Environmental Research 114 (2012) 1–11

Contents lists available at SciVerse ScienceDirect

Environmental Research
journal homepage: www.elsevier.com/locate/envres

Concentrations and correlations of disinfection by-products in municipal

drinking water from an exposure assessment perspective$
Cristina M. Villanueva a,b,c,n, Gemma Castaño-Vinyals a,b,c, Vı́ctor Moreno c,d,e,f,

Gloria Carrasco-Turigas a,b,c, Nuria Aragonés c,g, Elena Boldo c,g, Eva Ardanaz c,h, Estefanı́a Toledo h,i,
Jone M. Altzibar c,j, Itziar Zaldua j, Lourdes Azpiroz j, Fernando Goñi c,k, Adonina Tardón c,l, Antonio
J. Molina c,m, Vicente Martı́n c,m, Concepción López-Rojo c,n, José J. Jiménez-Moleón c,o, Rocı́o Capelo p,
Inés Gómez-Acebo c,q, Rosana Peiró r, Mónica Ripoll r, Esther Gracia-Lavedan a,b,c, Mark
J. Nieuwenhujsen a,b,c, Panu Rantakokko s, Emma H. Goslan t, Marina Pollán c,g, Manolis Kogevinas a,b,c,u
a
Centre for Research in Environmental Epidemiology (CREAL), Doctor Aiguader 88, 08003-Barcelona, Spain
b
IMIM (Hospital del Mar Research Institute), Doctor Aiguader 88, 08003-Barcelona, Spain
c
CIBER Epidemiologı́a y Salud Pública (CIBERESP), C/Melchor Fernández Almagro, 3–5, Pabellón 6, planta baja, 28029 Madrid, Spain
d
Catalan Institute of Oncology (ICO), Gran Via de l’Hospitalet 199-203, 08908 l’Hospitalet de Llobregat, Spain
e
Bellvitge Biomedical Research Institute (IDIBELL), Gran Via de l’Hospitalet, 199, 08908 Hospitalet de Llobregat, Spain
f
University of Barcelona (UB), Barcelona, Spain. Gran Via de les Corts Catalanes, 585, 08007 Barcelona, Spain
g
Cancer and Environmental Epidemiology Unit, National Centre for Epidemiology, Carlos III Institute of Health. C/Sinesio Delgado, 4, 28029 Madrid, Spain
h
Public Health Institute of Navarra, C/Leyre, 15 31003 Pamplona, Spain
i
Department of Preventive Medicine and Public Health, University of Navarra, Campus Universitario, 31009 Pamplona, Spain
j
Public Health Division of Gipuzkoa, Avda. de Navarra, 4 20013 Donostia, San Sebastián, Spain
k
Laboratory of Public Health of Gipuzkoa, Avda. de Navarra, 4 20013 Donostia, San Sebastián, Spain
l
University of Oviedo, Campus El Cristo s/n 33006 Oviedo, Spain
m
Department of Preventive Medicine and Public Health, IBIOMED, University of León, Campus Universitario, 24071 León, Spain
n
General Subdivision of Care Quality, Universidad de Murcia, Avda Teniente Flomesta, 5-30003, Murcia, Spain
o
Department of Preventive Medicine and Public Health, University of Granada, Avda. del Hospicio, s/n C.P. 18071 Granada, Spain
p
Centre for Research in Environment and Health (CYSMA), University of Huelva, Dr. Cantero Cuadrado, 6. 21071 Huelva, Spain
q
University of Cantabria, Avda. de los Castros, s/n 39005 Santander, Spain
r
Centre for Research in Public Health (CSISP), Avda. de Cataluña, 21, 46020 Valencia, Spain
s
National Institute for Health and Welfare (THL), PO Box 95, FI-70701 Kuopio, Finland
t
Cranfield Water Science Institute, Cranfield University, Bedfordshire MK43 0AL, United Kingdom
u
National School of Public Health, Athens, Greece Alexandras Avenue, 196, 11521 Athens, Greece

a r t i c l e i n f o abstract

Available online 20 March 2012 Although disinfection by-products (DBPs) occur in complex mixtures, studies evaluating health risks
Keywords: have been focused in few chemicals. In the framework of an epidemiological study on cancer in 11
Complex mixtures Spanish provinces, we describe the concentration of four trihalomethanes (THMs), nine haloacetic acids
Disifection by-products (HAA), 3-chloro-4-(dichloromethyl)-5-hydroxy-2(5H)-furanone (MX), four haloacetonitries, two halo-
Drinking water ketones, chloropicrin and chloral hydrate and estimate correlations. A total of 233 tap water samples
Exposure were collected in 2010. Principal component analyses were conducted to reduce dimensionality of
Exposure assessment DBPs. Overall median (range) level of THMs and HAAs was 26.4 (0.8–98.1) and 26.4 (0.9–86.9) mg/l,
respectively (N¼ 217). MX analysed in a subset (N ¼36) showed a median (range) concentration of 16.7
(0.8–54.1) ng/l. Haloacetonitries, haloketones, chloropicrin and chloral hydrate were analysed in a
subset (N¼ 16), showing levels from unquantifiable (o1 mg/l) to 5.5 mg/l (dibromoacetonitrile).
Spearman rank correlation coefficients between DBPs varied between species and across areas, being

$
Funding sources: This study is partly funded by the Instituto de Salud Carlos III FEDER (PI08/0533, PI08/1770), the ‘‘Acción Transversal del Cáncer del Consejo de
Ministros del 11/10/2007’’ and by the ‘‘CIBER Epidemiologı́a y Salud Pública’’.
n
Corresponding autor.
E-mail addresses: cvillanueva@creal.cat (C.M. Villanueva), gcastano@creal.cat (G. Castaño-Vinyals), v.moreno@iconcologia.net (V. Moreno),
gcarrasgo@creal.cat (G. Carrasco-Turigas), naragones@isciii.es (N. Aragonés), eiboldo@isciii.es (E. Boldo), me.ardanaz.aicua@cfnavarra.es (E. Ardanaz),
etoledo@unav.es (E. Toledo), epide11ss-san@ej-gv.es (J.M. Altzibar), i-zaldua@ej-gv.es (I. Zaldua), mambien2-san@ej-gv.es (L. Azpiroz), labora2-san@ej-gv.es (F. Goñi),
atardon@uniovi.es (A. Tardón), ajmolt@unileon.es (A. Molina), vicente.martin@unileon.es (V. Martı́n), concepcion.lopez7@carm.es (C. López-Rojo),
jjmoleon@ugr.es (J.J. Jiménez-Moleón), rocio.capelo@dbasp.uhu.es (R. Capelo), ines.gomez@unican.es (I. Gómez-Acebo), peiro_ros@gva.es (R. Peiró),
ripoll_monmae@gva.es (M. Ripoll), egracia@creal.cat (E. Gracia-Lavedan), mnieuwenhujsen@creal.cat (M.J. Nieuwenhujsen), panu.rantakokko@thl.fi (P. Rantakokko),
E.H.Goslan@cranfield.ac.uk (E.H. Goslan), mpollan@isciii.es (M. Pollán), kogevinas@creal.cat (M. Kogevinas).

0013-9351/$ - see front matter & 2012 Elsevier Inc. All rights reserved.
doi:10.1016/j.envres.2012.02.002
2 C.M. Villanueva et al. / Environmental Research 114 (2012) 1–11
highest between dibromochloromethane and dibromochloroacetic acid (rs ¼0.87). Principal component
analyses of 13 DBPs (4 THMs, 9 HAAs) led 3 components explaining more than 80% of variance. In
conclusion, THMs and HAAs have limited value as predictors of other DBPs on a generalised basis.
Principal component analysis provides a complementary tool to address the complex nature of the
mixture.
1. Introduction
Disinfection of drinking water and swimming pools is necessary
to prevent water-borne infections. However, unintended disinfection
by-products (DBPs) are formed by the reaction between organic
matter and the highly reactive disinfectants. DBPs are ubiquitous
contaminants of concern in drinking water since human exposure
has been associated with cancer and reproductive outcomes (Tardiff
et al., 2006; Villanueva et al., 2004). DBPs constitute a complex
mixture with around 600 identified chemicals, accounting for less
than 50% of total organic halogen (TOX) (Richardson et al., 2007).
Toxicity of DBPs differs among specific chemicals. Iodinated DBPs
occur in low concentrations but show an enhanced mammalian cell
cytotoxicity and genotoxicity as compared to their brominated and
chlorinated analogues (Richardson et al., 2007). Brominated com-
pared to chlorinated species, and the nitrogen-containing DBPs (N-
DBPs) compared to carbonaceous DBPs show higher toxicity across
different structural DBP classes (Plewa et al., 2008). The occurrence of
N-DBPs and Br-DBPs is generally low. However, levels of brominated
DBPs can be higher than chlorinated DBPs when bromide is present
in raw water. Human exposure to DBPs occurs in mixtures, and
among other factors, the internal dose differs by exposure route (Xu
and Weisel, 2003, 2005; Xu et al., 2002; Trabaris et al., 2011).
To deal with this complexity, epidemiological studies evaluat-
ing adverse health effects of chlorinated water or DBPs have relied
on index chemicals, generally trihalomethanes (THMs) and in a
lesser extent haloacetic acids (HAAs) and other DBPs (Hinckley
et al., 2005; Hoffman et al., 2008; Aggazzotti et al., 2004).
Trihalomethanes (THMs) and haloacetic acids represent the two
major classes of halogenated disinfection by-products on a weight
basis when chlorine is used, either as chlorine gas or hypochlorite
solution (Krasner et al., 2006). Four THMs (chloroform, bromodi-
chloromethane, dibromochloromethane and bromoform) and
5 HAAs (monochloroacetic acid [MCAA], dichloroacetic acid
[DCAA], trichloroacetic acid [TCAA], monobromoacetic acid
[MBAA], and dibromoacetic acid [DBAA]) are regulated in the
European Union (THMs), USA (THMs and HAAs) and other
countries (e.g., THMs in Australia).
The correlation between DBP constituents is complex and
strongly depends on raw water quality and the type of treatment.
Consequently, the use of a specific group of DBPs, such as THMs,
may result in the misclassification of exposure to the relevant
chemicals for health outcomes. In the framework of an epidemio-
logical study on cancer conducted in 11 provinces in Spain (MCC-
Spain, www.mccspain.org), we aim to describe the concentration
of 4 THMs, 9 HAA, 3-chloro-4-(dichloromethyl)-5-hydroxy-2(5H)-
furanone (MX), 4 haloacetonitries, 2 haloketones, chloropicrin
and chloral hydrate in 11 Spanish provinces, and estimate
correlations between them in order to obtain recommendations
oriented to exposure assessment in epidemiological studies.
2. Materials and methods
2.1. Study area and sampling points
A total of 233 tap water samples were collected in the MCC-Spain study areas:
Asturias, Barcelona, Cantabria, Gipuzkoa, Granada, Huelva, León, Madrid, Murcia,
& 2012 Elsevier Inc. All rights reserved.
Navarra and Valencia provinces. The specific study area within each province is
determined by the catching area of the 22 participating hospitals. Water source
was 100% surface in Asturias, Gipuzkoa, Madrid, and Valencia study municipa-
lities. Barcelona, Granada and León study municipalities had a varying fraction of
water from ground and surface sources. There are hundreds of treatment plants
and distribution systems in the study areas. Although specific treatment was not
available for all of them, chlorine was known to be used as primary disinfectant in
virtually all the study areas.
The number and location of sampling points was approximately defined
according to the population (person-years) that municipalities contributed to
the study (Appendix, Table A1). A total of 217 sampling points were defined (52 in
Barcelona, 34 in Madrid, 28 in Navarra, 18 in Asturias, 16 in León 14 in Gipuzkoa,
14 in Murcia, 11 in Granada, 10 in Huelva, 10 in Cantabria and 10 in Valencia). A
single sampling campaign was conducted in 2010 (from 19 April to 29 November)
in all locations except in Barcelona, where a replicated sampling was conducted in
16 points (on 15 March 2010), to analyse a different set of DBPs.
2.2. Water sample collection
Water samples were collected after leaving water running for 5 min approxi-
mately. Samples for MX measurements were collected in 500 ml polyethylene
bottles, acidified with HCl 1 M, quenched with ammonium sulphate (10 mg/
100 ml sample), and frozen. Samples for the analysis of the other DBPs were
collected in glass vials with Teflon-faced silicone septum screw caps. Vials were
filled in completely avoiding bubbles and loss of the quenching agent, kept
refrigerated (4 1C) and shipped immediately to the laboratory. Sample volume
and quenching agents differed slightly between the main sampling (N¼ 217) and
the replicate sampling (N ¼ 16). In the main sampling, vials for trihalomethane
analysis were 40 ml and contained 100 ml of sodium thiosulfate 10% and 35 ml of
4 N HCl. Vials for HAAs were 100 ml and contained 250 ml of ammonia chloride 4%
(parts/volume) in water. In the replicated sampling in Barcelona (N ¼16) all
containers were 40 ml. Vials for HAA measurements had ammonium chloride
(99%þ specified reagent grade) at 100 mg/l (according to USEPA Method 552.3).
Sample vials for other DBPs were prepared with either ammonium chloride
(haloacetonitriles, haloketones and chloropicrin) or sodium sulphite granules
(trihalomethanes and chloral hydrate), (99% þ specified reagent grade) at
100 mg/l. Vials also contained a phosphate buffer (granular) to adjust the pH to
4.5–5.5. (according to USEPA Method 551.1). The buffer contained 1% sodium
phosphate dibasic (Na2HPO4) and 99% potassium phosphate monobasic (KH2PO4).
2.3. Analytical measurement
The analysis of trihalomethanes and haloacetic acids in the 217 samples of the
main campaign followed a modified form of Standard Method 6232B (Eaton et al.,
2005). Trihalomethanes were measured within 72 h after reception through salted
liquid–liquid extraction with n-pentane and quantification by gas chromatography
(Agilent 5890, Santa Clara, CA, USA) with a HP-5 column (50 m  0.32 mm  1.05 mm,
Agilent) and electron capture detector. Nine haloacetic acids (MCAA, DCAA, TCAA,
MBAA, DBAA, bromochloro- [BCAA], bromodichloro- [BDCAA], dibromochloro-
[DBCAA], and tribromoacetic acid [TBAA]) were analysed within 1 week after
reception, using a modified form of USEPA Method 552.3 (USEPA, 2003). This
procedure involved salted liquid–liquid extraction with tert-butyl methyl ether at
pHr0.5, followed by methyl derivatization and gas chromatography (Agilent 5890)
with a HP-5MS column (30 m  0.25 mm  0.25 mm, Agilent) and negative chemical
ionization mass spectrometry detection (NCI-MS). Quantification limits were 0.4 mg/l
for trihalomethanes, 0.05 mg/l for TCAA, 0.3 mg/l for TBAA and DBAA, 0.5 mg/l for
MCAA and 0.1 mg/l for MBAA, DCAA, BCAA, BDCAA and DBCAA). The experimental
analysis was conducted in the Public Health Laboratory in Gipuzkoa (Spain), certified
by the National Acreditation Institution (ENAC 137/LE328).
MX was analysed in 36 samples (14 in Barcelona, 8 in Madrid, 4 in Gipuzkoa,
4 in Navarra, 2 in Asturias, 2 in León and 2 in Murcia). Measurement was based on
solid phase extraction (SPE) pretreatment followed by gas chromatography high
resolution mass spectrometry (GC-HRMS) that was modified from the method
previously published (Rantakokko et al., 2004). In brief, pH of 100 ml water
sample was adjusted to 2 with HCl and pumped with tubing pumps through a
train of Waters Sep-Pak Plus tC18 (traps impurities) and Waters HLB-Plus (traps
MX) SPE cartridges. MX was eluted from HLB-Plus with 4 ml of acetone that was
 C.M. Villanueva et al. / Environmental Research 114 (2012) 1–11 3

evaporated to dryness. Internal standard 13C-2,4,5-trichlorophenol in 125 ml of

isopropanol, 125 ml of 4% H2SO4-isopropanol, 100 ml of nonane, and 200 ml of
hexane were added to dried eluate. Samples were placed for 1 h to 85 1C to
isopropylate the analytes. After cooling 2 ml of ultrapure water (UPW) and 4 ml of
hexane were added, and samples were mixed. Hexane phase was separated,
washed with 1 ml of UPW, dried with Na2SO4, and evaporated to a final volume of
50 ml in nonane. Recovery standard PCB-30 was added to autosampler vials.
Isopropyl derivative of MX was analysed with gas chromatography (Hewlett
Packard 6890) coupled to high resolution mass spectrometry (Waters Autospec
Ultima). Column used was DB-5MS (Agilent Technologies: 30 m, i.d. 0.25 mm,
0.25 mm). Quantification limit for MX was 0.5 ng/l. The MX analyses were
conducted in the Chemical Exposure Unit of the National Institute for Health
and Welfare (THL), Kuopio, Finland. The laboratory is an accredited testing
laboratory No T077 in Finland (EN ISO/IEC 17025, more information at http://
www.mikes.fi).
In the 16 replicated samples collected in Barcelona, 4 trihalomethanes (chloro-
form, bromodichloromethane, dibromochloromethane and bromoform), 4 haloa-
cetonitriles (bromochloro-, dibromo-, dichloro-, and trichloroacetonitrile),
2 haloketones (1,1-dichloro- and 1,1,1-trichloropropanone), chloral hydrate, and
chloropicrin were extracted using a modified form of U.S. Environmental Protec-
tion Agency (USEPA) Method 551.1 (USEPA, 1995). This method involved salted
liquid-liquid extraction with solvent extracts analysed by gas chromatography
(GC) (Agilent 6890, Santa Clara, CA, USA) with micro electron capture detection
(mECD). Nine HAAs (monochloro-, monobromo-, dichloro-, trichloro-, bromo-
chloro-, dibromo-, bromodichloro-, dibromochloro-, and tribromoacetic acid were
analysed using a modified form of USEPA Method 552.3 (USEPA, 2003; Tung et al.,
2006). Briefly, the water samples were acidified and the protonated HAAs
extracted by liquid–liquid extraction into the organic solvent phase (methyl tert Fig. 1. Distribution of total trihalomethane, and total haloacetic levels in all areas
butyl ether, MTBE). The HAAs were then converted to their more volatile methyl (N¼ 217).
esters by heating with acidic methanol before neutralisation with sodium
sulphate. HAAs were converted to their methyl esters in order that they could
be analysed using gas chromatography with micro electron capture detection (GC-
mECD, Agilent 6890) for quantification. All samples were analysed in duplicate. León and Murcia, where the number of treatment plants and
The limit of detection for each DBP was 1 mg/l with the exception of MCAA, where distribution networks were largest.
the detection limit was 2 mg/l. All analytes were quantified using a ZB1-MS
Chlorine is being progressively substituted by alternative
column (Phenomenex UK Ltd, Macclesfield, Cheshire, UK, equivalent to DB1) and
at least 50% of the samples were analysed in duplicate using a ZB5-MS column
disinfectants to meet the regulations. The use of alternative
(Phenomenex UK Ltd, equivalent to DB5) for confirmation. Where duplicates were disinfectants such as ozone, chlorine dioxide, and chloramines,
analysed, the value reported was an average of the two measurements. These reduces the formation of the regulated THMs, HAAs, and TOX, but
analyses were conducted at the Cranfield Water Science Institute, Cranfield other DBPs of concern are formed at higher levels (Krasner et al.,
University (UK).
2006). For example, bromate is formed by ozonation when raw
water contains bromide. Chlorate and chlorite are by-products of
2.4. Statistical analysis chlorine dioxide, and chloramines may form iodinated THMs,
iodinated acids and nitrogenated by-products such as N-nitroso-
The concentration of chloroform, bromodichloromethane, dibromochloro- dimethylamine (NDMA) (Zhang et al., 2000; Richardson et al.,
methane and bromoform was added to generate the variable total THMs
(THM4). The same procedure was followed to create the sum of the 9 haloacetic
2008; Mitch and Sedlak, 2002; Choi and Valentine, 2002). Water
acids (HAA9). To avoid missing values, measurements below the quantification treatment included chlorination in all the study areas, although
limit (QL) were imputed QL/2. The distributions of variables were explored, treatment processes differed and included additionally alterna-
showing lack of normality. Spearman rank correlation coefficients (rs) were tive disinfectants in some water zones (such as chlorine dioxide,
calculated to evaluate the degree of correlation between DBPs. To correct for
ozone and chloramination). Our study did not aim and was not
between-area variability, partial correlations were calculated by using residuals
from linear regressions of chemicals adjusting for area. We used a factor analysis designed to evaluate by-products of ozonation, chlorine dioxide
and a principal component analysis (PCA) to reduce the information of the or chloramination, and we could not estimate the effect of
mixture. A factor analysis was first conducted to identify the number of factors different treatments on DBP levels. The change for alternative
with eigen-values above one. A PCA was followed, including the number of disinfectants in the study areas is recent, and chlorine has been
components identified in the factor analysis. Factor analysis and PCA are based on
orthogonal transformations of a set of correlated variables (DBPs in our case) to
used in the past during decades. As THMs are correlated with the
create a set of uncorrelated variables denominated principal components, that are total organic halogen content when chlorine is used as the main
linear combinations of the original DBP variables. disinfectant (Singer and Chang, 1989), we can expect that histor-
ical levels of chlorination by-products (e.g., trihalomethanes)
should be valid markers of chronic exposure when evaluating
3. Results and discussion cancer risk. However, this assumption does not necessarily apply
for health outcomes with a shorter latency periods (e.g., repro-
3.1. Total trihalomethane and haloacetic acid levels ductive outcomes), if alternative disinfectants are used.

Concentration of trihalomethanes and haloacetic acids differed
among areas (Fig. 1 and Appendix Table A2 and A3). Median
THM4 level was 26.4 mg/l (interquartile range 17.3–40.0 mg/l),
and median HAA9 was 26.4 mg/l (interquartile range 17.7–
34.2 mg/l). The median concentration of HAA9 was slightly higher
than THM4 concentration in some areas (Asturias, Granada, León,
Madrid and Navarra). All THM4 measurements were below
100 mg/l, in compliance with the European and the Spanish
maximum regulatory levels (Council Directive 98/83/EC, 1998;
Real Decreto num 1138/1990, 1990). Variability was largest in
3.2. Speciation of trihalomethanes and haloacetic acid levels
Proportion of brominated and chlorinated species varied
among areas (Fig. 2 and Appendix Tables A1 and A2). Madrid
showed the lowest while Valencia had the highest proportion of
brominated THMs and HAAs. Among the 9 HAAs, monochloracetic
acid and tribromoacetic acid occurred at lowest concentrations,
with 60% and 50% of samples below the quantification limit,
respectively. Dichloro- and trichloroacetic acids were the HAA
species occurring at highest concentrations (overall median of
4 C.M. Villanueva et al. / Environmental Research 114 (2012) 1–11

7.3 and 6.5 mg/l, respectively, overall). The sum of dihalo- and
trihaloacetic acid species showed similar concentrations
(median of 12.7 and 13.3 mg/l, respectively). The US EPA guideline
value for HAAs (60 mg/l for the sum of MCAA, DCAA, TCAA, MBAA,
and DBAA) was exceeded in 4 samples, ranging from 64.5 to
76.7 mg/l. These samples were located in 2 areas (Huelva and
León). The chlorine-bromine speciation depends on the bromide
content in the raw water. High levels of brominated species
have been traditionally observed in Barcelona, due to the salt
mine waste upstream in one of the rivers supplying the city
(Ventura and Rivera, 1985). Although the sources of bromide in
other areas are not described in the literature, salt intrusion could
be a source of bromide in coastal Mediterranean areas (e.g.,
Valencia, Murcia).

3.3. MX, haloacetonitriles, haloketones, chloropicrin and chloral

hydrate

MX was detected and quantified in all samples (N ¼36).

Fig. 2. Distribution of trihalomethane, haloacetic species and MX in all areas, (N¼ 217
for trihalomethanes and haloacetic acids, N¼ 36 for MX). Note: CHCl3 ¼ chloroform,
CHCl2Br¼ bromodichloromethane, CHClBr2 ¼dibromochloromethane, CHBr3 ¼ bromo-
form, MCAA¼monocloracetic acid, DCAA¼dichloroacetic acid, TCAA¼ trichloroacetic
acid, MBAA¼ monobromoacetic acid, DBAA¼ dibromoacetic acid, TBAA¼tribromoace-
tic acid, BCAA¼ bromochloroacetic acid, BDCAA¼bromodichloroacetic acid, DBCAA¼
dibromochloroacetic acid.
Median levels were 16.7 (interquartile range 2–32.1) ng/l (Fig. 2
and Appendix Fig. A1). These levels are in accordance to pre-
viously described levels in different settings, where MX occurs in
the order of magnitude of ng/l. A 4.0–79.9 ng/l range was
described for 88 water samples in Massachusetts (Wright et al.,
2002), and concentrations of 5.6 and 10 ng/l have been observed
in Barcelona (Romero et al., 1997). A median value of 20 ng/l and
maximum level of 310 ng/l has been described in the US (Krasner

Table 1
Description and correlations of trihalomethanes, haloacetic acids, haloacetonitriles, haloketones, chloropicrin and chloral hydrate (N ¼ 16). Concentrations are given
in mg/l.

Levels No. of samples Spearman rank correlation coefficients

Min Median Max Total 4QL Dibromoacetonitrile Total acetonitriles

Trihalomethanes N¼ 12 N ¼ 15
Chloroform 0.9 4.4 57.7 15a 15  0.38  0.22
Bromodichloromethane 1.4 4.7 18.9 15a 15  0.37  0.14
Dibromochloromethane 2.2 14.4 34.9 15a 15 0.47 0.87nnn
Bromoform 0.8 41.0 61.6 15a 15 0.58n 0.92nnn
Total trihalomethanes 41.6 70.9 127.5 15a 0.54 0.25

Haloacetic acids N¼ 12 N ¼ 16
Monochloracetic acid o2 o2 o2 16 0 – –
Dichloroacetic acid 0.7 1.8 18.0 16 16  0.27  0. 24
Trichloroacetic acid 1.5 3.1 5.0 16 16 0.57 0.44
Monobromoacetic acid o1 o1 o1 16 0 – –
Dibromoacetic acid 0.9 11.6 14.4 16 16 0.34 0.72nn
Tribromoacetic acid o1 4.0 5.7 16 13 0.66n 0.75nnn
Bromochloroacetic acid 0.8 1.7 20.1 16 16 0.15  0.43
Bromodichloroacetic acid 0.5 1.1 3.9 16 15 0.47  0.22
Dibromochloroacetic acid 0.5 2.3 4.2 16 15  0.02 0.73nn
Total haloacetic acids 18.9 29.0 53.1 16 0.43  0.27
Total dihaloacetic acids 8.5 17.2 41.5 16 0.60 0.76nnn
Total trihaloacetic acids 6.3 10.4 17.5 16 0.52  0.16

Haloacetonitriles
Dichloroacetonitrile o1 o1 o1 16 0
Dibromoacetonitrile o1 4.0 5.5 16 12
Bromochloroacetonitrile o1 1.1 2.3 16 8
Trichloroacetonitrile o1 o1 o1 16 0
Total haloacetonitriles 2.0 5.7 8.3 16
Haloketones
1,1 dichloropropanone o1 o1 o1 16 0
1,1,1 trichloropropanone 0.2 o1 2.2 16 4
Total haloketones 1.0 1.0 2.7 16
Chloropicrin o1 o1 o1 16 0
Chloral hydrate o1 o1 12.1 15a 4

QL. Quantification limit.

n
p value o 0.05.
nn
p valueo 0.01.
nnn
p value o 0.001.
a
One of the vials was broken and sample was lost.
 C.M. Villanueva et al. / Environmental Research 114 (2012) 1–11 5

CHCl3 vs.
MCAA 0.20
DCAA 0.78
TCAA 0.76
-0.52 MBAA
-0.69 DBAA
-0.71 TBAA
BCAA 0.23
BDCAA 0.74
-0.60 DBCAA
MX 0.65
CHCl2Br vs.
MCAA 0.39
DCAA 0.61
TCAA 0.53
-0.41 MBAA
-0.51 DBAA
-0.38 TBAA
BCAA 0.43
BDCAA 0.75
-0.22 DBCAA
MX 0.45
CHClBr2 vs.
-0.10 MCAA Fig. 4. Scatter plot between trihalomethanes and haloacetic acid analogues (same
number of Cl-, Br-substituents).
-0.44 DCAA
-0.47 TCAA
mutagenicity than MX (Richardson et al., 2007). The areas show-
MBAA 0.61
ing the higher MX levels corresponded to areas with higher
DBAA 0.71
proportions of chloroform and chlorinated HAAs, suggesting that
TBAA 0.73
the brominated analogues of MX, that were not measured in this
BCAA 0.26
study, could occur in areas with low MX but high brominated
-0.29 BDCAA
THMs and HAAs.
DBCAA 0.87
In the separate replicated sampling in Barcelona, the concen-
-0.52 MX
tration of haloacetonitriles, haloketones, chloropicrin and chloral
CHBr3 vs.
hydrate was low. Dibromoacetonitrile occurred at highest con-
-0.22 MCAA
centration (median 4.0 mg/l) while dichloroacetonitrile, trichlor-
-0.56 DCAA
oacetonitrile, 1-1-dichloropropanone and chloropicrin were
-0.57 TCAA
below the quantification limit in all samples (Table 1). These
MBAA 0.63
levels are similar to those described by Krasner et al. (2006) in a
DBAA 0.76
survey of 12 treatment plants in the US with different treatment
TBAA 0.79 procedures and disinfectants in waters exhibiting high total
BCAA 0.00 organic carbon and bromide. Similar levels have also been
-0.42 BDCAA
described in 116 water samples from different treatment plants
DBCAA 0.87 supplying Beijing area (China), where median levels of trichloro-,
-0.45 MX dichloro-, bromochloro-, dibromo- and total haloacetonitriles
-1.00 -0.50 0.00 0.50 1.00 were 0.11, 0.59, 0.88, 0.86, and 2.44 mg/l, respectively. Median
levels of dichloro-, trichloropropanone and total haloketones
Fig. 3. Partial Spearman rank correlation coefficients between trihalomethanes, were 0.72, 0.45 and 1.41 mg/l, respectively. Median levels of
haloacetic acids and MX (N ¼ 217). Note: CHCl3 ¼ chloroform, CHCl2Br¼ bromodi- chloral hydrate and chloropicrin were 0.50 and 0.21 mg/l, respec-
chloromethane, CHClBr2 ¼dibromochloromethane, CHBr3 ¼bromoform,
MCAA ¼monocloracetic acid, DCAA¼dichloroacetic acid, TCAA¼ trichloroacetic
tively (Wei et al., 2010). A previous report on chloral hydrate and
acid, MBAA ¼monobromoacetic acid, DBAA¼ dibromoacetic acid, TBAA¼ tribro- dichloroacetonitrile levels (Stevens et al., 1989) showed similar
moacetic acid, BCAA ¼ bromochloroacetic acid, BDCAA ¼ bromodichloroacetic acid, results compared to our study.
DBCAA ¼ dibromochloroacetic acid.
3.4. Correlations
et al., 2006). Despite these generally low levels, MX is a DBP of
concern due to the relatively highly mutagenicity (Kronberg and The partial Spearman rank correlation coefficient between
Christman, 1989). MX is carcinogenic in rats (Komulainen et al., THM4 and HAA9 overall (N ¼217) was 0.62 (p value o0.001).
1997; McDonald and Komulainen, 2005), and has been classified Correlations differed by area and ranged from  0.04 (p-value
as possibly human carcinogen by the international agency for 0.90, N ¼14) in Murcia to 0.99 (p-valueo0.001, N ¼11) in Gran-
research on cancer (IARC, 2004). The World Health Organisation ada. The partial Spearman rank correlation coefficients between
(WHO) has suggested a health-based value of 1.8 mg/l on the basis DBP constituents are shown in Table 1, Figs. 3 and 4 and
of the increase in cholangiomas and cholangiocarcinomas in Appendix Fig. A2. Partial correlations tend to be higher among
female rats (WHO, 2011). However, the WHO considered that analogues in terms of number and types of substituents, espe-
MX in drinking-water occurs at concentrations well below those cially bromodichloroacetic acid with bromodichloromethane (rs
of health concern, and no guideline value has been proposed. 0.75), and dibromochloroacetic acid with dibromochloromethane
Some brominated MX analogues show higher bacterial (rs 0.87) (Fig. 4). The correlation between chloroform and MX was
6 C.M. Villanueva et al. / Environmental Research 114 (2012) 1–11

Table 2 is usually sparse. As a general rule, brominated DBPs tend to be of

Principal component analysis of trihalomethanes and haloacetic acids more concern than chlorinated DBPs and use of total organic
(N ¼217).
bromine (TOBr) and total organic chlorine (TOCl) has been
Component 1 2 3 proposed (Zhang et al., 2000).
The use of principal component analysis has been proposed
Explained variance 0.52 0.27 0.09 as en exploratory method for data analysis suitable for complex
Component–variable correlation mixtures such as disinfection by-products (Feder et al., 2009).
Chloroform  0.24 0.31 0.20
Bromodichloromethane 0.06 0.46  0.32
The advantage of PCA is that generates a set of independent
Dibromochloromethane 0.34 0.19  0.03 variables (components) that are linear combinations of the
Bromoform 0.33  0.04 0.36 original variables. In our study, we show that the THM-HAA
Monocloracetic acid  0.21 0.20 0.39 mixture can be simplified in three main components explaining
Dichloroacetic acid  0.25 0.30 0.37
more than 80% of variance (Table 2). The first component
Trichloroacetic acid  0.26 0.27 0.36
Monobromoacetic acid 0.36 0.08 0.17 explained 52% of variance and correlated with DBP species
Dibromoacetic acid 0.37 0.01 0.24 dominated by bromine (dibromochloromethane, bromoforom,
Tribromoacetic acid 0.35  0.03 0.35 mono-, di- and tribromoacetic acids and dibromochloroacetic
Bromochloroacetic acid 0.19 0.42  0.15 acid). This component correlated negatively with the chlorinated
Bromodichloroacetic acid 0.03 0.47  0.27
Dibromochloroacetic acid 0.35 0.19  0.01
species (chloroform, mono-, di- and tribromoacetic acid).
The second component explained 27% of variance and showed
highest correlations with chloroform, dichlorobromomethane,
bromochloroacetic acid and bromodichloroacetic acid. The
third component explained 9% of variance and correlated both
moderate (rs 0.65), and negative with dibromochloromethane and with brominated species (bromoform, tribromoacetic acid),
bromoform. These trends are consistent with a previous study and chlorinated haloacetic acids (mono-, di-, trichloroacetic acid).
that explored correlations between MX and THMs (Suzuki and An alternative PCA using residuals from linear regressions of
Nakanishi, 1995). The authors observed a strong positive correla- chemicals adjusting for area was conducted, to control for
tion between CHCl3 and THM ratio vs. (MX/MXþbrominated MX). the effect of area. Results confirmed results shown in Table 2.
Poor or negative correlations were observed between MX/ Area-stratified PCA were also conducted, showing slightly
(MXþbrominated MX species) ratio and the other three trihalo- different results in some areas compared to the overall analysis
methanes/THM ratios. Since correlations vary between areas, our (not shown). However, most of the variance was explained by two
results confirm that universal correlations between DBPs do not components in all areas, one dominated by bromine- and the
exist, and that THM4 and HAA9 level involve different mixtures in other by chlorine- substituted DBPs. The ‘‘brominated’’ compo-
different water zones. Correlations within a water zone may vary nent explained the largest part of variability overall, although it
over time but our study was not designed to evaluate seasonal differed by area. In areas where bromine-based DBPs occurred at
variations. low concentrations, the chlorinated component explained most of
The THM- and HAA- correlation with haloacetonitriles, haloke- the variability.
tones, chloropicrin, chloral hydrate was hampered by the concen-
trations below quantification limits (Table 1). We limited our
calculations to pairs with at least 12 samples above the quantifica- 4. Conclusions
tion limit. Correlation between these specific DBPs has been poorly
explored in previous studies. Wei et al. (2010) measured 4 THMs, In summary, total and specific THMs and HAAs have limited
5 HAAs (MCAA, MBAA, DCAA, TCAA, DBAA), 4 haloacetonitriles value as predictors of other DBPs on a generalised basis. Reason-
(TCAN, DCAN, BCAN, DBAN) and 2 haloketones (DCP, TCP) in 116 able predictions should be limited to specific chemically related
samples of drinking water in Beijing (China). Correlations ranged DBPs, based on sufficient number of measurements and calcu-
from 0.22 (between total haloketones and chloral hydrate) lated specifically by water zone and probably time period (sea-
to 0.87 ((between THMs and total HANs). However, results are sons). In addition, correlations between single chemicals ignore
not completely comparable to ours since Wei et al. (2010) reported the complex nature of the mixture. Principal component analysis
Pearson correlation coefficients. Correlations among total trihalo- provides an interesting complementary tool to further explore the
methanes, 6 haloacetic acids (combined), mutagenicity (estimated relationships between different compounds.
by the Ames test), total organic carbon and total organic halogen
were evaluated by Schenck et al. (2009). In 18 finished water
samples from five different treatment plants, the highest correla- Acknowledgments
tion occurred between mutagenicity and the total organic halogen
content (r¼0.95). We thank the contribution of Pablo Fernández-Navarro, Javier
Garcı́a-Pérez, Esther Garcı́a-Esquinas (Cancer and Environmental
Epidemiology Unit, National Center for Epidemiology, Carlos III
Institute of Health, Madrid), Juan Alguacil, (Huelva University),
3.5. Mixtures Margarita Palau (Health Ministry of Spain), Remedios Pérez, José
Luis Rodrigo (Public Health Institute of Navarra) and the other
The occurrence of DBPs in mixtures of varying composition is members of the MCC-Spain study.
one main methodological challenge in the evaluation of health
effects, without a straightforward solution. This is particularly
problematic when evaluating chronic effects such as cancer, Appendix A
where exposure needs to be evaluated for decades and the
probability to misclassify exposure is higher since historical data See Tables A1–A3 and Figs. A1 and A2.
 C.M. Villanueva et al. / Environmental Research 114 (2012) 1–11 7

Table A1 Table A1 (continued )

Distribution of sampling locations and dates in the study areas. All samples were
collected in 2010. Study municipalities No. of samples Sampling day-month

Study municipalities No. of samples Sampling day-month Ayamonte 2 13-Jun

Barcelona
Barcelonaa
Barcelona
L’Hospitalet del Llobregata
L’Hospitalet del Llobregat
Badalonaa
Badalona
Cornella de Llobregat
Santa Coloma
El Prat del Llobregat
Sant Adria del Besós
Esplugues de Llobregat
Molins de Rei
Sant Feliu del Llobregat
Sant Boi del Llobregat
Sant Joan Despı́
Sant Vicenc- dels Horts
Valverde del camino 2 13-Jun
52
Cantabria 10 7,8-Jun
12 15-Mara
Santander 4 7-Jun
12 22-Apr
El Astillero 1 7-Jun
2 15-Mara
Camargo 1 7-Jun
5 29-Apr
Pontejos 1 7-Jun
2 15-Mara
Santiago de Heras 1 7-Jun
5 22-Apr
Somo 1 7-Jun
5 29-Apr
Lliencres 1 7-Jun
4 22-Apr
3 22-Apr Valencia 10
3 22-Apr Valencia 10 21-Jun
3 29-Apr
Total main sampling 217
3 29-Apr
Replicated samplesa 16
3 29-Apr
TOTAL 233
2 29-Apr
2 29-Apr a
Replicated sampling.
2 29-Apr

Madrid 34
Madrid 21 19,20,26, 27-Apr
Alcalá de Henares 3 26-Apr
Alcobendas 2 20-Apr
Colmenar viejo 2 20-Apr
San Sebastián de los Reyes 2 20-Apr
Alcorcón 2 26-Apr
Leganés 2 26-Apr

Navarra 28
Pamplona 12 17-May
Barañain 3 17-May
Burlada 3 24-May
Cizur 2 17-May
Alsasua 2 24-May
Aoiz 2 24-May
Olite 2 24-May
Tafalla 2 24-May

Asturias 18
Gijón 11 31-May
Villaviciosa 4 31-May
Carreño 3 31-May

León 16
León 6 10-May
Santa Marina del Rey 2 10-May
Astorga 1 10-May
La Bañeza 1 10-May
La Pola de Gordon 1 10-May
La Robla 1 10-May
Sahagún 1 10-May
San Andrés del Rabanedo 1 10-May
Valdefresno 1 10-May
Villaquejida 1 10-May

Gipuzkoa 14
Donostia-San Sebastián 8 7 Jun, 23-29 Nov
Lasarte 2 29 Nov
Pasaia 2 26 Oct
Errenteria 1 27 Oct
Irun 1 26 Oct

Murcia 14
Murcia 6 3-May
Archena 4 3-May
Molina de Segura 4 3-May

Granada 11
Granada 5 28-Jun
Armilla 1 28-Jun
La Zubia 1 28-Jun
Monachil 1 28-Jun
Moraleda de Zafayona 1 28-Jun
Santa Fé 1 28-Jun
Órgiva 1 28-Jun

Huelva 10
Huelva 6 13,14-Jun
 8
Table A2
Concentration of trihalomethanes (mg/l).

C.M. Villanueva et al. / Environmental Research 114 (2012) 1–11

Asturias Barcelona Cantabria Gipuzkoa Granada Huelva León Madrid Murcia Navarra Valencia All areas

CHCl3 Median 10.1 0.6 3.9 2.2 5.5 25.0 17.1 18.9 10.4 17.3 1.3 10.3
Range o 0.4–34.1 o 0.4–44.7 0.6–10.1 0.8–12.7 o0.4–13.6 14.8–69.8 o 0.4–40.3 9.8–38.7 1.0–34.6 9.5–40.0 1.0–1.6 o 0.4–69.8
No. of samples4 QL 16 (89%) 36 (69%) 10 (100%) 14 (100%) 9 (82%) 10 (100%) 10 (63%) 34 (100%) 14 (100%) 28 (100%) 10 (100%) 191 (88%)
CHCl2Br Median 4.0 2.4 5.2 4.3 2.1 14.4 1.5 2.3 13.1 3.9 4.8 3.2
Range o 0.4–6.3 o 0.4–12.5 0.7–8.3 2.0–8.3 o0.4–8.0 8.8–21.2 o 0.4–4.4 1.4–3.5 4.4–26.2 2.4–10.8 3.5–6.1 o 0.4–26.2
No. of samples4 QL 16 (89%) 46 (88%) 10 (100%) 14 (100%) 9 (82%) 10 (100%) 10 (63%) 34 (100%) 14 (100%) 28 (100%) 10 (100%) 201 (93%)
CHClBr2 Median 1.9 9.1 8.4 4.2 0.5 6.3 o 0.4 o0.4 20.2 0.6 14.8 2.2
Range 0.4–7.8 o 0.4–24.9 2.6–17.0 2.9–12.4 o0.4–2.5 1.1–11.2 o 0.4–1.4 o0.4–0.7 4.9–38.8 o 0.4–5.2 10.8–18.0 o 0.4–38.8
No. of samples4 QL 18 (100%) 49 (94%) 10 (100%) 14 (100%) 7 (64%) 10 (100%) 3 (19%) 3 (9%) 14 (100%) 26 (93%) 9 (90%) 164 (76%)
CHBr3 Median o 0.4 21.1 4.2 0.9 o0.4 o 0.4 o 0.4 o0.4 7.8 o 0.4 13.4 o 0.4
Range o 0.4–4.2 o 0.4–64.1 1.5–20.4 0.5–8.2 o0.4– o 0.4 o 0.4–1.7 o 0.4–3.1 o0.4– o 0.4 1.3–37.2 o 0.4–1.4 10.2–17.0 o 0.4–64.1
No. of samples4 QL 7 (39%) 44 (85%) 10 (100%) 14 (100%) 0 (0%) 2 (20%) 1 (6%) 0 (0%) 14 (100%) 3 (11%) 10 (100%) 105 (48%)
THM4 Median 16.9 44.8 23.1 12.4 8.2 48.6 19.5 21.8 53.9 22.8 34.9 26.4
Range 1.7–39.2 2.8–90.9 6.1–46.1 6.8–27 0.8–24.3 31.4–83.5 0.8–43.0 11.7–42.3 24.2–98.1 12.2–46.9 25.5–42.7 0.8–98.1
No. of samples 18 52 10 14 11 10 16 34 14 28 10 217

CHCl3 ¼ chloroform, CHCl2Br¼ bromodicloromethane, CHClBr2 ¼ dibromochloromethane, CHBr3 ¼ bromoform.

No. of samples 4QL: number of samples above the quantification limit.
Guideline values: European and Spanish legislation: 100 mg/l for THM4 [98/83/CE Directive; Real Decreto 140/2003]; USEPA: 80 mg/l for THM4 [Disinfectant and Disinfection Byproduct Rule, EPA 816-F-02-021].
Table A3
Concentration of haloacetic acids (mg/l).

Asturias Barcelona Cantabria Gipuzkoa Granada Huelva León Madrid Murcia Navarra Valencia All areas

MCAA Median, mg/l o0.5 o 0.5 o 0.5 o 0.5 o 0.5 1.0 1.0 1.1 o 0.5 1.5 o0.5 o0.5
Range, mg/l o0.5–1.0 o 0.5–1.0 o 0.5–0.9 o 0.5–1.1 o 0.5–1.1 0.7–2.3 o 0.5–1.5 o 0.5–2.3 o 0.5–1.8 o 0.5–2.8 o0.5–0.5 o0.5–2.8
No. of samples 4QL 5 (28%) 12 (23%) 2 (20%) 5 (36%) 1 (9%) 10 (100%) 9 (56%) 32 (94%) 6 (43%) 25 (89%) 1 (10%) 108 (50%)
DCAA Median 6.6 0.8 2.0 1.5 4.0 11.4 13.0 14.3 11.4 10.3 1.4 6.8
Range o0.1–16.1 o 0.1–10.8 0.4–9.3 0.7–7.8 0.2–9.3 9.0–37.3 o 0.1–38.0 8.4–21.9 2.1–18.2 2.9–17.3 0.8–1.7 o0.1–38.0
Samples 4QL 17 (94%) 48 (92%) 10 (100%) 14 (100%) 11 (100%) 10 (100%) 14 (88%) 34 (100%) 14 (100%) 28 (100%) 10 (100%) 210 (97%)

C.M. Villanueva et al. / Environmental Research 114 (2012) 1–11

TCAA Median 8.0 0.6 1.2 0.9 3.4 9.9 18.4 13.1 5.6 9.5 0.5 6.2
Range 0.1–25.9 o 0.05–17.1 0.4–8.3 0.3–6.8 0.1–6.9 6.1–36.7 o 0.05–34.7 5.4–23.2 0.6–20.8 5.5–21.6 0.3–0.7 o0.05–36.7
No. of samples 4QL 18 (100%) 50 (96%) 10 (100%) 14 (100%) 11 (100%) 10 (100%) 14 (88%) 34 (100%) 14 (100%) 28 (100%) 10 (100%) 213 (98%)
MBAA Median 0.1 0.8 0.3 0.2 o 0.1 0.3 o 0.1 o 0.1 0.9 o 0.1 0.9 0.2
Range o0.1–1.0 o 0.1–1.6 0.2–0.4 o 0.1–0.4 o 0.1–0.3 0.2–0.4 o 0.1–0.2 o 0.1–o 0.1 0.3–1.8 o 0.1–0.3 0.7–1.1 o0.1–1.8
No. of samples 4QL 12 (67%) 50 (96%) 10 (100%) 11 (79%) 1 (9%) 10 (100%) 2 (13%) 0 (0%) 14 (100%) 3 (11%) 10 (100%) 123 (57%)
DBAA Median 1.0 1.6 0.4 1.2 0.3 0.4 0.2 0.1 1.8 0.3 1.1 0.6
Range 0.1–3.4 0.3–13.7 1.6–4.5 0.3–4.7 0.1–0.7 0.3–2.0 o 0.05–2.1 o 0.05–0.2 1.2–12.3 0.1–1.6 4.9–8.0 o0.05–13.7
No. of samples 4QL 18 (100%) 52 (100%) 10 (100%) 14 (100%) 11 (100%) 10 (100%) 10 (63%) 17 (50%) 14 (100%) 28 (100%) 10 (100%) 194 (89%)
TBAA Median o0.3 5.0 1.1 o 0.3 o 0.3 o 0.3 o 0.3 o 0.3 2.4 o 0.3 3.9 o0.3
Range o0.3–1.6 o 0.3 10.8 0.6 4.7 o 0.3 1.1 o 0.3–o 0.3 o 0.3– o0.3 o 0.3–1.5 o 0.3–o 0.3 o 0.3–9.3 o 0.3–0.3 2.5–4.8 o0.3–10.8
No. of samples 4QL 6 (33%) 42 (81%) 10 (100%) 2 (14%) 0 (0%) 1 (10%) 1 (6%) 0 (0%) 12 (86%) 1 (4%) 10 (100%) 85 (39%)
BCAA Median 1.5 2.1 2.4 1.6 0.9 4.3 0.6 1.0 6.1 1.6 3.3 1.6
Range 0.2–2.1 o 0.1–5.8 0.8–3.9 0.4–4.3 0.2–3.6 3.8–5.3 o 0.1–3.7 0.5–1.6 4.2–9.8 1.1–4.2 2.3–3.9 o0.1–9.8
No. of samples 4QL 18 (100%) 49 (94%) 10 (100%) 14 (100%) 11 (100%) 10 (100%) 12 (75%) 34 (100%) 14 (100%) 28 (100%) 10 (100%) 210 (97%)
BDCAA Median 1.4 0.9 2.3 1.5 1.0 7.8 1.5 0.7 6.2 1.5 2.0 1.3
Range o0.1–2.4 o 0.1–4.6 0.4–3.3 0.7–4.6 o 0.1–3.0 3.6–9.0 o 0.1–5.0 0.4–1.3 2.6–9.1 0.9–3.1 1.5–2.5 o0.1–9.1
No. of samples 4QL 17 (94%) 47 (90%) 10 (100%) 14 (100%) 9 (82%) 10 (100%) 10 (63%) 34 (100%) 14 (100%) 28 (100%) 10 (100%) 203 (94%)
DBCAA Median 0.4 2.4 2.3 1.0 0.1 2.5 o 0.1 o 0.1 5.6 0.1 4.0 0.5
Range 0.2–2.2 o 0.1–5.6 0.8–5.2 0.5–2.2 o 0.1–0.5 0.4–2.8 o 0.1–0.6 o 0.1–0.2 1.5–8.3 o 0.1–0.8 2.8–4.9 o0.1–8.3
No. of samples 4QL 18 (100%) 49 (94%) 10 (100%) 14 (100%) 7 (64%) 10 (100%) 2 (13%) 34 (100%) 14 (100%) 16 (57%) 10 (100%) 153 (71%)
DXAA Median 8.7 11.2 7.8 4.9 5.0 17.3 13.7 15.5 22.7 11.9 11.2 11.8
Range 0.9–18.7 1.1–18.4 3.4–15.4 1.5–13.2 0.6–13.6 15.1–41.4 0.3–41.9 8.9–23.1 14.8–33.5 4.4–22.1 8.0–13.6 0.3–41.9
TXAA Median 10.1 11.4 7.7 4.1 4.7 21.0 20.3 14.2 24.8 11.0 10.3 13.3
Range 0.7–28.4 0.5–22.3 2.6–13.7 1.9–13.6 0.3–10.6 11.5–42.9 0.3–40.1 6.1–24.2 14.4–30.9 6.6–24.2 7.1–12.9 0.3–42.9
HAA9 Median 19.4 25.5 16.3 8.8 10.0 40.8 35.0 29.9 50.7 27.6 22.5 26.4
Range 1.9–42.5 1.9–38.5 6.4–30.4 4.7–27.7 1.2–25.4 27.9–86.9 0.9–83.5 18.3–48.0 30.6–66.7 15.–41.3 17.0–27.3 0.9–86.9
No. of samples 18 52 10 14 11 10 16 34 14 28 10 217

MCAA ¼monocloracetic acid, DCAA¼ dichloroacetic acid, TCAA¼trichloroacetic acid, MBAA ¼ monobromoacetic acid, DBAA ¼dibromoacetic acid, TBAA¼tribromoacetic acid, BCAA ¼ bromochloroacetic acid, BDCAA ¼ bromodi-
chloroacetic acid, DBCAA ¼dibromochloroacetic acid, DXAA¼ total dihaloacetic acids (S DCAA, DBAA, BCAA), TXAA¼total trihaloacetic acid (S TCAA, TBAAm DBCAAm BDCAA).
No. of samples4 QL: number of samples above the quantification limit.
Guideline values: US EPA: 60 mg/l for HAA5 [e.g., MCAA, DCAA, TCAA, MBAA, DBAA. Disinfectant and Disinfection Byproduct Rule, EPA 816-F-02-021]. Not regulated in the EU.

9
10 C.M. Villanueva et al. / Environmental Research 114 (2012) 1–11

50.0

40.0

30.0

20.0

10.0

0.0
Asturias Gipuzkoa Madrid Navarra
Barcelona León Murcia

Fig. A1. Levels of -chloro-4-(dichloromethyl)-5-hydroxy-2(5H)-furanone (MX), ng/l (N ¼16).

Fig. A2. Scatter plot of trihalomethanes and haloacetic acids with the same number of Cl-, Br-substituents, by area.
 C.M. Villanueva et al. / Environmental Research 114 (2012) 1–11
References
Aggazzotti, G., Righi, E., Fantuzzi, G., Biasotti, B., Ravera, G., Kanitz, S., Barbone, F.,
Sansebastiano, G., Battaglia, M.A., Leoni, V., Fabiani, L., Triassi, M., Sciacca, S.,
Collaborative Group for the Study of Chlorinated Drinking Waters and
Pregnancy, 2004. Chlorination by-products (CBPs) in drinking water and
adverse pregnancy outcomes in Italy. J. Water Health 2, 233–247.
Choi, J., Valentine, R.L., 2002. Formation of N-nitrosodimethylamine (NDMA) from
reaction of monochloramine: a new disinfection by-product. Water Res. 36,
817–824.
Council Directive 98/83/EC. Of 3 November 1998 on the quality of water intended
for human consumption. Official Journal of the European Communities , 1998.
Eaton, A.D., Franson, M.A.H., AWWA, W.E.F., 2005. Standard method 6232B,
trihalometanes and chlorinated organic solvents: liquid–liquid extraction gas
chromatographic method. In: Eaton, A.D., Franson, M.A.H., APHA (Eds.),
Standard Methods for the Examination of Water and Wastewater, 21st edn.,
pp. 6-41-4-46.
Feder, P.I., Ma, Z.J., Bull, R.J., Teuschler, L.K., Schenck, K.M., Simmons, J.E., Rice, G.,
2009. Evaluating sufficient similarity for disinfection by-product (DBP) mix-
tures: multivariate statistical procedures. J. Toxicol. Environ. Health A 72,
468–481.
Hinckley, A.F., Bachand, A.M., Reif, J.S., 2005. Late pregnancy exposures to
disinfection by-products and growth-related birth outcomes. Environ. Health
Perspect. 113, 1808–1813.
Hoffman, C.S., Mendola, P., Savitz, D.A., Herring, A.H., Loomis, D., Hartmann, K.E.,
Singer, P.C., Weinberg, H.S., Olshan, A.F., 2008. Drinking water disinfection by-
product exposure and fetal growth. Epidemiology 19, 729–737.
IARC, 2004. Some drinking water disinfectants and contaminants, including
arsenic. IARC Monographs on the Evaluation of Carcinogenic Risks to Humans,
Vol. 84. IARC Scientific Publications.
Komulainen, H., Kosma, V.M., Vaittinen, S.L., Vartiainen, T., Kaliste-Korhonen, E.,
Lotjonen, S., Tuominen, R.K., Tuomisto, J., 1997. Carcinogenicity of the drinking
water mutagen 3-chloro-4-(dichloromethyl)-5-hydroxy-2(5H)-furanone in
the rat. J. Nat. Cancer Inst. 89, 848–856.
Krasner, S.W., Weinberg, H.S., Richardson, S.D., Pastor, S.J., Chinn, R., Sclimenti, M.J.,
Onstad, G.D., Thruston Jr, A.D., 2006. Occurrence of a new generation of
disinfection byproducts. Environ. Sci. Technol. 40, 7175–7185.
Kronberg, L., Christman, R.F., 1989. Chemistry of mutagenic by-products of water
chlorination. Sci. Total Environ., 219–230.
McDonald, T.A., Komulainen, H., 2005. Carcinogenicity of the chlorination disin-
fection by-product MX. J. Environ. Sci. Health., Part C Environ. Carcinog.
Ecotoxicol. Rev. 23, 163–214.
Mitch, W.A., Sedlak, D.L., 2002. Formation of N-nitrosodimethylamine (NDMA)
from dimethylamine during chlorination. Environ. Sci. Technol. 36, 588–595.
Plewa, M.J., Wagner, E.D., Muellner, M.G., Hsu, K.M., Richardson, S.D., 2008.
Comparative mammalian cell toxicity of N-DBPs and C-DBPs. In: Karanfil, T.,
Krasner, S., Westerhoff, P., Xie, Y. (Eds.), In Occurrence, Formation, Health
Effects and Control of Disinfection by-Products in Drinking Water. American
Chemical Society, Washington, D.C., pp. 36–50.
Rantakokko, P., Yritys, M., Vartiainen, T., 2004. Matrix effects in the gas chromato-
graphic-mass spectrometric determination of brominated analogues of 3-
chloro-4-(dichloromethyl)-5-hydroxy-2(5H)-furanone. J. Chromatogr. A 1028,
179–188.
Real Decreto num 1138/1990. 14 Septiembre 1990, AGUAS. Reglamentación
técnico-Sanitaria para el abastecimiento y control de calidad de las potables
de consumo público. BOE 20 septiembre 1990 (num.226). 1990.
Richardson, S.D., Fasano, F., Ellington, J.J., Crumley, F.G., Buettner, K.M., Evans, J.J.,
Blount, B.C., Silva, L.K., Waite, T.J., Luther, G.W., McKague, A.B., Miltner, R.J.,
Wagner, E.D., Plewa, M.J., 2008. Occurrence and mammalian cell toxicity of
iodinated disinfection by products in drinking water. Environ. Sci. Technol. 42,
8330–8338.
Richardson, S.D., Plewa, M.J., Wagner, E.D., Schoeny, R., DeMarini, D.M., 2007.
Occurrence, genotoxicity, and carcinogenicity of regulated and emerging
11
disinfection by-products in drinking water: a review and roadmap for
research. Mutat. Res. 636, 178–242.
Romero, J., Ventura, F., Caixac, J., Rivera, J., Guerrero, R., 1997. Identification and
quantification of the mutagenic compound 3-chloro-4-(dichlormethyl)-2(5H)-
furanone (MX) in chlorine-treated water. Bull. Environ. Contam. Toxicol.,
715–722.
Schenck, K.M., Sivaganesan, M., Rice, G.E., 2009. Correlations of water quality
parameters with mutagenicity of chlorinated drinking water samples.
J. Toxicol. Environ. Health A 72, 461–467.
Singer, P.C., Chang, S.D., 1989. Correlations between trihalomethanes and total
organic halides formed during water treatment. J. Am. Water Works Assn. 81,
61–65.
Stevens, A.A., Moore, L.A., Miltner, R.J., 1989. Formation and control of non-
trihalomethane disinfection by-products. J. Am. Water Works Assn. 81, 61–65.
Suzuki, N., Nakanishi, J., 1995. Brominated analogues of MX (3-chloro-4-(dichlor-
omethyl)-5-hydroxy-2(5H)-furanone) in chlorinated drinking water. Chemo-
sphere 30, 1557–1564.
Tardiff, R.G., Carson, M.L., Ginevan, M.E., 2006. Updated weight of evidence for an
association between adverse reproductive and developmental effects and
exposure to disinfection by-products. Regul. Toxicol. Pharmacol. 45, 185–205.
Trabaris M., Laskin J.D., Weisel C.P.: 2011. Percutaneous absorption of haloaceto-
nitriles and chloral hydrate and simulated human exposures. J. Appl. Toxicol.
Tung, H.H., Unz, R.F., Xie, Y.F., 2006. HAA removal by GAC adsorption. J. Am. Water
Works Assn. 98, 107–112.
USEPA. Method 551.1: Determination of Chlorination Disinfection Byproducts,
Chlorinated Solvents, and Halogenated Pesticides/Herbicides in Drinking
Water by Liquid–Liquid Extraction and Gas Chromatography with Electron-
Capture Detection, Revision 1.0. National Exposure Research Laboratory,
USEPA Office of Research and Development, Cincinnati, OH. 1995.
USEPA. Method 552.3, 2003. Determination of haloacetic acids and dalapon in
drinking water by liquid–liquid microextraction, derivatization, and gas
chromatography with electron capture detection. EPA 815-B-03-002. Cincin-
nati, OH.
Ventura, F., Rivera, J., 1985. Factors influencing the high content of brominated
trihalomethanes in Barcelona’s water supply (Spain). Bull. Environ. Contam.
Toxicol. 35, 73–81.
Villanueva, C.M., Cantor, K.P., Cordier, S., Jaakkola, J.J., King, W.D., Lynch, C.F., Porru, S.,
Kogevinas, M., 2004. Disinfection byproducts and bladder cancer: a pooled
analysis. Epidemiology 15, 357–367.
Wei, J., Ye, B., Wang, W., Yang, L., Tao, J., Hang, Z., 2010. Spatial and temporal
evaluations of disinfection by-products in drinking water distribution systems
in Beijing, China. Sci. Total Environ. 408, 4600–4606.
WHO, 2011. Guidelines for Drinking-water Quality. Fourth Edition, online at:
http://whqlibdoc.who.int/publications/2011/9789241548151_eng.pdf
(accessed 1st September 2011).
Wright, J.M., Schwartz, J., Vartiainen, T., Maki-Paakkanen, J., Altshul, L., Harrington, J.J.,
Dockery, D.W., 2002. 3-Chloro-4-(dichloromethyl)-5-hydroxy-2(5H)-furanone
(MX) and mutagenic activity in Massachusetts drinking water. Environ. Health
Perspect. 110, 157–164.
Xu, X., Mariano, T.M., Laskin, J.D., Weisel, C.P., 2002. Percutaneous absorption of
trihalomethanes, haloacetic acids, and haloketones. Toxicol. Appl. Pharmacol.
184, 19–26.
Xu, X., Weisel, C.P., 2003. Inhalation exposure to haloacetic acids and haloketones
during showering. Environ. Sci. Technol. 37, 569–576.
Xu, X., Weisel, C.P., 2005. Dermal uptake of chloroform and haloketones during
bathing. J. Exposure Anal. Environ. Epidemiol. 15, 289–296.
Zhang, X., Echigo, S., Minear, R.A., Plewa, M.J., 2000. Characterization and
comparison of disinfection by-products of four major disinfectants. In: Barrett,
S.E., Krasner, S.W., Amy, G.L. (Eds.), Natural Organic Matter and Disinfection
By-Products. Characterization and Control in Drinking Water American Che-
mical Society, pp. 299–314.