2.1. BACKGROUND AND ENVIRONMENTAL EXPOSURES TO ATRAZINE IN THE UNITED STATES

Atrazine is a white, odorless powder (when pure) that is used as an herbicide to stop the growth of broadleaf and grassy weeds in crops such as corn, sugarcane, sorghum, pineapples, and macadamia nuts. It is not found naturally in the environment. It is moderately soluble in water, but is more soluble in organic solvents such as acetone, chloroform, and ethyl acetate. More than 37,000 tons of atrazine were used in agricultural and weed control settings in the United States in 1997.

Atrazine is released to the environment during its production and use, with the vast majority being released as a result of its application to soils as an herbicide. Atrazine that remains in the soil degrades with half-lives of a few weeks to several months. Atrazine may migrate out of the soil in surface runoff to streams, rivers, or lakes, or it may migrate deeper into the soil and become associated with groundwater. No significant degradation of atrazine has been observed in groundwater. Half-lives of atrazine in surface waters are generally >200 days. Relatively large amounts of atrazine may volatilize from the soils into the atmosphere. In the atmosphere, no direct photolysis degradation of atrazine is expected to occur, but oxidation in the presence of hydroxyl radicals is expected, with an estimated half-life of 14 hours. Most of the atrazine found in the atmosphere is expected to be sorbed to particulates; it can be transported significant distances in the atmosphere, and has been detected >180 miles from the nearest application site.

The general population may be exposed to atrazine found in water or air, but it is rarely found in foods. When the general population is exposed to atrazine, exposure levels are expected to be very low. Maximum seasonal and average atrazine concentrations of 61.6 and 18.9 ppb, respectively, were detected during a 1993–1998 monitoring program of community water systems in the United States. Air concentrations of atrazine vary with application season; concentrations usually range from just above the detection limit of approximately 0.03 to 0.20–0.32 μg/m³ during the application period. The concentrations of atrazine detected in foods were low (0.001–0.028 ppm) in the few samples where it was detected.

Populations residing near crops where atrazine is applied or hazardous waste disposal sites or manufacturing and processing plants may be exposed to higher than average levels of atrazine in ambient air or drinking water. As mentioned above, atrazine is mobile in soils and has been detected in a high percentage of the drinking water wells near crops where atrazine has been used. Atrazine has also been identified in at least 20 of the 1,636 hazardous waste sites that have been proposed for inclusion on the EPA National Priorities List (NPL). However, the number of sites evaluated for atrazine is not known.

2.2. SUMMARY OF HEALTH EFFECTS

Atrazine is a widely used herbicide. The general population, especially people living in the vicinity of farms, may be exposed to low levels in the air and drinking water. Occupational exposure is a concern for farmers, as several routes of exposure are likely to occur. The primary adverse health effects of atrazine exposure are reproductive/developmental effects following inhalation, oral, and dermal exposure. Data on the carcinogenicity of atrazine are inconclusive. Data regarding the health effects of atrazine in humans are limited to ecological, case-control, and cohort mortality cancer studies and reproductive/ developmental toxicity studies. The bulk of the available toxicity data is from oral exposure studies in animals.

The reproductive system and the developing organism are primary targets of atrazine toxicity. There was a possible association between atrazine use/exposure of male farmers and increased pre-term delivery, but not decreased fecundity. The lack of information on exposure levels and the concomitant exposure to other pesticides makes these studies inadequate to assess the contribution of atrazine to these effects. Several animal studies have shown that atrazine exposure disrupts estrus cyclicity and alters plasma hormone levels; these effects appear to be mediated by changes in the gonadal-hypothalamic-pituitary axis. Epidemiological studies, examining developmental end points, have found an association between Iowa communities exposed to atrazine in the drinking water and an increased risk of small for gestational age babies and other birth defects. Farm couples living year-round on farms in Ontario, Canada did not have altered sex ratios, and the risk of small for gestational age deliveries was not increased in relation to pesticide exposure. Developmental effects have been observed following pre-gestational, gestational, and lactational exposure of rat and rabbit dams or post-weaning exposure of rat pups to atrazine. The observed effects included post-implantation losses, decreases in fetal body weight, incomplete ossification, neurodevelopmental effects, and impaired development of the reproductive system.

A few epidemiology studies suggest evidence of a possible association between atrazine exposure and increased cancer risk, but many others do not, and the data are insufficient to adequately assess atrazine’s carcinogenic potential. The animal data associate atrazine with early onset of mammary tumors believed to be the result of atrazine-induced acceleration of reproductive senescence. It is unlikely that the mechanism by which atrazine induces mammary tumors in female Sprague-Dawley rats is operational in humans.

A limited number of animal studies have shown that atrazine exposure may affect other end points, including systemic effects, and damage to the heart, liver, and kidneys. The primary effects are discussed in greater detail below. The reader is referred to Section 3.2, Discussion of Health Effects by Route of Exposure, for additional information on the health effects resulting from exposure to atrazine.

Reproductive Effects. Results of a survey of farm couples in Ontario, Canada, conducted to assess reproductive effects of pesticides, indicated a weak to moderate association between atrazine use in the yard and an increase in preterm delivery. Other results from this survey indicated that atrazine was not associated with any decrease in fecundity as a result of effects on spermatogenesis.

Animal studies have shown that atrazine disrupts estrus cyclicity (i.e., irregular ovarian cycling and changes in the number and/or percentage of days in estrus and diestrus) and alters plasma hormone levels in rats and pigs. These effects appear to be mediated by changes in the gonadal-hypothalamic-pituitary axis that are species-, and even strain-, specific. In Sprague-Dawley rats, atrazine accelerates the normal process of reproductive senescence, which is initiated by a failure of the hypothalamus to release levels of gonadotropin releasing hormone (GnRH) that are adequate to stimulate the pituitary to release luteinizing hormone (LH). Without sufficient LH, ovulation does not occur, estrogen levels remain high, and persistent estrus results. In other strains of rats, atrazine causes elevated progesterone levels, which leads to pseudopregnancy and persistent diestrus. In pigs, atrazine exposure decreased serum estradiol-17β (E₂) concentrations resulting in a short-term delay in the onset of estrus.

The mechanism of reproductive senescence in humans does not involve disruption of hormonal regulation, but is initiated by depletion of ova in the ovaries, which ultimately results in decreased plasma estrogen levels. Therefore, disruption of the menstrual cycle or acceleration of reproductive senescence is not anticipated to occur in humans as a result of atrazine exposure. However, it is not known whether atrazine will cause other perturbations in the hypothalamus-pituitary-gonad axis results in reproductive effects in humans.

Atrazine exposure significantly reduced serum and intratesticular levels of testosterone and ventral prostrate and seminal vesicle weights in juvenile male Sprague-Dawley rats. Female Wistar and Sprague-Dawley rats had delayed vaginal opening, and delayed uterine growth was observed in female Wistar rats.

Developmental Effects. Atrazine exposure has been associated with developmental effects in both humans and animals. An association was found between Iowa communities exposed to an average of 2.2 μg/L atrazine in the drinking water in 1984–1990 and an increased risk of intrauterine growth retardation and cardiac, urogenital, and limb reduction defects. The results of a survey of farm couples living year-round on farms in Ontario, Canada indicate that the sex ratio was not altered and the risk of small for gestational age deliveries was not increased in relation to pesticide exposure (atrazine exposure level not available).

Developmental effects in response to oral exposure to atrazine have been demonstrated in laboratory animals. Studies have shown that gestational and peripubertal exposure to atrazine has an effect on reproductive development in rats and rabbits. The effects of gestational exposure to atrazine in rats and rabbits include increased post-implantation losses, full-litter resorptions, decreased live fetuses/litters, increased prenatal loss, decreased litter size, and reduced pup weights, which could be attributed to severe maternal toxicity. Atrazine exposure in rats is also associated with delayed vaginal opening, first estrus cycle, and uterine growth for female rats and decreased prostate weight, increased incidence and severity of inflammation of the lateral prostate, increased myeloperoxidase levels in the prostate, and increased total DNA in the prostate for male rats.

Atrazine has also been shown to have an effect on the development of the nervous system in rats. Mild neurobehavioral effects were observed in female offspring of Fischer rat dams exposed to atrazine pre-mating, including increased spontaneous activity level, and male offspring had improved performance (decreased latency and increased avoidance) in avoidance conditioning trials.

Other developmental effects include incomplete ossification of the skull, hyoid bone, teeth, forepaw metacarpals, and hindpaw distal phalanges in the offspring of exposed Sprague-Dawley rats, and nonossification of forepaw metacarpals and middle phalanges, hindpaw talus and middle phalanges, and patella in the offspring of exposed rabbits. No developmental effects were noted in a two-generation study in which rats were exposed to atrazine in the diet.

Cancer. The carcinogenic potential of atrazine has been investigated in a number of epidemiology studies, including cohort studies of workers at triazines manufacturing facilities, case-control studies of farmers using atrazine or triazines, and ecological studies of populations living in agricultural areas with high atrazine use and residents living in areas with atrazine-contaminated drinking water. In most of these studies, it is likely that the individuals were exposed to atrazine via several exposure routes. For example, in the studies of farmers, the likely exposure routes are inhalation during application of atrazine, dermal during handling and use of atrazine, and possible oral exposure due to contamination of groundwater. Epidemiological data are available for a number of types of cancers; however, for most cancer types, only one study investigated the possible association. The most widely studied cancer type is non-Hodgkin’s lymphoma. In general, case-control studies of farmers using atrazine (in some studies, data are only available for triazine exposures) found small elevations in the risk of developing non-Hodgkin’s lymphoma; typically, the odds ratios were ≤1.5 and the 95% confidence intervals included unity. A small number of cases is a common limitation of these studies. One study pooled the data from three other studies, representing farmers in four U.S. states, to increase the statistical power of the analyses. This study found an odds ratio of 1.4 (95% confidence interval of 1.1–1.8), suggesting a weak association between atrazine exposure and increased risk of non-Hodgkin’s lymphoma. To account for possible exposure to other pesticides that have been shown to induce non-Hodgkin’s lymphoma, the odds ratio was adjusted for exposure to 2,4-D and organophosphate insecticides. This adjustment resulted in a small decrease in the odds ratio (1.2, 95% confidence interval of 0.9–1.7). A cohort mortality study of workers at two triazines manufacturing facilities also found an increase in the risk of non-Hodgkin’s lymphoma (SMR=385; 95% confidence interval of 79–1,124). Collectively, these studies provide suggestive evidence of a possible association between atrazine exposure and non-Hodgkin’s lymphoma, but a causal relationship cannot be established.

Evidence on the possible association between atrazine exposure and increased risk of other cancer types is weak. Studies of farmers or possible agricultural workers did not find significant increases in the risk of multiple myeloma, leukemia, soft tissue sarcoma/carcinoma, or Hodgkin’s disease. Suggestive evidence between atrazine (or triazines) exposure and an increased risk of prostate cancer, breast cancer, and ovarian cancer have been reported. Although these data provide a suspicion of carcinogenicity, the limited number of investigations and study limitations preclude drawing conclusions regarding these cancer types.

The animal data suggest that the carcinogenicity of atrazine is species-, strain-, and sex-specific. Statistically significant earlier onset of mammary tumors or incidence of mammary tumors were observed in female Sprague-Dawley rats, but not in female Fischer 344 rats or in female CD-1 mice. An increase in mammary tumors was observed in male Fischer 344 rats; however, it is likely that the increased tumor incidence is due to increased lifespan of the atrazine-treated animals, as compared to the controls (aged Fischer 344 rats have a high rate of spontaneous mammary tumors). The early onset of mammary tumors in female Sprague-Dawley rats is believed to be the result of atrazine-induced acceleration of reproductive senescence. Both the failure to ovulate and the state of persistent estrus lead to constant elevated serum levels of endogenous estrogen, which could result in tumor formation in estrogen-sensitive tissues. Reproductive senescence in humans involves ovarian depletion and decreased serum estrogen levels instead of decreasing hypothalamic function and increased serum estrogen levels. It is unlikely that the mechanism by which atrazine induces mammary tumors in female Sprague-Dawley rats is operational in humans.

IARC has classified atrazine as “not classifiable as to its carcinogenicity to humans” (Group 3) based on inadequate evidence in humans and sufficient evidence in experimental animals.

2.3. MINIMAL RISK LEVELS