Chlorinated drinking water and cancer
COC statement COC/99/S2 - May 1999
Introduction
1. In the United Kingdom, North America, and many other countries,
chlorination has long been an important part of water treatment, intended
to ensure that drinking-water contains no microbes hazardous to human
health. In the mid-1970s, refinements in techniques of chemical analysis
resulted in the detection in drinking-water of traces of chemicals formed
when organic chemicals (such as those which may occur naturally in rivers,
lakes, reservoirs and other water sources) are subjected to chlorination.
In drinking- water, each of these chlorination byproducts (CBPs) is typically
present at a concentration below 1 part per billion (1 µg/l). Some
however, such as the trihalomethanes (THMs, ie chloroform, bromodichloromethane,
chlorodibromomethane and bromoform), are often present at concentrations
between 10 and 100 µg/l. Numerous CBPs have been identified, but
many have yet to be detected or characterised.
2. Some CBPs, including some of the THMs, are known to be carcinogenic
in laboratory mammals given doses far greater than human intakes from
drinking-water. Some CBPs are genotoxic in test systems, including the
bacterial "Mutagen X" (MX; 3-chloro-4-dichloromethyl-5-hydroxy-2(5H)-furanone).
There have been many epidemiological investigations into the possible
association between chlorination of drinking-water and cancer in humans
and also many experimental studies regarding the mutagenicity and carcinogenicity
of CBPs which have been considered in this statement.
Previous evaluations
3. In 1986, the Department of Health (DH) Committee on Medical Aspects
of the Contamination of Air, Soil and Water (CASW) reviewed the relevant
data on carcinogenicity, mutagenicity and epidemiology, and advised that
there was
no sound reason to conclude that the consumption of the byproducts
of chlorination, in drinking-water which has been treated and chlorinated
according to current practices, increases the risk of cancer in humans.
The effective disinfection of water supplies is clearly of great importance
in maintaining public health. In our opinion, modification of chlorination
processes which have proved effective over many years, or the replacement
of chlorination by other disinfectants, is not required by the available
data on cancer epidemiology, animal carcinogenicity and mutagenicity
in relation to chlorination byproducts in drinking-water.
4. In 1991, the DH Committee on Mutagenicity of Chemicals in Food, Consumer
Products and the Environment (COM) considered research on methods for
concentrating extracts of chlorinated drinking-water, and the mutagenicity
of these extracts, and of MX. The COM concluded that treated drinking-water
itself presents little risk in this regard, and that no further studies
on the mutagenic potential of these compounds were warranted.
5. In 1992, the DH Committee on Carcinogenicity of Chemicals in Food,
Consumer Products and the Environment (COC) evaluated further epidemiological
studies, and advised that
The conclusions of the 1986 CASW meeting were soundly based on the
data available at the time. Many of the studies considered were correlation
surveys which would have been difficult to interpret because of confounding
factors such as other chemicals in the water supply, and there was also
the problem of accurately determining the exposure of the population.
There was nothing in the more recent publications which would lead to
alteration of the 1986 conclusions. The work by Cantor (1) and Lynch
(17) was well conducted. The Cantor study was a case control study which
took account of confounding factors, and showed a weak association between
consumption of chlorinated drinking-water and an increase in bladder
cancer once smoking had been taken into account, but this was insufficient
to alter CASW's 1986 conclusions.
The Committee concluded that the 1986 conclusions of CASW were adequately
founded and that information from subsequent investigations did not
alter those conclusions. With regard to further epidemiological investigations
within the UK the Committee pointed out that it would be very difficult
to take account of consumption of chlorinated water in food, bottled
water and other beverages. The Committee could not recommend that further
epidemiological studies should be undertaken in the UK at the present
time.
6. COC also reviewed a meta-analysis (20) published in July 1992. The
authors estimated an overall relative risk (RR) of 1.15 (1.09-1.20) {95%
confidence interval, used throughout this statement} for all cancer sites
together, with statistically-significant elevated RRs for bladder cancer
(1.21 [1.09-1.34]) and rectal cancer (1.38 [1.01-1.87]) but not for the
other ten categories of cancer which were evaluated. COC considered that
the meta-analysis gave insufficient evidence for increased concern over
the carcinogenic effects of chlorinated drinking-water. It was noted that
no account had been taken of consumption of water other than tap water,
but that in fact bottled water and water used in food was often chlorinated
in the manufacturing plant. COC concluded that the meta-analysis did not
change its conclusions, but added a proviso that
the Committee could not recommend that further epidemiological studies
should be undertaken in the UK at the present time, unless a population
can be found with a distinctive exposure to chlorinated drinking-water.
7. In 1996, COC and COM considered the carcinogenicity and mutagenicity
data on the THMs, and COC advised that
The ratio between the lowest dose level giving rise to a carcinogenic
effect in animals and the likely human exposure level from drinking-water
for each of the four THMs considered by the Committee was in excess
of 10,000. Thus the levels of these THMs in drinking-water in the UK
are unlikely to provide a carcinogenic risk to humans.
New epidemiological studies
8. Twenty further relevant epidemiological studies (2-16, 18, 19, 21-23)
have been published since COC's 1992 evaluation. They include studies
on a wide range of cancers:
Type of Cancer
Reference Number
All Cancers combined
4,23
Brain and Nervous System
3,14,23
Oesophagus
14,23
Stomach
4,13,14,22,23
Liver
14,15,23
Gallbladder and Bile Ducts
14
Pancreas
9, 14-16, 23
Colon
4,7,8,13,14,23
Rectum/ Anus
4,7,8,10,13,14,21,23
Kidney
4,12-14,23
Bladder
2,4,6,10-14,19,21-23
Prostate
14,23
Testis
14
Ovary
4,14,23
Uterus
4,14,23
Breast
4,14,18,23
Lung
4,14,23
Skin
4,14
Soft Tissue
14,15
Thyroid
14
Leukaemia
5,14,15
Hodgkin's Lymphoma
14,15
Non-Hodgkin's Lymphoma
4,14,15
9. A mixture of case-control, cohort and ecological studies has been employed
to investigate the association between chlorinated drinking water and various
cancers. Most of the recent epidemiological studies were carried out in
North America. None were from the United Kingdom. The focus for case control
studies has been cancers at sites implicated in earlier epidemiological
studies, and for which there may be, theoretically, a higher exposure to
agents in drinking water.
10. Those carcinogenicity studies which have been performed on CBPs do not
identify any CBP, or group of CBPs, which appears likely to cause cancer
at these sites at the concentrations found in drinking-water. The Committee
reaffirmed its view that since bottled water products may contain chlorinated
water, it was not possible to identify an unexposed control group. In the
absence of an identified aetiological agent, or a precise means of measurement,
a number of different surrogates of exposure have been employed in these
studies including the following comparisons :
: chlorinated vs non-chlorinated water sources
: duration of time exposed to chlorinated water
: surface vs groundwater sources
: trihalomethane levels (total and individual substances)
: high organic content vs low organic content
: high level of estimated water mutagenicity vs low level
This consequently introduces uncertainty in exposure classification and
makes comparison between studies, and interpretation of individual studies,
more difficult.
11. In addition to these uncertainties, lifetime estimates of actual water
consumption cannot be ascertained with any certainty, and exposure to substances
occurring in drinking water via other routes (ie inhalation, dermal) or
from other sources (eg food) may also not be properly considered. Consequently
cancer epidemiological studies of chlorinated water suffer to a lesser or
greater degree from deficiencies of study design.
12. It is also not uncommon, in those studies where statistically significant
relative risks are observed, for these to be typically in the region of
2 or lower. Consequently, the strength of association between health outcomes
and measures of exposure is considered to be weak, and the elevated risks
may be within the range of uncertainty arising from possible confounding
factors.
13. Of the 20 recent studies, only 4 were particularly well conducted. These
comprised two case control studies dealing solely with bladder cancer, (2,
6) one case control study considering colon and rectal cancers (7) and a
prospective cohort study of postmenopausal women (4) which looked at many
different cancer sites including the bladder, colon and rectum. The remaining
studies were either ecological in nature or had other serious limitations
in design. Overall, however, all studies suffered to some extent from the
difficulty of assessing long term exposure to potential aetiological agents
in chlorinated drinking water. Additionally there was a lack of consistency
of effect across studies dealing with different cancer end points. Many
studies were also not directly comparable as they contained different measures
of assessing exposure to chlorinated drinking water.
Bladder cancer
14. Previous epidemiological studies have suggested associations between
bladder cancer and CBPs, and eleven of the twenty recent studies have
investigated this hypothesis. Five were case-control studies,(2, 6, 10-12,
19) two were cohort studies, (4 14) and four were ecological.(13, 21-23)
Most report some statistically significant elevated relative risks for
groups with the highest estimated duration or level of exposure, but the
associations are generally weak, with relative risks below 2. Exceptions
are found in subgroups in four of the case-control studies, but are not
consistent between studies. Thus, in two studies (2, 6) the relative risk
was confined to male smokers (respectively, odds ratios [ORs] of 2.3 for
more than 60 years of use of chlorinated water, and 3.2 for more than
40 years use of municipal water). This contrasts with another case-control
study (12) which found an elevated OR only in male non-smokers (OR 2.59
for 30 years exposure to drinking-water estimated as "substantially
mutagenic"), and with an earlier large case-control study (1) which
found associations primarily in non-smokers of both sexes. Members noted
that a new ecological study (23) of chlorination of drinking water and
cancer mortality in Taiwan had recently been published but agreed that
such studies were only useful in the generation of hypotheses and not
in respect of the evaluation of risk. A retrospective cohort study in
Finland (14) found an elevated relative risk for women only (1.48 [1.01-2.18])
but, as noted above, the same group's case-control study (12) found an
elevated OR in male smokers only. In another case-control study (10, 11)
the highest ORs (2.28-2.58) were seen in groups with 35 or more years
of unusually high consumption of water with estimated THM levels greater
than 50 µg/l.
15. These recent studies of bladder cancer do not show any consistent
dose-response relationship with estimated exposures to CBPs or THMs.
Colon and rectal cancers
16. Since the 1992 evaluation there have been 7 epidemiological studies
which have examined an association with cancer of the colon and 8 studies
investigating rectal cancer. Of these only two studies were considered
to be particularly well conducted, a case-control study of colon and rectal
cancers (7) and a prospective cohort study in postmenopausal women. (4)
Findings from these two studies were inconsistent; for cancer of the colon,
a moderately strong association with increasing duration of exposure was
found in the case-control study but no significant association was found
in the cohort study; conversely, for rectal cancer, a moderately strong
association was found in the cohort study but not in the case-control
study. Inconsistent findings were also evident in the other reviewed studies
of these sites.
Other sites
17. Studies of the other sites were not considered to be of good quality
and, although some elevated risks were identified, these studies overall
also failed to demonstrate any consistent association.
Conclusion
18. Overall, the further epidemiological studies fail to provide persuasive
evidence of a consistent relationship between chlorinated drinking-water
and cancer. It remains possible that there may be an association between
chlorinated drinking water and cancer which is obscured by problems such
as the difficulty of obtaining an adequate estimate of exposure to chlorination
by-products, misclassification of source of drinking water (including
the use of bottled water), failure to take adequate account of confounding
factors (such as smoking status), and errors arising from non-participation
of subjects.
We therefore consider that efforts to minimise exposure to chlorination
by-products remain appropriate, providing that they do not compromise
the efficiency of disinfection of drinking-water.
REFERENCES
1. Cantor KP, Hoover R, Hartge P, Mason TJ, Silverman DT, Altman R, Austin
DF, Child MA, Key CR, Marrett LD, Myers MH, Narayana AS, Levin LI, Sullivan
JW, Swanson GM, Thomas DB, West DW. Bladder cancer, drinking water source,
and tap water consumption: a case-control study. J Natl Cancer Inst 1987;
79: 1269-1279.
2. Cantor KP, Lynch CF, Hildesheim ME, Dosemeci M, Lubin J, Alavanja M,
Craun G. Drinking water source and chlorination byproducts I. Risk of
bladder cancer. Epidemiology 1998; 9: 21-28.
3. Cantor KP, Lynch CF and Hildesheim M. Chlorinated drinking water and
risk of glioma : a case control study Iowa, USA. Epidemiology 1996; 7(4)
: pS83 (T25).
4. Doyle TJ, Zheng W, Cerhan JR, Hong CP, Sellers TA, Kushi LH, Folsom
AR. The association of drinking water source and chlorination by-products
with cancer incidence among postmenopausal women in Iowa: a prospective
cohort study. Am J Public Health 1997; 87: 1168-1176.
5 Fagliano J, Berry M, Bove F and Burke T. Drinking water contamination
and the incidence of leukaemia: an ecologic study. Amer J Pub Health 1990
; 80 1209-1212
6. Freedman DM, Cantor KP, Lee NL, Chen LS, Lei HH, Ruhl CE, Wang SS.
Bladder cancer and drinking water: a population-based case-control study
in Washington County, Maryland (United States). Cancer Causes and Control
1997; 8: 738-744.
7. Hildesheim ME, Cantor K, Lynch CF, Dosemeci M, Lubin J, Alavanja M,
and Craun G. Drinking water and chlorination byproducts II. Risk of colon
and rectal cancers. Epidemiology 199 8; 9: 29-35.
8. Hoff G, Moen IE, Mowinkel P Rosef O, Nordbro E, Sauar J Vatn MH and
Torgrimsen T. Drinking water and the prevalence of colorectal adenomas
; an epidemiologic study in Telemark, Norway. Eur J Cancer Prev 1992 :
1 423-428.
9 Ijsselmuiden CB, Gaydos C, Feighner B, Novakoski WL, Serwadda D, Caris
LH and Comstock GW . Cancer of the pancreas and drinking water : a population
based case -control study in Washington County, Maryland. Am J Epidemiol
1992 ; 139: 836-42.
10 King W. Great Lakes water and your health. A summary of "Great
Lakes Basin Cancer risk Assessment : A case control study of cancers of
the bladder, colon and rectum" Health Canada publication. December
1995.
11. King WD, Marrett LD. Case-control study of bladder cancer and chlorination
by-products in treated water (Ontario, Canada). Cancer Causes and Control
1996; 7: 596-604.
12. Koivusalo M, Hakulinen T, Vartiainen T, Pukkala E, Jaakkola JJK, Tuomisto
J. Drinking water mutagenicity and urinary tract cancers: a population-based
case-control study in Finland. Am J Epidemiol 1998; 148: 704-12.
13. Koivusalo M, Jaakkola JJK, Vartiainen T, Hakulinen T, Karjalainen
S, Pukkala E, Tuomisto J. Drinking water mutagenicity and gastrointestinal
and urinary tract cancers: an ecological study in Finland. Am J Public
Health 1994; 84: 1223-1228.
14. Koivusalo M, Pukkala E, Vartiainen T, Jaakkola JJK, Hakulinen T. Drinking
water chlorination and cancer - a historical cohort study in Finland.
Cancer Causes and Control 1997; 8: 192-200.
15 Koivusalo M, Vartiainen T, Hakulinen T, Pukkala E and Jaakkola J. Drinking
water mutagenicity and leukaemia, lymphomas and cancers of the liver,
pancreas and soft tissue. Arch Environ Health 1995 ; 50 269-276.
16. Kukkula M and Lofroth G. Chlorinated drinking water and pancreatic
cancer. A population based case-control study. Eur J Pub Health 1997;
7: 297-301.
17. Lynch CF, Woolson RF, O'Gorman T and Cantor KP. Chlorinated drinking-water
and bladder cancer: effect of misclassification on risk estimates. Arch
Environ Health 1989; 44: 252-259.
18. Marcus PM, Savitz DA, Millikan RC and Morgenstern H. Female breast
cancer and trihalomethane levels in drinking water in North Carolina.
Epidemiol 1998; 9: 156-160.
19. McGeehin MA, Reif JS, Becher JC, Mangione EJ. Case-control study of
bladder cancer and water disinfection methods in Colorado. Am J Epidemiol
1993; 138: 492-501.
20. Morris RD, Audet AM, Angelillo IF, Chalmers TC, Mosteller F. Chlorination,
chlorination by-products and cancer: a meta-analysis. Am J Public Health
1992; 82: 955-963.
21. Savrin JE, Cohn PD. Comparison of bladder and rectal cancer incidence
with trihalomethanes in drinking water [Abstract]. Epidemiology 1996;
7(4 Suppl): S63.
22. Suarez-Varela MMM, Gonzales AL, Perez MLT, Caraco EF. Chlorination
of drinking water and cancer incidence. J Environ Pathol Toxicol Oncol
1994; 13:39-41.
You
are here: 