Triclosan

Carcinogenicity

Description of key information

Valid data on the carcinogenic potential of triclosan are available.
The following two studies were retained as key studies: 
- Two-year oral chronic toxicity/carcinogenicity study with rat, according to OECD TG 453 (Ciba-Geigy Corp 85152)
- 90-week oral carcinogenicity study with hamster, according  to OECDT TG 451 (Huntingdon Life Sciences Ltd CBG 756/972896)
In addition data from a 18-month carcinogenicity study with mouse conducted according to the OECD TG 451 are available, however only as secondary quotation (Pharmaco LSR 93-2260 cited in SCCP COLIPA No. P32)

Key value for chemical safety assessment

Carcinogenicity: via oral route

Endpoint conclusion

Dose descriptor:

NOAEL

48 mg/kg bw/day

Justification for classification or non-classification

Owing to the results obtained from the studies and discussed above, there is no need for classification of triclosan for carcinogenicity according to the EU Directive 67/548/EEC, and there is no need for classification according to the Annex VI of the CLP regulation.

Additional information

Triclosan showed no tumourigenic potential in lifetime cancer bioassays in rats and hamsters (Ciba-Geigy Corp 85152 and Huntingdon Life Sciences Ltd CBG 756/972896). In contrast, triclosan induced hepatic effects in a mouse carcinogenicity assay, starting at the lowest dose tested of 10 mg/kg bw/day, with liver tumour development observed starting at a dose of 30 mg/kg bw/day (Pharmaco LSR 93-2260 cited in SCCP COLIPA No. P32).

The doses tested in the 3 carcinogenicity assays were comparable, ranging from 12 to 127 mg/kg bw/day (for males) and 17 to 190 mg/kg bw/day (for females) in rats, from 12 to 250 mg/kg bw/day in hamsters of both sexes, and from 10 to 200 mg/kg bw/day in mice of both sexes.

Triclosan did not produce tumours in rats (doses of 12 to 127 or 17 to 190 mg/kg bw/day in males and females, respectively) and in hamsters (doses of 12 to 250 mg/kg bw/day). The NOAEL for the rat lifetime bioassay was determined to be 48 mg/kg bw/day, while the NOAEL for the hamster study was 75 mg/kg bw/day.

Survival was not altered by triclosan treatment in the rat study, whereas decreased survival was observed in both the hamster and the mouse studies.

In general, neither clinical chemistry nor haematology analyses showed any serious effects that could be attributed to triclosan treatment, except in the case of liver-related changes in the mouse study.

Rats showed no changes in liver function enzyme activity in plasma, including alanine (ALT) and aspartate aminotransferase (AST), indicative of hepatic damage. Liver effects in rats in the carcinogenicity study were limited to increased incidences of hepatocyte hypertrophy and hyaline-staining inclusions in high-dose males at early time points in the study. These findings were not observed in animals at the termination of the study.

As with rats, hamsters showed no alterations in liver function enzymes. The effects of triclosan in hamster liver, as observed in the carcinogenicity study, were limited to a slight decrease in organ weight in high-dose females. This finding was attributed to the decreased body weight gain at this dose level. In addition to the organ weight effects, slight, but significant, decreases in erythroid parameters and in triglycerides were measured in mid-dose females.

The only notable histopathological finding was an increased incidence of rarified hepatocytes in a few high-dose males.

An 18-month carcinogenicity study in mice was conducted for Colgate-Palmolive (Pharmaco LSR 93-2260 cited in SCCP COLIPA No. P32); since the original study is not available to us, the data referring to this study were taken the SCCP COLIPA review P 32. In contrast to the observations made in rats and hamsters, the mouse study revealed liver effects in this species which included increased activity of liver function enzymes in plasma; decreased blood cholesterol levels; significant increases in liver weights; increased incidences of hepatic nodules, discolorations, or masses; hepatocellular hyper­trophy; and, increased incidences of both hepatocellular adenomas and hepatocellular carcinomas.

Especially referring to the neoplastic findings, dose-dependent increases in liver adenomas and carcinomas, with increased incidence in male versus female mice were reported. The incidences of mice bearing at least 1 hepatic tumour (i.e., combined adenoma and carcinoma data) were 6, 10, 17, 32, 42 in males and 0, 1, 3, 6, 20 in females at doses of 0, 10, 30, 100, and 200 mg/kg bw/day, respectively. The NOEL for oncogenicity was considered to be 10 mg/kg bw/day, based on observations of liver tumours attributed to triclosan at doses of 30 mg/kg bw/d and higher. General liver toxicity was observed at all doses, including the lowest dose of 10 mg/kg bw/day, while increases in numbers of hepatic tumours were observed at doses of 30 mg/kg bw/day and higher.

 

Thus, in summary, triclosan produced hepatic effects and hepatic tumours in mice, but little evidence of toxicity and no tumours in rats. Hamsters showed increased liver toxicity relative to the rat, but no tumours.In assessing the data and interpreting the findings from all of the carcinogenicity studies, it was important to further evaluate the differences between the rodent species, specifically mice, rats, and hamsters. Liver biochemical, cell proliferation, and morphological responses to triclosan were investigated in a series of studies in all 3 rodent species. Triclosan showed peroxisome proliferator-type effects in the liver of mice (e.g., induction of large increases in peroxisomal fatty acid beta-oxidation, 11- and 12-hydroxylation of lauric acid, and levels of CYP4A proteins, together with increases in the numbers and size of peroxisomes), but not in rat or hamster livers at the doses tested.

It is notable that triclosan induced hepatic cell proliferation in the mouse, and not in the hamster or rat, in investigational studies of replicative DNA synthesis. Taking into account the results from these special investigations, sub-chronic toxicity data indicating an increased sensitivity in mice to triclosan’s hepatic effects, and pharmacokinetic data showing greater exposure levels to triclosan in mice compared to rats or hamsters, there is strong evidence that triclosan has peroxisome proliferator effects in mouse liver, but not in rat or hamster liver.

Given the association of peroxisome proliferation, cell proliferation, and tumour induction reported in the mouse, but no effects of these types in rats and hamsters, it was concluded that the mouse is uniquely sensitive to triclosan in the liver, thus, the triclosan-induced hepatic tumours in mice are considered to be a species-specific effect.

Importantly, it is generally accepted that chemicals that induce peroxisome proliferation and result in rodent hepatocarcinogenicity do not pose a health risk to humans. As such, the mouse is considered not to be a relevant or appropriate animal model for the evaluation of the tumourigenic potential of triclosan in humans.

As the hamster is the rodent species most comparable to humans based on pharmacokinetics (ADME) data and, therefore, the most appropriate model for the assessment of triclosan safety in humans, the lack of tumour development in this species provides the most conclusive non-clinical evidence for the absence of human carcinogenic potential.

This conclusion is supported strongly by the absence of effects of triclosan in a wide variety of in vitro and in vivo genotoxicity assays.

In addition, it is justifiable to state that the results of the studies in this section are considered representative of triclosan having the composition as was defined above in this document. To our judgement the dioxin and furan impurities that can occur in triclosan synthesis when present at concentrations above those found in the test substance used for these studies can be expected to have some level of effect on repeated dose endpoints and the resultant NOAELs determined for the respective study. Thus, in order to be represented by the animal testing included in this dossier the triclosan composition should have impurities equal to or below the values shown in the following Table 1.

Table 1.  Name of Substance	Composition
Triclosan: Phenol, 5-chloro-2-(2,4-dichlorophenoxy)-	>99.1% <100 % (w/w)
2,4-dichlorophenol	<10 mg/kg
3-Chlorophenol and 4-Chlorophenol, total	<10 mg/kg
2,8-Dichlorodibenzo-p-dioxin	<0.5 mg/kg
2,8-Dichlorodibenzofuran	<0.25 mg/kg
1,3,7-Trichlorodibenzo-p-dioxin	<0.25 mg/kg
2,4,8-Trichlorodibenzofuran	<0.5 mg/kg
2,3,7,8-Tetrachlorodibenzodioxin	ca. 0.0%(w/w)
2,3,7,8-Tetrachlorodibenzofuran	ca. 0.0%(w/w)
non-specified impurities	<0.8785%(w/w)

In summary, taking into account both the rat and hamster carcinogenicity studies, the mouse being an inappropriate model for assessing triclosan tumourigenic potential in humans, the overall NOAEL value from the lifetime bioassays was considered to be 75 mg/kg bw/day from the 95-week study in hamsters, as this species was judged to be the most relevant to humans based on pharmacokinetic (elimination, metabolism) data and, therefore, an appropriate animal model for the assessment of human safety. However, it should be noted that the NOEL of 48 mg/kg bw/day from the rat study represents, arguably, the most conservative value from carcinogenicity studies in rats or hamsters.