Journal of Exposure Analysis and Environmental Epidemiology (2004) 14, 466–472
r 2004 Nature Publishing Group All rights reserved 1053-4245/04/$30.00
www.nature.com/jea
Exposure assessment in epidemiologic studies of adverse pregnancy
outcomes and disinfection byproducts
WILL D. KING,a LINDA DODDS,b B. ANTHONY ARMSON,b ALEXANDER C. ALLEN,b DESHAYNE B. FELLb
AND CARL NIMRODc
a
Department of Community Health and Epidemiology, Queen’s University, Kingston, Ontario, Canada
Perinatal Epidemiology Research Unit, Departments of Obstetrics and Gynaecology and Pediatrics, Dalhousie University, Halifax, Nova Scotia, Canada
c
Department of Obstetrics and Gynaecology, Ottawa University, Ottawa, Ontario, Canada
b
A major challenge in studies that examine the association between disinfection byproducts in drinking water and pregnancy outcomes is the accurate
representation of a subject’s exposure. We used household water samples and questionnaire information on water-use behavior to examine several aspects
of exposure assessment: (i) the distribution and correlation of specific disinfection byproducts, (ii) spatial distribution system and temporal variation in
byproduct levels, and (iii) the contribution of individual water-use behavior. The level of specific trihalomethanes (THMs) and haloacetic acids (HAAs)
was determined for 360 household water samples in Eastern Ontario and Nova Scotia. Subjects were interviewed regarding tap water ingestion and
showering and bathing practices. In both provinces, total THMs correlated highly with chloroform (correlation coefficient (r) 40.95) and less so with
total HAAs (r ¼ 0.74 in Nova Scotia and r ¼ 0.52 in Ontario). The correlation between total THMs and bromodichloromethane was high in Nova Scotia
(r ¼ 0.63), but low in Ontario (r ¼ 0.26). The correlation was between THM level in individual household samples, and the mean THM level during the
same time period from several distribution system samples was 0.63, while a higher correlation in THM level was observed for samples taken at the same
location 1 year apart (r ¼ 0.87). A correlation of 0.73 was found between household THM level and a total exposure measure incorporating ingestion,
showering, and bathing behaviors. These results point to the importance of: measurement of different classes of byproducts; household rather than
distribution system sampling; and, incorporation of subject behaviors in exposure assessment in epidemiologic studies of disinfection byproducts and
adverse pregnancy outcomes.
Journal of Exposure Analysis and Environmental Epidemiology (2004) 14, 466–472. doi:10.1038/sj.jea.7500345
Published online 17 March 2004
Keywords: epidemiology, pregnancy, disinfection byproducts, environment, exposure.
Introduction
Most public drinking-water systems in North America use
chlorine compounds for disinfection. The reaction of chlorine
with organic material in source water results in the formation
of several classes of chemical byproducts. Epidemiologic
studies have examined the relationship between exposure to
disinfection byproducts and risk for a range of adverse
pregnancy outcomes. Although risk estimates are often small
in magnitude and inconsistent across studies, a moderate
degree of evidence exists for a relationship with small for
gestational age, neural tube defects, and fetal death (Bove
et al., 2002). Systematic reviews of the relationship between
disinfection byproducts and adverse reproductive outcomes
consistently identify non-differential misclassification of indivi-
1. Address all correspondence to: Dr. Will D. King, Department of
Community Health and Epidemiology, Abramsky Hall, Queen’s
University, Kingston, Ontario, Canada K7L 5H6, Tel.: þ 1-613-5336000 ext 74735. Fax: þ 1-613-5330-6686.
E-mail: [email protected]
Published online 17 March 2004
dual exposure to disinfection byproducts as the main limitation
influencing risk estimation (Reif et al., 1996, Nieuwenhuijsen
et al., 2000; Bove et al., 2002; Arbuckle et al., 2002). Exposure
assessment is complicated by the number of byproducts
formed, seasonal and distribution system variation in byproducts, and the influence of subject water-use behaviors.
Among disinfection byproducts, trihalomethanes (THMs)
generally occur in the highest concentration (Krasner et al.,
1989; Williams et al., 1997), are regulated, and routinely
monitored in many water supplies. As a result, THMs have
been the focus of several epidemiologic investigations of
adverse pregnancy outcomes (Shaw et al., 1991; Kramer
et al., 1992; Bove et al., 1995; Savitz et al., 1995; Gallagher
et al., 1998; Waller et al., 1998; Dodds et al., 1999; Klotz and
Pyrch, 1999). However, many other byproducts are potentially formed including haloacetic acids, acetonitriles, chlorinated ketones, and chlorophenols (Oliver and Lawence,
1979; Krasner et al., 1989). The biologic mechanisms
through which specific disinfection byproducts may confer
a deleterious effect on reproductive outcomes have not been
identified, and it is therefore not clear which byproducts
should be the focus of epidemiologic studies. In this regard,
King et al.
Exposure assessment of disinfection byproducts
exposure to THMs is often considered a proxy for byproduct
exposure and a better understanding of the concomitant
occurrence of other byproducts is necessary in the design and
interpretation of studies.
Assessment of disinfection byproduct exposure in
epidemiologic studies has most often consisted of a
combination of THM measurements from water samples
taken as a part of routine distribution system monitoring
of THM levels (Bove et al., 2002) and maternal residential
information. This approach does not usually take
into account spatial and seasonal variability in the concentration of disinfection byproducts within a distribution
system.
Exposure assessment is further complicated by the role of
subject water-use behavior in determining total exposure.
Meaningful exposure occurs through water ingestion, inhalation, and dermal absorption (Weisel and Jo, 1996; Backer
et al., 2000; Lynberg et al., 2001). Although several studies
have conducted maternal interviews to obtain information on
water consumption and detailed residence information (Shaw
et al., 1991; Savitz et al., 1995; Waller et al., 1998; Klotz and
Pyrch, 1999), most have based exposure assessment solely on
the maternal residence at delivery (Kramer et al., 1992;
Aschengrau et al., 1993; Bove et al., 1995; Kanitz et al.,
1996; Gallagher et al., 1998; Dodds et al., 1999; Kallen and
Robert, 2000; Yang et al., 2000; Jaakkola et al., 2001). Few
studies have accounted for exposure through showering and
bathing (Waller et al., 1998; Klotz and Pyrch, 1999), even
though THM exposure through a 10-min shower may equate
to the exposure received from consumption of 2 l of water
(Weisel and Jo, 1996).
We examined the variation in occurrence of specific
disinfection byproduct compounds across two distinct
geographic regions of Canada in order to provide insight
into methodological strategies and interpretation of findings
in epidemiologic studies. The distribution and correlation of
different classes of disinfection byproducts is presented,
spatial distribution system and temporal variation are
quantified, and the additional contribution of individual
water-use behavior to an exposure metric is examined.
Variation in the occurrence of specific disinfection byproducts was examined through the correlation of total and
specific THMs and haloacetic acids (HAAs) in the two
provinces. Temporal variation was examined through a
comparison with re-sampling 1 year later on a subset of the
original samples. Spatial variation within a distribution
system was examined in one large system where many
household samples were taken.
The contribution of water-use behaviors to a total
exposure metric was evaluated by comparison of subject
household THM level to a total exposure metric. Pearson
correlation coefficients were calculated for the continuous
representation of these variables and a kappa statistic was
used to compare the agreement between exposure quintiles
according to household THM level and the total exposure
metric. The measure of total daily THM exposure incorporated ingestion at home and work, home water-use behavior
(showering and bathing), and household and workplace
THM level (Dodds et al., 2004). This total exposure measure
assumed an equivalency of absorbed dose such that 1 l of
ingested water was equivalent to a 5-min shower (Weisel and
Jo, 1996) and a 15-min bath (Kerger et al., 2000). When use
of a carbon filter (for example, Brita) was indicated, we
applied a 50% reduction in THM intake through cold waterbased drinks (Consumer Reports, 1990). A reduction of
70% was applied to the THM ingestion estimate for boiled
hot water drinks (Kuo et al., 1997; Wu et al., 2001).
Results
Methods
The level of specific THM and HAA parameters was
determined for 360 household water samples from Nova
Scotia (n ¼ 146) and Eastern Ontario (n ¼ 214). The mean
and maximum levels of all disinfection byproducts were
higher in Nova Scotia than in Ontario (Table 1). The mean
level of total THMs was 62.7 mg/l in Nova Scotia samples
compared to 51.1 mg/l in Ontario. Total HAA level was
similar in Nova Scotia (mean ¼ 47.2 mg/l) and Ontario
(mean ¼ 44.2 mg/l). The byproduct level corresponding to
the 90th percentile of the distribution presented for each
byproduct demonstrates that more extreme values of each
compound were observed in Nova Scotia.
Household water samples were collected as part of a case–
control study of stillbirth and abruptio placentae conducted
in Nova Scotia and Eastern Ontario (Dodds et al., 2004). All
water samples were collected from subject residences using
identical procedures and tested at a single laboratory.
Household samples collected by subjects were stored in a
cooler delivered to the analysis laboratory by overnight
courier. A purge and trap gas chromotographic/mass
spectrometer method was used for THM analyses according
to standard methods (EPA Method 624).
Correlation Between Byproducts
Correlation between the three most prevalent THM compounds and HAA compounds is presented in Table 2. In
Nova Scotia, total THMs correlated highly with chloroform
(r ¼ 0.97) and moderately correlated with bromodichloromethane (BDCM, r ¼ 0.63) and total HAAs (r ¼ 0.74). In
Eastern Ontario, the correlation of total THMs with
chloroform was similar to that observed in Nova Scotia,
whereas the correlation of total THMs with BDCM
(r ¼ 0.26) and total HAAs (r ¼ 0.52) was weaker. Total
Journal of Exposure Analysis and Environmental Epidemiology (2004) 14(6)
467
King et al.
Exposure assessment of disinfection byproducts
Table 1. Distribution of THM and HAAs in residential water samples from Nova Scotia and Eastern Ontario
Nova Scotia (n ¼ 146)
Total trihalomethanes (THM)
Chloroform (CHCl3)
Bromodichloromethane (CHCl2Br)
Total haloacetic acids (HAA)
Dichloroacetic acid (DCAA)
Trichloroacetic acid (TCAA)
Bromochloroacetic acid (BCAA)
Eastern Ontario (n ¼ 214)
Mean
Standard deviation
90th percentile
Mean
Standard deviation
90th percentile
62.7
50.9
6.4
47.3
20.5
14.8
2.4
48.9
47.8
6
47.2
28.2
19.4
2.1
98
87
12
85
50
31
5
51.1
42.6
5.1
44.2
22.6
12.3
2
27.6
27.5
3.3
24.1
13.8
12.3
1.9
86
76
10
68
40
25
4
Table 2. Pearson correlation coefficients
Nova Scotia (n ¼ 146)
Total trihalomethanes (THM)
Chloroform (CHCl3)
Bromodichloromethane (BDCM)
Total haloacetic acids (HAA)
Dichloroacetic acid (DCAA)
Trichloroacetic acid (TCAA)
Bromochloroacetic acid (BCAA)
Eastern Ontario (n ¼ 214)
Total trihalomethanes (THM)
Chloroform (CHCl3)
Bromodichloromethane (BDCM)
Total haloacetic acids (HAA)
Dichloroacetic acid (DCAA)
Trichloroacetic acid (TCAA)
Bromochloroacetic acid (BCAA)
THM
CHCl3
BDCM
HAA
DCAA
TCAA
1
0.97
0.63
0.74
0.70
0.65
0.40
1
0.55
0.74
0.72
0.68
0.27
1
0.38
0.30
0.33
0.51
1
0.93
0.85
0.52
1
0.67
0.42
1
0.36
1
1
0.96
0.26
0.52
0.39
0.56
0.07
1
0.11
0.53
0.44
0.56
–0.16
1
0.14
–0.08
0.21
0.45
1
0.85
0.84
0.20
1
0.52
0.02
1
0.09
1
HAAs, trichloroacetic acid (TCAA), and dichloroacetic acid
(DCAA) were highly correlated in both provinces (r 40.85).
Bromochloroacetic acid was moderately correlated with
other HAAs in Nova Scotia (r ¼ 0.36–0.52) and weakly
correlated with other HAAs in Ontario (r ¼ 0.02–0.20).
Spatial Distribution System Variation
The existence of many samples (n ¼ 96) from different
households within one large distribution system facilitated
an evaluation of the spatial distribution of byproduct level
within a single system. The time period when household
water samples were collected (July 2000–August 2002) was
divided into 3-month periods. On average, nine samples were
available per period. The mean THM level within each
period was computed as a measure of the distribution system
average.
THM values for individual household samples ranged
from 13 to 84 mg/l and period distribution system averages
ranged from 30 to 56 mg/l. The THM level measured for
individual households is plotted against the distribution
468
BCAA
system average for the corresponding 3-month time period in
Figure 1. The correlation between household samples and the
period average was 0.63 and on average household samples
differed from the distribution system mean by 8.0 mg/l.
The variance in household samples increased with higher
mean THM levels. A log transformation of the individual
samples was used to improve the fit of the data with respect
to the constant variance assumption. The resulting correlation was similar to that observed in the untransformed data
(r ¼ 0.65).
Sampling 1-Year Post Exposure
In the case–control study which gave rise to this data,
sampling was conducted 1 year beyond a ‘‘critical’’ exposure
time period under the assumption that this time is
representative of relevant subject exposure (Figure 2). In
order to evaluate this approach to water sampling, a second
sample taken 1 year later was obtained for 24 of the study
samples. The 24 locations sampled represent 12 distribution
systems and total THM levels ranging from 14 to 141 mg/l.
Journal of Exposure Analysis and Environmental Epidemiology (2004) 14(6)
King et al.
Exposure assessment of disinfection byproducts
Figure 1. Comparison of household THM level to distribution system
3-month average (Pearson correlation (r) ¼ 0.63).
Contribution of Subject Water Use to the Total Exposure
Metric
We estimated an equivalency between major routes of
exposure (e.g. 1 l water of ingested water was equivalent to
a 5-min shower and a 15-min bath) and estimated total
THM exposure as the sum of the product of these exposures
and the subject’s residence and work THM level. Table 3
presents the mean and quartile distribution for the daily
contribution of each exposure route to the total exposure
metric in micrograms. The percent contribution of each
exposure route to the total metric is calculated as the mean
value for each route divided by the total exposure measure. A
positively skewed distribution was present for each exposure
route. This was most pronounced for water consumption at
work where less than half of subjects had exposures. On
average, exposure through showering accounted for 60% of
total exposure, and tap water consumption at home
accounted for 24% of exposure.
The influence of water-use behaviors on exposure assignment was evaluated through a comparison of household
THM level with the total exposure metric. Table 4 contrasts
the exposure quintiles for subjects based on household THM
level alone with a total exposure metric. The percentage of
the total subjects falling into the cross-classification is
presented. The total agreement of classification into the same
quintile in both measures is represented on the diagonal of
the table. Less than half the subjects (44.7%) were classified
in the same quintile for both exposure measures. The Kappa
value representing agreement beyond chance was 0.31. The
correlation between these two continuous measures was 0.73.
Discussion
Figure 2. Comparison of THM values for samples taken at the same
location 1 year apart (Pearson correlation (r) ¼ 0.90).
The correlation between original and 1-year post samples was
0.90 for total THMs with an average absolute difference of
9.8 mg/l.
Journal of Exposure Analysis and Environmental Epidemiology (2004) 14(6)
Measurement of exposure is a major challenge in studies of
adverse pregnancy outcomes and disinfection byproducts.
This study examined several sources contributing to exposure
misclassification, including the mixture of byproducts
formed, spatial distribution system and temporal variation
in byproduct levels, and the contribution of individual wateruse behaviors. The results of this analysis illustrate important
considerations for the design and interpretation of epidemiologic studies of disinfection byproducts.
Firstly, total THMs do not necessarily provide a good
proxy measure of exposure to specific byproducts, particularly brominated species. A similar level of THMs in the two
provinces studied was associated with a different mix of
specific disinfection byproducts. Recent expert panel reviews
have suggested that brominated species should be a focus of
epidemiologic studies, Environmental Protection Agency
(EPA), 1998; Mills et al., 1998). In Ontario, THMs were
poorly correlated with BDCM, making them a poor proxy
for this exposure. To date, few epidemiologic studies of
adverse birth outcomes have examined specific THM
469
King et al.
Exposure assessment of disinfection byproducts
Table 3. Contribution of different sources of exposure to the total exposure measure for THM exposure (n ¼ 360)
Exposure source
Daily mean (mg) (std. deviation)
Quartile distribution
25th
Consumption (home)
Consumption (work)
Shower
Bath
Total exposure
38.6
10.4
96.3
14.0
(38.8)
(20.4)
(93.7)
(28.6)
159.2 (126.03)
Percent of total exposure
(%)
50th
75th
11.5
0.0
37.8
0.0
28.0
0.0
75.0
3.5
51.2
8.8
128.0
14.2
24
7
60
9
79.5
126.7
202.5
100
Table 4. Percent of subjects (n ¼ 360) according to cross-classification in exposure quintiles of residential THM level and the total exposure metric
Total THM exposure (quintile)
Household THM (quintile)
1
2
3
4
5
1
2
3
4
5
13.3%
4.2%
1.4%
0.8%
0.3%
3.9%
8.9%
4.4%
2.2%
0.6%
2.2%
3.9%
5.0%
5.6%
3.3%
0.3%
3.1%
6.1%
6.1%
4.4%
0.0%
0.6%
3.6%
4.4%
11.4%
Kappa ¼ 0.31 (95% CI 0.25–0.37). Weighted kappa ¼ 0.52 (95% CI 0.46–0.58). Percent agreement ¼ 44.7%. Pearson r (continuous data) ¼ 0.73.
compounds or other byproduct compounds such as haloacetic acids (Nieuwenhuijsen et al., 2000; Bove et al., 2002).
The differences in byproduct levels and mixtures identified in
the two provinces also point to the utility of including unique
geographic areas in an epidemiologic study in terms of
providing meaningful contrasts in exposure.
Secondly, a large distribution system can have significant
spatial variation in disinfection byproduct levels within the
system, making household sampling a priority. Within one
large system serving approximately 500,000 people, the
correlation between the household THM level and corresponding system average for that 3-month period was only
0.63. Most studies to date that have examined the risk of
adverse pregnancy outcomes and exposure to disinfection
byproducts have based exposure assessment on THM
monitoring for a distribution system or on water treatment
and source characteristics (Nieuwenhuijsen et al., 2000; Bove
et al., 2002). This analysis demonstrates that considerable
misclassification of household byproduct levels may have
resulted from this strategy. However, our analysis considered
only one distribution system with total THM levels in the
range of 22–85 mg/l. In areas with temperature variations, the
level of THMs at a residence tends to follow a seasonal
pattern, with the highest levels observed in late summer and
the lowest in winter. Klotz and Pyrch (1999) took advantage
of this pattern in applying a measurement approach in a
retrospective study where household water samples were
470
collected 1 year after the exposure time of interest. We
evaluated this strategy in an analysis of THM measurements
for a small number of locations with re-samples taken 1 year
later. There was a high correlation between THM level in
samples taken 1 year apart, making this an appropriate
measurement strategy in case–control studies of reproductive
outcomes and disinfection byproduct exposures.
Lastly, our results point to the importance of incorporating subject behaviors into exposure metrics. Showering
accounted for 60% of subject total THM exposure, yet few
studies have incorporated showering behavior into exposure
measures (Waller et al., 1998; Klotz and Pyrch, 1999; Dodds
et al., 2004). As more people opt for bottled water or filter
their water, the contribution from showering and bathing will
be greater. Our comparisons of household THM levels to
total exposure measures incorporating ingestion, showering,
and bathing demonstrated that considerable misclassification
of exposure results from not accounting for individual wateruse behaviors. We observed a correlation of 0.73 between
household THM level and our total exposure measure.
Kelsey et al. provide an illustration of the potential
magnitude of attentuation induced by non-differential
misclassification of a continuous exposure measure. For
example, a correlation between an imperfect measure of
exposure and the true exposure of 0.80 would bias an odds
ratio of 2.0 for a difference in exposure of one standard
deviation to 1.74 (Kelsey et al., 1996).
Journal of Exposure Analysis and Environmental Epidemiology (2004) 14(6)
King et al.
Exposure assessment of disinfection byproducts
For all the byproducts evaluated, higher concentrations
were observed in Nova Scotia compared to Ontario samples.
The main factors affecting DBP formation are pH, contact
time, temperature and season, concentration and properties
of natural organic materials, concentration of chlorine and
residual chlorine, and concentration of bromide (Krasner
et al., 1989; Singer, 1993; Pourmogahaddas and Stevens,
1995; Williams et al., 1997). Variation in the bromide ion
concentration of source water is the primary determinant of
the relative percentage of brominated byproducts formed
(Krasner et al., 1989; Williams et al., 1997).
This analysis should be interpreted carefully with regard to
the assumptions of our total exposure measure. On the basis
of the Weisel and Jo (1996) investigation of chloroform
exposure through ingestion, inhalation, and dermal routes,
we assumed that the consumption of 1 l of tap water, a 5-min
shower, and a 15-min bath represented a similar internal
THM dose. However, the biologically effective dose
associated with different routes of exposure is likely to vary
for specific volatile byproducts and health effects of interest.
In addition, our total exposure metric did not incorporate
other potential sources of exposure such as bathing children,
washing dishes, washing clothes, and swimming.
In the study of rare reproductive events such as stillbirth
and congenital anomalies, both cohort and case–control
designs are limited in their ability to quantify individual
exposure to specific disinfection byproduct compounds. The
large number of subjects necessary for cohort studies limits
detailed subject water sampling and incorporation of subject
water-use behavior and case–control studies are limited by
the necessity for retrospective exposure assessment. As a
result, much of the previous research on the effects of
disinfection byproducts on adverse reproductive outcomes
has focused solely on total trihalomethane exposure and
using distribution system levels to represent household
exposure. This analysis demonstrates the importance of
measurement of different classes of disinfection byproducts,
household water sampling, and incorporation of subject
water-use behaviors in exposure assessment.
Acknowledgements
This research was funded by the Toxic Substance Research
Initiative, a research program managed jointly by Health
Canada and Environment Canada, and by the Canadian
Chlorine Coordinating Committee. LD and BAA are
supported by a Clinical Research Scholar Award from
Dalhousie University. We thank the study coordinators
Adelia Trenchard and Celine Zakos. We are grateful for
assistance from the Reproductive Care Program of Nova
Scotia, the Perinatal Partnership Program of Eastern and
Southeastern Ontario, and the Analytical Services Unit of
Queen’s University. We thank all the participating physicians
and hospitals.
Journal of Exposure Analysis and Environmental Epidemiology (2004) 14(6)
References
Arbuckle T.E., Hrudey S.E., Krasner S.W., Nuckols J.R.,
Richardson S.D., Singer P., Mendola P., Dodds L., Weisel C.,
Ashley D.L., Froese K.L., Pegram R.A., Schultz I.R.,
Reif J., Bachand A.M., Benoit F.M., Lynberg M., Poole C.,
and Waller K. Assessing exposure in epidemiologic studies
to disinfection byproducts in drinking water: report from
an international workshop. Environ Health Perspect 2002:
110(Suppl 1): 53–60.
Aschengrau A., Zierler S., and Cohen A. Quality of community
drinking water and the occurrence of late adverse pregnancy
outcomes. Archi Environ Health 1993: 48: 105–113.
Backer L.C., Ashley D.A., Bonin M.A., Cardinali F.L., Kieszak S.M.,
and Wooten J.V. Household exposures to drinking water disinfection
byproducts: whole blood trihalomethane levels. J Expos Anal
Environ Epidemiol 2000: 10: 321–326.
Bove F.J., Fulcomer M.C., Klotz J.B., Esmart J., Dufficy E.M., and
Savrin J.E. Public drinking water contamination and birth outcomes.
Am J Epidemiol 1995: 141(9): 850–862.
Bove F.J., Shim Y., and Zeitz P. Drinking water contaminants and
adverse pregnancy outcomes: a review. Environ Health Perspect
2002: 110(Suppl 1): 61–74.
Consumer Reports. Fit to drink?, January 1990 pp 27–43.
Dodds L., King W.D., Allen A.C., Armson B.A., Fell D., and Nimrod
C. Trihalomethanes in public water supplies and risk of stillbirth.
Epidemiology 2004 (in press).
Dodds L., King W., Woolcott C., and Pole J. Trihalomethanes in public
water supplies and adverse birth outcomes. Epidemiology 1999:
10(3): 233–237.
Environmental Protection Agency (EPA). EPA panel report and
recommendations for conducting epidemiological research on
possible reproductive and developmental effects of exposure to
disinfected drinking water, Research Triangle Park, NC, 1998.
Gallagher M.D., Nuckols J.R., Stallones L., and Savitz D.A. Exposure
to trihalomethanes and adverse pregnancy outcomes. Epidemiology
1998: 9: 484–489.
Jaakkola J.J.K., Magnus P., Skrondal A., Hwang B.-F., Becher G., and
Dybing E. Foetal growth and duration of gestation relative to water
chlorination. Occup Environ Med 2001: 58: 437–442.
Kallen B.A.J., and Robert E. Drinking water chlorination and delivery
outcomeFa registry-based study in Sweden. Reprod Toxicol 2000:
14: 303–309.
Kanitz S., Franco Y., Patrone V., Caltabellotta M., Raffo E., Riggi C.,
Timitilli D., and Ravera G. Association between drinking water
disinfection and somatic parameters at birth. Environ Health Perspect
1996: 104(5): 516–520.
Kelsey J.L., Whittemore A.S., Evans A.S., and Thompson W.D.
Methods in Observational Epidemiology. Oxford University Press,
New York, 1996.
Kerger B.D., Schmidt C.E., and Paustenbach D.J. Assessment of
airborne exposure to trihalomethanes from tap water in residential
showers and baths. Risk Anal 2000: 20: 637–651.
Klotz J.B., and Pyrch L.A. Neural tube defects and drinking water
disinfection byproducts. Epidemiology 1999: 10: 383–390.
Kramer M.D., Lynch C.F., Isacson P., and Hanson J.W. The
association of waterborne chloroform with intrauterine growth
retardation. Epidemiology 1992: 3: 407–413.
Krasner S.W., McGuire M.J., Jacaugelo J.G., Patania N.L., Reagen
K.M., and Aieta E.M. The occurrence of disinfection byproducts in
US drinking water. J Am Water Works Assoc 1989: 81: 41–53.
Kuo H.W., Chiang T.F., Lo I.I., Lai J.S., Chan C.C., and Wand J.D.
VOC concentration in Taiwan’s household drinking water. Science of
the Total Environ 1997: 208: 41–47.
471
King et al.
Lynberg M., Nuckols J.R., Langlois P., Ashley D., Singer P., Mendola
P., Wilkes C., Krapfl H., Miles E., Speight V., Lin B., Small L.,
Miles A., Bonin M., Zeitz P., Tadkod A., Henry J., and Forrester
M.B. Assessing exposure to disinfection byproducts in women of
reproductive age living in Corpus Christi, Texas, and Cobb county,
Georgia: descriptive results and methods. Environ Health Perspect
2001: 109(6): 597–604.
Mills C.J., Bull R.J., Cantor K., Reif J., Hrudey S., and an expert
working group. Health risk of drinking water chlorination byproduct: report of an expert working group. Chronic Diseases Canada
1998: 19(3): 91–102.
Nieuwenhuijsen M.J., Toledano M.B., Eaton N.E., Fawell J., and
Elliott P. Chlorine disinfection byproducts in water and their
association with adverse reproductive outcomes: a review. J Occup
Environ Med 2000: 57: 73–85.
Oliver B.G., and Lawence J. Haloforms in drinking water: a study of
precursors and precursor removal. J Am Water Works Assoc 1979: 7:
161–163.
Pourmogahaddas H., and Stevens A.A. Relationship between trihalomethanes and haloacetic acids with total organic halogen during
chlorination. Water Res 1995: 29: 2059–2062.
Reif J.S., Hatch M.C., Bracken M., Holmes L., Schwetz B., and Singer
P.C. Reproductive and developmental effects of disinfection
byproducts in drinking water. Environ Health Perspect 1996: 104:
1056–1061.
472
Exposure assessment of disinfection byproducts
Savitz D.A., Andrews K.W., and Pasore L.M. Drinking water and
pregnancy outcome in Central North Carolina: source, amount, and
trihalomethane levels. Environ Health Perspect 1995: 103: 592–596.
Shaw G.M., Malcoe L.H., Milea A., and Swan S.H. Chlorinated water
exposures and congenital cardiac anomalies (letter to the editor).
Epidemiology 1991: 2: 459–460.
Singer P.C. Formation and characterization of disinfection byproducts,
In: Cruan G.F. (ed.) Safety of Water Disinfection: Balancing
Microbial Risks. International Life Sciences Inst. Press, Washington
DC, 1993 pp 201–219.
Waller K., Swan S.H., DeLorenze G., and Hopkins B. Trihalomethanes
in drinking water and spontaneous abortion. Epidemiology 1998: 9:
134–140.
Weisel C.P., and Jo W.K. Ingestion, inhalation and dermal exposures to
chloroform and trichloroethene from tap water. Environ Health
Perspect 1996: 104(1): 48–51.
Williams D.T., LeBel G.L., and Benoit F.M. Disinfection byproducts in
Canadian drinking water. Chemosphere 1997: 34(2): 299–316.
Wu W.W., Benjamin M.M., and Korshin G.V. Effects of thermal
treatment on halogenated disinfection byproducts in drinking water.
Water Res 2001: 35(15): 3545–3550.
Yang C.Y., Cheng B.H., Tsai S.S., Wu T.N., Lin M.C., and Lin K.C.
Association between chlorination of drinking water and adverse
pregnancy outcome in Taiwan. Environ Health Perspect 2000: 108:
765–768.
Journal of Exposure Analysis and Environmental Epidemiology (2004) 14(6)