the White Volta River from Burkina Faso and serves eight communities with a total farmer population of 6000 as shown in Figure 1. The surface area of Vea reservoir is 405 ha with a maximum storage of 1.7 × 10⁷ m³ serving 21 km of main canal distance [15].

Three different fish species; tilapia (Oreochromis niloticus), sardi (Alestes baremose) and African catfish (Clarias gariepinus) were considered for this work, because they are the most frequently eaten in the study area. A total of seventy-five (75) fish samples, twenty-five (25) each were purchased from fishermen at the landing site of the reservoir with another twenty-five (25) sediment samples which were evenly taken from different parts of the reservoir.

The fish samples were first wrapped in aluminum foil and packed in labelled clean polyethylene bags as with the sediment samples and transported in thermos-insulated containers with ice packs to the laboratories of KNUST where they were stored at temperature below −10˚C until processed.

The sediment samples were air dried for one week. Stalks, stones and other debris were removed and each sample homogenized to maintain homogeneity. These samples were then ground and passed through a 2 mm sieve to remove the coarse sediment fraction and kept in plastic bags. The fish samples were identified at the Department of Fishery of KNUST. They were thawed, cleaned with distilled water and their scales sloughed. Their muscle tissues dissected, minced into pieces and homogenized. The extraction of pesticide residues in fish and in sediments was done according to the method developed by [16] with some modifications.

Pesticide grade acetonitrile, analytical grade sodium chloride, magnesium sulphate anhydrous, ethyl acetate, pesticide standards were obtained from BDH Laboratory Supplies, England. Reference Standards were certified by Dr. Ehrenstorfer Laboratories (GmbH Germany) and Silica and C18 were obtained from Phenomenex, USA.

Ten (10.0) g of homogenous fish sample was weighed into a 250 mL nalgene jar and 50 mL acetonitrile was added and macerate for 2 min. Centrifugation was done for 3 min at 3000 turns/min. The extract was then filtered through a filter paper into a 100 mL volumetric flask. Another (20 mL) acetonitrile was added to the residue, macerated for 1min and the dispersing element rinsed with acetonitrile into the jar and centrifuged for 3 min. at 3000 turns/min. The residue was filtered again into the 100 mL volumetric flask, and the jar rinsed for residue with acetonitrile and adjusted the filtrate to the 100 mL mark. Fifty (50) mL residue was transferred into a round bottom flask, and concentrated on the rotary evaporator to 2 mL for adsorption chromatography.

A Bond Elute C-18 was conditioned with a 1000 mg/6mL cartridge with (6 mL) acetonitrile and the sample extract were poured into the column and the column eluted with 10 mL acetonitrile. A Silica SPE cartridge was conditioned with 1000 mg/6mL which has 2 g of MgSO₄ on top with (6 mL) acetonitrile. The extract was poured unto the column and eluted with 10 mL acetonitrile. The extract was concentrated below 40˚C on the rotary evaporator to dryness. The filtrate was re-dissolved in 1 mL ethyl acetate and transferred into a 2 mL, vial for quantitation by GC-ECD.

Ten (10.0) g of homogenous sediment sample was put in a 100 mL separating flask and 10 mL acetonitrile added, corked and sonicated for 5 min. After that, a further 10 mL acetonitrile was added, corked and the flask place on the horizontal mechanical shaker and set to shake continuously for 30 min. The mixture was allowed to stand for 10 minutes to separate into layers. Ten (10) mL of the organic phase was pipetted into a round-bottomed flask and evaporated to 2 mL for extract clean-up.

A cartridge was conditioned with silica and acetonitrile (1000 mg/6mL) to have a 1 cm thickness layer of anhydrous magnesium sulphate on top with (10 mL) of acetonitrile. Two (2) mL of the extract was purified by loading onto the cartridge and the eluate collected into a flask. The cartridge was further eluted with 10 mL of acetonitrile and the filtrate concentrated below 40˚C on the rotary evaporator. The filtrate was re-dissolved in 1 mL ethyl acetate and the extract transferred into a 2 mL, standard vial prior to quantitation by GC-ECD.

The quality of analysis of pesticide residues was enhanced through solvent blanks and spikes. The solvent blanks were used to check for interferences from reagents, whilst the spike samples were used to determine recovery which is an indicator of method performance and accuracy. The triplicate and duplicate measurements of samples were used to confirm method precision. The recoveries ranged between 74% - 120% for organochlorine pesticide residues, 70% - 94% for organophosphorus pesticide residues and 73% - 100% for synthetic pyrethroid pesticide residues. Fortification level of 0.05 mg/kg was chosen for OCs in sediment samples and 0.01 mg/kg for fish samples based on the limit of determination of the pesticides being analysed. A gas chromatogram, Varian CP-3800 equipped with 63Ni electron capture detector, CTC Analytic Combi PAL autosampler, split-splitless injector, programme pneumatic control and a computer star workstation data processor were used for all classes of pesticides.

The estimated daily exposure (Ems) of individuals to pesticide residues from each fish species was determined using Equation (1);

Em = Cm × CR ∘ BW   (1)

where Cm, CR˚ and BW represent concentration of chemical contaminant in the muscle portion of fish (mg/kg), mean daily consumption rate of fish (kg/d) [17] and body weight of an individual consumer (kg) [18]. The estimated daily exposures are calculated for children and adults eating contaminated fish species of tilapia, catfish and sardi as shown in Table 1 for OPs, Table 2 for OCs and Table 3 for SPs.

Human health assessment of consumers of pesticide residue contaminated fish was characterized using health risk index (HI). The estimated HIs were obtained by dividing the estimated exposure by its corresponding reference dose by as shown in the Equation (2);

( HI ) = CDI RfD ≡ Em RfD   (2)

When HI is less than 1.0 it can be concluded with great certainty that there is essentially no probability of population or community level effect. However, if the ratio exceeds 1.0 then there is a potential for adverse effect of either carcinogenic or non-carcinogenic risk. Carcinogenic risk ( CR ∘ ) were estimated using the Equation (3);

CR ∘ = CDI × SF ≡ Em × SF (3)

where Em and SF represents the estimated exposure and slope factor.

For non-carcinogenic risk, the Equation (4);

RiskQuotient ( RQ ) = MEC PNEC   (4)

where MEC and PNEC represents Measured Environmental Concentration and Predicted No Effect Concentration. The results of RQ calculated values from PEC (MEC) and (PNEC) for the various pesticides residues are shown in Table 4 to estimate risk sediments pose to fish species.

The human health risk assessment of consumers of contaminated fish by pesticides is calculated for children and adults using the Hazard Indices. Table 5 and Table 6 show Hazard Indices of OPs and SPs respectively. There is no certainty of potential adverse effect of cancer or non-cancer effects. Table 7 however show Hazard Index values of OCs with potential risk of cancer or non-cancer effects from aldrin and p,p'-DDE for both children and adults from three fish species.

The benchmark concentration for carcinogenic effect was derived using USEPA cancer slope factor and the exposure concentration and also for non-carcinogenic effect. Risk assessments were conducted based on the concentrations of OC, OPs and SPs pesticides residues in fish tissues and in sediment. Hazard Indices (HIs) were calculated by dividing the average daily exposure by the reference dose concentrations for fish samples and Risk Quotients were by dividing measured environmental concentrations by predicted no effect concentrations [19]. A hazard ratio which is greater than unity indicates that the average exposure level exceeded the benchmark concentration and not acceptable.

From [20], Risk quotient method (RQ) was used to determine the risk sediments pose to aquatic organisms e.g. fish species. RQ is the ratio of the Predicted Environmental Concentration (PEC) to the Predicted No Effect Concentration (PNEC) [21]. The PNEC was estimated by dividing the LC₅₀ with an assessment factor (AF) of 100 [22]. AF is the total uncertainty factor which is from the product of the assumption that the least sensitive humans are 10 times more sensitive than the most sensitive animal species and the additional uncertainty factor of 10 is used to address differences in sensitivity among humans (this is from the assumption that the most sensitive human is 10 times more sensitive than the least sensitive human). This results in a total uncertainty factor of 100 as AF [23]. The results of RQ estimations from PEC (MEC) and (PNEC) [24] the various pesticides residues are calculated.

From [8] and [25] reported that RQ ≥ 1 indicates a high risk, while 0.1 ≤ RQ ≤ 1 indicates medium risk, and 0.01 ≤ RQ ≤ 0.1 indicates low risk. Concentrations of pesticide residues in sediment were generally higher than in fish samples. The highest concentration of β-HCH (1.207 mg·kg⁻¹) was in sediment and second in p,p'-DDE (1.026 mg·kg⁻¹). Similar results was obtained for p,p'-DDE residue which was predominant in the fate and assessment of persistent organic pollutants in water and sediment [26]. The RQ values of chlorpyrifos (0.0226) and malathion (0.0130) residues were both within the low risk but dimethoate (0.116) and chlorpyrifos (0.2956) were high within the medium risk category. Organochlorines in sediment recorded high concentrations yielding RQ of β-HCH (1.207) which is above the risk and is likely to cause adverse effects to fish species. p,p'-DDT (0.167) indicates medium risk and lindane (0.0318), dieldrin (0.0138) and endosulfan sulfate (0.0425) all indicate low risk levels. RQ values recorded for allethrin (0.0295), lambda-cyhalothrin (0.0161) and deltamethrin (0.0133) were all within the low potential risk level.

The [27] has defined acceptable risks for carcinogens as within the range 10⁻⁴ to 10^-6 excess lifetime cancer risk. Carcinogenic health risk was calculated for OCs pesticide residues because of their possible cause of cancer to humans. The cancer benchmark concentrations were derived using the oral slope factors (OSFs) of OCs pesticide residues [28]. From [29], individuals have up to a one in between 10,000 to 1,000,000 chance of not developing cancer in their lifetimes; anything short predisposes humans to cancer risk effects.

Table 8 shows that β-HCH, lindane and dieldrin did not record any cancer benchmark (CB). p,p'-DDT, p,p'-DDD, γ-chlordane and heptachlor recorded CB concentrations that were within acceptable cancer risk levels. Aldrin recorded cancer benchmark concentrations that were unacceptable for both children and adults. It was (8.09E−03; 3.47E−03) in Catfish, (3.49E−03; 1.49E−03) in tilapia and (2.86E−03; 1.22E−03) in sardi fish and p,p'-DDE (8.02E−04 ≡ 0.0010) in

Cancer Risk slope factors (OSF) in 1/(mg·kg⁻¹d⁻¹) were obtained from the USEPA’s Integrated Risk Information System (IRIS). ND = Not Available. Values in bold are cancer risk.

children. This indicate a probability of children and adults experiencing carcinogenic effects of aldrin and p,p'-DDE from ingestion of catfish or tilapia or sardi fish in their lifetimes. The cumulative risk posed to humans through fish samples is the sum of total risks from each individual exposure pathway in the type of fish ingested. The cumulative cancer risk of the individual pathways indicates a cancer risk for all of fish species [30].

The Risk Quotient (RQ) values presented in Table 4 are to determine the risk posed to aquatic organisms like fish. RQ ≥ 1 indicates a high risk, 0.1 ≤ RQ ≤ 1 indicates a medium risk and 0.01 ≤ RQ ≤ 0.1 indicates a low risk. β-HCH with 1.207 mg·kg⁻¹ indicates a high risk for organisms in the reservoir.

The Hazard Indices in Tables 5-7 showed that aldrin and p,p'-DDE recorded HI > 1. This establishes that there is health risk associated with lifetime consumption of catfish, tilapia and sardi from Vea reservoir. All other pesticides in catfish, tilapia and sardi showed no health hazard associated with their consumption as their Hazard Indices for all the detected residues were < 1 in spite of their presence in fish. The Oral Slope Factor (OSF) determines the chance of or not developing cancer in the lifetime of a person exposed to contaminated fish. Table 8 showed OSF of OCs where aldrin in catfish, tilapia and in sardi in children and adults and p,p'-DDE in catfish of children with over range benchmark concentrations of developing cancer in their lifetime.