Chlorine

Carcinogenicity

Description of key information

Chlorine reacts with water in aqueous solutions to form hypochlorous acid and/or hypochlorite. The toxicity of sodium hypochlorite following repeated or continuous administration has been extensively studied in a number of species following exposure through chlorinated water. Thus, for oral and dermal application a read across from sodium hypochlorite is scientifically justified.
Although slight and equivocal evidence is reported in two oral studies (NTP, 1992, and Soffritti et al., 1997) for leukaemia in female rats, it can be concluded that carcinogenicity is not a relevant endpoint for the administration of sodium hypochlorite by oral route. 
Carcinogenicity studies in rats and mice are available for the inhalation route. The incidence of neoplasia in these studies was not increased by exposure, indicating that inhaled chlorine in rodents is an upper respiratory tract toxicant but not a carcinogen. 
Long term animal carcinogenicity studies with chlorinated drinking water are presented for sodium hypochlorite. The conclusion is that the available animal studies are not sufficient to indicate a clear relationship between the oral administration of sodium hypochlorite in drinking water and cancer. 
No human data are available on carcinogenicity and the only data are related to chlorinated drinking water for which the epidemiological data are not sufficient to suggest a causal relationship between the use of chlorinated drinking water and increased cancer risk. 
The International Agency for Research on Cancer (IARC, 1991) has concluded that there is inadequate evidence for the carcinogenicity of sodium hypochlorite in animals and that sodium hypochlorite is not classifiable as to its carcinogenicity in humans (Group 3). This conclusion is still valid, taken into account the more recent available data and supported by the current CLP classification, Annex VI.

Justification for classification or non-classification

Taking into account all the available information, it can be concluded that carcinogenicity is not a relevant endpoint for the oral, dermal and inhalation route and sodium hypochlorite/chlorine is thus not classified cancerogenic according to 67/548/EEC and CLP (1272/2008/EC).

Additional information

Chlorine reacts with water in aqueous solutions to form hypochlorous acid and/or hypochlorite. The toxicity of sodium hypochlorite following repeated or continuous administration has been extensively studied in a number of species following exposure through chlorinated water. Thus, for oral and dermal application a read across from sodium hypochlorite is scientifically justified.

Animal data

The potential carcinogenicity of sodium hypochlorite has been examined in different strains of rats and in B6C3F1 mice (Kurokawa et al., 1986, Hasegawa et al., 1986, NTP, 1992; Soffritti et al., 1997) by long term oral administration in drinking water and in female Sencar mice, NMRI mice and in female ddN mice by skin application (Hayatsu et al., 1971, Pfeiffer, 1978, Kurokawa et al., 1984). Its potential carcinogenicty has also been studied in a multigeneration study by the oral administration of chlorinated drinking water to BDII (cPah albino) rats (Druckrey, 1968).

There are no reports of carcinogenicity studies by the inhalation route, although two year inhalation toxicity/carcinogenicity study on chlorine is available (Wolf, 1995)

Sodium hypochlorite has also been tested for tumour promoting effects in female Sencar mice following initiation with dimethylbenzanthracene (Kurokawa et al., 1984), in NMRI mice with benzo-pyrene (Pfeiffer, 1978) and in female ddN mice following initiation with 4-nitroquinoline 1-oxide (Hayatsu et al., 1971).

Oral administration

Groups of 50 male and 50 female B6C3F1 mice, 4-6 weeks old, were given 500 or 1000 mg/l (0.05 or 0.1%) sodium hypochlorite prepared from commercial product (14% available chlorine) in drinking water for 103 weeks. Groups of 73 male mice and 72 female mice were used as controls using distilled water. All surviving mice were sacrificed at week 106 at which time the survival rates were; control males, 48/73; low-dose, 39/50; high dose 37/50; control females, 56/72; low-dose 40/50; high-dose, 39/50. Dose related reductions in body weight occurred in both sexes. Combined tumour incidences of leukemias and malignant lymphomas and of adenomas and adenocarcinomas of the lung were very high in all groups (control and all treated) for both sexes. Also, a high incidence of hyperplastic nodules and hepatocellular carcinomas of the liver in males of all groups was noted. However, no statistically significant differences in tumour incidences were observed for any organ in treated animals. It was concluded that there was no effect upon tumour incidence in either male or female mice (Kurokawa et al., 1986)

Groups of 50 male and 50 female F344 rats, 7 weeks old, were given 0, 500 or 1000 mg/l (0, 0.05 or 0.1% - males) and 0, 1000 or 2000 mg/l (0, 0.1 or 0.2% - females) sodium hypochlorite prepared from a commercial product (14% as available chlorine), in drinking water for 104 weeks. The control groups were given distilled water. Survival rates at 112 weeks were high; control males, 30/50; low-dose, 26/50; highdose, 31/50; control females, 31/50; low-dose, 36/50; high dose, 35/50. Dosedependent inhibition of body weight increase was observed in both male and female rats. Drinking water intakes were comparable among treated and control groups. All three treated groups demonstrated relatively high incidences of tumours of the testis, pituitary, thyroid, lung, pancreas, uterus, mammary gland, spleen and subcutaneous tissue. Histologically, chromophobic adenomas of the pituitary, adenomas and adenocarcinomas of c-cells of the thyroid, adenomas of the lung, insulomas of the pancreas, fibroadenomas of the mammary gland and mononuclear cell leukemias were identified in both males and females. Also, a high incidence of interstitial cell tumours of the testis and fibromas in subcutaneous tissues in males and endometrial polyps of the uterus in females were observed. However, the occurrence of tumours at any site was not significantly greater in rats receiving sodium hypochlorite than in the controls. The proportions of low-dose and high-dose female rats with fibroadenomas of the mammary gland were significantly lower than that of controls. Similarly, the proportion of high-dose male rats with nodular hyperplasia of the liver was decreased. Haematological and serum biochemical analyses did not show significant dose-related changes of any parameters in either sex treated with sodium hypochlorite (Hasegawa et al., 1986, Kurokawa et al., 1986).

Groups of 70 male and 70 female F344 rats were given drinking water containing 0, 70, 140 or 275 ppm available chlorine for up to 2 years. Groups of 10 rats of each sex were pre-selected for evaluation at 14 or 15 and 66 weeks. Survival at 2 years of rats receiving chlorinated water was similar to that of the controls. After 2 years of exposure, the mean body weights of dosed male rats and high-dose female rats were slightly lower than those of their respective control groups. There was a dose related decrease in water consumption by rats. Water consumption by high-dose rats during the second year of these studies was 21% lower than that of control males and 23% lower than that of control females. The incidence of mononuclear cell leukemia in mid-dose, but not high-dose, female rats was significantly higher than that in controls (control, 8/50; low-dose, 7/50; mid-dose 19/50; high-dose 16/50). The proportion of female rats that died of leukemia before the end of the study and the mean time for observation of animals dying with leukemia were similar among all treated groups and controls. Although the marginal increase in leukemia incidence in the mid- and highdose female rats suggested a possible association with the administration of chlorinated water, the incidence of leukemia was not clearly dose related. There was no indication of reduced latency of leukemia and the incidence of leukemia in concurrent controls was less than the mean for historical controls. Furthermore, there was no supporting evidence of an effect in male rats. Thus the marginal increase in leukemia incidence in female rats was considered equivocal evidence of carcinogenic activity. There were no neoplasms or non-neoplastic lesions in male rats that were clearly associated with the consumption of chlorinated water (NTP, 1992).

In the same study, water containing 0, 70, 140 or 275 ppm available chlorine was also given to groups of 70 B6C3F1 mice of each sex for up to 2 years. Groups of 10 mice of each sex were pre-selected for evaluation at 14 or 15 and 66 weeks. Survival at 2 years of mice receiving chlorinated water was similar to that of the controls. Mean body weights oftreated mice were slightly lower than those of their respective control groups. There was a dose related decrease in water consumption by mice: water consumption by high-dose mice was 31% lower than that of controls for males and 26% lower for females. There were no neoplasms or non-neoplastic lesions in male or female mice that were clearly associated with the consumption of chlorinated water (NTP, 1992).

Four groups of 50 male and 50 female Sprague-Dawley rats each, 12 weeks old at the start of the study, received drinking water with added sodium hypochlorite, with concentrations of active chlorine of 750, 500 or 100 mg/l (treated groups) or tap water (active chlorine <0.2 mg/l) (control group), respectively, for 104 weeks. No group was treated with non-chlorinated water. In the male rats, a slight increase of total tumours was seen in all treated groups, compared to controls, but the effect was not doserelated. Among the treated female rats, an increased incidence of lymphomasleukemias (not specified) was observed, but the effect was not clearly dose-dependant. The control group (at <2mg/l of active chlorine) showed an unusually low incidence of leukemia (0 cases against 4 cases in male control group). Some unusual tumours in Sprague-Dawley rats were observed randomly but were not treatment related. Considering the conclusions not definitive, the authors suggested further studies for a quantitative assessment of cancer risk (Soffritti et al., 1997).

Chlorinated water, containing available chlorine at a periodically controlled level of 100 mg/l was well tolerated when given daily as drinking water over the whole lifespan (maximum of 2 years) to 236 BDII (cPah albino) rats for five generations (excluding F3 and F4). Following haematological tests and final necropsy of all animals, the results indicated that there was no difference in survival and in malignant tumour incidence in any generation group when compared to the untreated controls (Druckrey, 1968).

Summary of oral administration

Although slight and equivocal evidence is reported in two studies (NTP, 1992, and Soffritti et al., 1997) for leukaemia in female rats, it can be concluded that carcinogenicity is not a relevant endpoint for the administration of sodium hypochlorite by oral route.

Inhalation exposure

Groups of approximately 70 female and male F344/N rats and B6C3F1 mice were exposed to 0, 0.4, 1.0 or 2.5 ppm (0, 1.2, 3 or 7.5 mg/m³) chlorine gas for 6 hours/day 5 days/week (mice and male rats) or 3 alternate days/week (female rats) for 2 years, with an interim necropsy of rats performed at 12 months (10 rats/sex/concentration group). A complete necropsy was performed on all animals. Exposure-dependent lesions were confined to the nasal passage in all sex and species groups. Chlorine-induced lesions, which were most severe in the anterior nasal cavity, included respiratory and olfactory epithelial degeneration, septal fenestration, mucosal inflammation, respiratory epithelial hyperplasia, squamous metaplasia and goblet cell hypertrophy and hyperplasia, and secretory metaplasia of the transitional epithelium of the latera meatus. Intracellular accumulation of eosinophilic proteinaceous material was also a prominent response involving the respiratory transitional and olfactory epithelia and in some cases the squamous epithelium of the nasal vestibule. Many of these nasal lesions exhibited an increase in incidence and/or severity that was related to chlorine exposure concentration and were statistically-significantly increased at all chlorine concentrations studied. The incidence of neoplasia was not increased by exposure indicating that inhaled chlorine in rodents is an upper respiratory tract toxicant but not a carcinogen (CIIT 1993; Wolf et al., 1995). The LOAEL for respiratory irritation was 0.4 ppm.

Skin application

There is no dermal carcinogenicity study on sodium hypochlorite available.

In a mouse skin two-stage carcinogenesis model study, a control group of 40 ddN female mice, 5 weeks old, were given 60 topical applications of sodium hypochlorite (10% effective chlorine solution) (purity, vehicle and frequency of application unspecified). Another group of 40 female mice (group 2) were given 20 applications of 4-nitroquinoline 1-oxide, a potent carcinogen, (purity, vehicle and frequency of application unspecified); and a third group of 40 mice (group 1) were first given the carcinogen, painted on in 20 applications of 0.05 mg in 0.25% (w/v) benzene solution over the course of 50 days (1mg/mouse total), followed, six days later, by 45 applications of approx. 0.05 ml sodium hypochlorite solution (10% available chlorine), during a period of 245 days (frequency of application unspecified). No skin tumours occurred in the mice given sodium hypochlorite alone. However, skin tumours were seen in 9/32 mice given applications of sodium hypochlorite following initiating doses of 4-nitroquinolene 1-oxide and included one fibrosarcoma, three squamous cell carcinomas and five papillomas. No skin tumours occurred in mice given applications of 4-nitroquinolene 1-oxide only. The results of this study suggest that sodium hypochlorite might have the potential for co-carcinogenicity or tumourpromoting (Hayatsu et al., 1971).

A 1% solution of sodium hypochlorite applied to the skin twice weekly for 10 weeks prior to, or after treatment with 750 and 1500 μg of 3,4-benzopyrene, a potent carcinogen, reduced the number of NMRI mice developing skin tumours when compared with those receiving dermal applications of two doses of the carcinogen alone (Pfeiffer, 1978).

In another mouse skin two-stage carcinogenesis model study, a group of 20 female Sencar mice, 6 weeks old, were given a single topical application of 20 nmol (5μg) dimethylbenzanthracene (DMBA) in acetone, followed by applications of 0.2 ml of 1% sodium hypochlorite solution in acetone twice weekly for 51 weeks. A group of 15 female mice given a single application of DMBA followed by applications of acetone served as controls. The effective number of mice was 20; the number of survivors was not reported. A squamous cell carcinoma of the skin occurred in 1/20 mice treated with DMBA and sodium hypochlorite, whereas none occurred in the initiated controls. In the same test to verify the complete carcinogenic activity, another group of 20 female Sencar mice, 6 weeks old were given topical applications of 0.2 ml of a solution of 1% (10 g/l) sodium hypochlorite in acetone twice weekly for 51 weeks at which time the study was terminated. A group of 15 female mice given applications of acetone were used as controls. All animals survived to the end of the study. No skin tumours were observed in the treated or control groups (Kurokawa et al., 1984).

Summary of skin application

The available data show no carcinogenic effect due to topical application of sodium hypochlorite solution at different concentrations. There is an indication, supported by a poorly described study, of co-carcinogenicity using NaClO as a promoting agent. In the same study, 4-nitroquinolene 1-oxide alone did not show carcinogenicity, suggestive of methodological problems or unusual responses of the mouse strain used. Two other studies using different initiators did not show promoting effects of NaClO.

Human Data

There are no case reports or epidemiological studies of human carcinogenicity directly linked to administration of sodium hypochlorite. The only human data relating with sodium hypochlorite are in connection with its use to disinfect drinking water.

Drinking water epidemiological studies

The addition of sodium hypochorite and chlorine gas gives the same chlorine species in solution - i.e., an equilibrium mixture of mainly hypochlorous acid and hypochlorite anion. In this way much of the general population is exposed to hypochlorite via drinking water.

Several epidemiology studies are available to evaluate the carcinogenicity of chlorinated drinking water, but their use in sodium hypochlorite risk assessment is questionable considering the great differences in water quality in diverse geographical areas and the contribution of the chemical species present in water supplies and the derived reaction products. For an overall assessment of the effect of chlorinated drinking water on the general population, including sodium hypochlorite as the active chlorine source, specific activity should be assessed while the following text is only an initial contribution.

IARC reviewed in 1991 the available studies and gave prominence to the difficulties in the interpretation of the data for an evaluation of the carcinogenicity of chlorinated drinking water. In the studies performed there are several methodological inadequacies, many confounding variables, and no causal link between increased cancer risk and consumption of chlorinated drinking water. Some studies have reported a correlation between the higher risk of cancer of the urinary bladder and the long-term consumption of chlorinated drinking water. Evidence from other studies, although showing some degree of consistency, is severely compromised by the weaknesses outlined above.

IARC overall evaluation was that chlorinated drinking water and hypochlorite salts are not classifiable as to their carcinogenicity to humans and that there is inadequate evidence for the carcinogenicity of chlorinated drinking water and hypochlorite salts in humans (IARC, 1991).

Recent ecological studies reported a weak association between consumption of chlorinated drinking water and cancer of the colon and rectum in men and women (Flaten, 1992 in IPCS EHC 2000); no associations between chlorinated surface water supplies and bladder or stomach cancer were found in the Valencia Province, Spain (Suarez-Varela et al., 1994 in IPCS EHC 2000).

A study in Taiwan reported associations between use of chlorinated drinking water and cancer of the rectum, lung, bladder and kidney (Yang et al., 1998 in IPCS EHC 2000).

Some population-based case-control studies were conducted for bladder, colon, pancreatic, brain and cancer risk by different authors: King and Marrett (1996 cited in IPCS EHC 2000) conducted in residents between 25 and 74 years of age with a histologically confirmed primary cancer or carcinoma in situ of the bladder; they found higher relative risk estimats for a group of nonsmokers associated with many years of exposure to chlorinated surface water.

In a Colorado residents case-control study, McGeehin et al. (1993, cited in IPCS EHC 2000) suggested the hypothesis that prolonged (+30 years) exposure to chlorinated surface water is associated with an increased risk of bladder cancer, while Freedman et al. (1997, cited in IPCS EHC 2000) found a weak association between bladder cancer risk and the use of municipal chlorinated water, limited to cigarette smokers, in a population based case-control study in Maryland.

In a population-based case-control study for bladder, colon and rectal cancer, conducted by Cantor et al. (1998) in Iowa, an increased risk of bladder cancer in current and past male smokers and an association between brain cancer among men and increased duration of exposure to chlorinated surface water was found.

Hildesheim et al. (1998, cited in IPCS EHC 2000) suggested in a another, similar study in Iowa, the same association with bladder cancer and found a larger relative risk for rectal cancer among persons with low dietary fiber intake and longer duration exposure to chlorinated surface water sources compared to persons with high fiber diets and no exposure to chlorinated surface water.

In a population-based case-control for pancreatic cancer conducted by Ijsselmuiden et al.. (1992, cited in IPCS EHC 2000) in Maryland, no increased risk of pancreatic cancer was associated with the consumption of chlorinated drinkink water.

Summary of epidemiological studies

Although some studies reported small relative risks for colon and bladder cancer incidence for population consuming chlorinated drinking water for long periods of time, they are equivocal or insufficent to establish a causal relationship, considering the quality and the completeness of the studies and the interpretation of the available data and of the confounding factors.

Summary of carcinogenicity

Carcinogenicity studies in rats and mice are available for the inhalation route. The incidence of neoplasia in these studies was not increased by exposure, indicating that inhaled chlorine in rodents is an upper respiratory tract toxicant but not a carcinogen.

Long term animal carcinogenicity studies with chlorinated drinking water are presented for sodium hypochlorite. The conclusion is that the available animal studies are not sufficient to indicate a clear relationship between the oral administration of sodium hypochlorite in drinking water and cancer.

No human data are available on carcinogenicity and the only data are related to chlorinated drinking water for which the epidemiological data are not sufficient to suggest a causal relationship between the use of chlorinated drinking water and increased cancer risk.

The International Agency for Research on Cancer (IARC, 1991) has concluded that there is inadequate evidence for the carcinogenicity of sodium hypochlorite in animals and that sodium hypochlorite is not classifiable as to its carcinogenicity in humans (Group 3). This conclusion is still valid, taken into account the more recent available data and supported by the current CLP classification, Annex VI.