D

Va

b

a

ARRA

KITTAMQ

I

dmn(2i(

tiat2

mo

1d

International Journal of Hygiene and Environmental Health 215 (2011) 76– 85

Contents lists available at ScienceDirect

International Journal of Hygiene andEnvironmental Health

jou rn al h omepage: www.elsev ier .de / i jheh

eterminants of chlorination by-products in indoor swimming pools

incent Bessonneaua, Mickaël Derbezb, Michel Clémenta, Olivier Thomasa,∗

Environmental and Health Research Laboratory (LERES), French School of Public Health (EHESP), Avenue du Professeur Léon Bernard, 34043 Rennes, FranceUniversity Paris Est, Scientific and Technical Centre of Building (CSTB), French Indoor Air Quality Observatory (OQAI), Marne la Vallée, France

r t i c l e i n f o

rticle history:eceived 19 April 2011eceived in revised form 20 July 2011ccepted 25 July 2011

eywords:ndoor swimming poolrihalomethanerichloramineir samplingultivariate analysisuantile regression

a b s t r a c t

Water and air quality of fifteen swimming pools using chlorine disinfection was measured during eightsampling campaigns in Rennes, France. Concentrations of trihalomethanes (THMs) and trichloramine(NCl3) in air as well as concentrations of parameters, including total organic carbon (TOC), chlorine (free,combined and total), pH, Kjeldhal nitrogen (KN), chloride ions and THMs in water and air temperature,were measured. Water and air samples were collected during 3 h in the morning and afternoon, duringhigh and low attendance days, in summer and winter. Data analysis was perfomed by multivariate ordi-nary least square (OLS) and quantile (QR) regressions, from both data quality and other parameters suchas the ratio between the number of visitors and the volume of water available V/W, the proportion ofair re-circulated and the number of swimmers during sampling. The geometric means (GM ± 95% CIs)of the sum of THMs (TTHM) and NCl3 in air were 74.9 ± 6.7 �g/m3 and 190 ± 10 �g/m3, respectively. Asexpected, concentrations of TTHM in air were correlated to concentrations of TTHM in water. Concentra-tions of TTHM and NCl3 in air measured in winter were significantly higher than in summer, depicting thatlow air exchange rate in winter lead to accumulation of pollutants. NCl3 (especially at higher quantiles)

and TTHM concentrations in air were associated to the number of swimmers. NCl3 was also linked to airtemperature and pH. Moreover TTHM and NCl3 in air seemed to be affected by additional determinantsincluding the volume of hall and the ratio V/W. In a process of reducing exposure, the OLS regressionprovides useful information on key parameters influencing concentrations of THMs and NCl3 in indoorswimming pools. In addition, QR allows identifying factors associated to elevated levels of contamination.

ntroduction

Recreational waters, such as swimming pools waters requireisinfection for inactivation of microbial pathogens. Chlorine is theost used disinfectant. It is also known to react with organic and

itrogen compounds, forming numerous chlorination by-productsCBPs) (Richardson et al., 2010; Weaver et al., 2009; Zwiener et al.,007;). The primary source of organic and nitrogen compounds

nclude hair, saliva, sweat, urine and lotion, coming from bathersKim et al., 2002; LaKind et al., 2010).

Among CBPs, some compounds are highly volatiles and can beransferred from water to air. Trihalomethanes (THMs), includ-ng chloroform, dichlorobromomethane, dibromochloromethane

nd bromoform, and trichloramine (NCl3) are the most impor-ant volatiles CBPs found in swimming pool (Li and Blatchley,007; Weaver et al., 2009). Several studies have found an asso-

∗ Corresponding author. Tel.: +33 0 2 99 02 29 20; fax: +33 0 2 99 02 29 29.E-mail addresses: vincent.bessonneau@ehesp.fr (V. Bessonneau),

ickael.derbez@cstb.fr (M. Derbez), michel.clement@ehesp.fr (M. Clément),livier.thomas@ehesp.fr (O. Thomas).

438-4639/$ – see front matter © 2011 Elsevier GmbH. All rights reserved.oi:10.1016/j.ijheh.2011.07.009

© 2011 Elsevier GmbH. All rights reserved.

ciation between exposure to elevated THMs levels and adversehealth effects including irritations (eye, skin, nose and throat)(Fantuzzi et al., 2010; Font-Ribera et al., 2010), certain typeof cancers (Kogevinas et al., 2010; Morris et al., 1992; Morris,1995; Villanueva et al., 2007) and adverse reproductive effects(Hinckley et al., 2005; Nieuwenhuijsen et al., 2000). Addition-ally, a large part of the research regarding the contamination ofindoor swimming pools air has focused on the exposure to NCl3.NCl3 has been associated with asthma both in swimming poolworkers and in pool visitors (Bernard et al., 2006, 2009; Jacobset al., 2007; Lévesque et al., 2006; Massin et al., 1998; Thickettet al., 2002; Voisin et al., 2010; Weisel et al., 2008), and lungepithelium permeability in young children (Bernard et al., 2003,2007).

While swimming pool workers are exposed to THMs and NCl3mainly via inhalation route, the exposure of bathers by inhalation,dermal contact and ingestion is a reality (Erdinger et al., 2004).

Numerous studies have measured the concentrations of THMs

and NCl3 both in environmental samples of indoor swimming pools(water and air) and in the biological media (urine, blood or exhaledair), in order to assess the exposure of swimming pool workers andbathers. Table 1 describes the kind of media and the main outcomes

V.

Bessonneau et

al. /

International Journal

of H

ygiene and

Environmental

Health

215 (2011) 76– 8577

Table 1Description of previous study related to THMs and NCl3 measurements.

Chlorination by-products Experimental conditions Main outcomes Reference

CHCl3 Swimming pools n = 3 CHCl3 in blood was correlated to levels in water and air, number of swimmers, timeand intensity of swimming activity

Aggazzotti et al., 1990

Media: water, ambient air and bloodSubjects: 127 exposed and 40 non-exposed

CHCl3 Swimming pools n = 2 Significant correlation between CHCl3 in ambient and alveolar air Aggazzotti et al., 1993Media: water, environmental and exhaled air Influence of the swimming pattern on breath air levelsSubjects: 163 exposed and 77 non-exposed

CHCl3 Swimming pools n = 12 CHCl3 in ambient air was correlated to the number of swimmers and CHCl3 in water Aggazzotti et al., 1995Media: water, ambient and alveolar air, and blood Competitive swimmers were heavily exposedSubjects: 127 (102 competitive swimmers, 16non-competitive swimmers and 9 visitors)

Kinetic of CHCl3 elimination

Total trihalomethane Swimming pool n = 1 Detection of CHCl3, CHCl2Br and CHClBr2 in alveolar air Aggazzotti et al., 1998(CHCl3, CHCl2Br, CHClBr2, CHBr3) Media: water, blood, environmental and exhaled air Different elimination rate for CHCl3, CHCl2Br and CHClBr2

Subjects: 5 competitive swimmers

Total trihalomethane Swimming pools n = 4 Inhalation is the major route of exposure for workers Aprea et al., 2010(CHCl3, CHCl2Br, CHClBr2, CHBr3) Media: water, urine, personal and ambient air Ingestion and skin absorption may be important for bathers

Facilities studied: change rooms, offices, pool

NCl3 Study of acute effects Exposure to NCl3 may cause adverse effects on the lung epithelium of young children Bernard et al., 2003Swimming pools n = 1Media: ambient airSubjects: 29 (16 children and 13 adults)

Total trihalomethane Swimming pool n = 1 Significant correlation between THMs in water and in urine Caro and Gallego, 2007(CHCl3, CHCl2Br, CHClBr2, CHBr3) Media: water and urine Ingestion and skin absorption may be important for bathers

Subjects: 24 (14 workers and 10 swimmers)

Total trihalomethane Swimming pool n = 1 Correlation between CHCl3 ambient air/alveolar air, and between urine/alveolar air. Caro and Gallego, 2008(CHCl3, CHCl2Br, CHClBr2, CHBr3) Media: water, urine, ambient and breath air Breath air was more sensitive than urine

Subjects: 27 (15 workers and 12 swimmers)

CHCl3 Swimming pool n = 1 Skin absorption represented one third of the total exposure Erdinger et al., 2004Media: water, ambient air and blood Correlation between blood and ambient air concentrationsSubjects: 17 (swimmers and person walking aroundthe pool)

Total trihalomethane Swimming pools n = 5 Correlation between ambient and alveolar air Fantuzzi et al., 2001(CHCl3, CHCl2Br, CHClBr2, CHBr3) Media: water, ambient and alveolar air Pool attendants were higher exposed than employees working in others area

Subjects: 32 pool workers (16 men and 16 women)Facilities studied: poolside, reception and engine-room

78V

. Bessonneau

et al.

/ International

Journal of

Hygiene

and Environm

ental H

ealth 215 (2011) 76– 85

Table 1(Continued.)

Chlorination by-products Experimental conditions Main outcomes Reference

Total trihalomethane Swimming pools n = 20 Employees working at poolside experienced higher ocular and respiratory irritativesymptoms and cutaneous diseases than subjects working in others area

Fantuzzi et al., 2010

CHCl3, CHCl2Br, CHClBr2, CHBr3) Media: water, ambient and alveolar airSubjects: 133 workersFacilities studied: poolside, café and reception

Total trihalomethane Swimming pool n = 1 Lack of association between swimming and lung function tests Font-Ribera et al., 2010CHCl3, CHCl2Br, CHClBr2,CHBr3) and NCl3

Media: water and ambient air (TTHM and NCl3), andalveolar air (TTHM)

Slightly increase lung epithelium permeability after 40 min swimming

Subjects: 50 non-smokers

NCl3 Swimming pools n = 13 Development of a sampling method based on the reduction chloramines to chlorides Héry et al., 1995Media: ambient air NCl3 levels were influenced by water turbulences

A comfort limit value of 0.5 mg/m3 was proposedNCl3 Swimming pools n = 38 (6 where selected for

measurements of environmental conditions)NCl3 levels were associated to the number of swimmers, free chlorine and ceilingheight

Jacobs et al., 2007

Media: ambient air An excess risk for asthma in pool workersSubjects: 624 pool workers

NCl3 Swimming pools n = 7 Highest level of exposure was found in younger swimmers Lévesque et al., 2006Media: water and ambient air Irritative symptoms were correlated to NCl3 concentrationsSubjects: 600 swimmers (competitive swimmers andsoccer players)

Increase in respiratory complaints for NCl3 levels ≥ 0.37 mg/m3

CHCl3 andCHCl2Br

Swimming pool n = 1 Peak of alveolar CHCl3 was two times higher than the long term ambient levels Lindstrom et al., 1997Media: water, ambient and alveolar air Dermal absorption estimated at about 80%Subjects: 2 competitive swimmers (1 man and 1woman)

NCl3 Swimming pools n = 63 (46 public and 17 leisurecentre)

Higher concentrations was found in leisure centre and generally exceeded 0,5 mg/m3 Massin et al., 1998

Media: ambient air Lifeguards with exposure to NCl3 are at risk of developing acute irritant symptomsSubjects: 334 lifeguards

NCl3 Swimming pools n = 3 NCl3 can induce occupational asthma in swimming pool workers Thickett et al., 2002Media: ambient airSubjects: 3 (2 lifeguards and 1 swimming teacher)

NCl3 Swimming pools n = 4 Application of a method based on DPD/KI chemistry for NCl3 measurements Weng et al., 2011Media: ambient air NCl3 concentrations were associated with the number of bathers and the type of

swimming activity.

ygien

rephpcti

tf2pte

gl

M

S

spV

ompfititt2pmmn(cp

bwupcb

tepiwoswmtcbb

V. Bessonneau et al. / International Journal of H

eported in these studies. To our knowledge, only few studies havexamined parameters affecting concentrations of NCl3 above theools (Jacobs et al., 2007; Weng et al., 2011). Jacobs et al. (2007)ave found a positive association between NCl3 levels and severalool characteristics including the number of swimmers, the freehlorine content in water, the use of chlorine as disinfectant andhe height of ceiling. The number of bathers was also reported asnfluencing factor of THMs concentrations (Aggazzotti et al., 1995).

The French Indoor Air Quality Observatory (OQAI) has shownhat there is a significant lack of data on levels of exposure to disin-ection byproducts in the air of French indoor pools (Guillam et al.,007). This finding is particularly alarming because a lot of sensibleersons can be exposed like children for which swimming pools arehe fourth places of entertainment in terms of attendance (Guillamt al., 2010).

In this context, this study funding by OQAI aims at providing areater understanding of key parameters influencing the airborneevels of THMs and NCl3 in indoor swimming pools.

aterials and methods

tudy design

Fifteen swimming pools situated in Brittany, France wereelected according to (1) the ratio between the number of visitorser year and the total volume of water available from pools (ratio/W), and (2) the proportion of air re-circulated into the hall.

The ratio V/W was used as a proxy variable for assess the inputsf organic and nitrogen compounds in pools. The number of swim-ers is known to be a key parameter driving the contamination of

ool water by CBPs (Chu and Nieuwenhuijsen, 2002). Assuming axed number of swimmers, the lower is the volume of the pool,he higher the concentration of organic and nitrogen compoundsn water. Data concerning pool attendance being difficult to know,he number of schoolchildren (under 18 years old), evaluated byhe French Observatory of Indoor Air Quality (OQAI) (Guillam et al.,007) was chosen for attendance assessment. In order to obtain ariori significant differences in CBPs levels between indoor swim-ing pools, we defined two classes of swimming pools from theedian value (93 schoolchildren per cubic meter per year) of the

ational distribution of the ratio V/W: low contaminated poolsratio V/W <93 schoolchildren per cubic meter per year) and highontaminated pools (ratio V/W >93 schoolchildren per cubic meterer year).

The proportion of air re-circulated needed is a compromiseetween the comfort conditions for swimmers and swimming poolorkers, and the cost of heating fresh outdoor air to 28 ◦C, partic-larly during winter in temperate countries. Considering that thisarameter induces an accumulation of airborne pollutants, eachlass of contamination (ratio V/W) were subdivided in two groups,elow or higher than 50%.

The study was conducted from June to November 2009. To assesshe daily, weekly and seasonal variations in THMs and NCl3 lev-ls, eight sampling campaigns were conducted in each swimmingool: the morning and afternoon, in high and low attending day,

n summer and winter. For each sampling day, two air samplesere collected during 3 h for the largest pool, one in the length and

ne in the width, at lifeguards breathing zone (i.e. 1.5 m). A thirdample was taken inside the swimming area, at 0.25 m from theater surface (i.e. swimmer breathing zone). In order to assess theain determinants of air contaminations, several variables linked

o pools characteristics, ventilation system, and water quality wereollected both via questionnaire and via measurements. The num-er of swimmers was recorded during air sampling periods. At theeginning and at the end of air sampling, THMs, chlorine levels

e and Environmental Health 215 (2011) 76– 85 79

(free, combined and total), pH and water temperature were mea-sured. Once a day, at the beginning of the sampling period, totalorganic carbon, Kjeldahl nitrogen and chloride were determined inwater.

Air sampling and analysis

For measurements of NCl3, air samples were collected accordingto the method described by Héry et al. (1995). Briefly, air samplewas collected through a sampling cassette containing two quartzfibre filters both impregnated with 500 �L of a solution of sodiumcarbonate (50 g/L Na2CO3) and diarsenic trioxide (8 g/L As2O3)using a Gillian 3500 pump (GE Industry, Sensing, France) at 1 L/min.Then, after sampling, impregnated fibre filters were desorbedwith ultra pure water. NCl3 was reduced to chlorides ions, whichwere analyzed by ion chromatography (Dionex DX 120, DionexBV, Amsterdam, Netherlands; IonPac® AS19 hydroxide-selectiveanion-exchange column, ASRS ultra II self-regenerating suppres-sor; conductivity detector). The limit of quantification (LOQ) was30 �g/m3.

For the determination of THMs, air sample was dynami-cally collected through stainless thermal desorption tube (6 mmO.D. × 90 mm long, 5 mm I.D., Perkin-Elmer, Boston, MA, USA) con-taining 300 mg of Tenax® TA (Supelco, France) using a Gilian3500 low flow pump (GE Industry, Sensing, France) calibrated at10 mL/min. The pumping flow rate was set, before sampling session,using a Gilian Gilibrator 2 (GE Industry, Sensing, France). After sam-pling, sorbent tube was immediately capped with brass long-termstorage caps and stored at 6 ◦C for up to 2 months. After each sam-pling session, the pumping flow rate was measured again to ensurethat flow rate did not undergo changes. The analysis of chloroform,bromodichloromethane, dibromochloromethane and bromoformwas carried out with an automatic thermal desorption unit (ATDTurbomatrix 650, Perkin-Elmer, Boston, MA, USA) coupled witha capillary gas chromatograph (HP 6890, Hewlett-Packard, PabloAlto, CA, USA) and a mass spectrometer as detector (Agilent 5975C,Agilent technologies, Santa Clara, CA, USA). Identification of THMswas based on retention times matching. Quantification was per-formed in selected ion monitoring (SIM) mode and was conductedby external calibration method. Calibration solutions containingthe four THM at six concentration levels in the range of 0.1–25 mg/Lwere prepared in methanol by serial dilutions of a stock stan-dard solution of the four THM in methanol (100 mg/L, Promochem,Molsheim, France). Different aliquots of calibration solutions wereintroduced onto the adsorbent using a GC syringe in order to obtaina known amount of THM in six sorbent tubes (0.3, 0.5, 1, 5, 10,25 ng/tube). The analytical method provided LOQ of 0.5 �g/m3 forchloroform and 0.2 �g/m3 for other THMs.

Water sampling and analysis

For determination of THMs, pool samples were collected in250 mL brown bottles. Sodium thiosulfate was added to neutral-ize chloride and stop THMs formation, and bottles were sealed withscrew caps to avoid THMs volatilisation. For measurements of TotalOrganic Carbon (TOC), Kjeldahl Nitrogen (KN) and chloride ions,water samples were collected in 1 L plastic bottles. All pool sampleswere stored at 4 ◦C until analysis, immediately after collection.

THMs were extracted from 100 mL of a pool sample with pen-tane and subsequently analyzed using a gas chromatograph (HP6890, Hewlett-Packard, Pablo Alto, CA, USA) coupled to a massspectrometer (Agilent 5975C, Agilent technologies, Santa Clara, CA,

USA). Identification and quantification of THMs were performedas describe above for air samples. Quantification of field sampleswas also performed by external calibration method. Five differ-ent concentration levels of calibration solutions were prepared in

80 V. Bessonneau et al. / International Journal of Hygiene and Environmental Health 215 (2011) 76– 85

Table 2Descriptive characteristics of the swimming pools.

Valuesa Range of valuesb

Number of swimming pools 15Total volume of water (m3) 829.2 ± 841.9 75–3250Annual attending by school children (n) 48273 ± 42866 1000–174600Number of bathers during air sampling 187.5 ± 138.9 15–514Ratio V/W (number of school children/m3) 97.4 ± 79.9 11.3–318Number of pools with a ratio V/W <93 7 (47)Number of pools with a ratio V/W >93 8 (53)Volume of hall (m3) 6100 ± 5456 350–17565Proportion of air re-circulated (%) 44.9 ± 32.6 0–80Number of pools with a proportion of air re-circulated <50% 5 (33)Number of pools with a proportion of air re-circulated >50% 10 (67)Number of pools use disinfection by chlorine 15 (100)

nd per

ETaFo

cp(csG

dl1

9aBlbam

Dhr1

Q

mffaV

S

RLwnsrG

a Data are shown as number, arithmetic mean ± standard deviation, or number ab Range of values represents the minimum and the maximum.

vian water (5–150 �g/L for chloroform and 0.5–15 �g/L for otherHMs) from stock standard solutions (500 mg/L for chloroformnd 50 mg/L for other THMs in methanol, Promochem, Molshein,rance). LOQ were, respectively 5 and 0.5 �g/L for chloroform andther THMs.

Free and total chlorine were determined on site using theolorimetric method based on the addition of N,N′-diethyl-p-henylenediamine (DPD) reagent, according to NF ISO 7313-22000). Combined chlorine was calculated by subtracting freehlorine to total chlorine. Water temperature and pH were mea-ured using portable pH meter (HQ 40d, Hach lange, Düsseldorf,ermany).

TOC was determined by thermal oxidation coupled with infraredetection (Multi N/C 2100, Analytik Jena AG, Jena, Germany) fol-

owing acidification with hydrochloric acid solution (NF EN 1484,997). The LOQ was 0.2 mg/L.

KN was analyzed following mineralization with sulphuric acid8% and a Kjeldahl tablet containing 5 g of potassium sulphatend 50 mg of selenium as a catalyst (digestion block BUCHI 438,uchi Labortechnik AG, Flawil, Switzerland), and distillation (distil-

ation unit BUCHI 323, Buchi Labortechnik AG, Flawil, Switzerland)y titration (702 SM Titrino, Metrohm AG, Herisau, Switzerland)gainst hydrochloric acid according to NF EN 25663 (1994). Thisethod provided a LOQ of 0.48 mg/L.Chloride ions were analyzed by ion chromatography (Dionex

X 120, Dionex BV, Amsterdam, Netherlands; IonPac® AS19ydroxide-selective anion-exchange column, ASRS ultra II self-egenerating suppressor; conductivity detector) (NF EN ISO0304-1, 2009). The LOQ of this method was 2 mg/L.

uestionnaire

A questionnaire was proposed to be filled by swimming poolsanagers about swimming pool characteristics, including disin-

ection and ventilation system, the last information being neededor the assessment of air re-circulated rate. Additional data on thennual attendance were acquired for the calculation of the ratio/W.

tatistical analysis

Statistical analysis on the collected data was performed using-2.12.1 for Windows. Values below the LOQ were analyzed asOQ divided by square root of two. The Shapiro-Wilk normality testas used to test if data followed a Gaussian distribution. Variables

ot normally distributed, were log-transformed. For descriptivetatistic purpose, the geometric means (GM) of parameters wereeported with the 95% confidence interval (CI). Differences amongM values were evaluated by non-parametric Mann–Whitney

centage.

test for the comparison of 2 parameters, and by non-parametricKruskall–Wallis test for the comparison of 3 or more parameters.

Since no statistical difference in TTHM and NCl3 levels was foundbetween the three sampling points, the concentrations were aver-aged.

To assess associations between CBPs (TTHM and NCl3) levelsand other variables measured (ratio V/W, proportion of air re-circulated, season, volume of hall, number of bathers, chlorine (free,combined and total), pH, TOC, chloride ions, KN, ratio C/N, andTTHM in water), multivariate ordinary least square (OLS) and quan-tile regressions were used. Quantile regression (QR) was introducedby Koenker and Bassett (1978). It provides estimates either for themedian or for other quantiles, while OLS estimates the mean. Inaddition, QR is less sensitive to outliers. For multivariate analyses,response variables, TTHM and NCl3, were log-transformed. Vari-ables in OLS models were determined using backward elimination.Models were evaluated using adjusted R2, Akaike Information Cri-terion (AIC) and significance. Only variables selected in the finalOLS models were used in QR models.

Results

Swimming pools description

The swimming pool characteristics are presented in Table 2.All studied pools used chlorine as disinfectant. The total volumeof water and the volume of hall, respectively ranged from 75 to3250 m3 and from 350 to 17565 m3, indicating a wide variationof facilities sizes among studied swimming pools. A large variabil-ity in the numbers of swimmers recorded (from 15 to 514) wasalso found, due to the choice of sampling day (high versus lowexpected pool attending). The numbers of pools selected in eachclass of ratio V/W were roughly similar. Most of the studied swim-ming pools (67%) reported a proportion of air re-circulated greaterthan or equal to 50%.

Pool water quality

Levels of parameters related to swimming pools water qual-ity are listed in Table 3. The pH and temperature of water rangedfrom 6.72 to 8.02 and 26.8 ◦C to 32.6 ◦C, respectively. The geo-metric means (GM ± 95% CIs) of free and combined chlorine were1.33 ± 0.08 mg/L and 0.55 ± 0.04 mg/L, respectively. The concentra-tions of TOC ranged from 1.80 to 7.30 mg/L with a GM (± 95% CIs) of3.1 ± 0.15 mg/L. For KN, almost the half of data (44%) was below the

LOQ, indicating a low contamination of the pools by nitrogen con-taining compounds. The concentrations of TTHM in water rangedfrom 4.80 to 80.7 �g/L with a GM (±95% CIs) of 25.9 ± 2.3 �g/L. Chlo-roform was the predominant specie, representing 84% of TTHM,

V. Bessonneau et al. / International Journal of Hygiene and Environmental Health 215 (2011) 76– 85 81

Table 3Summary of the results for water samples.

Parameters N <LOQ (%) Geometric mean 95% CIs Minimum Maximum

Free chlorine (mg/L) 188 0 1.33 0.08 0.40 4.00Combined chlorine (mg/L) 188 0 0.55 0.04 0.05 1.70Total chlorine (mg/L) 188 0 1.97 0.09 1.00 5.00TOC (mg/L) 47 0 3.10 0.15 1.80 7.30KN (mg/L) 47 44 0.60 0.05 0.35a 2.50Chloride ions (mg/L) 46 0 219 14.0 103 614Chloroform (�g/L) 185 1.6 21.8 2.00 3.50a 72.6Dichlorobromomethane (�g/L) 185 0 2.60 0.25 0.60 15.3Dibromochloromethane (�g/L) 185 21 0.75 0.07 0.35a 3.80Bromoform (�g/L) 185 73.5 0.42a 0.03 0.35a 2.20TTHM (�g/L) 185 0 25.9 2.30 4.80 80.7Temperature (◦C) 180 29.3 0.20 26.8 32.6

7.46

w(

T

T(wm

1

F

atw(ot1(wdswwpGwm

Fm

the effect of this factor was highest at the lower quantiles anddecreased from the 50th percentiles, indicating a strong impact

pH 188

a Value below the LOQ.

hile the concentrations of other THMs were much lower: CHCl2Br11%), CHClBr2 (3%), CHBr3 (2%).

HMs and NCl3 levels in air

Fig. 1 provides the distribution of THMs and NCl3 levels in air.he TTHM levels ranged from 1.45 �g/m3 to 793 �g/m3 with a GM±95% CIs) of 74.9 ± 6.7 �g/m3. As for water the predominant specieas chloroform (88%), the concentrations of other species beinguch lower: CHCl2Br (9%), CHClBr2 (2%), CHBr3 (1%).The concentrations of NCl3 ranged from 20 �g/m3 to

260 �g/m3 with a GM (±95% CIs) of 190 ± 10 �g/m3.

actor determining TTHM and NCl3 levels in air

We examined variables affecting the concentrations of TTHMnd NCl3 in air of swimming pools. Given the air contamina-ion by CBPs is due to the volatilization from pool, the first stepas the study of relationship between TTHM in air and water

Fig. 2). Globally TTHM in water explained only 19% of the variancef levels measured in air. Considering the seasonality, concen-rations of TTHM measured in air during winter (GM ± 95% CIs;48.3 ± 25.1 �g/m3) were significantly higher than during summerGM ± 95% CIs; 78.1 ± 19.5 �g/m3) (p < 0.05). The opposite trendas shown for TTHM levels in water (GM ± 95% CIs; 30.1 ± 3.7 �g/Luring summer and GM ± 95% CIs; 23.3 ± 2.5 �g/L during winter,tatistically different at p < 0.05). Although few swimming poolsere investigated only during winter, the seasonal trend (p < 0.05)as not affect by eliminating concentrations measured in these

ools. For some swimming pools, no difference was shown betweenM of TTHM in water found during summer and winter (p > 0.05),hile the GM of TTHM in air increased 100-fold between sum-er and winter. For NCl3, the number of swimmers was used as

ig. 1. Boxplot of the distribution (minimum, percentiles 25th, 50th and 75th andaximum) of THMs and NCl3 concentrations in air.

0.03 6.72 8.02

predictor of concentrations in water. This parameter was weaklyassociated to NCl3 found in air, explaining 11% of the total variance.Plotting concentrations of NCl3 levels and the number of swim-mers depending on the season, no difference was found betweensummer and winter (p > 0.05).

For better understanding of parameters affecting CBPs levels inair, multivariate analyses were conducted. Concentrations of chlo-rine (free, combined and total), TTHM, TOC, temperature and pHwere assessed as water quality parameters; air temperature, vol-ume of hall and proportion of air re-circulated as swimming poolscharacteristics; the number of bathers during air sampling andthe ratio V/W as pollution inputs in pools; and time of sampling(i.e. morning and afternoon) and the season as temporal variabilityfactors. The KN parameter was excluded because 44% of the mea-surements were below the LOQ. Figs. 3 and 4 illustrate the OLSand QR results obtained for TTHM and NCl3 levels. The final OLSmodels explained 43% and 41% of the variance for TTHM and NCl3,respectively.

The concentrations of chlorine (free, combined and total), TOC,chloride ions and water temperature were not statistically signif-icant factors (p > 0.05). In addition, no daily variation was shown(p > 0.05).

The proportion of air re-circulated increased significantly(p < 0.05) TTHM levels in air for the lower quantiles (e.g. 25th and50th percentiles), while at upper quantiles this variable had noeffect. For NCl3, this factor was not significant.

The ratio V/W was associated to CBPs concentrations. For THMs,

on lowest concentrations but not on the highest ones. For NCl3,

Fig. 2. Scatter plot of THMs levels in air and water depending on the season. Valuesof R2 were 0.19, 0.20 and 0.30 for linear regression on global data, for linear regres-sion on data related to summer and for linear regression on data related to winter,respectively.

82 V. Bessonneau et al. / International Journal of Hygiene and Environmental Health 215 (2011) 76– 85

F The cop dicatet pondi

tq

lta

Fve

ig. 3. Plot of the coefficients of QR for the quantiles of the concentrations of TTHM.roportion of air re-circulated, and winter (reference = summer). The grey shading inhe estimate from the OLS regression and the dashed horizontal lines are the corres

his parameter seemed to affect lower quantiles as well as higheruantiles.

CBPs were strongly related to the season. As described ear-

ier, the levels of TTHM measured during winter exceededhose during summer. As shown in Figs. 3 and 4, for TTHMnd NCl3 the differences between hot and cold seasons were

ig. 4. Plot of the coefficients of QR for the quantiles of the concentrations of NCl3. The coolume of hall, pH and winter (reference = summer). The grey shading indicates the 95stimate from the OLS regression and the dashed horizontal lines are the corresponding 9

variates include log-transformed ratio V/W, TTHM in water, number of swimmers,s the 95th confidence intervals of the coefficient. The solid horizontal line representsng 95% CI.

lower at 25th percentiles and greatly increased at 90th per-centiles (140.3 �g/m3 for TTHM and 440 �g/m3 for NCl3 dur-ing summer and 429.9 �g/m3 for TTHM and 550 �g/m3 for

NCl3 during winter). The effects at the 90th percentiles arehigher than the upper confidence limit of the OLS models,

variates include log-transformed air temperature, ratio V/W, number of swimmers,th confidence intervals of the coefficient. The solid horizontal line represents the5% CI.

ygien

is

w

poToupt

lwaFtt

aco

D

i(r3bsp2Ric

i(aetaAtsies1RHctf2wiecpvb

V. Bessonneau et al. / International Journal of H

ndicating that OLS results underestimate the impact of sea-on.

The airborne levels of THMs were significantly associated withater concentrations, except for the 90th percentile.

The number of swimmers attending the pools during air sam-ling increased the levels of TTHM and NCl3. Effects of the numberf swimmers were statistically significant for lower quantiles ofTHM, while for the upper quantiles were not. For NCl3, the increasef bathers was strongly related to airborne concentrations, partic-larly at the 90th percentiles. The effect of this variable at the 90thercentiles is higher than the OLS result, indicating a strong rela-ionship between highest NCl3 levels and the number of swimmers.

The volume of swimming pools hall decreased significantly theevels of CBPs. This factor has no effect on lowest concentrations

hile significant effect was found at upper quantiles. The temper-ture of air was strongly associated to NCl3 levels but not to TTHM.or NCl3, the impact of air temperature increased along of the quan-iles. The largest effect was found at upper quantiles and was largerhan the OLS result.

The pH of water was significantly associated with NCl3 levels inir but not with TTHM. The increase of pH decreased the airborneoncentrations of NCl3. The effect of pH was particularly significantn highest levels, from 50th percentiles.

iscussion

The concentrations of free and combined chlorine measuredn water are consistent with the recommended values in FranceAFSSET, 2010). The GM of TOC found is lower than the valueseported by Chu and Nieuwenhuijsen (2002) which ranged from.3 to 12.9 mg/L. All concentrations of TTHM found in water wereelow the limit set by the French law (100 �g/L). Our data are in theame order of magnitude than those reported in other swimmingools (Chu and Nieuwenhuijsen, 2002: 132.4 �g/L; Fantuzzi et al.,001: 39.8 ± 21.7 �g/L; Font-Ribera et al., 2010: 45.4 ± 7.3 �g/L;ichardson et al., 2010: 49.6 ± 10.6 �g/L). We found a predom-

nance of chloroform because all swimming pools studied usedhlorine for disinfection.

The airborne concentrations of THMs found in our study aren agreement with data previously reported in the literatureAggazzotti et al., 1990, 1993, 1995, 1998; Aprea et al., 2010; Carond Gallego, 2007, 2008; Fantuzzi et al., 2001, 2010; Font-Riberat al., 2010; Richardson et al., 2010). Exposure to these concentra-ions levels have been associated to self-reported sneezing, itchynd red eyes among swimming pool workers (Fantuzzi et al., 2010).lthough the chloroform was the most abundant specie, exposure

o brominated species could induce genotoxic effects. A recenttudy has shown an increased response among various genotox-city biomarkers and exposure to brominated THMs (Kogevinast al., 2010). The concentrations of NCl3 measured are also con-istent with previous studies (Bernard et al., 2003; Héry et al.,995; Jacobs et al., 2007; Lévesque et al., 2006; Massin et al., 1998;ichardson et al., 2010; Thickett et al., 2002; Weng et al., 2011).owever, 5% and 17% of NCl3 concentrations measured exceed theomfort level of 500 �g/m3 proposed by Héry et al. (1995) andhe limit value of 300 �g/m3 recommended by the French Agencyor Food, Environmental and Occupational Health Safety (AFSSET,010), respectively. Exposure to mean concentrations of 500 �g/m3

as associated with an increase in lung epithelium permeabilityn swimmers (Carbonnelle et al., 2002). In addition, three work-rs (two lifeguards and one swimming instructor) exposed to NCl3

oncentrations ranging from 100 to 570 �g/m3 have reported occu-ational asthma (Thickett et al., 2002). Recently, a toxicity referencealue of 4 �g/m3, applicable for short and repeated exposure haseen proposed (Bonvallot et al., 2010).

e and Environmental Health 215 (2011) 76– 85 83

The main objective of this study was to examine variables affect-ing the concentrations of TTHM and NCl3 in air of swimming pools.The weak associations obtained from bivariate analysis betweenconcentrations of CBPs in air and water suggest that probablyother parameters affect the concentrations measured in air of theswimming pools studied. The results of the multivariate analysisperformed indicate that in our study other parameters should beconsidered as key determinants. The concentrations of TTHM in airwere associated to the proportion of air re-circulated, the ratio V/W,the season, the concentrations in water, the number of swimmersand the volume of swimming pool hall; while the concentrationsof NCl3 were associated to the ratio V/W, the season, the number ofswimmers, the volume of swimming pool hall, the temperature ofair and the pH of water. The results obtained for NCl3 models werediscussed cautiously because of the low specificity of the samplingmethod used (Weng et al., 2011). Some parameters related to CBPsformation in water including chlorine content (free, combined ortotal) and TOC are not correlated to airborne concentrations, proba-bly because chlorination kinetics are achieved in few hours. Despitewater temperature is known to affect the volatilization of CBPs fromwater to air, the lack of association between water temperature andCBPs in air is possibly due to the low variability of this parameter.In addition, no variation in airborne concentrations were shownbetween measurements performed in the morning and in the after-noon, probably because sampling periods were carried out from9 am to 12 pm, and from 1 pm to 4 pm, which did not take intoaccount the high pool attendance generally occurring after 4 pm(Weng et al., 2011).

Some parameters are proxy variable and present limitations.The proportion of air re-circulated was not measured but asked toswimming pools managers, due to difficulties for accessing the ven-tilation system. Thus, this variable cannot be unfortunately used, asexpected, to describe the accumulation of pollutants within indoorswimming pools. In addition the ratio V/W was calculated fromyearly pool attendance by schoolchildren less than 18 years oldand not from all visitors, including a bias.

The other determinants found in our study may be very helpful.In our case, the season depicts the ventilation within swimmingpools. During summer, the intake of air from outdoor via the ven-tilation system and the opening windows is more important thanduring winter, limiting the accumulation of airborne pollutants. Inother words, these findings indicate that increasing air exchangerate within indoor swimming pools could decrease significantlyair contamination. Although the average concentration of TTHMin air was clearly associated to the average concentration of thesecompounds in water, the highest airborne levels (from 90th per-centile: 255 �g/m3) were not, suggesting that the occurrence ofvery high concentrations in air could not be explained by THMsvolatilising from the pools to air. The number of swimmers had astrong impact of CBPs levels in air, which is consistent with pre-vious studies (Aggazzotti et al., 1990; Jacobs et al., 2007; Wenget al., 2011). For TTHM, the number of swimmers probably reflectsthe impact of activities such as swimming and playing, creatingwater turbulences and increasing the volatilisation of chemicalfrom water to air (Jacobs et al., 2007; Weng et al., 2011). In additionto water splashes, the increase of bathers increases the amount ofnitrogen-containing compounds in pools (Richardson et al., 2010),which probably explained the effects on upper quantiles of NCl3.The influence of the volume of swimming pool hall is consistentwith the fact that chemicals are more dispersed and diluted in largevolume of air. The temperature of indoor air was correlated withNCl3 concentrations, but not with TTHM concentrations. This result

is probably explained by the narrow range (from 19.2 to 28.4 ◦C)and the low variability of temperature measured (6.8%). Despitethe low variation in air temperature NCl3 levels were increased,which is consistent with its physical properties, NCl3 being more

8 ygiene

vuottwtcwTsfe

C

wFa1c1ihup

tbotttrwwedstptcciinbi

A

bttMdDEdat

4 V. Bessonneau et al. / International Journal of H

olatile than CHCl3, the most important THM found, and not sol-ble in water, leading to a high transfer efficiency. The influencef pH on NCl3 concentrations may be explained by the chlorina-ion processes. Sodium hypochlorite introduced in the pool leadso the formation of hypochlorous acid, the most disinfectant form,hich is dissociated to hypochlorite ion. The reaction of dissocia-

ion is driving by the water pH. The higher is the pH, the lower theoncentration of hypochlorous acid, decreasing CBPs formation inater (White, 1986). This factor was not significantly associated to

HMs, indicating that this result needs to be confirmed in furthertudies. However, the increase of pH could possibly promote theormation of THMs via the hydrolysis of precursors of THMs (Adint al., 1991; Hong et al., 2007).

onclusions

Most of the NCl3 levels measured in indoor swimming poolsere lower than the limit value of 300 �g/m3 recommended by the

rench Agency for Food, Environmental and Occupational Healthnd Safety. Though, no pool exceeded the French law limit of00 �g/L of TTHM in drinking and recreational water, high con-entrations of TTHM have been found in air (75th percentiles of32 �g/m3). Given the wide diversity of the population attend-

ng swimming pools (baby swimmers, lifeguard) and the negativeealth effects resulting to THMs exposure, there is a need to reg-late the levels of these substances in public indoor swimmingools.

Although the number of swimming pools studied is limited,he results of OLS and QR models for TTHM and NCl3 provides aetter understanding of parameters affecting the concentrationsf CBPs in air, and may be useful to develop recommendationso pools managers for decreasing exposure levels, particularly forhe highest ones. The results suggest significant seasonal varia-ions in CBPs concentrations, reflecting the high proportion of aire-circulated during winter. In temperate countries as France, theinter should be considered with a particular interest for assessingorkers’ and attendant’s exposure to CBPs. Additionally, CBPs lev-

ls may be drastically decreased by increasing air exchange rate. Asescribed in previous studies, the number of swimmers increasedignificantly CBPs concentrations, especially for NCl3, indicatinghat bathers’ hygiene remains a simple recommendation to reduceollution input in pools. In addition, elevation of air tempera-ure seemed to enhance the volatilisation of NCl3. This parameterould be considered for limiting CBPs emission, even if the users’omfort is important. Daily variations in CBPs levels were not signif-cant. Additionally, several variables associated to CBPs formationn water, such as concentrations of chlorine or TOC, were not sig-ificantly correlated to chemicals in air. Our sampling strategy maye improved by implementing a finer profiling of CBPs fluctuations

n indoor swimming pools.

cknowledgements

The French Observatory of Indoor Air Quality (OQAI) is fundedy the Ministries in charge of Housing, Environment and Health,he Environment and Energy Management Agency (ADEME) andhe Scientific and Technical Centre of Building (CSTB). We thank

s. Pauline Carillon, Emilie Surget and Gaëlle Raffy for her helpuring sampling campaign. We also thank Pr. Barbara Le Bot, Mr.ominique Verrey and the staff of the analytical service of the

nvironmental and Health Research Laboratory (LERES) for coor-ination and realization of samples analysis. We finally thank thenonymous reviewers for their valuable comments and suggestionso improve the quality of the paper.

and Environmental Health 215 (2011) 76– 85

References

Adin, A., Katzhendler, J., Alkaslassy, D., Rav-Acha, Ch., 1991. Trihalomethane forma-tion in chlorinated drinking water: a kinetic model. Water Res. 25, 797–805.

AFSSET, 2010. Evaluation des risques sanitaires liés aux piscines. Partie 1: piscinesréglementées. Edition scientifique juin.

Aggazzotti, G., Fantuzzi, G., Tartoni, P.L., Predieri, G., 1990. Plasma chloroform con-centrations in swimmers using indoor swimming pools. Arch. Environ. Health45, 175–179.

Aggazzotti, G., Fantuzzi, G., Righi, E., Tartoni, P., Cassinadri, T., Predieri, G., 1993.Chloroform in alveolar air of individuals attending indoor swimming pools. Arch.Environ. Health 48, 250–254.

Aggazzotti, G., Fantuzzi, G., Righi, E., Predieri, G., 1995. Environmental and biologi-cal monitoring of chloroform in indoor swimming pools. J. Chromatogr. A 710,181–190.

Aggazzotti, G., Fantuzzi, G., Righi, E., Predieri, G., 1998. Blood and breath analyses asbiological indicators of exposure to trihalomethanes in indoor swimming pools.Sci. Total Environ. 217, 155–163.

Aprea, M.C., Banchi, B., Lunghini, L., Pagliantini, M., Peruzzi, A., Sciarra, G., 2010.Disinfection of swimming pools with chlorine and derivatives: formation oforganochlorinated and organobrominated compounds and exposure of poolpersonnel and swimmers. Nat. Sci. 2, 68–78.

Bernard, A., Carbonnelle, S., Michel, O., Higuet, S., de Burbure, C., Buchet, J.P., Her-mans, C., Dumont, X., Doyle, I., 2003. Lung hyperpermeability and asthmaprevalence in schoolchildren: unexpected associations with the attendance atindoor chlorinated swimming pools. Occup. Environ. Med. 60, 385–394.

Bernard, A., Carbonnelle, S., de Burbure, C., Michel, O., Nickmilder, M., 2006. Chlori-nated pool attendance, atopy, and the risk of asthma during childhood. Environ.Health Perspect. 114, 1567–1573.

Bernard, A., Carbonnelle, S., Dumont, X., Nickmilder, M., 2007. Infant swimmingpractice, pulmonary epithelium integrity, and the risk of allergic and respiratorydiseases later in childhood. Pediatrics 119, 1095–1103.

Bernard, A., Nickmilder, M., Voisin, C., Sardella, A., 2009. Impact of chlorinated swim-ming pool attendance on the respiratory health of adolescents. Pediatrics 124,1110–1118.

Bonvallot, N., Glorennec, P., Zmirou, D., 2010. Derivation of a toxicity reference valuefor nitrogen trichloride as a disinfection by-product. Regul. Toxicol. Pharmacol.56, 357–364.

Carbonnelle, S., Francaux, M., Doyle, I., Dumont, X., De, B.C., Morel, G., Michel,O., Bernard, A., 2002. Changes in serum pneumoproteins caused by short-term exposures to nitrogen trichloride in indoor chlorinated swimming pools.Biomarkers 7, 464–478.

Caro, J., Gallego, M., 2007. Assessment of exposure of workers and swimmersto trihalomethanes in an indoor swimming pool. Environ. Sci. Technol. 41,4793–4798.

Caro, J., Gallego, M., 2008. Alveolar air and urine analyses as biomarkers of expo-sure to trihalomethanes in an indoor swimming pool. Environ. Sci. Technol. 42,5002–5007.

Chu, H., Nieuwenhuijsen, M.J., 2002. Distribution and determinants of tri-halomethane concentrations in indoor swimming pools. Occup. Environ. Med.59, 243–247.

Erdinger, L., Kühn, K.P., Kirsch, F., Feldhues, R., Fröber, T., Nohynek, B., Gabrio, T., 2004.Pathways of trihalomethane uptake in swimming pools. Int. J. Hyg. Environ.Health 207, 571–575.

Fantuzzi, G., Righi, E., Predieri, G., Ceppelli, G., Gobba, F., Aggazzotti, G., 2001. Occupa-tional exposure to trihalomethanes in indoor swimming pools. Sci. Total Environ.264, 257–265.

Fantuzzi, G., Righi, E., Predieri, G., Giacobazzi, P., Mastroianni, K., Aggazzotti, G., 2010.Prevalence of ocular, respiratory and cutaneous symptoms in indoor swimmingpool workers and exposure to disinfection by-products (DBPs). Int. J. Environ.Res. Public Health 7, 1379–1391.

Font-Ribera, L., Kogevinas, M., Zock, J.P., Gómez, F.P., Barreiro, E., Nieuwenhuijsen,M.J., Fernandez, P., Lourencetti, C., Pérez-Olabarría, M., Bustamante, M., Mar-cos, R., Grimalt, J.O., Villanueva, C.M., 2010. Short-term changes in respiratorybiomarkers after swimming in a chlorinated pool. Environ. Health Perspect. 118,1538–1544.

Guillam, M.T., Thomas, N., Nedellec, V., Derbez, M., Kirchner, S., Segala, C., 2007. Lespiscines couvertes en France: caractéristiques et qualité de l’air. Poll Atm. 196,329–335.

Guillam, M.T., Thomas, N., Nedellec, V., Derbez, M., Kirchner, S., Pédrono, G., Ségala,C., 2010. Children’s indoor leisure activities in France: time budget data forindoor air risk assessment. Hum. Ecol. Risk Assess. 16, 977–988.

Héry, M., Hecht, G., Gerber, J.M., Gender, J.C., Hubert, G., Rebuffaud, J., 1995. Exposureto chloramines in the atmosphere of indoor swimming pools. Ann. Occup. Hyg.39, 427–439.

Hinckley, A.F., Bachand, A.M., Reif, J.S., 2005. Late pregnancy exposures to disinfec-tion by-products and growth-related birth outcomes. Environ. Health Perspect.113, 1808–1813.

Hong, H.C., Liang, Y., Han, B.P., Mazumber, A., Wong, M.H., 2007. Modeling of tri-halomethane (THM) formation via chlorination of the water from DongjiangRiver. Sci. Total Environ. 385, 48–54.

Jacobs, J.H., Spaan, S., van Rooy, G.B.G.J., Meliefste, C., Zaat, V.A.C., Rooyackers, J.M.,Heederik, D., 2007. Exposure to trichloramine and respiratory symptoms inindoor swimming pool workers. Eur. Respir. J. 29, 690–698.

Kim, H., Shim, J., Lee, S., 2002. Formation of disinfection by-products in chlorinatedswimming pool water. Chemosphere 46, 123–130.

ygien

KK

L

L

L

L

M

M

M

N

N

N

N

V. Bessonneau et al. / International Journal of H

oenker, R., Bassett Jr., G.B., 1978. Regression quantiles. Econometrica 46, 33–50.ogevinas, M., Villanueva, C.M., Font-Ribera, L., Liviac, D., Bustamante, M., Espinoza,

F., Nieuwenhuijsen, M.J., Espinosa, A., Fernandez, P., DeMarini, D.M., Grimalt,J.O., Grummt, T., Marcos, R., 2010. Genotoxic effects in swimmers exposed todisinfection by-products in indoor swimming pools. Environ. Health Perspect.118, 1531–1537.

aKind, J.S., Richardson, S.D., Blount, B.C., 2010. The good, the bad, and the volatile:can we have both healthy and healthy people? Environ. Sci. Technol. 44,3205–3210.

évesque, B., Duchesne, J.F., Gingras, S., Lavoie, R., Prud’Homme, D., Bernard, E.,Boulet, L.P., Ernst, P., 2006. The determinants of prevalence of health com-plaints among young competitive swimmers. Int. Arch. Occup. Environ. Health80, 32–39.

i, J., Blatchley III, E.R., 2007. Volatile disinfection byproduct formation resultingfrom chlorination of organic-nitrogen precursors in swimming pools. Environ.Sci. Technol. 41, 6732–6739.

indstrom, A.B., Pleil, J.D., Berkoff, D.C., 1997. Alveolar breath sampling and analy-sis to assess trihalomethane exposures during competitive swimming training.Environ. Health Perspect. 105, 636–642.

assin, N., Bohadana, A.B., Wild, P., Héry, M., Toamain, J.P., Hubert, G., 1998. Respira-tory symptoms and bronchial responsiveness in lifeguards exposed to nitrogentrichloride in indoor swimming pools. Occup. Environ. Med. 55, 258–263.

orris, R.D., 1995. Drinking water and cancer. Environ. Health Perspect. (Suppl. 103),225–232.

orris, R.A., Audet, A.M., Angelillo, I.F., Chalmers, T.C., Mosteller, F., 1992. Chlorina-tion, chlorination by-products, and cancer: a meta-analysis. Am. J. Public Health82, 955–963.

F EN 1484, AFNOR, 1997. Analyse de l’eau. Lignes directrices pour le dosage ducarbone organique total (COT) et carbone organique dissous (COD).

F EN 25663, AFNOR, 1994. Qualité de l’eau. Dosage de l’azote Kjeldahl. Méthodeaprès minéralisation au sélénium.

F EN ISO 10304-1, AFNOR, 2009. Qualité de l’eau. Dosage des anions dissous

par chromatographie des ions en phase liquide. Partie 1: dosage du bromure,chlorure, fluorure, nitrate, nitrite, phosphate et sulfate.

F ISO 7313-2, AFNOR, 2000. Qualité de l’eau. Dosage du chlore libre et du chloretotal. Partie 2: méthode colorimétrique à la N,N-diéthylphénylène-1,4-diaminedestinée aux contrôles de routine.

e and Environmental Health 215 (2011) 76– 85 85

Nieuwenhuijsen, M.J., Toledano, M.B., Elliott, P., 2000. Uptake of chlorination disin-fection by-products; a review and a discussion of its implications for exposureassessment in epidemiological studies. J. Exp. Anal. Environ. Epidemiol. 10,586–599.

Richardson, S.D., DeMarini, D.M., Kogevinas, M., Fernandez, P., Marco, E., Lourencetti,C., Ballesté, C., Heederik, D., Meliefste, K., McKague, A.B., Marcos, R., Font-Ribera,L., Grimalt, J.O., Villanueva, C.M., 2010. What’s in the pool? A comprehensiveidentification of disinfection by-products and assessment of mutagenicity ofchlorinated and brominated swimming pool water. Environ. Health Perspect.118, 1523–1530.

Thickett, K.M., McCoach, J.S., Gerber, J.M., Sadhra, S., Burge, P.S., 2002. Occupationalasthma caused by chloramines in indoor swimming-pool air. Eur. Respir. J. 19,827–832.

Villanueva, C.M., Cantor, K.P., Grimalt, J.O., Malats, N., Silverman, D., Tardon, A.,Garcia-Closas, R., Serra, C., Carrato, A., Castano-Vinyals, G., Marcos, R., Rothman,N., Real, F.X., Dosemeci, M., Kogevinas, M., 2007. Bladder cancer and exposureto water disinfection by-products through ingestion, bathing, showering, andswimming in pools. Am. J. Epidemiol. 165, 148–156.

Voisin, C., Sardella, A., Marcucci, F., Bernard, A., 2010. Infant swimming in chlorinatedpools and the risks of bronchiolitis, asthma and allergy. Eur. Respir. J. 36, 41–47.

Weaver, W.A., Li, J., Wen, Y., Johnston, J., Blatchley, M.R., Blatchley, E.R., 2009.Volatile disinfection by-product analysis from chlorinated indoor swimmingpools. Water Res. 43, 3308–3318.

Weisel, C.P., Richardson, S.D., Nemery, B., Aggazzotti, G., Baraldi, E., Blatchley, E.R.,Blount, B.C., Carlsen, K.H., Eggleston, P.A., Frimmel, F.H., Goodman, M., Gordon,G., Grinshpun, S.A., Heederik, D., Kogevinas, M., LaKind, J.S., Nieuwenhuijsen,M.J., Piper, F.C., Sattar, S.A., 2008. Childhood asthma and environmental expo-sures at swimming pools: state of the science and research recommendations.Environ. Health Perspect. 117, 500–507.

Weng, S.C., Weaver, W.A., Afifi, M.Z., Blatchley, T.N., Cramer, J.S., Chen, J., BlatchleyIII, E.R., 2011. Dynamics of gas-phase trichloramine (NCl3) in chlorinated, indoorswimming pool facilities. Indoor Air, doi:10.1111/j. 1600-0668.2011.00710.x.

White, G.C., 1986. Handbook of Chlorination. Van Nostrand Reinhold Company, NewYork.

Zwiener, C., Richardson, S.D., De Marini, D.M., Grummt, T., Glauner, T., Frimmel, F.H.,2007. Drowning in disinfection byproducts? Assessing swimming pool water.Environ. Sci. Technol. 41, 363–373.