Environmental Research Section A 88, 182 ^187 (2002) doi:10.1006/enrs.2002.4331, available online at http://www/idealibrary.

com on

An Ecologic Study of Nitrate in Municipal Drinking Water and Cancer Incidence in Trnava District, Slovakia
Gabriel Gulis,*,1 Monika Czompolyova,* and James R. Cerhanw,2
*Public Health Institute inTrnava, Limbova 6, 917 09 Trnava, Slovak Republic; and wDepartment of Preventive Medicine and Environmental Health, The University of Iowa College of Medicine, Iowa City, Iowa 52242 Received July 24, 2001

Contamination of drinking water by nitrate is an evolving public health concern since nitrate can undergo endogenous reduction to nitrite, and nitrosation of nitrites can form N-nitroso compounds, which are potent carcinogens. We conducted an ecologic study to determine whether nitrate levels in drinking water were correlated with non-Hodgkin lymphoma and cancers of the digestive and urinary tracts in an agricultural district (Trnava District; population 237,000) of the Slovak Republic. Routinely collected nitrate data (1975^1995) for villages using public water supplies were computerized, and each village was categorized into low (0^10 mg/L), medium (10.1^20 mg/L), or high (20.1^50 mg/L) average levels of total nitrate in drinking water. Observed cases of cancer in each of these villages were ascertained through the district cancer registry for the time period 1986^1995. Standardized incidence ratios (SIRs) and 95% condence intervals (CI) for all cancer and selected cancer sites were calculated by indirect standardization using age- and sexspecic incidence rates from the entire district. For all cancer in women, SIRs increased from villages with low (SIR=0.87; 95% CI 0.72^0.95) to medium (SIR=1.07; 95% CI 1.00^1.13) to high (SIR=1.14; 1.06^1.22) levels of nitrate (P for trend o0.001); there was a similar trend for all cancer in men from low (SIR=0.90; 95% CI 0.81^ 0.99) to medium (SIR=1.08, 95% CI 1.02^1.16), but not for high (SIR=0.94; 0.88^1.02), nitrate levels (P for trend o0.001). This pattern in the SIRs (from low to high nitrate level) was also seen for stomach cancer in women (0.81, 0.94, 1.24; P for trend=0.10), colorectal cancer in women (0.64, 1.11, 1.29; P for trend o0.001) and men (0.77, 0.99, 1.07; P for trend=0.051), and non-Hodgkin lymphoma in women (0.45, 0.90, 1.35; P for trend=0.13) and men (0.25, 1.66, and 1.09; P for trend=0.017). There were no associations for kidney or bladder cancer.
1 Current address: Trnava University, Faculty of Health Care and Social Work, Department of Hygiene and Epidemiology, Hornopotocna 23, 91843 Trnava, Slovak Republic. 2 Current address: Department of Health Sciences Research, Mayo Clinic, 200 First Street SW, Rochester, MN 55905.

These ecologic data support the hypothesis that there is a positive association between nitrate in drinking water and non-Hodgkin lymphoma and colorectal cancer. # 2002 Elsevier Science (USA) Key Words: ecologic studies; drinking water; neoplasms; nitrate; Slovakia.

INTRODUCTION

The increasing contamination of ground water by nitrate, primarily from the widespread use of commercial fertilizers, is an evolving public health concern in many agricultural regions of the world. From the late 1950s through the late 1980s in Slovakia, cereal yields doubled, fertilizer application increased eight-fold, and nitrate concentration in groundwater under agricultural lands doubled (Benes et al., 1989; Trnovec et al., 1996). In Slovakia, drinking water nitrate standards are 50 mg/L total nitrate for adults and 15 mg/L for infants. While nitrate concentration for the majority of public drinking water supplies in Slovakia are below the adult limit, they are frequently over the infant limit. Nitrate per se is apparently nontoxic at levels below the Slovak standard, but the toxicity of nitrate greatly increases when bacteria commonly found in the upper gastrointestinal tract reduce it to nitrite. Nitrite can undergo nitrosation reactions in the gastrointestinal tract and bladder with amines and amides to give rise to N-nitroso compounds (Bruning-Fann and Kaneene, 1993; Forman, 1989; Magee and Barnes, 1967). N-nitroso compounds are some of the most potent known carcinogens (Tricker and Preussman, 1991) and can induce cancers in a variety of organs, including the stomach, colon, bladder, lymphatics, and hematopoietic system in many dierent animal models (Bogovski and Bogovski, 1981; Mirvish et al., 1987). In contrast to the extensive experimental data on these carcinogenic compounds, there is a relative decit of
182

0013 -9351/02 $35.00 r 2002 Elsevier Science (USA) All rights reserved.

NITRATE AND CANCER RISK

183

epidemiologic data addressing the association of nitrate in drinking water and cancer risk. Most epidemiologic studies have focused on gastric cancer, and the results have been extremely mixed (Cantor, 1997; Forman, 1989). Elevated risks have also been reported for cancer of the esophagus, nasopharynx, urinary bladder, prostate and non-Hodgkin lymphoma (NHL) (Cantor, 1997). However, most of the prior studies have used an ecologic design and have correlated cancer mortality, rather than incidence, with nitrate levels in drinking water at the place of residence of the decedent, thus not allowing for a latency period. Public drinking water levels have been nearly continuously monitored in Slovakia over the past several decades by the hygienical service, and there has been a population-based cancer registration system in place since the 1980s. This gave us the opportunity to conduct an ecologic study of municipal nitrate levels and cancer incidence in the Trnava District, a heavily agricultural district in southwestern Slovakia. We a priori focused on the gastrointestinal and urinary tracts since endogenous nitrosation can occur in these sites and experimental and limited epidemiologic data suggest that these may be associated with cancer risk (Cantor, 1997). We also evaluated risk of non-Hodgkin lymphoma since this cancer has also been linked to drinking water nitrate levels (Cantor, 1997).
MATERIALS AND METHODS

The district of Trnava (1389 km2) is located in southwestern Slovakia and has an ethnically homogenous population of over 237,000, of whom approximately 40% live in rural areas. This is an area of intense agricultural production, and the major crops are grain, barley, sugar beet, and corn.The population has been relatively stable with only a small amount of in and out migration over the time frame 1975^1995. The Public Health Institute in Trnava has constructed a computerized database for all public drinking water supplies in the District, many dating back to the implementation of regional groundwater monitoring in the early 1950s in Czechoslovakia (Benes et al., 1989). Groundwater monitoring was extensive and included routine analyses for microbiological and biological quality (e.g., total coliforms, psychrophyl bacteria, mesophyl bacteria, live organisms, dead organisms) and chemical and physical quality (nitrates, nitrites, ammonia cations, calcium, chlorides, absorbance at 245 nm, chlorine, color, taste and odor, turbidity). Since monitoring was part of a regulatory environment, laboratory quality assurance/quality control mechanisms were in place, including periodic

interlaboratory sample validation. Selected water quality data were abstracted and entered into the software program VYDRA, which also included multiple quality control checks to ensure data integrity. For this report, we used nitrate data from 1975 through 1995, which covered all persons using municipal water supplies in the Trnava district. This represented water supply systems serving 60 villages or cities (furthermore termed villages). We excluded all areas that were not on a municipal supply, which was about 20% of the total population of the district. During this time period, approximately eight samples per year were collected; we averaged these values to calculate the nitrate exposure for each of the villages. Obligatory notication of cancer cases and deaths was initiated in Slovakia in the early 1950s, but it was not until 1968 that the registration system was upgraded to cover western Slovakia (including the Trnava district) and structures were implemented to reliably collect detailed and complete data on all cancer patients (Plesko et al., 1993). In 1980, the system was expanded to cover the entire country. Currently, the cancer registry is computerized, uniquely identies each cancer patient, and collects demographic data (age at diagnosis, date of birth, residence), tumor site and morphology using the International Classication of Diseases for Oncology (ICD-O), details of diagnostic procedures, clinical and TNM staging of the cancer, details on treatment, and long-term follow-up of the patient, including cause of death coded according to ICD-9. The Slovak cancer data has been included in Volumes V (1978^1982), VI (1983^1987), and VII (1988^1992) of Cancer Incidence in Five Continents (Parkin et al., 1997) and has contributed to multiple studies of international patterns of cancer (Plesko et al., 1993), attesting to the quality of the data. Each village in the district with a municipal water supply was a priori categorized into one of three levels of average nitrate for the years 1975^1995: r10, 10.1^20, and 420 mg/L total nitrate.These cutpoints were chosen to provide roughly equal exposure groups. The cancer experience (observed number of cancer cases) for each village for the years 1986^1995 was determined using the Trnava District Cancer Registry, supplemented by data from the National Cancer Registry, and was based on residence at the time of cancer diagnosis. Standardized incidence ratios (SIRs) and 95% condence intervals (CI) for all cancer and selected cancer sites were calculated by indirect standardization using age (10 -year) and calendar year strata and sex-specic incidence rates from the entire district (Clayton and Hills, 1993). A trend test in the SIRs across levels of nitrate exposures was also calculated (Clayton and Hills, 1993).

184
RESULTS

GULIS, CZOMPOLYOVA, AND CERHAN

Table 1 presents the distribution of average nitrate levels in the 60 villages served by municipal water supplies. There were no water supplies above the Slovak maximum limit of 50 mg/L of total nitrate. However, 33% of the population on a municipal supply had an average nitrate level over 20 mg/L nitrate. The distribution of people living in urban areas (420,000 inhabitants) versus rural villages across the three exposure categories was relatively homogenous (data not shown). The three larger urban communities of districts Trnava, Hlohovec, and Piestany belong to three dierent categories. For total cancer, there was a positive association between nitrate level and cancer incidence (Table 2). For villages with nitrate levels r10 mg/L, there was a decit of cancer (SIR=0.88; 95% CI 0.82^0.95), while for villages with nitrate levels of 10.1^20 mg/L, there was an excess of cancer (SIR=1.07; 95% CI 1.03^1.13). For the highest level of nitrate exposure (420 mg/L), the excess was slight for all cancer (SIR=1.03), which was not statistically signicant (95% CI 0.97^1.08). However, the trend test in the SIRs across these levels was statistically signicant (Po0.001). The SIRs for villages at the low and middle levels of nitrate were similar for males and females (Table 2), but for villages at the highest level of nitrate exposure there was an excess cancer risk among women (SIR=1.38; 95% CI 1.28^1.47) but not men (SIR=0.94; 95% CI 0.88^1.02). The trend tests for both men and women were each statistically signicant (Po0.001). There was a positive association between nitrate level and cancer of all digestive organs (P for trend o0.001) and non-Hodgkin lymphoma (P for trend o0.05). The positive association for digestive tract cancer was mainly due to colorectal cancer, as the positive association for stomach cancer was much weaker and failed to achieve statistical signicance. For colorectal cancer, the SIRs were 0.71 (95% CI 0.57^ 0.88), 1.05 (95% CI 0.92^1.20), and 1.18 (95% CI 1.04^ 1.34) across increasing levels of nitrate, and the pattern was stronger for women. For non-Hodgkin lym-

phoma, the SIRs were 0.36 (95% CI 0.11^1.11), 1.26 (95% CI 0.82^1.93), and 1.22 (95% CI 0.76^1.96) across nitrate levels, but there were dierent patterns in men and women. Among men, the highest SIR was seen in the villages with nitrate levels of 10.1^20 mg/L (SIR=1.66; 95% CI 0.97^2.87); in villages exposed to 420 mg/L, the SIR was only 1.09 (0.52^2.28). In women, there was a clearer trend in the SIR from low (SIR=0.45; 95% CI 0.11^1.80) to medium (SIR=0.90; 95% CI 0.45^1.81) to high (SIR=1.35; 95% CI 0.72^2.50) nitrate levels. There were no associations between nitrate level and incidence of kidney or bladder cancer.
DISCUSSION

Our a priori hypotheses were that nitrate in municipal water would be positively associated with cancer incidence for the digestive organs (including stomach and colorectal cancer), the urinary tract organs (kidney and bladder), and non-Hodgkin lymphoma. We found positive associations for all digestive organs, which was stronger in women and was mainly accounted for by colorectal cancer. Specically, there were statistically signicant, positive associations for colorectal cancer in women, weaker positive associations for colorectal cancer in men and stomach cancer in women (neither statistically signicant), and no association for stomach cancer in men.We also found a positive association for non-Hodgkin lymphoma in both men and women, although the trend in women did not reach conventional statistical signicance. Finally, we found no evidence for a positive association with either kidney or bladder cancer. There are several strengths to this study. Our exposure data were based on routinely collected nitrate data over a 20 -year exposure period. This allowed us to take into account a latency period. The nitrate levels were determined as part of a regulatory environment and included quality controls. Abstraction and computerization of these data were also under quality control mechanisms. There was a relatively wide range of nitrate exposure for a regulated municipal

TABLE 1
Distribution of Villages (Number, Population) byAverage Level of Nitrate (1975^1995), Trnava District, Slovakia Nitrate level (mg/L) 0^10.0 10.1^20 20.1^50.0 Total Mean (median), levels (mg/L) 4.5 (3.9) 15.5 (15.7) 35.5 (29.7) Number of villages 16 18 26 60 Males Population 21,047 44,133 31,161 96,341 Percentage 21.8% 45.8% 32.3% Females Population 21,769 46,265 33,479 101,513 Percentage 21.4% 45.6% 33.0% Total Population 42,816 90,398 64,640 197,854 Percentage 21.6% 45.7% 32.7%

NITRATE AND CANCER RISK

Non-Hodgkin lymphoma (200+2002) Females 2 4.5 0.45 0.11^ 1.80 8 8.8 0.90 0.45^ 1.81 10 7.4 1.35 0.72^ 2.50 2.3 0.13

185

supply, but a homogenous distribution of nitrate levels in urban versus rural areas, which should limit the potential for urban^rural bias. A further strength was the use of cancer incidence data (rather than mortality data from death certicates) from an established cancer registration system that has been included in Cancer Incidence in Five Continents since 1978. Incidence data, rather than mortality data, are generally more desirable for the study of cancer etiology, in part because mortality data are a function of both incidence and survival. Also, the use of residence at time of diagnosis (which was available in the cancer registry) is preferred to the use of residence at time of death (from death certicates), since the latter may not reect the usual residence of the decedent.The relative stability of the population studied (in terms of both geographic region and age groups) also adds to the validity of study, since residence at time of diagnosis is likely to be highly correlated with residence during the majority of exposure to municipal water. Finally, incidence data from a cancer registry are subject to more quality controls than are the use of death certicate data. The major limitations of the study are the small sample size for some of the cancer sites and the lack of adjustments for potential confounders, which are dicult to incorporate into an ecologic study even if they are known. We also did not have actual water consumption patterns in the communities, which would inuence actual exposure to nitrate and other sources of nitrates including diet, smoking, and certain medications. From a scientic perspective, the ecologic study design is a relatively weak form of evidence, since an association at the ecologic level may not hold at the level of the individual. Another potential limitation is that we excluded persons who were not on a municipal water supply, the very supplies that often have the highest levels of nitrate.While this would not aect the internal validity of our study results, it does limit our ability to fully dene the dose-response relation between nitrate and cancer incidence. Finally, we did not have information on pesticide contamination of the municipal water supplies, which could potentially confound our results. While pesticide use increased in this part of Slovakia during the time frame of the study, two lines of evidence suggest that pesticides are unlikely to explain our results. First, all of the wells serving municipal supplies were at least 50 m deep, and wells at this depth are unlikely to have signicant pesticide contamination in Slovakia. Second, absorbance data, an overall measure of organic contamination in drinking water, remained low over the time frame of the study, arguing against major pesti-

Males Males Females Males Females Males Females Males Females Males Females All Average nitrate level (mg/L) 1975^1995 Males Females All All All All All All

Standardized Incidence Ratios (SIR) and 95% Condence Intervals by Average Level of Nitrate in Municipal Drinking Water Supply for All Cancer and Selected Cancer Subsites, Trnava District, Slovakia, 1986^1995

Kidney (189)

Bladder (188)

Cancer site (ICD-9 Code)

Colon and rectum (153^154)

TABLE 2

Stomach (151)

All Digestive organs (150^159)

10.1^20

420

w2 trend P value

0^10

759 380 379 157 861.5 421.9 438.0 200.4 0.88 0.90 0.87 0.78 0.82^ 0.81^ 0.72^ 0.67^ 0.95 0.99 1.96 0.92 Obs. 1793 876 917 380 Exp. 1668.5 808.4 858.7 379.9 SIR 1.07 1.08 1.07 1.00 95% CI 1.03^ 1.02^ 1.00^ 0.90^ 1.13 1.16 1.13 1.10 Obs. 1499 682 817 395 Exp. 1461.9 722.1 593.4 347.9 SIR 1.03 0.94 1.38 1.13 95% CI 0.97^ 0.88^ 1.28^ 1.03^ 1.08 1.02 1.47 1.25 22.4 12.0 96.2 15.8 o0.001 o0.001 o0.001 o0.001 Obs. Exp. SIR 95% CI

All cancers (140^208)

89 106.6 0.83 0.68^ 1.03 195 202.3 0.96 0.84^ 1.11 198 186.7 1.06 0.92^ 1.21 3.8 0.051

68 93.6 0.73 0.57^ 0.92 185 177.2 1.04 0.90^ 1.20 207 161.8 1.28 1.12^ 1.47 19.9 o0.001

41 26 42.7 24.1 0.96 1.08 0.71^ 0.73^ 1.30 1.59 70 37 80.6 45.6 0.87 0.81 0.69^ 0.59^ 1.10 1.12 81 41 72.7 42.8 1.08 0.96 0.87^ 0.70^ 1.35 1.30 2.4 1.8 0.012 0.18

15 18.5 0.81 0.48^ 1.34 33 35.0 0.94 0.67^ 1.33 40 32.1 1.24 0.91^ 1.70 2.7 0.10

80 113.3 0.71 0.57^ 0.88 226 215.2 1.05 0.92^ 1.20 232 196.6 1.18 1.04^ 1.34 16.7 o0.001

46 59.8 0.77 0.57^ 1.03 113 113.5 0.99 0.82^ 1.19 113 105.0 1.07 0.89^ 1.29 3.8 0.051

34 43 30 53.3 35.4 25.5 0.64 1.22 1.18 0.46^ 0.90^ 0.82^ 0.89 1.64 1.68 113 68 53 101.5 67.2 48.5 1.11 1.01 1.09 0.93^ 0.80^ 0.83^ 1.34 1.28 1.43 119 57 43 92.1 61.5 44.5 1.29 0.93 0.97 1.08^ 0.71^ 0.71^ 1.55 1.20 1.30 16.1 2.0 1.3 o0.001 0.16 0.25

13 9.8 1.32 0.76^ 2.28 15 18.5 0.81 0.49^ 1.34 14 17.2 0.82 0.48^ 1.38 2.3 0.13

18 14 22.2 14.2 0.81 0.99 0.51^ 0.58^ 1.29 1.67 45 29 42.9 27.3 1.05 1.06 0.78^ 0.74^ 1.40 1.53 39 26 37.1 23.5 1.05 1.11 0.77^ 0.75^ 1.44 1.62 1.0 0.4 0.32 0.53

4 8.0 0.50 0.19^ 1.32 16 14.9 1.07 0.66^ 1.75 13 13.0 0.96 0.55^ 1.64 2.1 0.15

3 8.4 0.36 0.11^ 1.11 21 16.7 1.26 0.82^ 1.93 17 13.9 1.22 0.76^ 1.96 5.3 0.021

1 4.0 0.25 0.03^ 1.79 13 7.8 1.66 0.97^ 2.87 7 6.4 1.09 0.52^ 2.28 5.7 0.17

186

GULIS, CZOMPOLYOVA, AND CERHAN

cide contamination in the municipal water supplies of this region. Our nding of a positive association with colorectal cancer has not been previously reported and is in contrast to ecologic data from the United States (Geleperin et al., 1976), Canada (Thouez et al., 1981), Denmark (Jensen, 1982), and Spain (Morales-Suarez-Varela et al., 1995). However, the US and Spanish studies were based on mortality data, and the Canadian study was based on cancer incidence data that was not complete and did not allow for a latency period. A single cohort study reported a lack of a dose-response association between nitrate in municipal drinking water and risk of colon cancer and the presence of a suggestive inverse association with rectal cancer (Weyer et al., 2001). Most (Fandrem et al., 1993; Fraser et al., 1989; Hagmar et al., 1991) but not all (Rafnsson and Gunnardsdottir, 1990) studies of nitrate-based fertilizer workers report no association with colorectal cancer; the single positive study found elevated SMRs for the large intestine (SMR=160; 95% CI 19^578) and the rectum (SMR=164; 95% CI 4^913), but these estimates were based on two and one cases, respectively, and lacked precision. The large intestine is of interest since nitrosation reactions can occur here (Bruning-Fann and Kaneene, 1993), and Nnitroso compounds induce colon cancer in animal models (Eichholzer and Gutzwiller, 1990). Further evaluation of this association is warranted at the individual level. Our lack of compelling data for a positive association between municipal levels of nitrate and stomach cancer incidence adds to an already extremely mixed literature. Of the ecologic studies, positive, null, and inverse results are about equal (Cantor, 1997; Forman, 1989). Of studies evaluating risk in individuals, two reported positive associations (Cuello et al., 1976;Yang et al., 1998) and two reported no association (Rademacher et al., 1992; van Loon et al., 1998). Most (Al-Dabbagh et al., 1986; Fandrem et al., 1993; Hagmar et al., 1991; Rafnsson and Gunnardsdottir, 1990), but not all (Bulbulyan et al., 1996), studies of workers involved in the manufacture of nitrate-based fertilizers have reported no elevation in risk of gastric cancer. The nding of a positive association between municipal water nitrate levels and non-Hodgkin lymphoma incidence is consistent with two ecologic studies from the central United States (Isacson, 1988; Weisenburger, 1993), while a study from Denmark (Jensen, 1982) found no association for lymphatic or hematopoietic cancers (separate results not reported). A single prospective cohort study did not nd an association with NHL (Weyer et al., 2001). In contrast, in a populationbased case^control study (157 cases, 527 controls) con-

ducted by Ward et al. (1996), long-term consumption (1947^1979) of community water in Nebraska (United States) with average nitrate levels  4 mg/L nitrate^ nitrogen (17.7 mg/L nitrate), compared to o1.6 mg/L (7.1 mg/L nitrate), was positively associated with NHL risk (OR=2.0; 95% CI 1.1^3.6), and there was evidence of a dose-response with increasing nitrate levels (P for trend=0.03). The association was similar in men and women, was somewhat stronger among persons consuming lower levels of vitamin C and carotene, and was not inuenced by dietary intake of nitrate. Thus, most of the epidemiologic data support an association between nitrate in drinking water and risk of non-Hodgkin lymphoma. These ndings, coupled with experimental data that show that N-nitrosoureas can induce lymphomas in rats (Mirvish et al., 1990, 1987; Ogiu and Odashima, 1982), suggest that this may be an important research direction to better understand this cancer that is rapidly rising in Western populations. Our nding of no association for bladder cancer is consistent with ecologic studies conducted in Canada (Thouez et al., 1981), France (Leclerc et al., 1991), and Spain (Morales-Suarez-Varela et al., 1995); there are no data available for kidney cancer per se. A single cohort study from the midwestern United States reported no association with kidney cancer and a strong positive association with bladder cancer (Weyer et al., 2001). Follow-up studies of workers involved in the manufacture of nitrogen-based fertilizers have found no association with bladder cancer (Al-Dabbagh et al., 1986; Fraser et al., 1989; Hagmar et al., 1991), although the number of cases in these studies was small. An association between nitrate intake and bladder cancer is biologically plausible (PrestonMartin and Correa, 1989), since under normal circumstances, approximately 70% of orally ingested nitrate is excreted in the urine (Bartholomew and Hill, 1984) and endogenous nitrosation can occur in the bladder (Bruning-Fann and Kaneene, 1993). In addition, nitrosation by-products rapidly appear in the urine after oral ingestion of nitrate in drinking water (Mirvish et al., 1992), and N-nitroso compounds are carcinogenic in the bladder in animal models (Preston-Martin and Correa, 1989). However, the limited human data from ecologic studies do not suggest an association with bladder or kidney cancer. In summary, we found positive associations between municipal levels of nitrate and risk of colorectal cancer and non-Hodgkin lymphoma. Further follow-up of these observations at the individual level is warranted, given the increasing levels of nitrates in the water supplies in the agricultural regions of central Europe and other locations.

NITRATE AND CANCER RISK

187

ACKNOWLEDGMENTS
This research was funded by the National Cancer Institute for Environmental Health Sciences through the University of Iowa Environmental Health Sciences Research Center NIEHS/NIH P30 ES05605. Dr. Cerhan was supported in part by a National Cancer Institute Preventive Oncology Academic Award (K07 CA64220). We thank Mary Jo Janisch for her assistance in preparing the manuscript.

Magee, P. N., and Barnes, J. M. (1967). Carcinogenic nitroso compounds. Adv. Cancer Res. 10, 163^246. Mirvish, S. S., Grandjean, A. C., Moller, H., Fike, S., Maynard, T., Jones, L., Rosinsky, S., and Nie, G. (1992). N-nitrosoproline excretion by rural Nebraskans drinking water of varied nitrate content. Cancer Epidemiol. Biomarkers Prev. 1, 455^461. Mirvish, S. S., Weisenburger, D. D., Joshi, S. S., and Nickols, J. (1990). 2 -Hydroxyethylnitrosourea induction of B cell lymphoma in female Swiss mice. Cancer Lett. 54, 101^106. Mirvish, S. S., Weisenburger, D. D., Salmasi, S., and Kaplan, P. A. (1987). Carcinogenicity of 1-(2-hydroxyethyl)-1-nitrosourea and 3 nitroso-2-oxazolidinone administered in drinking water to male MRC^Wistar rats: Induction of bone, hematopoietic, intestinal, and liver tumors. J. Natl. Cancer Inst. 78, 387^393. Morales-Suarez-Varela, M. M., Llopis-Gonzalez, A., and TejerizoPerez, M. L. (1995). Impact of nitrates in drinking water on cancer mortality in Valencia, Spain. Eur. J. Epidemiol. 11, 15^21. Ogiu, T., and Odashima, S. (1982). Induction of rat leukemias and thymic lymphoma by N-nitrosoureas. Acta Pathol. Jpn. 32, 223^235. Parkin, D. M., Whelan, S. L., Feday, J., and Young, J. (1997).Cancer Incidence in Five Continents, Vol.VII. IARC, Lyon. Plesko, I., Vlasak, V., Kramarova, E., and Obsitnikova, A. (1993). The role of the registry in the study of relation between cancer and environment experiences from Slovakia. Cent. Eur. J. Public Health 1, 19^24. Preston-Martin, S., and Correa, P. (1989). Epidemiological evidence for the role of nitroso compounds in human cancer. Cancer Surv. 8, 459^473. Rademacher, J. J., Young, T. B., and Kanarek, M. S. (1992). Gastric cancer mortality and nitrate levels in Wisconsin drinking water. Arch. Environ. Health 47, 292^294. Rafnsson, V., and Gunnardsdottir, H. (1990). Mortality study of fertiliser manufacturers in Iceland. Br. J. Ind. Med. 47, 721^725. Thouez, J.-P, Beauchamp, Y., and Simard, A. (1981). Cancer and the physicochemicals quality of drinking water in Quebec. Soc. Sci. Med. 15D, 213^223. Tricker, A. R., and Preussmann, R. (1991). Carcinogenic N-nitrosamines in the diet: Occurrence, formation, mechanisms and carcinogenic potential. Mutat. Res. 259, 277^289. Trnovec,T., Kross, B. C., and Ginter, E. (1996). Perspectives on rural environmental health in Central Europe. Int. J. Occup. Environ. Health 2, 125^134. van Loon, A. J. M., Botterweck, A. A. M., Goldbohm, R. A., Brants, H. A. M., van Klaveren, J. D., and van den Brandt, P. A. (1998). Intake of nitrate and nitrite and the risk of gastric cancer: A prospective cohort study. Br. J. Cancer 7, 129^135. Ward, M. H., Mark, S. D., Cantor, K. P.,Weisenburger, D. D., CorreaV|llasenore, A., and Zahm, S. H. (1996). Drinking water nitrate and the risk of non-Hodgkins lymphoma. Epidemiology 7, 465^471. Weisenburger, D. (1993). Potential health consequences of groundwater contamination of nitrates in Nebraska. Nebr. Med. J. 78, 7^10. Weyer, P. J., Cerhan, J. R., Kross, B. C., Hallberg, G. R., Kantamneni, J., Breuer, G., Jones, M. P., Zheng, W., and Lynch, C. F. (2001). Municipal drinking water nitrate level and cancer risk in older women: The Iowa womens health study. Epidemiology 11, 327^338. Yang, C.-Y., Cheng, M.-F.,Tsai, S.-S., and Hsieh,Y.-L. (1998). Calcium, magnesium, and nitrate in drinking water and gastric cancer mortality. Jpn. J. Cancer Res. 89, 124^130.

REFERENCES
Al-Dabbagh, S., Forman, D., Bryson, D., Stratton, I., and Doll, R. (1986). Mortality of nitrate fertiliser workers. Br. J. Ind. Med. 43, 507^515. Bartholomew, B., and Hill, M. J. (1984).The pharmacology of dietary nitrate and the origin of urinary nitrate. Food Chem. Toxicol. 22, 789^795. Benes,V., Pekny,V., Skorepa, J., and Vrba, J. (1989). Impact of diuse nitrate pollution sources on groundwater qualityFSome examples from Czechoslovakia. Environ. Health Perspect. 83, 5^24. Bogovski, P., and Bogovski, S. (1981). Animal species in which Nnitroso compounds induce cancer. Int. J. Cancer 27, 471^474. Bruning-Fann, C. S., and Kaneene, J. B. (1993).The eects of nitrate, nitrite and N-nitroso compounds on human health. Vet. Hum. Toxicol. 35, 521^538. Bulbulyan, M. A., Jourenkova, N. J., Boetta, P., Astashevsky, S.V., Mukeria, A. F., and Zaridze, D. G. (1996). Mortality in a cohort of Russian fertilizer workers. Scand. J.Work Environ. Health 22, 27^33. Cantor, K. P. (1997). Drinking water and cancer. Cancer Causes Control 8, 292^308. Clayton, D., and Hills, M. (1993).Statistical Methods in Epidemiology. Oxford Univ. Press, Oxford. Cuello, C., Correa, P., Haenszel,W., Gordillo, G., Brown, C., Archer, M., and Tannenbaum, S. (1976). Gastric cancer in Columbia I. Cancer risk and suspect environmental agents. J. Natl. Cancer Inst. 57, 1015^1020. Eichholzer, M., and Gutzwiller, F. (1990). Dietary nitrates, nitrites, and N-nitroso compounds and cancer risk: A review of the epidemiologic evidence. Nutr. Rev. 56, 95^105. Fandrem, S. I., Kjuus, H., Andersen, A., and Amlie, E. (1993). Incidence of cancer among workers in a Norwegian nitrate fertiliser plant. Br. J. Ind. Med. 50, 647^652. Forman, D. (1989). Are nitrates a signicant risk factor in human cancer? Cancer Surv. 8, 443^458. Fraser, P., Chilvers, C., Day, M., and Goldblatt, P. (1989). Further results from a census based mortality study of fertiliser manufacturers. Br. J. Ind. Med. 46, 38^42. Geleperin, A., Moses, V. J., and Fox, G. (1976). Nitrate in water supplies and cancer. Ill. Med. J. 149, 251^253. Hagmar, L., Bellander, T., Andersson, C., Linden, K., Attewell, R., and Moller, T. (1991). Cancer morbidity in nitrate fertilizer workers. Int. Arch. Occup. Environ. Health 63, 63^67. Isacson, P. (1988).Proceedings of Technical Workgroup, Agricultural, Occupational and Environmental Health: Policy Strategy for the Future. Iowa City. Jensen, O. M. (1982). Nitrate in drinking water can cancer in northern Jutland, Denmark, with special reference to stomach cancer. Ecotoxicol. Environ. Saf. 9, 258^267. Leclerc, H., V|ncent, P., and Vandevenne, P. (1991). Nitrates in the drinking water and cancer. Ann. Gastroenterol. Hepatol. (Paris) 27, 326^332.