Languages:
Home » Water Disinfectants » Level 3 » Question 5

Water Disinfectants & disinfectant by-products

5. What are the risks posed by disinfectants and their by-products?

  • 5.1 What tolerable daily intakes and guideline values has the WHO set?
    • 5.1.1 TDI for water disinfectants
    • 5.1.2 TDI for chlorine by-products
    • 5.1.3 TDI for chlorine dioxide by-products
    • 5.1.4 TDI for ozonation by-products
  • 5.2 How much disinfectants and by-products are we exposed to?
  • 5.3 Are there uncertainties in assessing exposure?
  • 5.4 Have epidemiological studies been able to evaluate the risks?
    • 5.4.1 Uncertainties of epidemiological data
    • 5.4.2 Epidemiological studies

The source document for this Digest states:

It should be noted that the use of chemical disinfectants in water treatment usually results in the formation of chemical by-products, some of which are potentially hazardous. However, the risks to health from these by-products at the levels at which they occur in drinking-water are extremely small in comparison with the risks associated with inadequate disinfection. Thus, it is important that disinfection not be compromised in attempting to control such by-products.

Source & ©: IPCS "Environmental Health Criteria for Disinfectants and disinfectant by-products ,
EHC 216, Chapter 1: Summary, section 1.5

5.1 What tolerable daily intakes and guideline values has the WHO set?

    • 5.1.1 TDI for water disinfectants
    • 5.1.2 TDI for chlorine by-products
    • 5.1.3 TDI for chlorine dioxide by-products
    • 5.1.4 TDI for ozonation by-products

5.1.1 TDI for water disinfectants

The source document for this Digest states:

1) Chlorine

A WHO Working Group for the 1993 Guidelines for drinking-water quality considered chlorine. This Working Group determined a Tolerable Daily Intake (TDI) of 150 µg/kg of body weight for free chlorine based on a no-observed-adverse-effect level (NOAEL) of approximately 15 mg/kg of body weight per day in 2-year studies in rats and mice and incorporating an uncertainty factor of 100 (10 each for intra- and interspecies variation). There are no new data that indicate that this TDI should be changed.

2) Monochloramine

A WHO Working Group for the 1993 Guidelines for drinking-water quality considered monochloramine. This Working Group determined a TDI of 94 µg/kg of body weight based on a NOAEL of approximately 9.4 mg/kg of body weight per day, the highest dose tested, in a 2-year bioassay in rats and incorporating an uncertainty factor of 100 (10 each for intra- and interspecies variation). There are no new data that indicate that this TDI should be changed.

3) Chlorine dioxide

The chemistry of chlorine dioxide in drinking-water is complex, but the major breakdown product is chlorite. In establishing a specific TDI for chlorine dioxide, data on both chlorine dioxide and chlorite can be considered, given the rapid hydrolysis to chlorite. Therefore, an oral TDI for chlorine dioxide is 30 µg/kg of body weight, based on the NOAEL of 2.9 mg/kg of body weight per day for neurodevelopmental effects of chlorite in rats.

Source & ©: IPCS "Environmental Health Criteria for Disinfectants and disinfectant by-products ,
EHC 216, Chapter 1: Summary, section 1.5

5.1.2 TDI for chlorine by-products

The source document for this Digest states:

4) Trihalomethanes

Cancer following chronic exposure is the primary hazard of concern for this class of DBPs. Because of the weight of evidence indicating that chloroform can induce cancer in animals only after chronic exposure to cytotoxic doses, it is clear that exposures to low concentrations of chloroform in drinking-water do not pose carcinogenic risks. The NOAEL for cytolethality and regenerative hyperplasia in mice was 10 mg/kg of body weight per day after administration of chloroform in corn oil for 3 weeks. Based on the mode of action evidence for chloroform carcinogenicity, a TDI of 10 µg/kg of body weight was derived using the NOAEL for cytotoxicity in mice and applying an uncertainty factor of 1000 (10 each for inter- and intraspecies variation and 10 for the short duration of the study). This approach is supported by a number of additional studies. This TDI is similar to the TDI derived in the 1998 WHO Guidelines for drinking-water quality, which was based on a 1979 study in which dogs were exposed for 7.5 years.

Among the brominated THMs, BDCM is of particular interest because it produces tumours in rats and mice and at several sites (liver, kidneys, large intestine) after corn oil gavage. The induction of colon tumours in rats by BDCM (and by bromoform) is also interesting because of the epidemiological associations with colo-rectal cancer. BDCM and the other brominated THMs are also weak mutagens. It is generally assumed that mutagenic carcinogens will produce linear dose-response relationships at low doses, as mutagenesis is generally considered to be an irreversible and cumulative effect.

In a 2-year bioassay, BDCM given by corn oil gavage induced tumours (in conjunction with cytotoxicity and increased proliferation) in the kidneys of mice and rats at doses of 50 and 100 mg/kg of body weight per day, respectively. The tumours in the large intestine of the rat occurred after exposure to both 50 and 100 mg/kg of body weight per day. Using the incidence of kidney tumours in male mice from this study, quantitative risk estimates have been calculated, yielding a slope factor of 4.8 × 10-3 [mg/kg of body weight per day]-1 and a calculated dose of 2.1 µg/kg of body weight per day for a risk level of 10-5. A slope factor of 4.2 × 10-3 [mg/kg of body weight per day]-1 (2.4 µg/kg of body weight per day for a 10-5 risk) was derived based on the incidence of large intestine carcinomas in the male rat. The International Agency for Research on cancer (IARC) has classified BDCM in Group 2B (possibly carcinogenic to humans.

DBCM and bromoform were studied in long-term bioassays. In a 2-year corn oil gavage study, DBCM induced hepatic tumours in female mice, but not in rats, at a dose of 100 mg/kg of body weight per day. In previous evaluations, it had been suggested that the corn oil vehicle may play a role in the induction of tumours in female mice. A small increase in tumours of the large intestine in rats was observed in the bromoform study at a dose of 200 mg/kg of body weight per day. The slope factors based on these tumours are 6.5 × 10-3 [mg/kg of body weight per day]-1 for DBCM, or 1.5 µg/kg of body weight per day for a 10-5 risk, and 1.3 × 10-3 [mg/kg of body weight per day]-1 or 7.7 µg/kg of body weight per day for a 10-5 risk for bromoform.

These two brominated THMs are weakly mutagenic in a number of assays, and they were by far the most mutagenic DBPs of the class in the GST-mediated assay system. Because they are the most lipophilic THMs, additional concerns about whether corn oil may have affected their bioavailability in the long-term studies should be considered. A NOAEL for DBCM of 30 mg/kg of body weight per day has been established based on the absence of histopathological effects in the liver of rats after 13 weeks of exposure by corn oil gavage. IARC has classified DBCM in Group 3 (not classifiable as to its carcinogenicity to humans). A TDI for DBCM of 30 µg/kg of body weight was derived based on the NOAEL for liver toxicity of 30 mg/kg of body weight per day and an uncertainty factor of 1000 (10 each for inter- and intraspecies variation and 10 for the short duration of the study and possible carcinogenicity).

Similarly, a NOAEL for bromoform of 25 mg/kg of body weight per day can be derived on the basis of the absence of liver lesions in rats after 13 weeks of dosing by corn oil gavage. A TDI for bromoform of 25 µg/kg of body weight was derived based on this NOAEL for liver toxicity and an uncertainty factor of 1000 (10 each for inter- and intraspecies variation and 10 for the short duration of the study and possible carcinogenicity). IARC has classified bromoform in Group 3 (not classifiable as to its carcinogenicity to humans).

5) Haloacetic acids

The induction of mutations by DCA is very improbable at the low doses that would be encountered in chlorinated drinking-water. The available data indicate that DCA differentially affects the replication rates of normal hepatocytes and hepatocytes that have been initiated. The dose-response relationships are complex, with DCA initially stimulating division of normal hepatocytes. However, at the lower chronic doses used in animal studies (but still very high relative to those that would be derived from drinking-water), the replication rate of normal hepatocytes is eventually sharply inhibited. This indicates that normal hepatocytes eventually down-regulate those pathways that are sensitive to stimulation by DCA. However, the effects in altered cells, particularly those that express high amounts of a protein that is immunoreactive to a c-Jun antibody, do not seem to be able to down-regulate this response. Thus, the rates of replication in the pre-neoplastic lesions with this phenotype are very high at the doses that cause DCA tumours to develop with a very low latency. Preliminary data would suggest that this continued alteration in cell birth and death rates is also necessary for the tumours to progress to malignancy. This interpretation is supported by studies that employ initiation/promotion designs as well.

On the basis of the above considerations, it is suggested that the currently available cancer risk estimates for DCA be modified by incorporation of newly developing information on its comparative metabolism and modes of action to formulate a biologically based dose-response model. These data are not available at this time, but they should become available within the next 2-3 years.

The effects of DCA appear to be closely associated with doses that induce hepatomegaly and glycogen accumulation in mice. The lowest-observed-adverse-effect level (LOAEL) for these effects in an 8-week study in mice was 0.5 g/litre, corresponding to approximately 100 mg/kg of body weight per day, and the NOAEL was 0.2 g/litre, or approximately 40 mg/kg of body weight per day. A TDI of 40 µg/kg of body weight has been calculated by applying an uncertainty factor of 1000 to this NOAEL (10 each for inter- and intraspecies variation and 10 for the short duration of the study and possible carcinogenicity). IARC has classified DCA in Group 3 (not classifiable as to its carcinogenicity to humans).

TCA is one of the weakest activators of the peroxisome proliferator activated receptor (PPAR) known. It appears to be only marginally active as a peroxisome proliferator, even in rats.

Furthermore, treatment of rats with high levels of TCA in drinking-water does not induce liver tumours. These data strongly suggest that TCA presents little carcinogenic hazard to humans at the low concentrations found in drinking-water.

From a broader toxicological perspective, the developmental effects of TCA are the end-point of concern. Animals appear to tolerate concentrations of TCA in drinking-water of 0.5 g/litre (approximately 50 mg/kg of body weight per day) with little or no signs of adverse effect. At 2 g/litre, the only sign of adverse effect appears to be hepatomegaly. Hepatomegaly is not observed in mice at doses of 0.35 g of TCA per litre in drinking-water, estimated to be equivalent to 40 mg/kg of body weight per day.

In another study, soft tissue anomalies were observed at approximately 3 times the control rate at the lowest dose administered, 330 mg/kg of body weight per day. At this dose, the anomalies were mild and would clearly be in the range where hepatomegaly (and carcinogenic effects) would occur. Considering the fact that the PPAR interacts with cell signalling mechanisms that can affect normal developmental processes, a common mechanism underlying hepatomegaly and the carcinogenic effects and developmental effects of this compound should be considered.

The TDI for TCA is based on a NOAEL estimated to be 40 mg/kg of body weight per day for hepatic toxicity in a long-term study in mice. Application of an uncertainty factor of 1000 (10 each for inter- and intraspecies variation and 10 for possible carcinogenicity) to the estimated NOAEL gives a TDI of 40 µg/kg of body weight. IARC has classified TCA in Group 3 (not classifiable as to its carcinogenicity to humans).

Data on the carcinogenicity of brominated acetic acids are too preliminary to be useful in risk characterization. Data available in abstract form suggest, however, that the doses required to induce hepatocarcinogenic responses in mice are not dissimilar to those of the chlorinated acetic acids. In addition to the mechanisms involved in the induction of cancer by DCA and TCA, it is possible that increased oxidative stress secondary to their metabolism might contribute to their effects.

There are a significant number of data on the effects of dibromoacetic acid (DBA) on male reproduction. No effects were observed in rats at doses of 2 mg/kg of body weight per day for 79 days, whereas an increased retention of step 19 spermatids was observed at 10 mg/kg of body weight per day. Higher doses led to progressively more severe effects, including marked atrophy of the seminiferous tubules with 250 mg/kg of body weight per day, which was not reversed 6 months after treatment was suspended. A TDI of 20 µg/kg of body weight was determined by allocating an uncertainty factor of 100 (10 each for inter- and intraspecies variation) to the NOAEL of 2 mg/kg of body weight per day.

6) Chloral hydrate

Chloral hydrate at 1 g/litre of drinking-water (166 mg/kg of body weight per day) induced liver tumours in mice exposed for 104 weeks. Lower doses have not been evaluated. Chloral hydrate has been shown to induce chromosomal anomalies in several in vitro tests but has been largely negative when evaluated in vivo. It is probable that the liver tumours induced by chloral hydrate involve its metabolism to TCA and/or DCA. As discussed above, these compounds are considered to act as tumour promoters. IARC has classified chloral hydrate in Group 3 (not classifiable as to its carcinogenicity to humans).

Chloral hydrate administered to rats for 90 days in drinking-water induced hepatocellular necrosis at concentrations of 1200 mg/litre and above, with no effect being observed at 600 mg/litre (approximately 60 mg/kg of body weight per day). Hepatomegaly was observed in mice at doses of 144 mg/kg of body weight per day administered by gavage for 14 days. No effect was observed at 14.4 mg/kg of body weight per day in the 14-day study, but mild hepatomegaly was observed when chloral hydrate was administered in drinking-water at 70 mg/litre (16 mg/kg of body weight per day) in a 90-day follow-up study. The application of an uncertainty factor of 1000 (10 each for inter- and intraspecies variation and 10 for the use of a LOAEL instead of a NOAEL) to this value gives a TDI of 16 µg/kg of body weight.

7) Haloacetonitriles

Without appropriate human data or an animal study that involves a substantial portion of an experimental animal's lifetime, there is no generally accepted basis for estimating carcinogenic risk from the HANs.

Data developed in subchronic studies provide some indication of NOAELs for the general toxicity of dichloroacetonitrile (DCAN) and dibromoacetonitrile (DBAN). NOAELs of 8 and 23 mg/kg of body weight per day were identified in 90-day studies in rats for DCAN and DBAN, respectively, based on decreased body weights at the next higher doses of 33 and 45 mg/kg of body weight per day, respectively.

A WHO Working Group for the 1993 Guidelines for drinking-water quality considered DCAN and DBAN. This Working Group determined a TDI of 15 µg/kg of body weight for DCAN based on a NOAEL of 15 mg/kg of body weight per day in a reproductive toxicity study in rats and incorporating an uncertainty factor of 1000 (10 each for intra- and interspecies variation and 10 for the severity of effects). Reproductive and developmental effects were observed with DBAN only at doses that exceeded those established for general toxicity (about 45 mg/kg of body weight per day). A TDI of 23 µg/kg of body weight was calculated for DBAN based on the NOAEL of 23 mg/kg of body weight per day in the 90-day study in rats and incorporating an uncertainty factor of 1000 (10 each for intra- and interspecies variation and 10 for the short duration of the study). There are no new data indicating that these TDIs should be changed.

LOAELs for trichloroacetonitrile (TCAN) of 7.5 mg/kg of body weight per day for embryotoxicity and 15 mg/kg of body weight per day for developmental effects were identified. However, later studies suggest that these responses were dependent upon the vehicle used.

No TDI can be established for TCAN.There are no data useful for risk characterization purposes for other members of the HANs.

8) MX

The mutagen MX has recently been studied in a long-term study in rats in which some carcinogenic responses were observed. These data indicate that MX induces thyroid and bile duct tumours. An increased incidence of thyroid tumours was seen at the lowest dose of MX administered (0.4 mg/kg of body weight per day). The induction of thyroid tumours with high-dose chemicals has long been associated with halogenated compounds. The induction of thyroid follicular tumours could involve modifications in thyroid function or a mutagenic mode of action. A dose-related increase in the incidence of cholangiomas and cholangiocarcinomas was also observed, beginning at the low dose in female rats, with a more modest response in male rats. The increase in cholangiomas and cholangiocarcinomas in female rats was utilized to derive a slope factor for cancer. The 95% upper confidence limit for a 10-5 lifetime risk based on the linearized multistage model was calculated to be 0.06 µg/kg of body weight per day.

Source & ©: IPCS "Environmental Health Criteria for Disinfectants and disinfectant by-products ,
EHC 216, Chapter 1: Summary, section 1.5

5.1.3 TDI for chlorine dioxide by-products

The source document for this Digest states:

Chlorite

The primary and most consistent finding arising from exposure to chlorite is oxidative stress resulting in changes in the red blood cells. This end-point is seen in laboratory animals and, by analogy with chlorate, in humans exposed to high doses in poisoning incidents. There are sufficient data available with which to estimate a TDI for humans exposed to chlorite, including chronic toxicity studies and a two-generation reproductive toxicity study. Studies in human volunteers for up to 12 weeks did not identify any effect on blood parameters at the highest dose tested, 36 µg/kg of body weight per day. Because these studies do not identify an effect level, they are not informative for establishing a margin of safety.

In a two-generation study in rats, a NOAEL of 2.9 mg/kg of body weight per day was identified based on lower auditory startle amplitude, decreased absolute brain weight in the F1 and F2 generations, and altered liver weights in two generations. Application of an uncertainty factor of 100 (10 each for inter- and intraspecies variation) to this NOAEL gives a TDI of 30 µg/kg of body weight. This TDI is supported by the human volunteer studies.

10) Chlorate

Like chlorite, the primary concern with chlorate is oxidative damage to red blood cells. Also like chlorite, 0.036 mg of chlorate per kg of body weight per day for 12 weeks did not result in any adverse effect in human volunteers. Although the database for chlorate is less extensive than that for chlorite, a recent well conducted 90-day study in rats identified a NOAEL of 30 mg/kg of body weight per day based on thyroid gland colloid depletion at the next higher dose of 100 mg/kg of body weight per day. A TDI is not derived because a long-term study is in progress, which should provide more information on chronic exposure to chlorate.

Source & ©: IPCS "Environmental Health Criteria for Disinfectants and disinfectant by-products” ,
EHC 216, Chapter 1: Summary, section 1.5

5.1.4 TDI for ozonation by-products

The source document for this Digest states:

11) Bromate

Bromate is an active oxidant in biological systems and has been shown to cause an increase in renal tumours, peritoneal mesotheliomas and thyroid follicular cell tumours in rats and, to a lesser extent, hamsters, and only a small increase in kidney tumours in mice. The lowest dose at which an increased incidence of renal tumours was observed in rats was 6 mg/kg of body weight per day.

Bromate has also been shown to give positive results for chromosomal aberrations in mammalian cells in vitro and in vivo but not in bacterial assays for point mutation. An increasing body of evidence, supported by the genotoxicity data, suggests that bromate acts by generating oxygen radicals in the cell.

In the 1993 WHO Guidelines for drinking-water quality, the linearized multistage model was applied to the incidence of renal tumours in a 2-year carcinogenicity study in rats, although it was noted that if the mechanism of tumour induction is oxidative damage in the kidney, application of the low-dose cancer model may not be appropriate. The calculated upper 95% confidence interval for a 10-5 risk was 0.1 µg/kg of body weight per day.

The no-effect level for the formation of renal cell tumours in rats is 1.3 mg/kg of body weight per day. If this is used as a point of departure from linearity and if an uncertainty factor of 1000 (10 each for inter- and intraspecies variation and 10 for possible carcinogenicity) is applied, a TDI of 1 µg/kg of body weight can be calculated. This compares with the value of 0.1 µg/kg of body weight per day associated with an excess lifetime cancer risk of 10-5.

At present, there are insufficient data to permit a decision on whether bromate-induced tumours are a result of cytotoxicity and reparative hyperplasia or a genotoxic effect.

IARC has assigned potassium bromate to Group 2B (possibly carcinogenic to humans).

Source & ©: IPCS "Environmental Health Criteria for Disinfectants and disinfectant by-products” ,
EHC 216, Chapter 1: Summary, section 1.5

5.2 How much disinfectants and by-products are we exposed to?

The source document for this Digest states:

Disinfectant doses of several milligrams per litre are typically employed, corresponding to doses necessary to inactivate microorganisms (primary disinfection) or doses necessary to maintain a residual in the distribution system (secondary disinfection).

A necessary ingredient for an exposure assessment is DBP occurrence data. Unfortunately, there are few published international studies that go beyond case-study or regional data.

Occurrence data suggest, on average, exposure to about 35-50 µg of total THMs per litre in chlorinated drinking-water, with chloroform and BDCM being the first and second most dominant species. Exposure to total HAAs can be approximated by a total HAA concentration (sum of five species) corresponding to about one-half of the total THM concentration (although this ratio can vary significantly); DCA and TCA are the first and second most dominant species. In waters with a high bromide to TOC ratio or a high bromide to chlorine ratio, greater formation of brominated THMs and HAAs can be expected. When a hypochlorite solution (versus chlorine gas) is used, chlorate may also occur during chlorination.

DBP exposure in chloraminated water is a function of the mode of chloramination, with the sequence of chlorine followed by ammonia leading to the formation of (lower levels of) chlorine DBPs (i.e., THMs and HAAs) during the free-chlorine period; however, the suppression of chloroform and TCA formation is not paralleled by a proportional reduction in DCA formation.

All factors being equal, bromide concentration and ozone dose are the best predictors of bromate formation during ozonation, with about a 50% conversion of bromide to bromate. A study of different European water utilities showed bromate levels in water leaving operating water treatment plants ranging from less than the detection limit (2 µg/litre) up to 16 µg/L. The brominated organic DBPs formed upon ozonation generally occur at low levels. The formation of chlorite can be estimated by a simple percentage (50-70%) of the applied chlorine dioxide dose.

Source & ©: IPCS "Environmental Health Criteria for Disinfectants and disinfectant by-products” ,
EHC 216, Chapter 1: Summary, section 1.5

5.3 Are there uncertainties in assessing exposure?

The source document for this Digest states:

Uncertainties of water quality data

A toxicological study attempts to extrapolate a laboratory (controlled) animal response to a potential human response; one possible outcome is the estimation of cancer risk factors. An epidemiological study attempts to link human health effects (e.g., cancer) to a causative agent or agents (e.g., a DBP) and requires an exposure assessment.

The chemical risks associated with disinfected drinking-water are potentially based on several routes of exposure: (i) ingestion of DBPs in drinking-water; (ii) ingestion of chemical disinfectants in drinking-water and the concomitant formation of DBPs in the stomach; and (iii) inhalation of volatile DBPs during showering. Although the in vivo formation of DBPs and the inhalation of volatile DBPs may be of potential health concern, the following discussion is based on the premise that the ingestion of DBPs present in drinking-water is the most significant route of exposure.

Human exposure is a function of both DBP concentration and exposure time. More specifically, human health effects are a function of exposure to complex mixtures of DBPs (e.g., THMs versus HAAs, chlorinated versus brominated species) that can change seasonally/temporally (e.g., as a function of temperature, nature and concentration of NOM) and spatially (i.e., throughout a distribution system). Each individual chemical disinfectant can form a mixture of DBPs; combinations of chemical disinfectants can form even more complex mixtures. Upon their formation, most DBPs are stable, but some may undergo transformation by, for example, hydrolysis. In the absence of DBP data, surrogates such as chlorine dose (or chlorine demand), TOC (or ultraviolet absorbance at 254 nm [UVA254]) or bromide can be used to indirectly estimate exposure. While TOC serves as a good surrogate for organic DBP precursors, UVA254 provides additional insight into NOM characteristics, which can vary geographically. Two key water quality variables, pH and bromide, have been identified as significantly affecting the type and concentrations of DBPs that are produced.

An exposure assessment should first attempt to define the individual types of DBPs and resultant mixtures likely to form, as well as their time-dependent concentrations, as affected by their stability and transport through a distribution system. For epidemiological studies, some historical databases exist for disinfectant (e.g., chlorine) doses, possibly DBP precursor (e.g., TOC) concentrations and possibly total THM (and, in some cases, THM species) concentrations. In contrast to THMs, which have been monitored over longer time frames because of regulatory scrutiny, monitoring data for HAAs (and HAA species), bromate and chlorite are much more recent and hence sparse. However, DBP models can be used to simulate missing or past data. Another important consideration is documentation of past changes in water treatment practice.

Source & ©: IPCS "Environmental Health Criteria for Disinfectants and disinfectant by-products” ,
EHC 216, Chapter 1: Summary, section 1.5

5.4 Have epidemiological studies been able to evaluate the risks?

    • 5.4.1 Uncertainties of epidemiological data
    • 5.4.2 Epidemiological studies

5.4.1 Uncertainties of epidemiological data

The source document for this Digest states:

Even in well designed and well conducted analytical studies, relatively poor exposure assessments were conducted. In most studies, duration of exposure to disinfected drinking-water and the water source were considered. These exposures were estimated from residential histories and water utility or government records. In only a few studies was an attempt made to estimate a study participant's water consumption and exposure to either total THMs or individual species of THMs. In only one study was an attempt made to estimate exposures to other DBPs. In evaluating some potential risks, i.e., adverse outcomes of pregnancy, that may be associated with relatively short term exposures to volatile by-products, it may be important to consider the inhalation as well as the ingestion route of exposure from drinking-water. In some studies, an effort was made to estimate both by-product levels in drinking-water for etiologically relevant time periods and cumulative exposures. Appropriate models and sensitivity analysis such as Monte Carlo simulation can be used to help estimate these exposures for relevant periods.

A major uncertainty surrounds the interpretation of the observed associations, as exposures to a relatively few water contaminants have been considered. With the current data, it is difficult to evaluate how unmeasured DBPs or other water contaminants may have affected the observed relative risk estimates.

More studies have considered bladder cancer than any other cancer. The authors of the most recently reported results for bladder cancer risks caution against a simple interpretation of the observed associations. The epidemiological evidence for an increased relative risk of bladder cancer is not consistent -- different risks are reported for smokers and non-smokers, for men and women, and for high and low water consumption. Risks may differ among various geographic areas because the DBP mix may be different or because other water contaminants are also present. More comprehensive water quality data must be collected or simulated to improve exposure assessments for epidemiological studies.

Source & ©: IPCS "Environmental Health Criteria for Disinfectants and disinfectant by-products” ,
EHC 216, Chapter 1: Summary, section 1.5

5.4.2 Epidemiological studies

The source document for this Digest states:

Epidemiological studies must be carefully evaluated to ensure that observed associations are not due to bias and that the design is appropriate for an assessment of a possible causal relationship. Causality can be evaluated when there is sufficient evidence from several well designed and well conducted studies in different geographic areas. Supporting toxicological and pharmacological data are also important. It is especially difficult to interpret epidemiological data from ecological studies of disinfected drinking-water, and these results are used primarily to help develop hypotheses for further study.

Results of analytical epidemiological studies are insufficient to support a causal relationship for any of the observed associations. It is especially difficult to interpret the results of currently published analytical studies because of incomplete information about exposures to specific water contaminants that might confound or modify the risk. Because inadequate attention has been paid to assessing exposures to water contaminants in epidemiological studies, it is not possible to properly evaluate the increased relative risks that were reported. Risks may be due to other water contaminants or to other factors for which chlorinated drinking-water or THMs may serve as a surrogate.

Source & ©: IPCS "Environmental Health Criteria for Disinfectants and disinfectant by-products” ,
EHC 216, Chapter 1: Summary, section 1.5


FacebookTwitterEmail
Themes covered
Publications A-Z
Leaflets