Chlorination of drinking water and cancer mortality in Taiwan.
ABSTRACT Chlorination has been the major strategy for disinfection of drinking water in Taiwan. An ecologic epidemiological study design was used to examine whether chlorination of drinking water was associated with cancer risks. A "chlorinating municipality" (CHM) was defined as one in which more than 90% of the municipality population was served by the chlorinated water while an "nonchlorinating municipality" (NCHM) was one in which less than 5% of the municipality population was served by chlorinated water. Age-adjusted mortality rates for cancer during 1982-1991 among the 14 CHMs were compared to rates among the 14 matched NCHMs with similar urbanization level and sociodemographic characteristics. The results of this study suggest a positive association between consumption of chlorinating drinking water and cancer of the rectum, lung, bladder, and kidney. Although these findings must be interpreted with caution because of limitations in the ecological study design, their public health significance should not be disregarded because chlorination of water is so widely practiced in Taiwan.
- [show abstract] [hide abstract]
ABSTRACT: The incidence of head and neck, oesophagus and lung cancer between 1981 and 1985 was studied in Eastern Austria for an urban-rural division. In males, rural rates of oral cavity, oropharynx and oesophagus tumours were higher than urban rates. For lung tumours, urban rates slightly exceeded rural rates. In females, the incidence of oral cavity, oropharynx, larynx, hypopharynx, oesophagus and lung cancer showed an urban predominance, steepest for head and neck and oesophagus cancers. Cancer of the oral cavity, pharynx, larynx, oesophagus and lung had a high male preponderance.European Journal of Cancer 02/1991; 27(1):83-5. · 5.06 Impact Factor
- [show abstract] [hide abstract]
ABSTRACT: To examine whether chlorination of drinking water was associated with cancer of the digestive or other organs, an ecological epidemiological study using nationwide incidence data from the Cancer Registry of Norway was carried out. On two geographical levels (counties and municipalities), both for men and women, chlorination of drinking water was associated with an increased incidence of cancer of the colon and rectum. After adjusting for potential confounding variables, also measured on a geographical basis, the associations were still significant at the county level (adjusted for population density, income, education, fat and fibre intake etc.), but not at the municipality level. The observed associations are weak, chlorination being associated with a 20-40% increase in colorectal cancer rates. Due to inherent methodological limitations in ecological studies like the present one, causal interpretations should be made with great care. Thus, although the results give some support to the hypothesis that drinking water chlorination is associated with colorectal cancer, they do not provide strong evidence of a causal relationship.International Journal of Epidemiology 03/1992; 21(1):6-15. · 6.98 Impact Factor
- [show abstract] [hide abstract]
ABSTRACT: Individual epidemiological investigations into the association between chlorination by-products in drinking water and cancer have been suggestive but inconclusive. Enough studies exist to provide the basis for a meaningful meta-analysis. An extensive literature search was performed to identify pertinent case-control studies and cohort studies. Consumption of chlorinated water, surface water, or water with high levels of chloroform was used as a surrogate for exposure to chlorination by-products. Relative risk estimates were abstracted from the individual studies and pooled. A simple meta-analysis of all cancer sites yielded a relative risk estimate for exposure to chlorination by-products of 1.15 (95% CI: 1.09, 1.20). Pooled relative risk estimates for organ-specific neoplasms were 1.21 (95% CI: 1.09, 1.34) for bladder cancer and 1.38 (95% CI: 1.01, 1.87) for rectal cancer. When studies that adjusted for potential confounders were pooled separately, estimates of relative risks did not change substantially. The results of this meta-analysis suggest a positive association between consumption of chlorination by-products in drinking water and bladder and rectal cancer in humans.American Journal of Public Health 08/1992; 82(7):955-63. · 3.93 Impact Factor
ENVIRONMENTAL RESEARCH, SECTION A 78, 1—6 (1998)
ARTICLE NO. ER973823
Chlorination of Drinking Water and Cancer Mortality in Taiwan
Chun-Yuh Yang,* Hui-Fen Chiu,- Ming-Fen Cheng,‡ and Shang-Shyue Tsai‡
*School of Public Health, Kaohsiung Medical College, Kaohsiung, Taiwan; -Department of Pharmacology, Kaohsiung Medical College,
Kaohsiung, Taiwan; and ‡School of Public Health, Kaohsiung Medical College, Kaohsiung, Taiwan
Received J uly 30, 1997
Chlorination has been the major strategy for dis-
infection of drinking water in T aiwan. An ecologic
epidemiological study design was used to examine
whether chlorination of drinking water was asso-
ciated with cancer risks. A ` ` chlorinating municipal-
ity'' (CHM) was de®ned as one in which more than
90% of the municipality population was served by
the chlorinated water while an ` ` nonchlorinating
municipality'' (NCHM) was one in which less than
5% of the municipality population was served by
chlorinated water. Age-adjusted mortality rates for
cancer during 1982–1991 among the 14 CHMs were
compared to rates among the 14 matched NCHMs
with similar urbanization level and sociodemo-
graphic characteristics. T he results of this study
suggest a positive association between consump-
tion of chlorinating drinking water and cancer of
the rectum, lung, bladder, and kidney. Although
these ®ndings must be interpreted with caution be-
cause of limitations in the ecological study design,
their public health signi®cance should not be disre-
garded because chlorination of water is so widely
practiced in T aiwan.
? 1998 Academic Press
K ey Words: chlorination; drinking water; cancer;
INT R ODUCT ION
The economy and effectiveness of chlorine in kill-
ing waterborne organisms has made water chlorina-
tion a tremendous public health success worldwide.
Chlorination has been the major strategy for the
disinfection of drinking water in Taiwan. It is cur-
rently added toapproximately 75.8% of the nation's
A number of epidemiologic studies have been con-
ducted which examine the possible associations be-
tween consumption of chlorinated drinking water
and cancer mortality or incidence (Page et al., 1976;
Kuzma et al., 1977; Cantor et al., 1978,1987;Wilkins
and Comstock, 1981; Young et al., 1981,1987; Bean
et al., 1982; Gottlieb et al., 1982; Lawrence et al.,
1984; Carpenter and Beresford, 1986; Cech et al.,
1987; Zierler et al., 1988; Flaten, 1992; Morris et al.,
1992; McGeehin et al., 1993). Thesestudies consider
a wide range of populations and regions but have
been mainly carried out in the U.S. Most studies
have shown positive associations between chlorin-
ated drinking water and colorectal and bladder can-
cer. This has been attributed to trihalomethanes
(THMs), a carcinogenic organic halogenated byprod-
uct of water chlorination (Reuber, 1979; Dunnick
and Melnick, 1993; IARC, 1987).
The present study was carried out because few
epidemiological studies havebeen conducted outside
the U.S. (Carpenter and Beresford, 1986; Flaten,
1992). The study reported here was designed to ex-
plorefurther theassociation between cancer mortal-
ity and the use of chlorinated water.
MAT E R IAL S AND ME T HODS
Selection of Study Municipalities
Taiwan is divided into 361 administrative dis-
tricts, which will bereferred toherein as municipali-
ties. They are the units that were subjected to
statistical analysis. Excluded from theanalysis were
30 aboriginal townships and 9 islets which had dif-
ferent lifestyles and living environments and Taipei
city (including 12 municipalities) because of its dis-
tinctly more urban character and higher population
than other Taiwan municipalities. This elimination
of unsuitable municipalities left 310 municipalities
for the analysis.
The current Taiwan water system is rather
simple. Residents obtain their drinking water either
from the public drinking water supply systems ser-
ved by the Taiwan Water Supply Corporation
(TWSC) or from nonmunicipal sources. The major
sources of municipal water supplies are almost all
Copyright ? 1998 by Academic Press
All rights of reproduction in any form reserved.
This article is copyright protected and is for viewing only.
To download, a copy must be purchased online.
Edited by Foxit Reader
Copyright(C) by Foxit Software Company,2005-2007
For Evaluation Only.
surface waters and are often treated with chlorine.
The nonmunicipal sources are mainly privately
owned wells (groundwater) and are often un-
In this study, an individual municipality was clas-
si® ed as a chlorinating municipality (CHM) if more
than 90% of the municipality population was served
by chlorinated water. In all, 156 of the 310 munici-
palities satis® ed this criterion. A nonchlorinating
municipality (NCHM) was de® ned as one in which
less than 5% of the municipality population was
served by chlorinated water, i.e., more than 95% of
the residents of these 15 municipalities obtained
their drinking water from unchlorinatedwater sour-
ces. In all, 15 municipalities satis® ed this de® nition.
These 15 NCHMs provide a unique opportunity to
investigate the issue of chlorination.
Mortality from cancer has been found to vary be-
tween regions in Taiwan (DOH/ROC, 1993), and
several studies have examined the variation in can-
cer rates across urbanization gradients (Greenberg,
1983; Miller et al., 1987; Swoboda and Friedal,
1991). Totake intoaccount the possibleconfounding
effect resulting from differing urbanization levels,
the urbanization level of the nonchlorinating mu-
nicipalities should be the same as that of the
chlorinating municipalities. The assignment
urbanization levels was based on the urban—rural
classi® cation of Tzeng and Wu (1986). This urbaniz-
ation index has been applied toour previous studies
(Yang et al., 1996,1997a). Each municipality in
Taiwan (n"361) was assigned to an urbanization
category from 1 to8. Municipalities with thehighest
urbanization score, such as the Taipei metropolitan
area, were classi® ed in category 1, while mountain-
ous areas with the lowest score were assigned to
More speci® cally, each NCHM was matched with
one CHM with the same urbanization level. Among
the 15 NCHMs, one was excluded as there was no
appropriate municipality for matching. If a NCHM
had more than one appropriate CHM for matching,
a random sampling method was used to select one
Data Collection and Mortality Analysis
Information concerning both thenumber of deaths
and the midyear population by sex, age, and calen-
dar year during 1982—1991 was obtained from the
Bureau of Vital Statistics of the Taiwan Provincial
Department of Health, which is in charge of the
death registration system in Taiwan. The Interna-
tional Classi® cation of Disease, Injury, and Causes
of Death [9th revisions (ICD-9)] is used to code the
cause of death, and the system has been completely
computerized since 1972.
100,000 population were calculated for males and
femalesfor each municipality of thetwochlorination
groups, 1982—1991. As the age distribution was not
similar among the municipalities of the two chlori-
nation groups, theage-standardizedrates werecom-
puted by the direct method, using the world
population in 1976 as the standard population
(Waterhouse et al., 1976).
The age-standardized rates for various cancer
sites were calculated ® rst among residents in the
individual chlorinating and nonchlorinating munici-
pality. The mean age-standardized rates in all
CHMs and all NCHMs were then calculated. The
ratios of the mean age-standardized mortality rates
from various malignant neoplasms for CHMs and
the mean rates from all NCHMs represent the rela-
tive cancer risk in the CHMs compared to the
NCHMs (standardizedrateratio, SRR). For theSRR
the null hypothesis (H?: SRR"1) was tested, and
the 95% con® dence interval of the SRR was cal-
culated according tothe method of Rothman (1986).
R E SUL T S
The sociodemographic characteristics of the CHMs
and NCHMs weregenerally similar except for a high-
er population and a higher percentage of population
using the chlorinated water among CHMs (Table 1).
Average annual age-adjusted cancer mortality
rates per 100,000 population and ratios of the age-
adjusted mortality rates (SRR) for 1982—1991 by
cancer site and sex for the CHMs and NCHMs are
listed in Tables 2 and 3. A signi® cant relationship
was observed for cancers of the rectum, lung, blad-
der, and kidney in both males and females and for
cancer of the liver in males. Mortality rates for can-
cers of the esophagus, stomach, colon, pancreas,
prostate, brain, breast, cervix uteri and uterus, and
ovary werenot associatedwith theuseof chlorinated
Themajor ® ndings of this study suggest that there
was a signi® cant association between municipal
(chlorinated) drinking water in Taiwan and mortal-
ity from certain cancers. Before evaluating the
YANG ET AL.
T ABL E 1
Some Characteristics of T wo Groups of T aiwan Munici-
palities, Grouped According to Chlorination Practice
Total population (1989)
Population density (per Km2)
Percentage of population served
by chlorinating water
cProfessional, technical, administrative,superintendents, cleri-
cal, sales, and service workers as a percentage of total employed
(aged 15 and over) population.
dProducers, transportation operators, and laborers as a per-
centage of total employed population.
eFarmers, loggers, grazers, ® shermen, hunters, and related
workers as a percentage of total employed population.
meaning of these ® ndings, consideration ® rst must
be given to the design of the study.
Mortality data have been widely used togenerate
epidemiologic hypotheses, despite their inherent
limitations (Morgenstern, 1982). The completeness
and accuracy of the death registration should be
evaluated before any conclusion based on the mor-
tality analysis is made. In Taiwan, it is mandatory to
register all deaths at local household registration
T ABL E 2
Mean Annual Age-Adjusted Mortality R ates per 100,000
Population and R atios of Age-Adjusted Mortality R ates
(SR R ), 1982–1991, among Males in Chlorinating Municipal-
ities (CHMs) to T hose in Nonchlorinating Municipalities
(NCHMs) by Cancer Site
14 CHMs14 NCHMsSRR (95% CI)a
All sites (140—208)
a95% con® dence interval.
T ABL E 3
Mean Annual Age-Adjusted Mortality R ates per 100,000
Population and R atios of Age-Adjusted Mortality R ates
(SR R ), 1982–1991, among F emales in Chlorinating Munici-
palities (CHMs) to those in Nonchlorinating Municipali-
ties (NCHMs) by Cancer Site
Cancer site (ICD 9) 14 CHMs14 NCHMs SRR (95% CI)a
All sites (140—208)
a95% con® dence interval.
of® ces andsincethehouseholdregistration informa-
tion is veri® ed annually through a door-to-door sur-
vey, the death registration is very
Although causes of death may be misdiagnosed or
misclassi® ed this problem has been minimized
through an improvement in theveri® cation andclas-
si® cation of causes of death in Taiwan since 1972.
Furthermore, malignant neoplasms have been re-
ported to be one of the most unequivocally classi® ed
causes of death in Taiwan (Chen and Wang, 1990).
Because of their fatal outcome, it is believed that in
recent years in Taiwan, all cancer cases from the
studied municipalities have had access to medical
care regardless of geographical location. The com-
pleteness and accuracy of death certi® cate registra-
tion is thus believed to be comparable.
Problems inherent in aggregate studies, including
the ` ` ecologic fallacy,'' are well known. However, the
degree towhich this fallacy is a problem varies from
study tostudy. It was a distinct problem in the early
water-cancer studies when associations between
consumption of surface water and rates of cancer
were looked for by comparing the proportion of
county or parish residents supplied by surfacewater
sources with cancer mortality rates for the total
county or parish. In our study, theeffectsof drinking
water chlorination on cancer mortality were investi-
gated using an ` ` extreme points contrast'' in order
to maximize the inherent power of the design
WATER CHLORINATION AND CANCER
(Miettinen, 1985; Rothman, 1986). This method was
applied to our study of cancer mortality and resi-
dence near petrochemical industries (Yang et al.,
1997b). The percentage of the population served by
chlorinated water in the CHMs and NCHMs were
96.1 and 1.5%, respectively. Also, the municipalities
selected for this study were rural municipalities and
it is likely to preclude much of the resident's budget
being allocated to bottled water, thus reducing the
likelihood of water coming from a source other than
the home. In line with this assumption, we expect
that persons living in the CHMs do in fact drink
water from the public supply and residents living in
NCHMs doin fact drink water from theprivatewells
(nonchlorinated water). Thus, the importance of one
problem associated with ecologic analyses is reduced.
Migration is especially important in cancer stud-
ies, since the latency period is probably very long.
The migration that does occur will result in a reduc-
tion of the strength of the geographical association
between thediseaseand thestudiedfactor (Polissar,
1980;Bentham, 1988). Taiwan's population is rather
stationary compared with those of most other West-
ern countries. It was reportedthat morethan 90% of
rural residents lived in the same municipality in
which they were born for their entire life (Wu et al.,
1989). Thus, the migration problem is probably
Sincethemeasureof effect in this study is mortal-
ity rather than incidence, migration during the in-
terval between cancer diagnosisand death must also
be considered. During this period, the cancer diag-
nosis may in¯uence a decision to migrate and pos-
sibly introduce bias. If there is a different trend
toward migration between the CHMs and NCHMs
due toproximity tomedical care, for example, a spu-
rious association between chlorinated water and
cancer death would result. Since each NCHM was
matched with one CHM with the same urbanization
level, this possibility should be minimized.
Potential exposure to industrial pollution for the
population not necessarily working in the cancer
risk industrial plants but living nearby may be an-
other confounding factor. In this study, we used the
percentage of a municipality's total population em-
ployed in the chemical and petrochemical industries
as an indicator of a resident's exposure to air emis-
sions from industrial plants (Yang et al., 1997b). The
workers employed in chemical and petrochemical
industrial plants constituted only 0.36% of the
CHMs' population, while for the NCHMs this value
was 0.39% (MOE/ROC, 1989). This result suggests
that industrial pollution was unlikely to have an
effect on cancer mortality.
The bladder and the rectum both serve a similar
physiological function, storing concentrated ex-
cretory products. One might speculate that the epi-
thelial tissueat both sites is exposedtohigh levels of
chlorination by-products and is therefore at in-
creased risk for the development of neoplasia. Our
study is in accordance with many past studies (Page
et al., 1976; Cantor et al., 1978, 1987; Kuzma et al.,
1977; Moriris et al., 1992; Flaten, 1992; McGeehin et
al., 1993). The lung is a biologically plausible target
organ, since it is a major excretory route of ingested
chloroform (Fry et al., 1972) and a site of consider-
able enzymatic metabolism of toxic compounds
(Becker, 1975). Our ® nding is consistent with pre-
vious studies (Cantor et al., 1978; J olley et al., 1978).
Kidney and liver have been suggested as target or-
gans on the basis of experimental animal study
(Reuber, 1979), and cancer of these organs has been
reported tobe associated with the use of chlorinated
water (Wilkins et al., 1979). The same holds for
cancer of the kidney in both males and females in
Taiwan and cancer of the liver for males. It seems
biologically implausible for chlorinated by-products
liketrihalomethanes toaffect cancer risk for one sex
only. Also, a sex-speci® ceffect in this direction is not
consistent with the hypothesis that women may be
more routinely exposed to domestic water sources
than men. Alcohol drinking has been found to be
associatedwith liver cancer (Yu et al., 1983; Oshima
et al., 1984; Tsukuma et al., 1990; Chen et al., 1991).
There is unfortunately no information available on
alcohol consumption patterns for individual munici-
palities. If alcohol drinking were more prominent in
the CHMs, one would expect liver cancer excesses in
both male and female residents. Our results, how-
ever, indicate that excess liver cancers were re-
stricted to men. Since the CHMs and NCHMs were
reasonably homogeneous for several socioeconomic
indicators, there is no reason to expect sex differ-
ences in alcohol drinking between theCHMs andthe
NCHMs. Therefore, the possibility that this is
a chance ® nding should be considered.
The associations between exposure to chlorinated
water and mortality from rectum, lung, bladder, and
kidney cancer for both sexes were signi® cant in this
study. These results were not readily explained by
confounding due to degree of urbanization, socio-
demographic characteristics, or industrial pollution.
The most important potential confounders not ad-
justed for in this study are diet and smoking. If
consumption of a diet high in fat and low in ® ber or
smoking rates were associated with consumption of
chlorinatedwater, onecould arguethat theobserved
association is confounded by dietary factors or
YANG ET AL.
smoking. There is unfortunately no information
available on the smoking and dietary patterns for
individual study municipalities. Since the CHMs
and NCHMs were reasonably homogeneous for sev-
eral socioeconomic indicators, this possibility would
have tended to be diminished. In addition, if diet
were responsible for the observed association with
rectal cancer, we would expect tosee the same asso-
ciation with colon cancer (Morris et al., 1992). Also,
there is no reason to expect differences in smoking
patterns between CHMs and NCHMs in the present
study. Therefore, the marked differences in these
associationstend torefutethecontention that smok-
ing and diet are explanatory factors for observed
In conclusion, the results give some support to
indications from other epidemiological studies that
chlorination of drinking water may be associated
with cancer of therectum, lung, bladder, andkidney.
Due to inherent methodological limitations in eco-
logical studies like the present one, the results could
not provide suf® cient evidence to establish a causal
relationship.However, thepublichealth signi® cance
of a cancer risk associated with consumption of
chlorinated drinking water may be substantial. Our
® ndings are in no way intended to suggest that the
disinfection of drinking water with chlorine should
be abandoned. However, it should not be forgotten
that the primary public health concern for drinking
water supplies is still waterborne infectious disease
transmission, against which chlorine provides very
effectiveprotection.Therefore, these® ndings should
provide an impetus to identify, develop, and imple-
ment disinfection strategies that are not associated
with adverse health effects.
ACK NOWL E DGME NT S
This study was partly supported by a grant from the National
Science Council, Executive Yuan, Taiwan (NSC-87-2314-B-037-
R E F E R E NCE S
Bean, J . A., Isacson, P., andHausler, W. J . (1982). Drinking water
and cancer incidencein Iowa. I. Trends and incidenceby source
of drinking water and size of municipality. Am. J . Epidemiol.
Becker, F. (1975). ` ` Cancer 1, Etiology: Chemical and Physical
Carcinogenesis.'' Plenum, New York.
Bentham, G. (1988). Migration and morbidity: Implications for
geographical studies of disease. Soc. Sci. Med. 26, 49—54.
Breslow, N. E., and Day, N. E. (1980). ` ` Statistical Methods in
Cancer Research: The Analysis of Case-Control Studies.'' Inter-
national Agency for Research on Cancer, Lyon.
Cantor, K. P., Hoover, R., Mason, T. J ., and McCabe, L. J . (1978).
Association of cancer mortality with halomethanes in drinking
water. J . Natl. Cancer Inst. 61, 979—985.
Cantor, K. P., Hoover, R., and Hartge, P. (1987). Bladder cancer,
drinking water sources, and tap water consumption: A case-
control study. J . Natl. Cancer Inst. 79, 1269—1279.
Carpenter, L. M., and Beresford, S. A. A. (1986). Cancer mortality
and type of water source: Findings from a study in the UK. Int.
J . Epidemiol. 15, 312—319.
Cech, I., Holguin, A. H., Littell, A. S., Henry, J . P., and O'Connell,
J . (1987). Health signi® cance of chlorination byproducts in
drinking water: The Houston experience. Int. J . Epidemiol. 16,
Chen, C. J ., Liang, K. Y., Chang, A. S., and Chang, Y. C. (1991).
Effects of hepatitis B virus, alcohol drinking, cigarette smoking
and familial tendency on hepatocellular carcinoma. Hepatology
Chen, C. J ., and Wang, C. J . (1990). Ecological correlation be-
tween arsenic level in well water and age-adjusted mortality
from malignant neoplasms. Cancer Res. 50, 5470—5474.
Department of Health, Taiwan (ROC) (1993). ` ` The Atlas of Can-
cer Mortality in Taiwan.'' Department of Health, Taipei.
Dunnick, J . K., and Melnick, R. L. (1993). Assessment of the
carcinogenic potential of chlorinatedwater: Experimental stud-
ies of chlorine, chloramine, and trihalomethanes. J . Natl. Can-
cer Inst. 85, 817—822.
Flaten, T. P. (1992). Chlorination of drinking water and cancer
incidence in Norway. Int. J . Epidemiol. 21, 6—15.
Fry, B. J ., Taylor, T., and Hathway, D. E. (1972). Pulmonary
elimination of chloroform and its metabolites in man. Arch. Int.
Pharmacodyn. Ther. 196, 98—111.
Gladen, B., and Rogan, W. (1979). Misclassi® cation and the de-
sign of environmental studies.Am. J . Epidemiol. 109, 607— 616.
Gottlieb, M. S., Carr, J . K., and Clarkson, J . R. (1982). Drinking
water and cancer in Louisiana: A retrospectivemortality study.
Am. J . Epidemiol. 116, 652—667.
Greenberg, M. R. (1983) ` ` UrbanizationandCancer Mortality:The
UnitedStatesExperience, 1950—1970. Oxford Univ. Press, New
International Agency for Research on Cancer (IARC) (1987).
` ` Monograph on Chloroform.'' International Agency for Re-
search on Cancer (IARC), Lyon.
J olly, R. L., Gorchev, H., and Hamilton, D. H. (1978). ` ` Water
Chlorination: Environmental Impact and Health Effects,''
Vol. 2. Ann Arbor Science, Ann Arbor.
Kuzma, R. J ., Kuzema, C. M., and Buncher, C. R. (1977). Ohio
drinking water source and cancer rates. Am. J . Public Health
Lawrence, C. E., Taylor, P. R., Trock, B. J ., and Reilly, A. A.
(1984). Trihalomethanes in drinking water and human colorec-
tol cancer. J . Natl. Cancer Inst. 72, 563—568.
McGeehin, M. A., Reif, J . S., Becher, J . C., Mangione, E. J . (1993).
Case-control study of bladder cancer and water disinfection
methods in Colorado. Am. J . Epidemiol. 138, 492—501.
Miller, M. K., Stokes, C. S., and Clifford, W. B. (1987). A compari-
son of the rural-urban mortality differential for deaths from all
causes, cardiovascular disease and cancer. J . Rural Health 3,
Miettinen, O. S. (1985). ` ` Theoretical Epidemiology.'' Wiley, New
WATER CHLORINATION AND CANCER