Case study of complaints on drinking water quality

Case Study of Complaints on Drinking Water QualityRelationship to Copper Content?

FERNANDO PIZARRO,*,1 MAGDALENA ARAYA,1

MARCELA VÁSQUEZ,2 GUSTAVO LAGOS,3 MANUEL OLIVARES,1

MARCO A. MÉNDEZ,1 BÁRBARA LEYTON,1 ARTURO REYES,3

VICTORIA LETELIER,3 AND RICARDO UAUY 1,4

1Instituto de Nutrición y Tecnología de los Alimentos,Universidad de Chile, Santiago, Chile; 2Facultad de Ciencias de la Salud, Universidad de Talca, Talca, Chile; 3Facultad de

Ingeniería, Pontificia Universidad Católica de Chile; and 4London School of Hygiene and Tropical Medicine London

University, London, UK

Received May 12, 2006; Revised June 19, 2006; Accepted July 20, 2006

ABSTRACTSeveral families of Talca city, Chile complained to health authorities

for what they attributed to consumption of copper (Cu)-contaminateddrinking water. We assessed the situation 6–12 mo after the initiation ofcomplaints by characterizing the symptoms reported, the chemistry ofdrinking water, and the Cu concentration in stagnant drinking water. Aftercompleting a census, 1778 households accepted participation and were cat-egorized as follows: category 1, Cu plumbing for tap water and dwellersreporting health complaints (HC); category 2, Cu plumbing for tap waterand dwellers reporting no HC; category 3, plastic plumbing for tap waterand dwellers reporting no HC. Questionnaires recorded characteristics ofhouseholds and symptoms presented by each member of the family in thelast 3 mo. The Cu concentration in drinking water was measured in a sub-sample of 80 homes with Cu pipes. In category 1, participants presentedsignificantly more abdominal pain, diarrhea, and/or vomiting (gastroin-testinal [GI] symptoms) in comparison to category 3 and to categories 2plus 3. The stagnant Cu concentrations measured in drinking water in allhouses studied were below the US Environmental Protection Agencyguideline value (<1.3 mg Cu/L). In summary, data obtained by interviewssuggested that individuals in some areas of Talca city were suffering moreGI symptoms potentially related to Cu excess, but measurement of Cu

Biological Trace Element Research 131 Vol. 116, 2007

© Copyright 2007 by Humana Press Inc.All rights of any nature, whatsoever, reserved.0163-4984/(Online) 1559-0720/07/11602–131 $30.00

*Author to whom all correspondence and reprint requests should be addressed.

concentration in stagnant tap waters ruled out the association between Cuexposure and GI symptom reports at the time of this study. Thedose–response curves for GI symptoms and Cu exposure now availablewere crucial in the analyses of results.

Index Entries: Copper pipe; PVC pipe; drinking water; gastrointesti-nal symptoms; diarrhea.

INTRODUCTION

Acute copper (Cu) intoxication results primarily from consumption ofcontaminated water or from heating foods using Cu containers (1,2). Theconcentration of this metal in drinking water can be influenced by multi-ple variables such as the natural mineral content, water pH, dissolvedoxygen, presence of oxidizing agents, complexing compounds or ions, andthe plumbing system (3,4). Although the Cu content in water from publicsupplies is generally low; higher levels have been reported as result of cor-rosion of the infrastructure of Cu piping (5).

In November 2001, acting on a request made by the EnvironmentalHealth Service of the Maule Region of Chile we investigated the origin ofhealth complaints related to drinking water quality reported by consumersin specific districts of Talca (a city of 175,000 inhabitants 450 km south ofSantiago, Chile). Complaints included gastrointestinal (GI) symptomssuch as vomiting, diarrhea, abdominal pain, and other nonspecific subjec-tive symptoms reported by the population of discrete geographic districts.In recent years, we have systematically studied the effects of acute Cuexposure in healthy volunteers and reported the dose response for GIsymptoms in relation to the Cu content of water (6–10).

Drinking water consumed in Talca originates from suppliers whoobtain water from local wells. In order to investigate the possibility thatthe Cu content of drinking water was responsible for the reported con-sumer complaints, we established an assessment protocol aimed at (1)characterizing reported symptoms, analyzing their frequency and distri-bution and (2) determining the characteristics of drinking water in thestudy area and the level of exposure to Cu based on the concentration intap water and in water as consumed. This report provides an account ofour findings, testing the hypothesis of a potential association betweensymptoms and Cu exposure from drinking water, by analyzing prevalenceof GI symptoms by level of exposure to Cu in drinking water.

METHODS

Characterization of the Potential for Exposure and Complaints

Households were categorized according to the type of piping system(Cu or plastic [PVC]) and the presence or absence of complaints in the

132 Pizarro et al.

Biological Trace Element Research Vol. 116, 2007

given geographic area: category 1, two districts with Cu plumbing for tapwater where dwellers reported health complaints; category 2, two districtswith Cu plumbing for tap water where dwellers reported no health com-plaints; category 3, a district with plastic (PVC) plumbing for tap waterwhere dwellers reported no health complaints. Socioeconomic status andtype of housing were similar in all districts. Homes in each district werevisited, and after providing information about the purpose of study, thehead of household was invited to participate in the survey. A total of 3110households were visited and 2009 agreed to participate and provided con-sent for the study. Similar rates of participation were observed to be inde-pendent of the type of pipes and the presence or absence of symptoms.Reasons for not agreeing to participate in the study were that no residentsat home at the time of contact (n = 706), rejection upon being informed ofthe study (n = 302), and inability to complete the answer questionnaire (n= 93). In addition, 203 households were excluded from the analysesbecause the water source was a private well at the residence. Thus, thesample population constituted 1778 families: 682 corresponded to category1 of potential exposure/symptoms and 700 and 396 in categories 2 and 3,respectively.

The families had an average of 3.8 ± 1.4, 3.6 ± 1.7, and 4.2 ± 1.6 sub-jects/household for categories 1, 2, and 3, respectively. Family memberswith chronic GI illness and pregnant women were excluded from theanalyses in order to eliminate confounding factors. The total number ofsubjects in the study sample was 6782, which correspond to 3.3% of thepopulation of Talca. The number of subjects was 2613, 2515, and 1654 forcategories 1, 2, and 3, respectively.

Assessment of Symptoms

During the first weekend of July 2002, the homes located in the 3 cat-egories were visited by 70 trained interviewers. Participants who con-sented to participate in the survey answered a questionnaire and agreed toreceive one additional visit for water sampling. The questionnaire wasanswered by each member of the family except for children younger than12 yr of age, for whom the mother or caretaker answered on their behalf.The head of the household answered the questions about characteristics ofthe home (year of construction, years living in the home, type of tap waterpiping, change of piping system within recent years, etc.). Each subject ortheir parent reported demographic data (sex, age, and time spent at home),presence of chronic diseases, and information on self-perception of symp-toms experienced during the last 3 mo in the following order: high bloodpressure, heartburn, nausea, arthritis, abdominal pain, chest pain, aller-gies, headache, diarrhea, asthma, bronchitis, palpitations, cough, vomit-ing, and emotional stress.

Based on our previous studies, we defined selective GI symptomsrelated to Cu exposure as the dependent variable (6–10). These were the

Cu in Drinking Water 133


presence or absence of vomiting, diarrhea, and abdominal pain. Nauseawas excluded from the definition because it is difficult to objectively defineit and it might be subtle and, therefore, more likely to be forgotten overtime. Other symptoms considered to be unrelated to Cu exposure in drink-ing water were interspersed in the list in order to mask the primary pur-pose of the study to the participants.

Assessment of Copper Exposure

The Cu-exposure assessment, conducted about 6 mo after completingthe survey for symptoms, confirmed that drinking water in all urban house-holds originated from a common underground network fed by multiplelocal wells; the water was transported through iron cast tubes until itreached the houses, where PVC or Cu pipes distributed water to the taps. Inorder to relate the presence of symptoms and Cu exposure in drinkingwater, a subsample of 80 homes with Cu pipes (categories 1 and 2) was ran-domly selected: 40 from areas where GI symptoms were frequently reported(2 or more family members reporting symptoms) and 40 from families thatreported no GI symptoms (at least 4 family members) and who lived closestto the index cases formerly described. One home of the no-GI-symptomgroup was excluded from the water analysis because they refused to partic-ipate in the study. Water sampling and analyses followed previously estab-lished protocols (11). Samples were obtained from the cold water kitchen tapin the early morning at least 8 h after the last use (stagnant water).

Stagnant (overnight, 8 h, 125 mL) and running (2-min flush, 1 L) watersamples were collected from the houses and stored at 4°C. Water samplesfrom each house were analyzed for Cu concentrations, pH, alkalinity, andtemperature. Cu concentrations were measured using a Perkin-Elmer Ana-lyst 100 atomic absorption spectrophotometer. Alkalinity, pH, and temper-ature measurements were made on site using portable instruments fromWTW (Weilheim, Germany). The sample handling and analysis of param-eters of the drinking water were carried out according to the US Environ-mental Protection Agency (EPA) and WHO (12–14)

Statistical Analyses

Data were analyzed by using the SAS statistical package (SAS 8.0E;SAS Institute Inc, Cory, NC). The Student’s t-test and two-way analysis ofvariance (ANOVA) were carried out to compare continuous variables(when ANOVA demonstrated statistical significance, group differenceswas assessed based on Scheffe’s post hoc test).

Univariate analyses (χ2 test and Fisher exact test) were used to iden-tify categorical variables associated with GI symptoms (diarrhea, vomit-ing, and/or abdominal pain). The variables were categorized according tothe following: category (1 vs 2 vs 3), age (<12 vs ≥12 yr), sex (male vsfemale), presence of Cu pipes for the cold kitchen tap (yes vs no), presenceof new Cu pipes because of replacement in piping system (yes vs no),intake of plain water (<0.6 vs ≥0.6 L/d), consumption of bottled water

134 Pizarro et al.


(<0.2 vs =0.2 L/d), year of construction of home (<1996 vs ≥1996), timespent at home (<16 vs ≥16 h/d), and first household member to get up inthe morning (yes vs no). A backward stepwise multiple logistic regressionmodel using factors influencing GI symptoms (p < 0.05) was applied toidentify significant predictors associated with GI symptoms. All p-valueswere two-sided and statistical significance was considered at p < 0.05.

RESULTS

Households

The main characteristics of household assessed appear in Table 1. Themean date of home construction was significantly different among the



Table 1Characteristics of Households

Note: Values within a column with different superscript letters are significantly different, p< 0.05 (Scheffe’s post hoc test).

136 Pizarro et al.


Table 2Characteristics of Subjects

Note: Values within a column with different superscript letters are significantly different, p< 0.05 (Scheffe’s post hoc test).

* Chi square test.

three categories. In category 2, houses were located in Talca center; onaverage, they were built 36.8 yr prior to the study, whereas houses in cat-egories 1 and 3 were built 8.9 and 6.6 yr before, respectively). Seventy-fivepercent of the homes in category 1 were built after 1996 versus 31.2% and38.1% of those in categories 2 and 3, respectively (p < 0.001).

Almost all homes in the three categories were connected to publicwater, sewer, and electricity systems. Ninety-four percent of the homes incategories 1 and 2 had Cu pipes for the cold water kitchen tap versus29.4% of the homes in category 3. Thirteen percent of the homes in the cat-egory 3 had replaced plastic (PVC) with Cu piping within the last 2 yr ver-sus 16.6% and 6.7% in categories 2 and 1, respectively.

Subjects

Table 2 shows the characteristics of the 2613, 2515, and 1654 subjectsincluded in categories 1, 2, and 3, respectively. In all three categories, themajority of the population studied were women, although the proportionof women was statistically higher in category 2. Subjects in category 3were significantly younger than in other categories; 31.1%, 24.3%, and13.8% of subjects in categories 3, 1, and 2, respectively were younger than12 yr of age. Because the age distribution was statistically different amongcategories, age was used as a covariate in the ANOVA analyses for thevariables time spent at home and water consumption.

Individuals in category 1 spent significantly more time away fromhome and presented a lower proportion of people with chronic diseasesthan those from other categories. Individuals in category 3, in which 70%of homes had PVC pipes, consumed significantly more plain water thanpeople of categories 1 and 2, in which homes had Cu pipes (ANOVA, p <0.001; Scheffé test, p < 0.05). Two-way ANOVA showed a significant effectof category (p < 0.001) and sex (p < 0.03) on plain water consumption.However, the interaction between category and sex was not statisticallysignificant.

Complaints

Figure 1A shows the percentage of subjects who reported having oneor more episodes of GI symptoms (abdominal pain, diarrhea, and/or vom-iting) during the last 3 mo, by category. Men in category 1 presented sig-nificantly higher percentages of GI symptoms than those of category 3.Women of category 1 presented significantly higher percentages of GIsymptoms than women in other categories. There were no differencesbetween categories 2 and 3 (no GI complaint). When category 3 was com-pared to categories 1 and 2 (homes with mainly Cu pipes), reports of GIsymptoms in category 3 was twofold less than in the other categories (6.6%vs 12.3%, χ2 = 41.7, p < 0.0001). No differences among groups wereobserved for frequency of symptoms unrelated to the GI tract (Fig. 1B).



138 Pizarro et al.


Fig. 1. Percentage of subjects (by gender and category) who reported one ormore episodes of GI symptoms. Men of category 1 had significantly higher per-centages of GI symptoms than those of category 3. Women of category 1 showedsignificantly higher percentages of GI symptoms than women in other categories.

Fig. 2. Percentage of subjects (by age and category) who reported one ormore episodes of GI symptoms. Children <12 yr in category 1 were significantlyhigher than in categories 2 and 3. Participants 12 to 64 yr of category 1 showedmore GI symptoms than those in categories 2 and 3, and subjects in category 2 pre-sented more GI symptoms than those in category 3.

Analysis of GI symptoms by category and age (Fig. 2) showed that theprevalence of complaints reported by the mothers of children <12 yr of agein category 1 was significantly greater than in categories 2 (p < 0.006) and3 (p < 0.001). Assessment of participants 12–64 yr of age (70% of total ineach group) showed that in category 1, these presented more GI symptomsthan those in categories 2 (p < 0.002) and 3 (p < 0.001); subjects in category



Table 3Percentage of Subjects with Symptoms in the Last 3 mo

by Type of Cold Water Pipes Used in the Kitchen

2 presented more GI symptoms than those in category 3 (p < 0.001). Therewere no differences among categories for subjects older than 65 yr.

Data analyzed by type of pipes for the cold water kitchen tap, inde-pendent of category, showed that subjects living in homes with Cu pipesreported a significantly higher percentage of subjects with abdominal painand diarrhea than subjects living in households with PVC piping systems(Table 3). Subjects living in homes with Cu piping showed a higher fre-quency of episodes of allergies, bronchitis, and emotional stress.

The univariate analysis showed that age (χ2 test, p < 0.001), sex (χ2 test,p < 0.03), category (χ2 test, p < 0.001), presence of Cu pipes for the coldwater kitchen tap (Fisher test, p < 0.001), consumption of bottled water(Fisher test, p < 0.001), and year of construction of home (Fisher test, p <0.0001) were related to GI symptoms.

Backward stepwise multiple logistic regression model analysis usingthe significant variables obtained from the univariate assessment , <12 yrliving in the house, female sex, home built during or after 1966, and con-sumption of <200 mL of bottled water were identified as risk factors for GIsymptoms (Table 4).

Studies of Copper Content in Stagnant Water

No members reported GI symptoms in the 39 homes included in thegroup “without GI symptoms”; instead, reports of abdominal pain, diar-rhea, and vomiting in the group “with GI symptoms” were 30.9%, 20.8%,

140 Pizarro et al.


Table 4Factors Predicting GI Symptoms: Multivariate Logistic Regression

Note: OR-odds ratio and 95% confidence interval.

and 5.1%, respectively. Subjects in this latter group also presented signifi-cantly more symptoms unrelated to Cu included as part of the maskingstrategy (episodes of joint pain, allergies, headaches, bronchitis, and cough).

Figure 3 shows the cumulative frequency distributions of stagnantwater Cu concentrations and running water Cu concentrations for the 80household sampled. It can be observed that the 90th percentile concentra-



Fig. 3. Cumulative frequency distributions of stagnant water Cu concentra-tions and running water Cu concentrations for the 80 households sampled.

Table 5Water Composition of Stagnant Water Sampled

in Households With and Without GI Symptoms Reported

a Units in milligrams per liter unless otherwise stated.

tion was 0.93 mg/L, with a highest concentration of 1.15 mg/L and a min-imum of about 0.03 mg/L. The stagnant Cu concentrations measured indrinking water in all households studied were below the US EPA Maxi-mum Contaminant Level Goal (MCLG) (13). The mean stagnant water Cuconcentration was 0.50 ± 0.32 mg/L and running water Cu was 0.06 ± 0.05mg/L. The water composition of stagnant water of household groups“with” and “without” GI symptom reports is shown in Table 5. The pHand temperature of stagnant water did not show differences between thegroups “with” and “without” GI symptom reports. Cu exposure throughdrinking water was calculated at 0.78 ± 0.75 and 0.60 ± 0.50 mg/d forhouseholds “with” and “without” GI symptoms, respectively (p < 0.01).

When data were analyzed by categories 1, 2, and 3, the mean pH instagnant water in category 1 was significantly lower than in those fromcategory 2 (6.64 ± 0.11 vs 6.92 ± 0.26 respectively, p < 0.01). The Cu con-centration in stagnant tap water samples was also significantly different,being lower in category 2 than in category 1 (0.16 ± 0.13 vs 0.68 ± 0.36,respectively, p < 0.01); 33% and 100% of homes from category 1 and 2 had<0.5 mg Cu/L, respectively.

DISCUSSION

This case study illustrates the difficulties encountered in assessingreal-life situations in which a potential toxic effect of Cu is suspected. Peo-ple living in Talca attributed their health complaints to consumption ofdrinking water contaminated with Cu. This provoked a public alarm, lead-ing to local newspapers and television reports on these matters (Diario ElCentro, October 21–24, 2001); public discussion led to a consensus ascrib-ing the presence of symptoms (GI and non-GI symptoms) to the high Cuexposure through drinking water. Some neighbors stopped consuming tapwater and were concerned about using tap water in the shower or formouth washing. This situation led the Environmental Health Service ofTalca to request this assessment of the problem. Unfortunately, owing tothe logistics of securing financial support, obtaining pertinent authoriza-tions, preparing the protocols, and training of interviewers, the health sur-vey was performed approx 8 mo after the complaints had occurred and thewater sampling 6 mo after the survey. In addition, all participants wereaware of the problem and had personal views and apprehensions aboutthe problem. These factors constitute significant limitations in the inter-pretation of these data, yet, despite the problems of time lag in completionof the various steps of this study, we were able to determine an associationbetween exposure to Cu and GI symptoms.

Data obtained by questionnaire support the hypothesis that symptomswere related to Cu, as individuals in category 3 (in which 70% of house-holds had PVC piping and residents knew this) had significantly higherconsumption of water (p < 0.001) and lower frequency of GI symptom

142 Pizarro et al.


reported (p < 0.008). Moreover, analysis of the GI symptoms reported by thetype of pipes used in the kitchen cold water tap showed that subjects usingCu piping systems reported significantly more GI symptoms, independentof the category of exposure/symptoms established in the study. A potentialconfounder in this observation is the fact that individuals with Cu pipingsystems in their homes also reported significantly more symptoms unre-lated to Cu intake, such as allergies, bronchitis, and emotional stress, andthat all homes assessed were built earlier (>2 yr earlier), which would makeCu leaching from pipes less likely to contribute to the Cu in drinking water.In contrast to data obtained from the survey, studies of stagnant tap watershow that GI symptoms were unlikely to be related to Cu exposure. Mea-surements of stagnant waters revealed Cu concentrations well below 1.2mg/L, and on the basis of mean daily water consumption, exposurethrough drinking water was 0.78 ± 0.75 and 0.60 ± 0.50 mg/d in familiesreporting and not reporting health complaints, respectively.

Acute GI responses to Cu exposure have now been characterized,defining 2 mg Cu/L as the NOEL (No Observed Effects Level) for copperin water (8,10,15). The dose–response curve that led to this figure wasbased on double-blind, randomized, controlled clinical trials that assessedadult women and men living in a variety of conditions and cultural set-tings; these studies served to support maintaining the current WHOguideline of 2 mg Cu/L as the safe level of Cu in drinking water (15). Cuconcentrations in stagnant tap waters below the range that induce acute GIadverse effects do not explain the increase of complaint reports at the timeof assessment.

Unfortunately, the health survey and water sampling were performedseveral months after the complaints appeared, limiting our capacity toreach a firm conclusion; this problem has previously been faced by otherinvestigators under similar situations. Because Cu content of drinkingwater might vary over time, a second sampling in the area and householdswas performed 6 mo (i.e., contrasting season) after the first evaluation. Itrevealed minimal changes in Cu concentration (0.50 ± 0.32 mg/L in thefirst measurements vs 0.570 ± 0.360 mg/L in the second measurementsmade 6 mo later).

A well-known study performed in 1993 in Nebraska (USA) also usinga retrospective approach showed a high risk of developing diarrhea in agroup that had been exposed to concentrations of Cu >3 mg/L (16). In asecond stage of this study, a case-control design assessed those subjects liv-ing in the area who reported GI illness during the 2 wk before the second-phase evaluation. As shown in the present study, these subjects wereinterviewed using an ad hoc standard questionnaire, and analyses wereperformed on the drinking water in their homes. The Cu concentrationsfound in drinking water were markedly different from those obtained pre-viously, many homes (with a prior Cu concentration greater than 3 mg/L)presenting much lower concentrations in the second assessment. Thesediscrepancies were attributed to gathering the 1994 water samples during



the 3 wk after the illness report, whereas the 1993 samples had been col-lected at least 7 mo after the sickness report, giving rise to serious doubtsof the validity of these results.

These studies demonstrate the need to address public concern using asystematic evaluation rather than favoring rapid answers. In our case studyof Talca, laboratory experiments of Cu pipe leaching made with the mostaggressive water compositions found in Talca (alkalinity: 90 mg/L asCaCO3, pH 6,3, SO4 = 13 mg/L, Cl = 5.2 mg/L, Ca = 12 mg/L; sample stag-nation time: 24 h) suggested that the water Cu concentration with newlyinstalled Cu pipes ranged between 2 and 3 mg/L, but this value woulddecrease to less than 1 mg/L in about 5 wk, because of the natural agingprocess of a Cu pipe. This suggests that at the time when complaints wereoccurring the water Cu concentration was likely below 1 mg/L, possiblyvery similar to the measurements made many months after the complaints.

In summary, public concern led the formal assessment of a situationin Talca; interviews of the individuals potentially suffering the effects ofCu exposure (that at the same time were possibly biased in the analysis ofthe situation) suggest that Cu exposure was related to the complaints;however, objective measurements of Cu concentration in stagnant tapwater rules out the association between Cu exposure and GI symptomreported at the time of this study. Finally, it seems worthwhile emphasiz-ing that the existence of the dose–response curve for GI symptoms versusCu exposure was crucial to assess the results obtained. Regions where nat-ural water is aggressive and pH tends to be acidic need to be closely mon-itored to assure that the range of variability of Cu content in drinkingwater stays within the international norms.

ACKNOWLEDGMENTS

We thank to the students at the Health Science School of the Univer-sity of Talca for their participation in this study as interviewers. This studywas supported in part by an unrestricted grant from International CopperAssociation (ICA) Research Program and by CINUT (Corporación paraApoyo de la Investigación Científica en Nutrición).

REFERENCES

1. National Research Council, Drinking Water and Health, Vol. 3 National Academy Press,Washington DC (1980).

2. F. Pizarro, M. Olivares, V. Gidi, and M. Araya, The gastrointestinal tract and acuteeffects of copper in drinking water and beverages, Rev. Environ. Health 14, 231–238(1999).

3. National Research Council, Copper in Drinking Water, National Academy Press, Wash-ington, DC (2000).

144 Pizarro et al.


4. US Environmental Protection Agency, Effect of pH, DIC, Orthophosphate and Sulfate onDrinking Water Cuprosolvency, EPA 600/R-95/085, Office of Research and Development,Washington DC (1995).

5. A. M. Dietrich, D. Glindemann, F. Pizarro, et al., Health and aesthetic impacts of cop-per corrosion on drinking water, Water Sci. Technol. 49, 55–62 (2004).

6. F. Pizarro, M. Olivares, R. Uauy, P. Contreras, A. Rebelo, and V. Aidi, Acute gastroin-testinal effects of graded levels of copper in drinking water, Environ. Health Perspect.107, 117–121 (1999).

7. F. Pizarro, M. Olivares, M. Araya, V. Gidi, and R. Uauy. Gastrointestinal effects associ-ated with soluble and insoluble copper in drinking, Environ. Health Perspect. 109,949–952 (2001).

8. M. Olivares, M. Araya, F. Pizarro, and R. Uauy, Nausea threshold in apparently healthyindividuals who drink fluids containing graded concentration of copper, Regul. Toxicol.Pharmacol. 33, 271–275 (2001).

9. M. Araya, M. Olivares, F. Pizarro, M. Gonzalez, H. Speisky, and R. Uauy, Gastrointesti-nal symptoms and blood indicators of copper load in apparently healthy adults under-going controlled copper exposure, Am. J. Clin. Nutr. 77, 646–650 (2003).

10. M. Araya, M. Olivares, F. Pizarro, A. Llanos, G. Figueroa, and R. Uauy, Community-based randomized double-blind study of gastrointestinal effects and copper exposurein drinking water, Environ. Health Perspect. 112, 1068–1073 (2004).

11. G. Lagos, Aging of copper pipes by drinking water, J. Am. Water Works Assoc. 93, 94–105(2001).

12. US Environmental Protection Agency, Seminar Publication: Control of Lead and Copper inDrinking Water, EPA 625/R-93/001, Office of Research and Development, WashingtonDC (1993).

13. US Environmental Protection Agency, Methods for the Determination of Metal, EPA600/491/010, 1991, Office of Research and Development, Washington DC (1991)(Revised 1994, revision 5.4).

14. World Health Organization, Guidelines for Drinking-Water Quality, 2nd ed., Addendumto Volume 2, Health Criteria and other Supporting Information, WHO/EOS/98-1, WHO,Geneva (1998).

15. M. Araya, B. Chen, L. M. Klevay, et al., Confirmation of an acute no-observed-adverse-effect and low-observed-adverse-effect level for copper in bottled drinking water in amulti-site international study, Regul. Toxicol. Pharmacol. 38, 389–399 (2003)

16. S. D. Buchanan, R. Diseker, J. Daniel, T. Floodman, and T. Sinks, Evaluating gastroin-testinal irritation among humans from copper in drinking water, Report Epi-E94-73(1994).



Case study of complaints on drinking water quality

Documents

Transcript of Case study of complaints on drinking water quality