Human health risk assessment of lead in drinking water

280

Int. J. Environment and Pollution, Vol. 31, Nos. 3/4, 2007

Human health risk assessment of lead in drinking water: a case study from Port-au-Prince, Haiti Evens Emmanuel*, Ruth Angerville and Osnick Joseph Laboratoire de Qualité de l’Eau et de l’Environnement, Université QuisqueyaBP 796, Port-au-Prince, Haiti Fax: 509 221 4211 E-mail: [email protected] E-mail: [email protected] *Corresponding author

Yves Perrodin Laboratoire des Sciences de l’Environnement, Ecole Nationale des Travaux Publics de l’Etat, Rue Maurice Audin, 69518 Vaulx-en-Velin, France Abstract: In Latin America and the Caribbean (LAC), human intoxication to lead is considered as an important public health issue. In Port-au-Prince, concentrations of lead ranging from 40 µg/L to 90 µg/L, greater than the threshold value (10 µg/L) for drinking water, were measured in groundwater and drinking water. This study aims to assess human health risks generated by exposure to lead in the Port-au-Prince water supply. Two sampling campaigns were performed between April 2004 and December 2004 on different structures of the public water supply. A significant lead concentration of 250 µg/L, greater than the threshold value, had been detected in a water tank. Risk of deterioration of the psychological development of children exposed to these waters was calculated. These results require monitoring in order to control the human health risk by lead in Port-au-Prince’s drinking water. Keywords: lead; human health; risk; drinking water. Reference to this paper should be made as follows: Emmanuel, E., Angerville, R., Joseph, O. and Perrodin, Y. (2007) ‘Human health risk assessment of lead in drinking water: a case study from Port-au-Prince, Haiti’ Int. J. Environment and Pollution, Vol. 31, Nos. 3/4, pp.280–291. Biographical notes: Evens Emmanuel is a Professor and Dean of the Faculty of Sciences and Engineering, and the Director of the Laboratory of Water and Environment Quality (LAQUE) of Université Quisqueya in Haiti. He is also the responsible for the Graduate School of Ecotoxicology, Environment and Management of Water. He holds a Master of Sciences in Sanitary Engineering of Universidad San Carlos de Guatemala and a PhD of National Institute of Applied Sciences of Lyon. He is Professor of Water and Wastewater Treatment. His research field is methodological approach of human health risk assessment and ecological risk assessment of water and wastewater.

Copyright © 2007 Inderscience Enterprises Ltd.

Human health risk assessment of lead in drinking water

281

Ruth Angerville is an Environmental Engineer, and she received Master of Sciences in Ecotoxicology, Environment and Water Management of Quisqueya University (UniQ) in Haiti. She is doing a PhD in the field of ecological risks assessment in the Laboratory of Environmental Sciences (LSE) of the National School of State Public Works of State in France, and in the Laboratory of Water and Environment Quality (LAQUE) of Université Quisqueya in Haiti, under the direction of Professors Yves Perrodin and Evens Emmanuel. Osnick Joseph is an Environmental Engineer, and he received Master of Sciences in Ecotoxicology, Environment and Water Management of Université Quisqueya in Haiti. He is a PhD student of National Institute of Applied Sciences of Lyon. His working on the use of agricultural by-products, such as sugar cane bagasse, in water and waste water treatment, under the direction of Professors Rémy Gourdon and Evens Emmanuel. Yves Perrodin is a Professor and Director of the Laboratory of Environmental Sciences (LSE) of the National School of State Public Works (ENTPE) in France. Holder of a Master of Sciences in Microbiology and a PhD of National Institute of Applied Sciences of Lyon, his general field of research relates to health and ecological risks assessment of scenarios in urban environment. He assumes responsibilities in several scientific networks intervening on the topic of ‘Water and Health’. He teaches the methods of human health and ecological risks assessment in several French Grande École as well as in universities in developing countries (Haiti, Cameroon).

1

Introduction

The presence of lead in environmental compartments constitutes an ecological and human health concern, since heavy metals are not degraded biologically like organic pollutants (Angerville et al., 2004). Lead particularly targets the nervous system, blood and kidney (INERIS, 2002). Concentrations higher than 10 µg/L in drinking water constitute important public health problems owing to metal absorption and bioaccumulation (Chiron et al., 2003). Indeed, drinking water is one of the major sources of human exposure to lead (INERIS, 2002; Fertmann et al., 2004). A guideline value of 10 µg/L was adopted for drinking water in order to minimise human exposure to lead (OMS, 1996). This new threshold value was also adopted by the European Union in (1998), allowing for transitional regulations until 2013 (European Commission, 1998). Human beings are continuously exposed to lead from natural as well as anthropogenic sources (Christensen, 1995). Lead and its inorganic by-products in drinking water (oral exposure) are not carcinogenic (ATSDR, 1999). However, long-term lead exposure may generate irreversible functional and morphological renal changes (Christensen, 1995), distal motor neuropathy and possibly seizures and coma (Robson, 2003). Infants and small children are more sensitive to the effects of lead, which moreover is transported through the placenta to the foetus (Yule and Rutter, 1985; Milman et al., 1988; Shennan and Boyd, 1988; Bellinger et al., 1990; Dietrich et al., 1990; Goyer, 1990; Cleymaet et al., 1991; Christensen, 1995). Lead accumulation in foetuses and small children might cause developmental disruption in terms of neurological impairment characterised by a decrease of cognitive faculties, which can be reversible or not, evaluated by psychomotor tests such as the verbal IQ (intellectual quotient) test

282

E. Emmanuel, R. Angerville, O. Joseph and Y. Perrodin

(Académie des Sciences, 1998). The period when IQ is most affected is from birth to about four years of age (Watt et al., 2000). In the Americas, lead has been a well-known contaminant for populations exposed to road traffic. During the Miami Summit of 1994, the Presidents of American and Caribbean countries chose to switch from leaded to unleaded gasoline. Information reported on lead characterisation in urban areas of Latin America and the Caribbean (LAC) did not mention the magnitude of the lead intoxication problem in Haiti (OPS/OMS, 1996). However, the fact that rainwater, polluted by atmospheric lead particles originating from industrial activities and motor vehicles (Cabrera et al., 1995), feeds the karstic aquifers of Port-au-Prince (Butterlin, 1960; Simonot, 1982; Desreumaux, 1987), therefore leads to the assumption that the groundwater is subject to lead pollution (Denić-Jukić and Jukić, 2003). In addition, significant lead concentrations (up to 1670 µg/L) were measured in wastewater discharged by paint factories in Port-au-Prince (Carré, 1997). Lead concentrations ranging from 40 µg/L to 90 µg/L were measured in the drinking water of Port-au-Prince (Emmanuel et al., 2004). This study aims to evaluate the human health risk caused by chronic exposure to lead from Port-au-Prince’s public water supply.

2

Theoretical aspects of human health risk assessment

Human health risk assessment is the activity or process that evaluates the effects of a chemical’s toxic properties on human beings exposed to it (NRC, 1983). This process is structured in four stages: hazard identification, dose-response assessment, exposure assessment and risk characterisation. In the first stage, the known or potential health effects associated with a particular agent are identified. Hazard identification requires the establishment of a causal relation between the appearance of one or several adverse effects on a living organism after its exposure to a chemical substance, according to the scenario (route, intensity, duration) considered within the framework of the evaluation (Bonvallot and Dor, 2002). The dose-response relationship, the specific exposure route (oral, dermal or inhalation) and the link between the amount of substance in contact with the organism results in the appearance of the toxic effect (INVS, 2000). The European Commission Guide on Risk Assessment for New and Existing Substances states that the objective of the third stage, ‘exposure assessment’, is to predict the concentration profile or dose of a substance to which the receptor will be exposed (European Commission, 1995). For human health exposure assessment this involves evaluating occupational, consumer and environmental exposure. Finally, risk characterisation is a synthesis of the dose-response assessment and the exposure assessment. It consists in linking the data from the exposure assessment with the dose-response assessment. If no carcinogenic substances, such as lead in drinking water (ATSDR, 1999), are present, the risk is expressed as a quotient of danger (QD) resulting from the ratio ‘MDI/ADI’. This risk ratio is purely qualitative. When the ‘D’ quotient value is greater than 1, the risk is considered significant, becoming even more so as the quotient increases. Conversely, the lower the quotient is below 1, the weaker the risk.


3

283

Materials and methods

3.1 Presentation of study area In this human health risk assessment of lead in drinking water, an area containing four boreholes, a water tank of 4600 m3 and five domestic water taps were selected in the north of Port-au-Prince as an experimental site (Figure 1). Domestic water taps C1 and C2 were directly connected to the pumping pipe of the four boreholes (upstream of the water tank), while domestic water taps D1, D2 and D3 were connected to the water supply network (downstream of the water tank). Figure 1

Different sampling points from borehole and domestic tap water facilities

The boreholes and water tank supply water to a population of over 90,000. Ninety percent of this population were adults and children over five years and 10% were children under five years (IHSI, 2003). Nine thousand people of this estimated population could be considered as children under five years and were thus the most sensitive subgroup to lead intoxication (Banks et al., 1997; Bellinger et al., 1987; Needleman and Gatsonis, 1990; Winneke et al., 1990; Goyer, 1996; Académie des Sciences, 1998). The only source of lead exposure considered in this study was the drinking water provided by the public water supply of Port-au-Prince.

3.2 Sampling Two sampling events were conducted. The first event occurred from 16 to 22 April 2004 during the rainy season, while the second was conducted from 7 to 9 December 2004 during the dry season. Water samples from the water tank were collected by a telescopic perch. Conductivity, total dissolved solids (TDS), temperature and pH were measured directly on site after sampling. All water samples for lead characterisation were kept in a 1-L glass flask at 4°C until analysis. Storage conditions of samples were in accordance

284


with Eaton et al. (1995). Only lead was analysed in the samples collected during December 2004.

3.3 Physicochemical analysis pH: The pH of collected water samples was measured using a WTW pH 340 ION pH meter fitted with reference and pH electrodes. Electric Conductivity (EC) and Total Dissolved Solids (TDS): EC and TDS were measured on the sampling sites using a WTW–LF 330 multipurpose potentiometer coupled with specific electrodes. Chlorides (Cl): This parameter was determined by the Mohr method, consisting in proportioning chlorides with silver nitrate and potassium chromate. In the presence of silver nitrate (AgNO3), Cl ions are mobilised to form cerargyrite. When all the chloride ions precipitate as AgCl, silver nitrate reacts with potassium chromate to form a brick red precipitate. Knowing the concentration of AgNO3 (Co = 10–2 M) in 100 ml of solution (E = 100 ml), the volume necessary to obtain equivalence (Ve), the concentration of Cl ions in the solution is given by the formula: (Cl –) = Co × Ve/E. Water samples from the first collection batch were filtered at 0.45 µm and treated with nitric acid (HNO3) at pH < 2.0. Lead measurements in pre-treated water samples were carried out by flame atomic absorption spectrophotometry (F-AAS) according to the method described by Eaton et al., (1995). Lead concentrations in the December samples were determined according to the ISO 11 885 protocol (AFNOR, 1999) using ICP-AES (Inductively Coupled Plasma-Atom Emission Spectroscopy) ULTIMA II from Jobin Yvon. The detection principle of the F-AAS method functions by the absorption of light emitted by a ‘lead’ lamp. As for the detection principle of ICP-AES, it obtains a spectrum characteristic of lead lines following atomisation in argon plasma. The intensity of these lines is proportional to the quantity of atoms present in solution (AFNOR, 1999).

3.4 Risk characterisation The risk generated by lead in the drinking water of Port-au-Prince was characterised as a function of equation (1), which expressed the ratio between the exposure assessment (MDI) and the dose-response assessment (ADI): R=

R: MDI: ADI:

MDI ADI

(1)

risk maximum daily intake dose-response assessment.

For a chemical substance and a given exposure pathway, the general equation for estimating MDI, controlled by exposure vector i, is (INVS, 2000): MDIi = Ci × Qi × ER × ET/BW × WT

(2)

where MDI (maximum daily intake) is the proportion of substance absorbed per day of exposure, Ci is the concentration of the toxic substance in the polluted medium i,


285

Q the quantity of this vector brought into daily contact with the organism by the route considered (expressed in L/day for water media), ER is the exposure rate (without unit), ET is the exposure time (in years), BW is the body weight (in kg) and WT the time of weighing. TP is the duration (in years) over which the amount is weighed. In this formula, by convention, the time of weighing is identical to the exposure time (TW = EW) to reach the threshold: MDI is close to an annual average without consideration of the total period of exposure. To estimate the maximum daily intake via drinking water, a normal default assumption is 2 L/day for adults and 0.75 L/day for bottle-fed infants (Fawell and Young, 1999). Body weights (BW) of 70 kg and 5 kg were attributed to adults and infants (less than five years), respectively. To interpret the risk resulting from the ratio expressed by equation (1), three levels were considered:

4

R1

H: high.

Results and discussion

The highest concentrations obtained for the physicochemical characterisation of the samples from the two events are summarised in Table 1. In all the samples of the two events (April and December, 2004), pH was always in the alkaline range (7.26–7.83). The values for chlorides were lower than 250 mg/L. All the results obtained for the pH and chlorides were within acceptable limits for the drinking water guidelines (OMS, 1996). Table 1 Sampling points A1 A2 A3 A4 B C1 C2 D1 D2 D3

Highest concentrations obtained for selected physiochemical parameters from different water sampling points pH 7.83 7.26 7.38 7.36 7.31 7.29 7.47 7.40 7.31 7.44

Temperature (°C) 28.2 29.3 29.0 29.0 29.0 29.6 28.3 29.1 29.7 28.4

Conductivity (µS/cm) 960 1297 1004 1089 1250 1178 1197 1252 1121 1007

TDS (mg/L) 960 1238 960 1069 1205 1122 1203 1126 1121 1006

Cl– (mg/L) 30.49 27.82 23.33 18.49 31.66 20.83 22.99 21.49 26.33 19.16

Hardness (mg/L) 187 251 204 218 240 234 238 234 233 227

The conductivity values indicated high mineralisation of the samples. All the values were higher than 400 µS/cm, i.e., the threshold value for drinking water (ERB, 1999). The same observation could be made for the values obtained for TDS. Indeed, all the TDS concentrations were higher than 500 mg/L, the value suggested for fresh water (Desjardins, 1988). Furthermore, a value of 0.9 was obtained for the ratio between

286


TDS and conductivity. Theoretically, this ratio is an empirical factor that can vary from 0.55 to 0.9 depending on the soluble components of the water and on the measurement temperature. Relatively high factors may apply to saline or boiler waters, whereas lower factors may apply where considerable hydroxide or free acid is present (Eaton et al., 1995). Since pH in this study varied from 7.26 to 7.83, it is obvious that the salinity of the samples studied was very high. Table 2 shows the variations in lead concentrations from the water samples collected in April 2004. Generally, the presence of lead in drinking water can originate from industrial discharges, mines and foundries or deterioration of old lead piping (Eaton et al., 1995). Water samples from the boreholes for the first sampling events were lead-free, whereas during the same season, high lead concentrations were observed in the water samples from domestic taps and tanks. These observations can be explained by different factors. The lead concentrations of these analysed samples could be owing to the existence of equipment, lead piping or joints welded onto the network. On the other hand, the values obtained for lead determination using the water samples taken in April 2004 were below the detection limit of the equipment. However, this information does not mean that the groundwater is free of lead pollution. The lead detection limit of the equipment (Perkin-Elmer) is 15 µg/L, which is higher than the threshold value (10 µg/L). A human health risk to the consumers of drinking water from points ‘C’ may exist for any value not detected by the equipment and for which lead concentration would be higher than the threshold value. To verify this hypothesis, the presence of lead in water samples taken in December 2004 was determined using ICP-AES whose detection limit is lower than the threshold value (10 µg/L). Table 3 summarises the results obtained for lead from water samples collected during December 2004. Variations in lead concentrations between the two sampling events are illustrated in Figure 2. Table 2

Variation of lead concentrations in April 2004 water samples from different sampling points

Sampling points B C2 D3 Others Table 3

Means (µg/L) 245 45 185 –

Minima (µg/L) 240 38 180 –

Maxima (µg/L) 250 50 190