This paper summarizes the epidemiological evidence for adverse health effects associated with disinfection by-products (DBPs) in drinking water and describes the potential mechanism of action.
There appears to be good epidemiological evidence for a relationship between exposure to DBPs, as measured by trihalomethanes (THMs), in drinking water and bladder cancer, but the evidence for other cancers including colorectal cancer is inconclusive and inconsistent. There appears to be some evidence for an association between exposure to DBPs, specifically THMs, and little for gestational age/intrauterine growth retardation and, to a lesser extent, pre-term delivery, but evidence for relationships with other outcomes such as low birth weight, stillbirth, congenital anomalies and semen quality is inconclusive and inconsistent. Major limitations in exposure assessment, small sample sizes and potential biases may account for the inconclusive and inconsistent results in epidemiological studies. Moreover, most studies have focused on total THMs as the exposure metric, whereas other DBPs appear to be more toxic than the THMs, albeit generally occurring at lower levels in the water.
The mechanisms through which DBPs may cause adverse health effects including cancer and adverse reproductive effects have not been well investigated. Several mechanisms have been suggested, including genotoxicity, oxidative stress, disruption of folate metabolism, disruption of the synthesis and/or secretion of placental syncytiotrophoblast-derived chorionic gonadotropin and lowering of testosterone levels, but further work is required in this area.
Disinfection of drinking water has led to major improvements in public health in developed countries since its introduction in the first half of the twentieth century. It has now been more than 30 years since the discovery that by-products can be formed in small quantities as part of the chlorination process (Rook 1974). Disinfection by-products (DBPs) are formed when water is disinfected, and natural organic matter, bromide and iodide in the water react with chlorine, chlorine dioxide, chloramines and/or ozone (Bichsel & von Gunten 2000; Zhang et al. 2000). Their formation and occurrence depend on many factors including disinfectant type(s) and dose(s), type(s) of treatment, pH, temperature, contact time(s) with disinfectant(s), water source, amount and character of natural organic matter and bromide and iodide levels (Reckhow & Singer 1985; Stevens et al. 1989; Amy et al. 1991; Singer 1994). Up to 600 DBPs have been identified (Richardson 1998; Richardson et al. 2007), and these chemicals differ considerably in their physico-chemical properties (e.g. volatility). Different mixtures of by-products may exist in different locations depending on the various factors mentioned earlier, making it more difficult to ascertain the risk, if any, of health effects in relation to specific DBPs and mixtures of DBPs, as well as to compare the findings from different epidemiological studies.
Trihalomethanes (THMs) are the most commonly formed group of DBPs. These are volatile DBPs, and individuals may be exposed not only through ingestion but also through inhalation and dermal absorption during activities such as showering, bathing and swimming (Weisel & Jo 1996; Nieuwenhuijsen et al. 2000a). For non-volatile DBPs such as the haloacetic acids (HAAs), ingestion is thought to be the main route of exposure. However, dermal adsorption has also been examined for such DBPs (Kim & Weisel 1998). Recent modelling of THM uptake suggested that swimming may lead to the highest levels in the blood (Whitaker et al. 2003). Uptake of DBPs through showering, bathing and swimming was associated with an increased risk of bladder cancer in a recent Spanish epidemiology study (Villanueva et al. 2007).
In this paper, we first summarize the epidemiological evidence regarding health effects associated with exposure to DBPs, particularly for reproductive outcomes, and briefly describe the main mechanisms proposed for the action of these compounds.
2. Epidemiological studies examining health effects related to exposure to chlorination disinfection by-products
The health effects of DBPs in drinking water have been a concern since DBPs were first reported in the 1970s. According to a review by the IPCS (2000): ‘more studies have considered bladder cancer than any other cancer. The authors of the report caution against a simple interpretation of the observed associations. The epidemiological evidence for an increased relative risk for bladder cancer is not consistent—different risks are reported for smokers and non-smokers, for men and women, and for low and high water consumption. Risk may differ among various geographic areas because the DBP mix may be different or because other water contaminants are also present’. For example, as part of an improved exposure assessment (Amy et al. 2005) for two well-conducted bladder cancer epidemiology studies (King & Marrett 1996; Cantor et al. 1998), substantial differences in the mixture of DBPs were found within and between one US state and one Canadian province (e.g. amount of brominated DBPs, relative proportion of THMs to HAAs, relative proportion of di- and trihalogenated HAAs). A recent pooled analysis by Villanueva et al. (2004), which provided quantitative information on THM exposure, confirmed some of the gender differences. For men, there was an exposure–response relationship between THM intake and bladder cancer, but there was no relationship for women (table 1). For other cancers, the evidence is much weaker. Some studies have suggested an association between DBPs and colorectal cancers, whereas others have not (Wilkins & Comstock 1981; Young et al. 1981, 1987; Doyle et al. 1997, Koivusalo & Vartiainen 1997; Hildesheim et al. 1998; King et al. 2000a; Bove et al. 2007). Furthermore, there is little evidence for an association between exposure to DBPs and other cancers such as liver, kidney, brain, lung and breast cancer, lymphomas, cancer of the pancreas, but the number of studies is small (IPCS 2000). A recent report suggested an association between THMs and skin cancer, but further work needs to be conducted (Karagas et al. 2008).
(b) Reproductive outcomes
Reproductive health outcomes should be easier to study than cancer because of the shorter relevant exposure period. Among others, congenital anomalies, stillbirth, spontaneous abortion, birth weight, prematurity and semen quality have been the focus of investigation. Various thorough reviews have been conducted and have concluded that the relationship between DBP exposure and reproductive health outcomes remains unclear, mainly owing to limitations in the exposure assessment in most studies (Reif et al. 1996; IPCS 2000; Nieuwenhuijsen et al. 2000b; Gevecker Graves et al. 2001; Bove et al. 2002; Tardiff et al. 2006).
A number of studies have found statistically significant positive associations between THMs and neural tube defects (NTDs), one of the most studied groups of congenital anomalies (Bove et al. 1995; Klotz & Pyrch 1999; Dodds & King 2001), whereas other studies have not found statistically significant associations
(Dodds et al. 1999; Magnus et al. 1999; Källen & Robert 2000; Hwang et al. 2002; Shaw et al. 2003; Nieuwenhuijsen et al. 2008) (table 2). Klotz & Pyrch (1999) found a statistically significant association between total THM (TTHM) levels in the water and NTDs, but not with levels of haloacetonitriles and HAA. Also, the effects were most pronounced in offspring from women who did not take supplementary vitamins, but these findings were not confirmed by the Shaw et al. (2003) study. Cedergren et al. (2002), Hwang et al. (2002) and Chisholm et al. (2008) found significant associations between chlorinated water, levels of TTHM above 10 μg l−1 and high levels (i.e. 130 μg l−1 or more) of THMs and cardiovascular congenital anomalies, respectively, but other studies did not find such an association (Bove et al. 1995; Dodds et al. 1999; Magnus et al. 1999; Källen & Robert 2000; Dodds & King 2001; Shaw et al. 2003; Nieuwenhuijsen et al. 2008).
Few studies have been published on chlorinated water and respiratory congenital anomalies, but two studies found a significant positive association (Aschengrau et al. 1993; Hwang et al. 2002), whereas two did not (Chisholm et al. 2008; Nieuwenhuijsen et al. 2008).
Similarly, for urinary tract defects, three studies reported statistically significant positive associations (Aschengrau et al. 1993; Magnus et al. 1999; Hwang et al. 2002), while one did not (Nieuwenhuijsen et al. 2008) and another one showed almost statistically significant effects (Chisholm et al. 2008; odds ratio (OR)=1.40, 95% confidence interval (CI): 0.98–1.99).
Evidence for risk of hypospadias is also inconclusive. There was no association with THM or HAA concentrations or proxies (Källen & Robert 2000; Luben et al. 2007; Hwang et al. 2008); however, estimates of actual THM ingestion were associated with increased risk of hypospadias (Luben et al. 2007).
In a meta-analysis, Hwang & Jaakkola (2003) reported evidence for an effect of exposure to chlorination by-products on the risk of neural tube and urinary system defects, but results for respiratory system, major cardiac and oral cleft defects were heterogeneous and inconclusive. The exposure index used was, however, fairly crude, without levels of DBPs being taken into account. Since the meta-analysis was published in 2003, the largest study published to date by Nieuwenhuijsen et al. (2008) was conducted, which was larger than all previous studies combined and which reported no association between THMs and cleft palate/lip, abdominal wall, major cardiac, neural tube, urinary and respiratory defects, except for a restricted set of anomalies with isolated defects, which appears to be due to a more reliable means of case identification. There were excess risks in the highest exposure categories of TTHMs (i.e. 60 μg l−1 or more) for ventricular septal defects and the highest exposure category of bromoform (i.e. 4 μg l−1 or more) and a subset of major cardiovascular defects and gastroschisis (Nieuwenhuijsen et al. 2008). In the meta-analysis by Hwang et al. (2008), the summary OR for ventricular septal defects (OR 1.59, 95% CI: 1.21, 2.07) for high versus low exposure to DBPs was statistically significant, but the exposure categories in the individual studies were inconsistent (different levels of THMs, or chlorination as a proxy), rendering the results difficult to interpret.
Only a few studies have assessed the relationship between DBPs and spontaneous abortion. The Californian study has attracted the most attention since its authors found a statistically significant association between TTHMs (i.e. 75 μg l−1 or more), especially for bromodichloromethane (BDCM) (i.e. 18 μg l−1 or more)—together with a high consumption of water (five glasses or more per day)—and spontaneous abortion (Waller et al. 1998). The ORs were even larger after re-analysis when restricting it to subjects for whom exposure had been characterized with greater confidence (Waller et al. 2001). However, in a study trying to replicate these results with substantially improved exposure assessments—including a study site with high-bromide water—Savitz et al. (2006) found no evidence for an association between a number of DBPs and spontaneous abortion, nor did they find any such association in an earlier study (Savitz et al. 1995).
A number of Canadian studies and one English study found statistically positive associations between DBPs and stillbirth (Dodds et al. 1999, 2004; King et al. 2000b; Toledano et al. 2005). However, the case–control study by Dodds et al. (2004) did not show a monotonic relationship between THM levels and stillbirth, and they did not find an association between HAAs and stillbirth (King et al. 2005).
Studies on pre-term delivery have generally shown no statistically significant associations with DBPs (Kramer et al. 1992; Bove et al. 1995; Savitz et al. 1995; Gallagher et al. 1998; Wright et al. 2003; Aggazzotti et al. 2004; Hinckley et al. 2005; Lewis et al. 2007; Yang et al. 2007), with the exception of the study by Yang et al. (2000a) and Yang (2004), who found a statistically significant increased risk. Wright et al. (2004) and Jaakkola et al. (2001) found a statistically significant decreased risk of pre-term delivery.
Study results on (term) low birth weight (LBW) have been mixed, with some studies reporting statistically significant associations (Bove et al. 1995; Gallagher et al. 1998; Källen & Robert 2000; Lewis et al. 2006) and others not (Kramer et al. 1992; Savitz et al. 1995; Kanitz et al. 1996; Dodds et al. 1999; Yang et al. 2000a, 2007; Jaakkola et al. 2001; Wright et al. 2003; Yang 2004; Toledano et al. 2005). Hinckley et al. (2005) found no association with THMs, but did for some specific HAAs. Studies on small for gestational age (SGA) and/or intrauterine growth retardation (IUGR) showed some more consistent results, and a good proportion of them have found statistically significant associations (Kramer et al. 1992; Bove et al. 1995; Wright et al. 2003, 2004), while some did not (Dodds et al. 1999; Porter et al. 2005; Yang et al. 2007; Hoffman et al. 2008b) (table 3). Aggazzotti et al. (2004) found some effects with by-products of chlorine dioxide. Wright et al. (2004) found statistically significant associations with THMs and a measure of mutagenicity, but not with HAAs or the chlorinated furanone 3-chloro-4-(dichloromethyl)-5-hydroxy-2-(5H)-furanone (MX) (table 4). Infante-Rivard (2004) found that the association between THMs and IUGR was modified by a metabolic polymorphism, with newborns with the CYP2E1 (G1259C) variant at high risk.
Two small case–control studies have investigated the relationship between DBPs and semen quality (Fenster et al. 2003; Luben et al. 2007). Halogenated acetic acids have been found to cause testicular damage in rats through disruption of spermatogenesis and motility, with the brominated analogues being the strongest toxicants (Smith et al. 1989; Toth et al. 1992; Linder et al. 1994a,b, 1995, 1997a,b). The results of the two epidemiology studies were inconclusive, with inconsistent evidence across various measures of semen quality and DBP exposure. Fenster et al. (2003) found that TTHM levels were not associated with decrements in semen quality. Per cent normal morphology decreased and per cent head defects increased at higher levels of a THM ingestion metric compared with the lowest level, although there were no monotonic dose–responses, and at this level, they observed a small decrease in per cent morphologically normal sperm. BDCM exposure was inversely related to linearity (a motility parameter). Luben et al. (2007) studied the relation between exposure to classes of DBPs and sperm concentration and morphology, as well as DNA integrity and chromatin maturity, but found no association—or consistent pattern—of increased abnormal semen quality with elevated exposure to THMs or HAAs.
MacLehose et al. (2008) investigated time to pregnancy in relation to DBP exposure, but found little evidence for a relationship. Joyce et al. (2008) investigated the effect of DBPs on pre-laboural rupture of membranes but found no relationship.
Very few studies have examined the gene–environment interaction and/or the presence of susceptible groups. Infante-Rivard (2004) found that newborns with a high-metabolism CYP2E1 gene variant who experienced pregnancy average exposures of more than 29.4 μg l−1 for TTHMs were at much higher risk (OR=13.2, 95% CI: 1.19–146.7) of IUGR compared with those without this CYP2E1 variant, but found no indication that MTHFR C677T modified the effect of exposure to chloroform and risk of foetal growth in humans. A study investigating NTDs, isolated cleft lip palate with or without cleft palate, also found no evidence that MTHFR C677T modified the effect of TTHM exposure (Shaw et al. 2003). Lewis et al. (2006) reported an increased risk of term LBW in non-Caucasians associated with second trimester exposure to THMs greater than the increased risk found for Caucasians and non-Caucasians combined.
The major limiting factor in these studies has often been crude exposure assessment, with the exception of some of the more recent studies. The use of ecological water supply zone estimates as an exposure index may result in exposure misclassification (Whitaker et al. 2003), which likely biases the measures of effect towards the null. Furthermore, while ingestion has generally been the primary exposure route of interest, uptake through showering, bathing and swimming could be considerable, specifically for THMs owing to their volatility and dermal adsorption, but these routes have only been considered in a few studies (e.g. Savitz et al. 2006; Luben et al. 2007, 2008; Villanueva et al. 2007; Hoffman et al. 2008a,b; MacLehose et al. 2008). Combining information on individual water use with water supply zone estimates would provide more detailed exposure assessment, but the individual information should be evaluated for measurement error because within-subject variability in questionnaire data may be substantial (Forssén et al. in press) and attenuate risk estimates. Furthermore, exposure estimates have been based primarily on maternal residence at birth. This ignores any exposure that occurs outside the home, e.g. in the workplace, and also ignores the possibility that a mother may have moved her residence during her pregnancy. Exposure assessment based on maternal residence at birth may, therefore, result in exposure misclassification.
In addition, studies from countries including Scandinavia (Magnus et al. 1999; Källen & Robert 2000; Jaakkola et al. 2001; Cedergren et al. 2002; Hwang et al. 2002) and Taiwan (Yang et al.2000a,b, 2007; Yang 2004; Hwang et al. 2008) have generally shown low levels of DBPs with a small range, making the assessment of risks more difficult owing to both a higher probability of exposure misclassification and a smaller difference in exposure between dose groups. Furthermore, Cedergren et al. (2002) (in Sweden) and Chisholm et al. (2008) (in Australia) found significant associations between levels of TTHM above 10 μg l−1 and high levels greater than or equal to 130 μg l−1 of THMs and cardiovascular congenital anomalies, respectively, but the low exposure group in the latter study (i.e. less than 60 μg l−1) represented higher levels of THM exposure than that of the cases in the former study (i.e. more than 10 μg l−1), making the comparison more difficult. Moreover, other studies did not find such an association, such as in Magnus et al. (1999) (in Norway), in which the average level of TTHMs for chlorinated water was 9.4 μg l−1. In the latter study, exposure assessment was based on whether the mothers received unchlorinated or chlorinated water, albeit with relatively low levels of THMs (on average) in the latter group. Thus, neither group had high exposure to THMs (on average). Where seasonal variability in DBPs has not been taken into account, further errors in the exposure assessment are likely.
Particularly for reproductive epidemiological studies, the sample sizes have often been insufficient to produce robust results, especially for congenital anomalies and, to a lesser extent, for stillbirth, semen quality and other outcomes, but there are exceptions. For example, studies on SGA/IUGR by Dodds et al.(1999), Wright et al. (2003, 2004) and Hinckley et al. (2005), on congenital anomalies by Hwang et al. (2002, 2008) and Nieuwenhuijsen et al. (2008) and on stillbirth by Toledano et al. (2005) provide sufficiently large numbers of cases to create various exposure categories with more robust risk estimates, which could improve the overall assessment of risk.
The retrospective and registry-based nature of many of the reproductive studies has meant that information on potential confounders, and other risk factors for birth outcomes, such as maternal smoking and alcohol consumption, have often been lacking.
On the whole, epidemiological studies have used TTHMs as a proxy for total DBP load, but TTHMs are not necessarily a good proxy measure. Some studies have examined individual (brominated) THM species (e.g. Waller et al. 1998; King et al. 2000b; Dodds & King 2001; Shaw et al. 2003; Dodds et al. 2004; Wright et al. 2004; Nieuwenhuijsen et al. 2008). In addition, some studies have investigated other DBPs such as HAAs and/or MX (e.g. Klotz & Pyrch 1999; Wright et al. 2004; Hinckley et al. 2005; Porter et al. 2005; Savitz et al. 2006; Luben et al. 2007, 2008; Hoffman et al. 2008; MacLehose et al. 2008) and/or total organic halides (TOX) (e.g. Savitz et al. 2006; Hoffman et al. 2008). The metabolism of different DBP species varies (IPCS 2000), the toxicity of different DBP classes varies, specific DBPs in a particular class have substantially different toxicities (e.g. Hunter et al. 2006) and the relationship of THMs to that of other DBPs (e.g. HAAs and TOX) varies, so it is insufficient to use TTHMs as a proxy for DBPs as a whole. Investigation of the relation between non-THM by-products and reproductive outcomes is required in order to help elucidate the specific DBPs driving the associations observed. A detailed assessment of the DBP mixture (including speciation within different DBP classes) is necessary to explain any observed epidemiological results. Wright et al. (2004) is a good example of a study in which different DBP classes were examined, as well as specific DBPs within these classes (table 4).
Furthermore, outcomes such as spontaneous abortion, foetal growth restriction or congenital anomalies have not been defined well and/or are difficult to study. Previous epidemiological studies have used a variety of outcomes as proxies for foetal growth restriction: terms LBW, IUGR and SGA. There are some limitations to these measures. LBW is rather crudely defined—the fixed criterion of birth weight below 2500 g takes no account of population-specific birth weight distributions (Wilcox 2001). Somewhat confusingly, the terms IUGR and SGA have been used interchangeably in the literature, and criteria for IUGR/SGA diagnosis have varied, some studies using the 5th and some the 10th percentile of gestational specific weight according to a standard population growth chart as a cut-off point. These measures fail to distinguish between those babies who are constitutionally small and those who are pathologically small (i.e. growth restricted). Some small but normally grown babies will fall below the cut-off point, and some growth-restricted babies will reach a weight above the cut-off point. Therefore, a proportion of infants are misclassified, and in epidemiological studies, this may bias any association towards the null. There is evidence to show that the use of customized foetal growth charts, which take into account factors such as maternal height and ethnicity, significantly reduce the proportion of false-positive and false-negative diagnoses of foetal growth restriction, compared with using standard population growth charts (Gelbaya & Nardo 2005; Gardosi 2006), but these are poorly developed at present.
Congenital anomalies have often been analysed either as one group or in main categories, e.g. neural tube, major heart and abdominal defects, owing to the small number of cases in each study. These anomalies, however, are generally heterogeneous with respect to both phenotype and presumed aetiology. Nieuwenhuijsen et al. (2008) showed that focusing on isolated subcategories may result in different findings. Furthermore, in some countries, registration of congenital anomalies may occur up to 1 year after the birth (e.g. in Taiwan), which will improve the completeness of the registry by including cases, such as hypospadias, that are more difficult to identify at birth.
Investigation of gene–environment interaction and/or the effects on susceptible groups has been limited (e.g. Shaw et al. 2003; Infante-Rivard 2004). Preliminary studies suggest that certain groups may be more susceptible to the influence of DBPs (Lewis et al. 2006), and thus these effects may be masked in studies that only look at the population in general.
The mechanisms through which DBPs may cause adverse health effects, including cancer and adverse reproductive effects, are not well investigated. Several mechanisms have been suggested that involve genotoxicity, oxidative stress, disruption of folate metabolism, disruption of the synthesis and/or secretion of placental syncytiotrophoblast-derived chorionic gonadotropin and lowering of testosterone levels.
Richardson et al. (2007) reviewed 30 years of research on the occurrence, genotoxicity and carcinogenicity of 85 DBPs. Of these, 11, including THMs, are currently regulated by the United States Environmental Protection Agency and 74 are considered emerging DBPs owing to their moderate occurrence levels and/or toxicological properties. Sixty-eight of the 85 DBPs reviewed were considered genotoxic, including the regulated brominated THMs, where the THMs are generally at higher levels in drinking water. In general, the brominated DBPs are more genotoxic and carcinogenic than chlorinated compounds, and iodinated DBPs are the most genotoxic (Plewa et al. 2008). Moreover, certain nitrogenous DBPs were found to be more genotoxic than the regulated carbonaceous DBPs (i.e. THMs and HAAs) (Plewa et al. 2008). Recently, Ross & Pegram (2004) reported GSTT1-1-dependent covalent binding of brominated THMs to DNA and formation of deoxyguanosine adducts in vitro. Because there is structural similarity among the brominated THMs and evidence for common pathways of bioactivation (DeMarini et al. 1997; Pegram et al. 1997), the findings of Ross & Pegram (2004) support the idea that glutathione (GSH) conjugation of tribromomethane may lead to the formation of DNA-reactive metabolites in the liver, and perhaps even more likely in the colons, of rodents and humans.
(b) Oxidative stress
There is evidence that maternal oxidative stress during pregnancy may play an important role in adverse foetal development (Scholl & Stein 2001; Meek et al. 2002; Myatt & Cui 2004; Kim et al. 2005; Min et al. 2006). For example, increased concentrations of oxidative stress biomarkers (8-OH-dG and MDA) observed in the urine of pregnant women have been associated with decreased birth weight (Scholl & Stein 2001; Kim et al. 2005). In late gestation, an increase in oxidative stress is observed in pregnancies complicated by IUGR, pre-eclampsia and diabetes, and this is associated with increased trophoblast apoptosis and alterations to placental vascular reactivity (Myatt & Cui 2004). There is also evidence to suggest that exposure to DBPs can cause oxidative stress in human cells. An in vitro study on human hepatoma (HepG2) cells reported that increasing chloroform dose resulted in decreasing GSH, which induces oxidative stress (Beddowes et al. 2003). Another in vitro study on human HepG2 cells found that when exposed to chlorinated drinking water, MDA increased and GSH decreased in a dose-dependent manner, indicating oxidative stress (Yuan et al. 2005).
Cytochrome P-450E1 (CYP2E1) is the primary enzyme involved in the metabolism of low doses of chloroform (Meek et al. 2002), and Tomasi et al. (1985) showed that chloroform metabolism generates free radicals. Chloroform is oxidatively metabolized and decomposed to electrophilic phosgene, which is highly reactive and will bond to cell components including proteins, phospholipid polar heads and reduced GSH (Gemma et al. 2003).
Other compounds such as trichloroethanol, trichloroacetic acid and dichloroacetic acid have all been shown to induce lipid peroxidation, a biomarker of oxidative stress, presumably via a free radical mechanism (Larson & Bull 1992; Ni et al. 1996).
Polymorphisms in pro-inflammatory cytokines (i.e. tumour necrosis factor (TNF)) have been associated with pre-term births (Crider et al. 2005; Engel et al. 2005a), whereas polymorphisms in anti-inflammatory cytokines (i.e. interleukin (IL)-4) have been associated with SGA outcomes (Engel et al. 2005b). Animal studies have shown that both pro-inflammatory (TNF, IL-6 and IL-8) and anti-inflammatory (IL-10 and transforming growth factor) cytokines have been affected by exposure to carbon tetrachloride, a haloalkane similar to chloroform, and to phosgene, a metabolite of chloroform (Sciuto et al. 2003; Weber et al. 2003).
The observed associations between various adverse birth outcomes and markers of oxidative stress, and the associations between exposure to DBPs and their metabolites and markers of oxidative stress, suggest the possibility that DBPs may act on foetal growth via the oxidative stress mechanism.
(c) Folate metabolism
One suggested mechanism by which DBPs could cause cancer and adverse birth outcomes is the interference of folate metabolism by DBPs. Folate and folic acid are the forms of the B vitamin that are involved in the synthesis, repair and functioning of DNA and required for the production and maintenance of cells (Kamen 1997). Folate plays an important role for cells that are undergoing rapid turnover such as tissues in the colon and the developing foetus. Folate is involved in the synthesis of methionine, an essential amino acid. Low levels of folate have been associated with several forms of cancer and congenital anomalies such as NTDs. Furthermore, defects in the methionine–homocysteine metabolic pathway, which can be the result of low folate levels and result in elevated homocysteine levels, may be a contributing factor for abruptio placentae (Ray & Laskin 1999). Both chloroform and TCAA have been found to inhibit the vitamin B12-dependent methionine biosynthesis pathway. Inhibition of this pathway can lead to vitamin B12 deficiency and consequently folate deficiency. Alston (1991) found that chloroform inhibited methionine biosynthesis in cell culture. Dow & Green (2000) showed that trichloroacetic acid interacts with vitamin B12, probably by a free radical mechanism, inhibiting both the methylmalonyl CoA and methionine salvage pathways in rats. As a result of the latter, a secondary folate deficiency develops owing to the ‘methyl folate trap’, leading to a major impairment in formate metabolism. Geter et al. (2005) showed that rats exposed to bromoform and fed a no-folate diet had significant increases in aberrant crypt foci (putative precursor lesions in the development of colon cancer) when compared with rats exposed to bromoform and fed a normal diet.
(d) Chorionic gonadotropin disruption
Chen et al. (2003) showed that the THM BDCM reduced the secretion of immunoreactive and bioactive chorionic gonadotropin in primary cultures of human trophoblasts and thus appears to target human placental trophoblasts. Trophoblasts are the sole source of chorionic gonadotropin during normal human pregnancy; a decrease in the amount of this bioactive hormone could have adverse effects on pregnancy outcome, including those leading to growth retardation. Chen et al. (2004) reported that BDCM directly inhibits the morphological differentiation of mononucleated placental cytotrophoblast cells to multi-nucleated syncytiotrophoblast-like colonies in vitro. Syncytiotrophoblast formation was inhibited in a dose-dependent manner and was accompanied by no loss of cell viability.
Potter et al. (1996) showed that all of the THMs reduced serum testosterone in rats treated with 1.5 mmol kg−1 by oral gavage for 7 days. The finding that male F-344 rats treated with THMs had decreased circulating concentrations of testosterone also raises the question as to whether THMs may produce androgenic deficiency in male rats.
The various mechanisms described earlier begin to provide plausible biological pathways through which DBPs may cause adverse health effects, including cancer and adverse reproductive effects. However, they are clearly still in their infancy and further research is required to provide more definitive evidence for causal biological mechanisms.
There appears to be good epidemiological evidence for an association between chlorination by-products, as measured by THMs, in drinking water and bladder cancer, but the evidence for other cancers including colorectal cancer appears to be inconclusive and inconsistent. There appears to be some evidence for a relationship between chlorination by-products and SGA and IUGR and, to a lesser extent, pre-term delivery, but evidence for other outcomes such as LBW, stillbirth, congenital anomalies and semen quality appears to be inconclusive and inconsistent. Major limitations in exposure assessment, small sample sizes and potential biases may account for the inconclusive and inconsistent results in epidemiological studies. Moreover, most studies have focused on TTHMs as the exposure metric, whereas some emerging DBPs appear to be more toxic than the THMs. The mechanisms through which DBPs may cause adverse health effects, including cancer and adverse reproductive effects, have not been well investigated to date.
One contribution of 12 to a Theme Issue ‘Emerging chemical contaminants in water and wastewater’.
- © 2009 The Royal Society