Birth defects. The effects of exposure to nitrates in drinking water on the incidence of birth defects have been evaluated in several epidemiologic studies (Table 1) (Arbuckle et al. 1988; Cedergren et al. 2002; Croen et al. 2001; Fan and Steinberg 1996; Fan et al. 1987). However, the results from epidemiologic studies addressing this topic are equivocal.
In a case-control study of Mexican-American women, Brender et al. (2004) examined nitrosatable drug exposure and the occurrence of neural tube defects (NTDs) in relation to dietary nitrites and nitrates. They examined 184 cases of NTD-affected pregnancies from the Texas Neural Tube Defects Projects and 225 women with normal live births. All participants were interviewed to obtain detailed dietary information, periconceptional medication use, and drinking water source. The water sources of 110 women (43 cases and 67 controls) were also tested for nitrate. Nitrosatable drug use was reported as a risk factor for having an NTD-affected pregnancy [odds ratio (OR) = 2.7; 95% confidence interval (CI), 1.4-5.3]. Among those who had their water tested for nitrate, drinking water nitrate level ≥ 3.5 mg/L nitrate-N was associated with having an NTD-affected pregnancy (OR = 1.9; 95% CI, 0.8-4.9). The risk estimate increased drastically (OR = 14; 95% CI, 1.7-660) for women who took nitrosatable drugs and had nitrate levels ≥ 3.5 mg/L nitrate-N in their drinking water source. The authors concluded that because the level of nitrate in the water sampled was relatively low, and women were not asked about frequency and amount of water consumed, the amount of nitrate in the water directly contributed to the increased risk observed among women who used nitrosatable drugs (Brender et al. 2004).
A study of 71,978 infants born from 1982 through 1996 was conducted in a Swedish county served by 80 municipal water systems (Cedergren et al. 2002). The study assessed the possible association between mothers' preconception or early pregnancy exposure to chlorination by-products and nitrate in public drinking water and incidence of congenital cardiac defects. The study population was identified through the Swedish Birth Registry and was limited to infants whose mothers used the municipal water system and had addresses for the preconception or early pregnancy period that could be geocoded. The Registry of Congenital Malformations provided information on cardiac defects. Additional data on the pregnancy, delivery, and newborn health were obtained from medical records and the hospital discharge registry. Exposure assessment was ascertained by using a geographic information system to link the study subjects to specific water supplies. Groundwater as a source of drinking water was reported as a potential risk factor for cardiac defects (adjusted OR = 1.31; 95% CI, 1.09-1.57). A very small and not statistically significant excess risk for cardiac defects was found to be associated with levels ≥ 2 mg/L nitrate-N compared with those
A case-control study in California investigated the potential association between maternal exposure to nitrates in drinking water and diet before pregnancy and the risk of NTDs in the mothers' infants (Croen et al. 2001). Case infants (538) with NTDs (both live and stillborn singleton births) born from 1989 through 1991 were selected from California's birth defects program. Control infants (539) were live births with no malformations selected from each area birth hospital for the same time period. Exposure assessment was done through interviews with the mothers, which included a detailed beverage and dietary questionnaire, and the state's community water systems data for sources serving their preconception addresses. The authors found an increased risk for NTDs among babies born to mothers living in areas where the drinking water nitrate level was above the MCL compared with those in areas below the MCL (OR = 2.7; 95% CI, 0.76-9.3). This association did not change after adjusting for other dietary nitrate intake (OR = 1.9; 95% CI, 0.73-4.7). Increased risk with increasing levels of nitrates was observed. Risk estimates were higher among groundwater users; however, other risk factors (e.g., Hispanic ethnicity, young age, low socioeconomic status, and no vitamin use) for NTDs were also more common among groundwater users.
In this study, the authors examined risk separately for anencephaly and spina bifida (Croen et al. 2001). Increased risk for anencephaly in babies was associated with their mother living in an area where the nitrate in drinking water was above the MCL (OR = 4.0; 95% CI, 1.0-15.4). This association was not substantially altered when adjusted for dietary intake of nitrate. A doubling in risk for anencephaly in babies whose mothers lived in areas where the nitrate level in groundwater was ≥ 5 mg/L compared with
Fan and Steinberg (1996) summarized the studies conducted on maternal exposure to nitrates in drinking water and birth defects by Scragg et al. (1982) and Dorsch et al. (1984) in South Australia, Arbuckle et al. (1988) in New Brunswick, Canada, and Bove et al. (1992) in New Jersey. The study by Dorsch et al. (1984) suggested an increased risk of bearing a child with a congenital malformation among women whose homes were served by water with a nitrate concentration > 5 ppm. Although the effects of nitrate cannot be discounted, the finding that the risks associated with multiple defects were increased suggests possible multiple risk factors. The study by Arbuckle et al. (1988) reported a protective relationship for users of public and spring water sources compared with private sources, but the ORs were not statistically significant. Five water samples exceeded the Canadian MCL (44 ppm) for nitrate. Bove et al. (1992) reported a positive association with water contamination and NTDs but cautioned that the study did not provide sufficient evidence of causation for any of the contaminants in question.
These studies assessed exposure to nitrates in drinking water on the basis of the source of the water. The lack of individual exposure assessment--whether the women actually drank the water--is a limitation in some investigations. The presence of other substances, such as pesticides, toxic metals, and chemicals (e.g., chlorinated solvents and chlorinated disinfection by-products), in private and public water systems may be correlated with the presence of nitrate. Some of these constituents are reported as risk factors for congenital malformations (Bove et al. 1992, 2002).
Spontaneous abortions. Aschengrau et al. (1989) investigated the quality of community drinking water and the occurrence of spontaneous abortions among a group of women in the Boston, Massachusetts, area. The population, selected from 1976 through 1978, consisted of women who lived in Massachusetts during their pregnancy, lived in a town with a public water supply, and were admitted to a specific community hospital. Cases were 286 women who had a miscarriage during their first 27 weeks of pregnancy, and controls were 1,391 women who had live births. The women were interviewed to obtain demographic and behavioral information, and water quality data were obtained from public records of routine analyses of public tap water. All levels of nitrate were below the MCL. A negative association was reported for any detectable level of nitrate and the occurrence of spontaneous abortions. The report noted that risk estimates may have been diluted by the measurement, recording, and classification of exposure, because this information was obtained indirectly from public water supply records of the communities where the women lived at the time the spontaneous abortions occurred (Aschengrau et al. 1989).
Gelperin et al. (1975) evaluated data from 1959 through 1966 on infant and fetal deaths in 16 Illinois communities, nine of which had nitrate levels ranging from 43 ppm to 123 ppm (NO3) in their water supply. Communities were grouped into three categories, consistently having high nitrate (above the MCL), having high nitrate in spring only, or not having high nitrate levels. The community water supply data provided nitrate levels. No significant increase in fetal deaths was found in areas consistently reporting nitrate levels above the MCL in their water compared with other areas.
Skrivan (1971) measured methemoglobin levels in the blood of pregnant women over a 2-year period to evaluate whether the observed mean values of methemoglobin would be similar in women who experienced spontaneous abortions and in those who experienced term delivery. The mean values in women before spontaneous abortion did not differ significantly from the mean values among women with term deliveries (Skrivan 1971).
An earlier evaluation of methemoglobin levels in pregnancy reported a relationship between methemoglobinemia and miscarriages in humans (Schmitz 1961). The study tested methemoglobin levels in 25 women in their first trimester of pregnancy. Higher levels were observed in women who spontaneously aborted or who threatened abortion in the first trimester. The report noted nitrates and nitrites as the most common methemoglobin inducers and concluded that high maternal methemoglobin levels are possibly related to miscarriages.
Case reports. A report on a cluster of spontaneous abortions in LaGrange, Indiana, cited nitrate-contaminated water from private wells as the possible cause (CDC 1996). The cases included a 35-year-old woman who experienced four consecutive miscarriages and a 37-year-old and a 20-year-old who each experienced one miscarriage. All three women lived within 1 mile of each other and were in the first trimester of pregnancy at the time of the miscarriages. Testing of the wells serving the homes of the women found nitrate to be the only elevated contaminant. The wells had nitrate levels over the MCL, with reported levels of 19.0 mg/L, 26 mg/L, and 19.2 mg/L nitrate-N for the three women, respectively. Although these incidents of spontaneous abortion may have been related to the ingestion of nitrate contaminated drinking water, other possible explanations such as genetic defects in the fetuses and cluster by chance could not be ruled out (CDC 1996).
Other reproductive effects. Besides birth defects and prenatal mortality, reproductive toxicity includes less readily observed effects that may be influenced by chronic low-level exposure to a toxic substance. These effects include sterility, intrauterine growth restriction, premature birth, and complications of pregnancy. Several of these outcomes have been addressed in epidemiologic studies of the potential effects of nitrate exposure on reproductive health (Table 1). Hypothetically, the oxidation of hemoglobin to methemoglobin, which limits the oxygen-carrying capacity of the blood, may interfere with the course and outcome of pregnancy (NAS 1981; Tabacova and Balabaeva 1993; Tabacova et al. 1997, 1998).
Bukowski et al. (2001) conducted a population-based case-control study on singleton births that occurred from 1991 through 1994 to mothers who resided in Prince Edward Island, Canada, and who used municipal and private water systems. The study examined the potential impact of groundwater nitrate exposure on prematurity and intrauterine growth restriction (IUGR). The study included 210 cases of IUGR, 336 cases of premature births, and 4,098 controls that were identified through a Reproductive Care Program (RCP) database. The authors developed a nitrate level exposure map using data on public and private wells collected from 1990 through 1993. Premature birth was defined as birth at
Bukowski et al. (2001) found a significant relationship between IUGR and higher nitrate levels. Using the lowest exposure category (median level of ≤ 1.3 mg/L nitrate-N) as the comparison group, an adjusted OR showed an excess risk for higher exposure categories with median levels 3.1 mg/L (OR = 2.31; 95% CI, 1.47-3.64) and 4.3 mg/L (OR = 2.56; 95% CI, 1.44-4.45). An increased but not significant risk for the highest exposure group of 5.5 mg/L (OR = 1.34; 95% CI, 0.31-3.99) was observed, as well as a significant dose-response association between nitrate exposure and prematurity. The authors reported excess risk estimates for the exposure categories with median levels of 3.1 mg/L (OR = 1.82; 95% CI, 1.23-2.69), 4.3 mg/L (OR = 2.33; 95% CI, 1.46-3.68), and 5.5 mg/L (OR = 2.37; 95% CI, 1.07-4.80). The dose-response relationship is fairly consistent and suggestive; however, assessing exposure to nitrate based on ecologic classification makes it difficult to interpret the findings exclusively in terms of water nitrate exposure (Bukowski et al. 2001).
In a study in Bulgaria, Tabacova et al. (1997) investigated the association between maternal exposure to environmental nitrogen sources and subsequent complications of pregnancy. The study was done in an area known to have nitrate levels in drinking water ranging from 8 to 54 mg/L nitrate-N (depending on the season and the water supply source), as well as in the food supply and oxides of nitrogen in ambient air. The study, which included pregnant women at ≥ 24 weeks' gestation, analyzed blood methemoglobin levels as markers of exposure to nitrogen compounds. Personal interviews with the women furnished medical and lifestyle histories. Only 16% of the study group experienced a normal pregnancy. The others experienced complications, including anemia (67%), threatened spontaneous abortion/premature labor (33%), and toxemia (23%), with some women having more than one of these diagnoses. Among women with pregnancy pathology, methemoglobin levels were 0.1-11.2% of total hemoglobin, compared with 0.4-2.8% in women with normal pregnancy. The mean methemoglobin levels for all categories of pregnancy complications were above the physiologic normal level of 2%, with the highest mean level occurring in the toxemia group. The report suggests an increased risk for pregnancy complications from exposure to nitrogen compounds. However, the lack of individual exposure assessment and the presence of several potential sources of exposure prevent implicating drinking water as a single source (Tabacova et al. 1997).
Tabacova et al. (1998) also evaluated mother-infant pairs in an area of Bulgaria with chronic exposure to pollution by nitrogen compounds, including drinking water nitrate levels of 8-54 mg/L nitrate-N in municipal water system, 13-400 mg/L nitrate-N in wells, and nitrogen oxides in ambient air. They examined methemoglobin levels as markers for exposure to nitrogen compounds, measures of oxidative stress, and the health status of infants at birth. Fifty-one mother-infant pairs were recruited from a local hospital; interviews were conducted with the mothers, and clinical records were reviewed for information on demographics, lifestyle, and medical factors. Maternal and cord blood were tested for methemoglobin and markers of oxidative stress. More than half of maternal blood and almost half of cord blood had methemoglobin levels > 2%. The study indicated a strong association between maternal methemoglobin and cord blood methemoglobin. Maternal and cord blood methemoglobin levels were higher in cases of abnormal birth outcomes (preterm birth, fetal distress, and low birth weight) than in cases of normal birth outcomes. The authors also reported an association between methemoglobin levels in cord blood and adverse birth outcome. Overall, the results show that nitrate exposure may have a role in adverse reproductive effects. However, the small sample size, multiple sources of exposure, and the lack of individual exposure assessment make the findings difficult to interpret in terms of nitrate exposure from drinking water per se. Other pollutants potentially associated with abnormal birth outcomes were not discussed (Tabacova et al. 1998).
Fan et al. (1987; Fan and Steinberg 1996) discussed the study by Super et al. (1981) that evaluated infant morbidity in a rural area of southwest Africa known to have wells with high levels of nitrates. Super et al. (1981) found no association between the incidence of premature birth or the size of infant at birth and living in an area with nitrate levels > 20 mg/L nitrate-N compared with areas with levels ≤ 20 mg/L. No association was found between birth weight, current or previous premature births, stillbirths, and spontaneous abortions and nitrate levels > 20 mg/L nitrate-N. However, increased deaths among infants previously born to mothers in regions with nitrate levels > 20 mg/L nitrate-N were reported.