Final Assessment

3. Human Health

3.1 Toxicology Profile of Triclosan

Reviews of the triclosan toxicological database conducted by the US EPA (2008b), the Australian Department of Health and Ageing (NICNAS 2009), which was adopted by the Organisation for Economic Co-operation and Development (OECD) at the Screening Information Data Set (SIDS) Initial Assessment Meeting (SIAM) 30 in April 2010 (OECD 2011), and the EU SCCP (2009) and SCCS (2011) were used to inform Health Canada's human health hazard evaluation. Where appropriate, secondary review references are cited. Additional review of pivotal toxicological studies was undertaken by Health Canada when deemed necessary. A review of additional toxicological studies investigating the effects of triclosan on thyroid hormones presented by the US EPA Office of Research and Development to the US FIFRA Scientific Advisory Panel (US EPA 2011a), was also considered. Furthermore, more recently published studies and reviews since the publication of the preliminary assessment (up to April 2015) were considered and incorporated into the assessment when determined relevant for risk assessment purposes.

3.1.1 Metabolism and Toxicokinetics

Data available on the absorption, distribution, metabolism and elimination of triclosan in mice, rats, hamsters, rabbits, dogs and baboons suggest that there are interspecies differences in the clearance profile.

Oral metabolism studies conducted in hamsters with radiolabelled triclosan showed that 60-80% of the radioactivity was excreted in the urine, while 12-35% was excreted with feces. Compared to the low dose, administration of a single high dose or repeat dose resulted in a shift towards urine elimination and a decrease in fecal elimination. Radioactivity in fecal material was primarily parent, suggesting little metabolism prior to limited biliary excretion. Intravenous and oral administration at low doses resulted in similar patterns of elimination in male and female hamsters. At terminal sacrifice, following a single or repeated oral dose, negligible low residues were found in organs, and low amounts were noted in blood. In fact, residues at terminal sacrifice were lower following repeated dosing in comparison to single dosing, suggesting an increased clearance rate. The major urinary metabolite detected after oral and intravenous administration in hamsters was the glucuronide conjugate of triclosan, while the major fecal metabolite was parent triclosan in all oral dose groups. Distribution patterns in the orally and intravenously dosed animals were similar between the single- and repeated-dose groups, with the highest residual radioactivity found in the kidney, liver, lung and plasma. No organ demonstrated accumulation of triclosan with the highest levels of triclosan equivalent in the plasma 7 days after dosing. Urinary excretion was also found to be a major route of elimination following oral, intravenous and intraduodenal administration in rabbits and oral administration in baboons. The major urinary metabolite in the baboon was a glucuronide conjugate (US EPA 2008b).

Following oral administration of radiolabelled triclosan in mice, rats and dogs, triclosan was rapidly absorbed and eliminated primarily through the feces via biliary excretion. Following intravenous administration in dogs, feces expressed about 60% of unchanged parent, suggesting an efficient biliary excretion. Urinary excretion was secondary to that in the gastrointestinal tract. This excretory pattern was consistent following either intravenous or intraduodenal administration in these species. Following repeated oral administration in the mouse and rat, triclosan concentrations were higher in the liver than in plasma, supporting the liver as a target organ. In fact, liver toxicity is noted to be a consistent finding in the rodent database (see below). Triclosan was found to be metabolized in rats to both glucuronide and sulfate conjugates. Although different ratios of the individual glucuronide and sulfate conjugates were observed among species, no unique species-specific metabolites have been identified to date. Repeated high-dose administration of triclosan was also shown to change the ratio of these two metabolites in hamsters, mice and monkeys, with the sulfate shown to predominate following chronic oral administration (SCCP 2009). Primary excreted compounds in the urine following single oral exposures in mice included the unmetabolized parent compound and two parent conjugates (sulfate and glucuronide conjugates of triclosan); fecal excretion was primarily that of the free parent compound, as only small amounts of glucuronide were detected, and no sulfate was detected. In addition, four conjugated metabolites (M5, M6, M8 and M9) accounting for 5% of the administered dose were detected in kidney, plasma and liver extracts in the mouse. The major biliary product in the rat was the glucuronide conjugate, with unmetabolized parent compound contributing up to 30% of residues. The major urinary metabolite in the rat after oral and intravenous administration was the glucuronide conjugate of triclosan. In the rat, the parent compound could be detected in the brain, indicating that triclosan crosses the blood-brain barrier (US EPA 2008b).Whole-body autoradiography studies in the mouse and rat showed the presence of two peak concentrations in the plasma following single or repeated dosing, indicating enterohepatic circulation. As such, these species with significant enterohepatic circulation would experience an enhanced or prolonged local exposure to triclosan in the liver and gastrointestinal tract (SCCP 2009). Consistent with this, liver toxicity was noted to be the most consistent finding in the rodent database.

In humans, triclosan is rapidly absorbed and distributed, with plasma levels increasing rapidly within 1-4 hours. Following oral and dermal administration, absorbed triclosan is nearly totally converted to glucuronic and sulfuric acid conjugates due to a pronounced first-pass effect, with only trace amounts of the parent compound detected in the plasma. Elimination is rapid, with a terminal plasma half-life of 21 hours (SCCP 2009). Similar to baboons, hamsters, monkeys and rabbits, the major route of excretion is via the kidneys (24-83%, according to Sandborgh-Englund et al. 2006), with the majority of the compound appearing as the glucuronide conjugate. Unlike the excretion pattern noted in rodents, excretion of triclosan in the feces represents a smaller portion of the administered dose (10-30%), and triclosan is present in the feces as the free unchanged compound in humans. The human oral and dermal data provide no evidence of bioaccumulation potential (SCCP 2009).

There is sufficient evidence that the toxicokinetics of triclosan are different in humans and rodents; however, the interspecies differences are difficult to quantify based on the available toxicokinetic data. Data examining area under the plasma concentration versus time curve (AUC) and maximum concentrations in plasma (C_max) in rodents were typically generated with doses 10-fold higher or more than those used in humans. In general, C_max values were lower in humans than in rodents, but AUC data were more variable, depending on the dosing regimen as shown below:

For a single oral dose of 2 mg/kg body weight (bw) per day in rats and mice, AUC values ranged from 63.9 µg·h/mL (rats) to 166 µg·h/mL (mice), and C_max values ranged from 4.77 µg/mL (rats) to 19.48 µg/mL (mice); single oral doses ranging from 0.017 to 0.17 mg/kg bw per day in adult humans yielded AUC values from 0.2 to11.2 µg·h/mL and C_maxvalues from 0.023 to 0.974 µg/mL (SCCP 2009).
For repeated doses of 2 mg/kg bw per day (14 days) in rats, an AUC value of 77.4 µg·h/mL and a C_max value of 4.49 µg/mL were reported. In adult humans, an AUC value of 219 µg·h/mL and a C_max of 0.878 µg/mL were reported after daily swallowing of a dental slurry containing 0.3 mg/kg bw per day for 14 days. Similar doses in toothpaste (expelled after brushing) resulted in an AUC of 34 µg·h/mL and a C_max of 0.146 µg/mL in adult humans (SCCP 2009).

In dermal absorption studies, triclosan was shown to be relatively well absorbed through the skin in all tested species. In vivo systemic absorption in humans following dermal application of products containing triclosan ranged from 11% to 17%, depending on the formulation, applied dose, duration of exposure, type of skin and skin occlusion (Maibach 1969; Stierlin 1972; Queckenberg et al. 2010). In vitro dermal absorption studies using human skin and various formulations containing triclosan showed dermal absorption values ranged from 7% to 30% (Moss et al. 2000; SCCP 2009).

In the in vivo dermal absorption studies in rats, the extent of dermal absorption was much more variable ranging from 4% to 93%, depending on formulation, applied dose and duration of exposure (Black and Howes 1975; Chun Hong et al. 1976; Ciba-Geigy 1976a; Moss et al. 2000; SCCP 2009). Lower absorption ranging from 4% to 28% was reported with triclosan in shampoo, soap suspension or a cream formulation. Higher absorption was observed with triclosan in an aqueous solution or in petroleum jelly (SCCP 2009). In addition, the US EPA reported in vivo dermal absorption in rabbits of up to 48% of an applied dose (US EPA 2008b).

3.1.2 Acute Toxicity

Technical triclosan was non-toxic via oral and dermal routes and of moderate toxicity via the inhalation route in rats. It was moderately irritating to the rabbit eye and mildly to moderately irritating to the rabbit skin. Triclosan is not considered a skin sensitizer based on the results from a guinea pig test (US EPA 2008b).

3.1.3 Subchronic Toxicity

In a 28-day dietary study, exposure of MAGf[SPF] mice (five of each sex per dose) to technical triclosan at a dose of 6.48 or 135.59 mg/kg bw per day in males and 8.25 or 168.78 mg/kg bw per day in females resulted in no effects on mortality, body weight or feed consumption. A no-observed-adverse-effect level (NOAEL) of 6.48 mg/kg bw per day (males) and 8.25 mg/kg bw per day (females) was established based on changes in clinical chemistry (increases in alkaline phosphatase, alanine aminotransferase and aspartate aminotransferase activities; significant decrease in globulin fraction) and liver pathology (an increased incidence of liver cell necrosis, hemosiderosis of Kupffer cells in the vicinity, cytoplasmic vacuoles in hepatocytes, liver cell hypertrophy) observed at the lowest-observed-adverse-effect level (LOAEL) of 135.59 mg/kg bw per day for males and 168.78 mg/kg bw per day for females (US EPA 2008b).

In a 90-day toxicity study, CD-1 mice (15 of each sex per dose) were exposed to triclosan (99.7% a.i.) in the diet at a dose of 0, 25, 75, 200, 350, 750 or 900 mg/kg bw per day. Treatment-related effects were observed at all dose levels in a dose-related manner, as evidenced by clinical pathology, organ weight changes and increased incidence or severity of histopathological lesions (especially of the liver). A statistically significant and generally dose-related reduction in measures of oxygen-carrying capacity, including reduced red blood cells, hemoglobin and hematocrit, was noted in all dose groups, reaching a level of toxicological significance at a dose of 200 mg/kg bw per day. Lower dose groups demonstrated adaptive changes in measures of red blood cells, with deficits less than 10% change from control values. Supporting evidence of a toxicological effect on the hematopoietic system was noted as a regenerative response in the spleen by an increased severity (but not incidence) of splenic hematopoiesis at doses of 200 mg/kg bw per day and greater in males and 750 mg/kg bw per day and greater in females. Statistically significant but not dose-related increases in enzymes indicative of liver injury included aspartate aminotransferase at 750 mg/kg bw per day and above, alanine aminotransferase at 350 mg/kg bw per day and above (males) and 750 mg/kg bw per day and above (females) and alkaline phosphatase (not dose related) at 200 mg/kg bw per day and above (males) and 900 mg/kg bw per day (females). An increase in triglyceride level was observed in males at 350 mg/kg bw per day and above and in females at 750 mg/kg bw per day and above. A decrease in cholesterol level (statistically significant, but not dose related) was reported at 25 mg/kg bw per day and above (NICNAS 2009; SCCP 2009). Given the known increase in peroxisomal fatty acid β-oxidation in mice exposed to triclosan, this is not unexpected (SCCP 2009). At 25 mg/kg bw per day, a slight increase in liver/gallbladder weights in females (7% and 9%, absolute and relative to brain, respectively) was not considered significant; no change in liver/gallbladder weights in males was reported at this dose. Absolute and relative liver/gallbladder weights increased 1.3- to 3.0-fold at 75 mg/kg bw per day and above in both sexes, and the increases were statistically significant. A slight increase in the number of animals with liver lesions (vacuolization observed in 2/15 males and 1/15 females; individual cell necrosis observed in 3/15 females) was observed at 25 mg/kg bw per day (Trutter 1993). This dose level was considered a LOAEL by other agencies (NICNAS 2009; SCCP 2009). Based on the observation that there was no increase in the severity of liver lesions when compared with the control group at this dose level, but a further increase in the incidence of liver lesions (including an increase in both incidence and severity of vacuolization) observed at 75 mg/kg bw per day and above, a NOAEL of 25 mg/kg bw per day was established by Health Canada for this study.

In a 90-day oral study, Sprague-Dawley rats (25 of each sex per dose) received triclosan (purity not reported) at a dietary concentration of 0, 1000, 3000 or 6000 ppm, equivalent to 0, 65, 203 and 433 mg/kg bw per day in males and 0, 82, 259 and 555 mg/kg bw per day in females. A statistically significant decrease in relative spleen weight (11-12%) and increase in relative kidney weight (12-17%) were seen at the middle dose and above in males and females, respectively. A statistically significant and dose-dependent decrease in cholesterol level in the presence of mild liver centrilobular cytomegaly was observed in males at the middle dose and above. A NOAEL of 1000 ppm (equivalent to 65 and 82 mg/kg bw per day for males and females, respectively) was established based on histopathological changes in the liver observed at the LOAEL of 3000 ppm, equivalent to 203 and 259 mg/kg bw per day in males and females, respectively (US EPA 2008b; NICNAS 2009).

In a 91-day study, Beagle dogs (three of each sex per group) were administered daily gelatin capsules containing triclosan at a dose of 0, 25, 50, 100 or 200 mg/kg bw per day. Limited hematology, clinical biochemistry and urinalysis investigations were undertaken, together with a limited histopathological examination. One female died at 25 mg/kg bw per day, two males at 100 mg/kg bw per day and four animals (two females and two males) at 200 mg/kg bw per day. Diarrhea was seen in animals at 25 mg/kg bw per day and above, and the severity and frequency increased with dose. Emesis was also seen in some animals at all doses. Body weight changes were not determined. Hematology and clinical chemistry assessment revealed a number of "abnormal" values in individual animals at 25 mg/kg bw per day and above suggestive of liver dysfunction, as were urinalysis findings of bile salts and polymorphonuclear leukocytes in the urine at all doses. Statistically significant and dose-related increases in combined male and female relative organ weights were seen only in the pancreas (35-50%), kidneys (38-44%) and adrenals (12-29%) at 100 mg/kg bw per day and above. However, histopathological changes were seen in only one of these organs, the kidney. At necropsy, focal interstitial nephritis (a kidney disorder in which the spaces between the kidney tubules become swollen or inflamed) was seen in one female at 100 mg/kg bw per day and in one male and one female at 200 mg/kg bw per day. Additionally, "unusual" Kupffer cell activation, bile retention and/or necrosis were seen in the liver of one female, two males and two animals of each sex at 25, 100 and 200 mg/kg bw per day, respectively. In addition, pathological fat was seen in the liver of one or more male and female animals at all doses. Severe liver damage was associated with bone marrow hyperplasia and was seen in one female at 25 mg/kg bw per day, one male and one female at 50 mg/kg bw per day, two males and two females at 100 mg/kg bw per day and two females at 200 mg/kg bw per day. All of these histopathological changes were absent in control animals. Since clinical signs of toxicity, liver damage and enhanced hematopoietic activity were observed at the lowest dose tested (LOAEL of 25 mg/kg bw per day), a NOAEL was not established (NICNAS 2009; SCCP 2009).

In a 90-day study, Beagle dogs (four of each sex per group) were administered triclosan in the diet at a dose equivalent to 0, 5, 12.5 or 25 mg/kg bw per day. No deaths or effects on body weight gain, feed consumption or water consumption were seen. Pasty to thin feces were observed occasionally in all groups and were considered not treatment related. Compared with controls, no treatment-related effects were seen in hematology, clinical chemistry or urinalysis parameters at the top dose, the only dose level examined. No treatment-related histological findings or effects on organ weight were seen at any dose level. Thus, the NOAEL was determined to be 25 mg/kg bw per day in this 90-day study (NICNAS 2009). SCCP (2009) did not establish a NOAEL for this study, as the highest dose did not produce any treatment-related effects.

In a 90-day oral toxicity study, Beagle dogs were administered daily gelatin capsules containing triclosan at a dose of 0, 12.5, 25, 50 or 100 mg/kg bw per day. Body weight gain in females at 12.5 mg/kg bw per day was significantly lower in relation to untreated controls, but body weight decrements were not observed at higher doses in either sex. There were treatment-related morphological changes in the livers (including focal acidophilic to granular degeneration of the cytoplasm of hepatocytes) of most animals in the 25, 50 and 100 mg/kg bw per day dose groups. One male receiving 100 mg/kg bw per day died after 23 days on test, and another 100 mg/kg bw per day male was sacrificed in extremis after 26 days. One female receiving 50 mg/kg bw per day was sacrificed in extremis after 57 days. Each of the three animals that died or was sacrificed during the study displayed weight loss, anorexia, lethargy and symptoms of jaundice 3-5 days prior to death. Upon autopsy, histopathological examination of tissues revealed that the jaundice was a result of hepatotoxicity. A NOAEL of 12.5 mg/kg bw per day was established based on treatment-related liver morphology changes observed at the LOAEL of 25 mg/kg bw per day (US EPA 2008b).

In a 13-week study, Syrian Golden hamsters (15-20 of each sex per group) were administered triclosan in the diet at a dose equivalent to 0, 75, 200, 350, 750 or 900 mg/kg bw per day. Additional groups of 10 animals of each sex receiving 0, 75, 350 or 900 mg/kg bw per day were sacrificed at week 7 of exposure. No treatment-related deaths were reported in the study. Polyuria (increased urination; statistically significant and dose related) was observed at 350 mg/kg bw per day and above. A slight to moderate increased incidence of blood in urine, which was statistically significant, was reported at 200 mg/kg bw per day and above, along with statistically significant decreases in urine specific gravity (2-3%) and osmolarity (31-65%). Increased coagulation times and statistically significant changes in red blood cell morphology were reported at 750 mg/kg bw per day and above. Statistically significant increases in relative liver (21-36%) and brain weights (14-38%) were observed at 750 mg/kg bw per day in the absence of histopathological changes. Dose-related nephrotoxicity (tubular casts, basophilia and dilation) was reported at 350 mg/kg bw per day and above. Significant increases in the incidence and severity of erosion to the stomach were seen at 750 mg/kg bw per day and above. Consequently, a NOAEL of 75 mg/kg bw per day was established, based on effects on urinalysis parameters together with blood in the urine in both sexes at the LOAEL of 200 mg/kg bw per day (NICNAS 2009). The SCCP considered 75 mg/kg bw per day to be a no-observed-effect level (NOEL) (SCCP 2009). It is interesting to note the absence of liver histopathology in hamsters, which is consistent with the apparent differences in toxicokinetic between hamsters and mice or rats.

In a 90-day dermal toxicity study, Sprague-Dawley rats (10 of each sex per group) were exposed to triclosan in propylene glycol by dermal application at a dose level of 10, 40 or 80 mg/kg bw per day for 6 hours/day during the study. An additional group of 10 animals of each sex per group received 80 mg/kg bw per day for 90 days followed by a 28-day recovery period. Dermal irritation was observed at the application site in all treated animals. Minor adaptive changes in hematology parameters (decrease in red blood cells, hemoglobin and hematocrit) in males and decreased triglyceride (males) and cholesterol levels (males and females) were noted at 80 mg/kg bw per day. Also, an increased incidence of occult blood in the urine (2/9 males vs. 0/10 controls, 3/9 in recovery males, 1/10 in recovery females) and a slight focal degeneration of cortical tubules (3/10 males vs. 1/10 controls) were observed at 80 mg/kg bw per day (Trimmer 1994). The NOAEL of 40 mg/kg bw per day established by the US EPA (2008b) was accepted by Health Canada. A NOAEL of 80 mg/kg bw per day (excluding dermal irritation) was determined by other jurisdictions (NICNAS 2009; SCCP 2009).

In a 21-day inhalation toxicity study, rats (nine of each sex per dose) were exposed (nose only) to triclosan (purity not reported) 5 days/week for 2 hours/day at a dose level of 0, 3.21, 7.97 or 24.14 mg/kg bw per day for males and 0, 4.51, 9.91 or 30.81 mg/kg bw per day for females. Twelve high-dose animals (five males and seven females) died during the course of the study. For females, a NOAEL of 4.51 mg/kg bw per day was established based on treatment-related effects, including slightly decreased body weight, body weight gain, feed consumption and thrombocytes, as well as increased leukocytes and alkaline phosphatase activity and a slightly increased incidence of respiratory irritation, observed at the next dose (LOAEL of 9.91 mg/kg bw per day). In males, treatment-related effects (decreased thrombocytes (platelets) and total serum proteins, increased alkaline phosphatase activity) were observed at the lowest dose tested (Ciba-Geigy 1974). Although the US EPA established a LOAEL of 3.21 mg/kg bw per day based on the above-mentioned effects in males, given a shallow dose-response curve for the measured endpoints, Health Canada determined that the observed effects were minor, and a NOAEL of 3.21 mg/kg bw per day was established.

3.1.4 Reproductive Toxicity

In a two-generation reproduction study in the rat, triclosan (99% a.i.) was administered to Sprague-Dawley rats (25 of each sex per dose) in the diet at a dose of 15, 50 or 150 mg/kg bw per day for 10 weeks prior to mating and through postnatal day (PND) 21 for both generations. No treatment-related effects were seen on mortality, clinical signs or estrous cyclicity. In the F0 generation, there were no significant decreases in parental body weight during pre-mating. Body weight in high-dose F0 females during lactation was significantly decreased on PND 7 (statistically significant). An increased incidence of liver discoloration in 50 and 150 mg/kg bw per day parental F0 males was observed at necropsy, but no histopathological assessment was undertaken of any organs. No effects on reproductive performance were found in the F0 generation. Pups of the F0 generation (Fl pups) showed statistically significant decreases in mean body weight on PNDs 14 and 21 at the 150 mg/kg bw per day dose. Slightly increased pup mortality was observed on PNDs 0-3 in high-dose pups, resulting in a decreased viability index (82% compared with 90% in controls), as well as an increased incidence of dilated renal pelvis at the 150 mg/kg bw per day dose in Fl pups. In Fl parental animals, significantly lower group mean body weights were observed during pre-mating at the 150 mg/kg bw per day dose (statistically significant). Gestational body weights in high-dose Fl females were significantly decreased by 12% during the period of gestation, with a significant negative trend for gestational days 1, 7, 14 and 20. There were no differences in number of pregnant animals, mean gestation duration or mean precoital (pairing to insemination) interval in Fl females. In pups of the Fl parental generation (F2 pups), a slight increase in number of pups found dead or missing was observed at 150 mg/kg bw per day (84% compared with 87% in controls), as well as a statistically significant, but slight (less than 10%), decrease in mean body weights in both sexes compared with controls. The weaning index was decreased at the high dose in F2 pups, and total litter deaths were increased.

A parental NOAEL of 50 mg/kg bw per day was established based on reduced mean body weight observed at the LOAEL of 150 mg/kg bw per day. A reproductive/developmental NOAEL of 50 mg/kg bw per day was established based on reduced pup weights and reduced pup viability in both generations at the LOAEL of 150 mg/kg bw per day) (US EPA 2008b). Similar findings were reported by NICNAS (2009) and SCCP (2009).

The association between triclosan exposure and male reproductive parameters was also examined in the following studies:

In a published male pubertal study by Zorilla et al. (2009), triclosan (99.5% a.i.) was administered daily by oral gavage to weanling male Wistar rats (8-10 per group) at doses of 0, 3, 30, 100, 200, and 300 mg/kg bw per day for 31 days. No visible signs of toxicity were observed in any of treated animals following exposure to triclosan. Triclosan did not affect the age of onset of preputial separation (PPS) at any of the doses evaluated. Triclosan exposure did not significantly affect ventral prostate, seminal vesicle, levator ani plus bulbocavernosus (LABC), epididymal or testicular weights. A significant decrease in the serum testosterone level (60%) was observed at 200 mg/kg bw per day but not at 300 mg/kg bw per day. The serum and pituitary luteinizing hormone (LH) and prolactin (PRL) were not different from controls. Histological evaluation did not reveal any significant treatment-induced lesions or alterations in either testes or epididymides following triclosan exposure. The study authors measured the effect of triclosan on EROD activity as a surrogate to monitor for dioxin contamination (as dioxins activates the aryl hydrocarbon receptor, AhR). Consistent with previous reports for triclosan, no increase in hepatic EROD activity was observed following exposure to triclosan suggesting that triclosan was not contaminated with dioxins.

In a 90-week study, triclosan (99.5% a.i.) was administered in the diet to male hamsters (70 per group) (more details in section 3.1.6 Chronic toxicity) at doses of 0, 12.5, 75, and 250 mg/kg bw per day. A significant increased incidence of absent spermatozoa and abnormal spermatogenic cells and reduced numbers of spermatozoa in the epididymides was observed at a dose of 250 mg/kg bw per day in males that died and those that were sacrificed at the end of the study. An increased incidence of partial depletion of one or more generations of germ cells within the testis was also observed.

In a published study by Kumar et al. (2009), triclosan (98% a.i.) in phosphate buffer saline was administered via intubation to the male Wistar rats (8 per dose) at doses of 0, 5, 10, and 20 mg/kg bw per day for 60 days. Administration of triclosan caused a significant decrease in the weight of testis and sex accessory tissues (SATs) at 10 and 20 mg/kg bw per day. A statistically significant decrease in the activity of both the testicular steroidogenic enzymes (1-HSD and 17-HSD) was observed at two higher dose levels in the in vitro assay. A statistically significant decrease in the serum LH (38.5%), FSH (17%), cholesterol (35%), pregnenolone (31%), and testosterone (41%) levels was reported in males treated with a dose of 20 mg/kg bw per day. Several histopathological abnormalities were observed in cauda epididymis (CE), ductus deference and prostate from rats treated at the highest dose. Further, a 34% decrease in the daily sperm production (DSP) per gram of testis was reported in males at 20 mg/kg bw per day as compared to the control. However, concerns regarding the potential dioxin contamination in the triclosan used in this study were raised by others (Axelstad et al. 2013). As well, there are some concerns regarding the low dosing volumes used in this study.

In a published study by Lan et al. (2013), triclosan (analytical grade) was administered via gavage in corn oil to five-week-old male Sprague-Dawley rats (8 per dose) at doses of 0, 10, 50 or 200 mg/kg bw per day for eight weeks. A statistically significant dose-responsive decrease in daily sperm production and dose-responsive increase in sperm abnormalities were observed at 50 mg/kg bw per day. Decreases in sperm production at 50 mg/kg bw per day compared to control level was approximately 20%. At 200 mg/kg bw per day, the reduction in sperm production was 46% compared with the control group. Sperm abnormalities (1000 sperm examined per dose group) included increased numbers of abnormal sperm heads and tails, reduced hook (banana head), and bent flagella in the mid (~66%) and high (~86-90%) dose groups relative to controls. Statistically significant decreases in both final body weight and ventral prostate gland weight were also observed at 200 mg/kg bw per day. Minor changes in the cauda epididymis at a high dose of triclosan included vacuolated and exfoliated epithelial cells and detached stereocilia from the epithelium. The kinetics of triclosan in the plasma of reproductive organs of male rats was also investigated. While it did not appear to accumulate in the testes or prostate, the authors hypothesized that triclosan could potentially accumulate to some degree in the epididymides based on the epididymial kinetic parameters showing that triclosan had a longer half-life, an increase mean retention time and lower clearance in this organ compared with plasma. No histopathology or organ weight measurements were reported for the liver therefore it could not be determined if the effects in the sperm parameters were secondary to liver injury.

In a published developmental study by Axelstad et al (2013), triclosan (99% a.i) was administered via gavage to Wistar rats from gestation day (GD) 7 to postnatal day (PND) 16 at doses of 0, 75, 150, and 300 mg/kg bw per day. No effects on anogenital distance, nipple retention, prostate weight or prostate histopathology were observed following exposure to triclosan. Given that these endpoints are typically affected by perinatal exposure to anti-androgenic chemicals, the study authors concluded that triclosan exposure at the tested dose levels did not affect male reproductive development.

A recent epidemiological study on men from Nanjing, China, examined 877 idiopathic infertile men and 713 fertile controls between 2005-2010 for an association between triclosan and other phenols and male infertility (Chen et al. 2013). Urinary concentrations of triclosan were measured from single samples along with semen samples obtained from study participants on the same day. Semen analysis included semen volume, sperm concentration and sperm number per ejaculate. No evidence of an association between triclosan urinary level and these semen parameters was observed, although other phenols evaluated in this study did appear to be associated with idiopathic (of unknown cause) male infertility (i.e., 3-tert-octylphenol, 4-n-octylphenol, and 4-n-nonylphenol; Chen et al. 2013). Based on this one study, the limited epidemiological data do not suggest an association between exposure to triclosan and an adverse effect on sperm production in humans.

A recent retrospective study examining urinary concentrations of triclosan in 1699 Canadian women recruited between 2008 and 2011 reported that women in the highest quartile of triclosan levels (greater than 72 ng/ml measured in the first trimester) reported a longer time to pregnancy (TTP) based on responses to a questionnaire (Vélez et al 2015). Mean maternal age was 32.8 years, more than half of the women had had at least one prior pregnancy, and 15% were obese or active smokers during the preconception period, all factors associated with TTP. Further, two thirds of the women had university degrees that may be associated with postponed childbirth. After statistical modelling accounted for maternal and paternal age, smoking, education, body mass index (BMI), and household income, increased TTP for the higher quartile of triclosan exposure was maintained. Factors such as exposures of the male partner and other lifestyle parameters that could also affect TTP were not considered and would need further investigation. Furthermore, since this was a pregnancy-based TTP study, women who were infertile and/or did not have access to infertility treatment were excluded by design from the study (Vélez et al 2015). It should be noted that results from animal studies do not show any treatment-related effects in number of pregnant animals, mean gestation duration or mean precoital (pairing to insemination, equivalent to TTP) intervals after exposure to high levels of triclosan.

The available animal studies provide conflicting results with respect to the examined reproductive endpoints, namely testicular weight, sex accessory organ weights, serum testosterone and LH levels. Further, when signs of testicular toxicity were reported, these effects were observed either at low doses of 20-50 mg/kg bw per day (Kumar et al. 2009 and Lan et al. 2013) or high doses 200-300 mg/kg bw per day (Zorilla et al. 2009 and 90-week study in hamsters). Although differences exist in the strain of rat or species used, design and the duration of each study, these differences may not be sufficient to explain the discrepancies in results between studies. However, it is possible that discrepancies in the observed effects could reflect the presence of impurities in the test substance used in each of the studies. For example, in the study by Zorilla et al. (2009), testicular toxicity occurred only after exposure to high doses of triclosan and the test substance was free of dioxin contamination as proved by measuring of EROD activity. Similar, in a 90-day study in the hamster, in which a technical grade triclosan was used, effects on reproductive parameters were observed only at the highest dose tested. For both Kumar et al. (2009) and Lan et al (2013) studies reporting triclosan effects on reproductive parameters at low doses, it is unknown if dioxins contamination was present. However, concerns regarding the potential dioxin contamination in the triclosan used by Kumar et al. (2009) were raised before (SCCS 2011; Axelstad et al. 2013) and dosing volume was extremely low.

Although, triclosan effects on sperm were not measured in the available 2-generation reproduction study in the rat, notwithstanding that rat fertility is generally resilient to modest reduction in sperm count, there is no evidence of infertility or impaired reproductive performance. Further, no correlation between triclosan urinary levels and semen parameters was observed in one available human epidemiology study (Chen et al. 2013); suggesting that human exposure to triclosan does not result in adverse effects on semen parameters.

3.1.5 Developmental Toxicity

In a prenatal developmental toxicity study in rabbits, triclosan (100% a.i.) was administered by gavage to pregnant female New Zealand White rabbits (18 per group) on gestational days 6-18 at a dose level of 0, 15, 50 or 150 mg/kg bw per day. Signs of maternal toxicity at the high dose (150 mg/kg bw per day) consisted of statistically significant decreases in body weight and feed consumption and statistically significant decreases in body weight gain over the period of treatment. A maternal NOAEL of 50 mg/kg bw per day was established based on decreased body weight gain and feed consumption during treatment observed at the LOAEL of 150 mg/kg bw per day. There were no statistically significant differences in the mean number of resorptions or the resorption/implant ratio between the control and treatment groups. Fetal body weights of both sexes were comparable between the control and treatment groups. No treatment-related external, visceral or skeletal malformations or variations were observed in fetuses. A developmental NOAEL of 150 mg/kg bw per day, the highest dose tested, was confirmed by Health Canada in accordance with that established by the US EPA (US EPA 2008b; NICNAS 2009; SCCP 2009).

In a prenatal developmental toxicity study in rats, triclosan (99.8% a.i.) was administered by gavage to pregnant female Wistar rats (30 rats per group, 60 per group in controls) on gestational days (GD) 6-15 at a dose level of 30, 100 or 300 mg/kg bw per day. At 300 mg/kg bw per day, maternal toxicity consisted of transient diarrhea, statistically significant decreases in body weight gain during treatment, and reduced feed consumption and increased water consumption from onset of treatment through gestation. Based on these findings, a maternal NOAEL of 100 mg/kg bw per day (LOAEL of 300 mg/kg bw per day) was established. There was no evidence of prenatal toxicity at any dose level in this study; therefore, a developmental NOAEL of 300 mg/kg bw per day, the highest dose tested, was established (US EPA 2008b; NICNAS 2009; SCCP 2009).

In a developmental toxicity study in mice, triclosan (99% a.i.) was administered via the diet to 25 CD-1 (ICR)BR female mice at a target dose level of 0, 10, 25, 75 or 350 mg/kg bw per day from GD 6-15. The maternal toxicity appeared to be minor, with liver weight increases (7% and 17% absolute and relative to brain weight, respectively; statistically significant) and 1 out of 25 dams with a tan-coloured liver at 75 mg/kg bw per day. The NOAEL of 25 mg/kg bw per day for maternal toxicity may represent a marginal NOAEL in view of these findings. Developmental effects were noted at 350 mg/kg bw per day as a statistically significant increased incidence of variations (characterized as irregular ossification of the phalanges). Irregular ossification of interfrontal bones (an extra bone between the frontal bones of the skull) was reported at 75 mg/kg bw per day; however, the biological significance of this finding was unclear, and incidences were within historical control ranges (NICNAS 2009). Fetal weight was decreased by 14% and 18%, respectively, at the 75 and 350 mg/kg bw per day target dose levels. The decreased fetal body weight at 75 mg/kg bw per day was considered treatment related, and a developmental NOAEL of 25 mg/kg bw per day was confirmed by Health Canada in accordance with that established by the US EPA (2008b). NICNAS (2009) determined the NOAEL to be 75 mg/kg bw per day.

3.1.6 Chronic Toxicity

In a 1-year toxicity study, triclosan was administered to baboons (seven of each sex per dose) by capsule at a dose level of 0, 30, 100 or 300 mg/kg bw per day. Signs of vomiting were reported at 100 mg/kg bw per day (one female on day 196, one male on day 341) and at 300 mg/kg bw per day (one male on day 17). Failure to eat was reported at 100 mg/kg bw per day and above. Dose-related increases in incidences of diarrhea (4-6 hours after dosing or during the night) occurred within the first 90 days of exposure in 1 out of 14 animals at 30 mg/kg bw per day, in 7 out of 14 animals at 100 mg/kg bw per day and in all animals at the top dose. Statistically significant increases in mean relative kidney and liver weights were reported at 300 mg/kg bw per day and in mean absolute brain weight from 30 mg/kg bw per day (no treatment-related histopathological changes observed) (NICNAS 2009). At necropsy, an effect on the lining of the stomach was observed at the high dose. As seen in other studies, intragastric administration of triclosan via either gavage or capsule appears to cause irritation and/or enteritis, which confounded the interpretation of the study results. A systemic NOAEL of 30 mg/kg bw per day was established based on clinical signs of toxicity observed at the LOAEL of 100 mg/kg bw per day (US EPA 2008b; NICNAS 2009). The SCCP considered 30 mg/kg bw per day to be a NOEL (SCCP 2009).

In a chronic toxicity/carcinogenicity study conducted in male and female Sprague-Dawley rats (85 of each sex per dose), triclosan (99% a.i.) was administered for 104 weeks in the diet at a dose of 0, 300, 1000 or 3000 ppm (equivalent to 0, 15.3, 52.4 or 168.0 mg/kg bw per day in males and 0, 20.0, 66.9 or 217.4 mg/kg bw per day in females, according to US EPA 2008a). An additional satellite group of animals (20 of each sex) received triclosan in the diet at 415.0 mg/kg bw per day (males) or 519.3 mg/kg bw per day (females) for 52 weeks.

No treatment-related effects on mortality, clinical toxicity, ophthalmology, urinalysis or gross pathology were observed at any dose level tested. No carcinogenic potential was demonstrated for triclosan in this study. Slightly but significantly decreased erythrocyte counts were observed in males at the middle (8%) and high doses (11%) at week 78 and at all doses (10%, 14% and 11%) by the end of the study (week 104) compared to controls. Hemoglobin concentrations at the high dose level (6%) and hematocrit at the middle and high dose levels (9%) were decreased in males at week 78, but these effects were not statistically significant at week 104 and were below 10%, and they were therefore considered adaptive. Erythrocyte counts were decreased in females at 66.9 mg/kg bw per day and above at week 78 (8% at the middle dose and 6% at the high dose), but were not statistically significant at week 104 and were below 10%, and they were therefore considered adaptive. It should be noted that hematology parameters in control animals (both male and female) dropped by 8-23% from week 13 to week 104. Minor changes in alanine aminotransferase and aspartate aminotransferase activities were noted in males at a dose of 168 mg/kg bw per day, but the changes never reached levels of biological significance. Slight changes in clinical chemistry (triglycerides, blood urea nitrogen and glucose) were noted (dosed females only) at the earliest test period of week 13. From week 26 onward, the female clinical chemistry results were comparable to those for controls, suggesting that effects noted in subchronic testing may be transient and that animals can compensate adequately with prolonged dosing. Histopathology findings were limited to 7 out of 85 males with hepatocellular hypertrophy and 12 out of 85 males with chronic progressive renal calculi (kidney stones), a common aging disease in rats. Between two and five males or females (out of 85 per group) demonstrated hepatocellular necrosis, determined to be not related to treatment by a pathology working group. The SCCP (2009) considered the NOAEL to be 12-17 mg/kg bw per day based on changes in hematology. However, these changes were considered toxicologically insignificant, and a NOAEL of 52.4 mg/kg bw per day was established based on significant decreases in body weight in male and female rats and non-neoplastic changes of the liver in males at the LOAEL of 168.0 mg/kg bw per day (US EPA 2008b). Similar findings were reported by NICNAS (2009).

In an 18-month carcinogenicity bioassay, triclosan was administered to CD-1 mice (50 of each sex per dose) in the diet at a dose level of 0, 10, 30, 100 or 200 mg/kg bw per day. An additional group of mice (20 of each sex per dose) was exposed for 6 months. There were no significant signs of clinical toxicity at any dose level and no significant effects of treatment on group mean body weight, feed consumption, ophthalmology or urinalysis. A dose-related increase in the activities of alanine aminotransferase and alkaline phosphatase was observed in male and female mice at 100 mg/kg bw per day and above in both the 6-month and 18-month dose groups. Significant decreases in both albumin and total protein levels were observed in males at 6 months and in females at 18 months at doses of 100 mg/kg bw per day and above. Serum cholesterol level was markedly reduced at all doses, including the 10 mg/kg bw per day dose, but the decrease was not considered to be adverse at this dose in the absence of frank liver toxicity. Treatment-related hematological effects included increased reticulocyte count in males and platelet count in males and females at 200 mg/kg bw per day. Mean liver weights (absolute and relative) were increased in both male and female mice at 30 mg/kg bw per day and above at 18 months and at 100 mg/kg bw per day and above at the 6-month interim sacrifice. A dose-related increase in severity of hepatocellular hypertrophy was observed in both male and female mice at 30 mg/kg bw per day and above. A statistically significant increase in the incidence of hepatocellular adenoma and/or carcinoma was observed in male and female mice at 100 mg/kg bw per day and above. The incidence was dose related in both sexes. The combined incidence of adenoma and carcinoma was 12%, 20%, 34%, 64% and 84% for males and 0%, 2%, 6%, 12% and 40% for females at 0, 10, 30, 100 and 200 mg/kg bw per day, respectively. The incidence of adenoma/carcinoma combined exceeded the historical control incidence (17% for males, 1% for females) at 10 mg/kg bw per day but became statistically significant at 30 mg/kg bw per day for males and at 100 mg/kg bw per day for females. Consequently, a NOAEL of 10 mg/kg bw per day was established, based on an increased incidence of liver neoplasms in males and females at the LOAEL of 30 mg/kg bw per day (US EPA 2008b). The SCCP did not establish a NOAEL for this study based on findings of liver effects at all doses and considered triclosan a peroxisome proliferator in mouse liver (SCCP 2009).

In a chronic toxicity/carcinogenicity study in the Bio F1D Alexander Syrian hamster, triclosan (99.5% a.i.) was administered in the diet to 70 animals of each sex per group at a target dose level of 0, 12.5, 75 or 250 mg/kg bw per day for up to 90 weeks. No treatment-related clinical signs of toxicity were observed during the first 80 weeks of the study. After this time, high-dose males showed deterioration in their general clinical condition, with signs of lethargy, hunched posture, pallor, thin appearance and unsteady gait. High-dose males had an increase in mortality after week 80, which correlated with their deteriorating condition. A statistically significant decrease was seen in body weight gain in males receiving 250 mg/kg bw per day at the end of the study (compared with controls), and a slight, although statistically significant, decrease (3%) was seen in feed consumption in females at 250 mg/kg bw per day (NICNAS 2009). At terminal sacrifice, no dose- or treatment-related gross findings were observed in males. However, in the control, low-dose, mid-dose and high-dose female groups, white nodules in the forestomach, pale kidneys and irregular cortical scarring of the kidney were observed in some animals. Microscopically, a statistically significant increase in the incidence of nephropathy was observed in high-dose males and females as compared with control animals and was considered the main factor contributing to death in animals that died before study termination. In males tested with the high dose of triclosan, statistically significant increases in the incidences of absent spermatozoa and abnormal spermatogenic cells and reduced numbers of spermatozoa were observed. An increased incidence of partial depletion of one or more generations of germ cells within the testis was also observed. The incidence of lesions in the stomach was significantly increased in high-dose males and females at termination (focal atypical hyperplasia of the fundic region in males, statistically significant increases in distended gastric glands with or without debris in females). No evidence of potential carcinogenicity of triclosan was observed in this study. A NOAEL of 75 mg/kg bw per day was established, based on decreased body weight gain, increased mortality (males), nephropathy and histopathological findings in the stomach and testes at the LOAEL of 250 mg/kg bw per day (US EPA 2008b; NICNAS 2009).

No chronic dermal toxicity study was available at the time of the assessment report.

3.1.7 Genotoxicity

Triclosan has been tested for genotoxic activity in several assays, including two bacterial reverse mutation tests, an in vitro mammalian cell gene mutation test, in vitro mammalian chromosomal aberration tests, a mammalian bone marrow chromosomal aberration test and an unscheduled deoxyribonucleic acid (DNA) synthesis assay in mammalian cells in culture.

Triclosan was negative at all doses in both bacterial reverse mutation tests with and without metabolic activation (dose levels ranging from 0.005 to 5000 µg/plate) and the in vitromammalian cell gene mutation test (dose levels ranging from 1 to 25 µg/mL), with and without metabolic activation. Nonactivated triclosan was found to induce a dose-related increase in the yield of cells with abnormal chromosome morphology in the in vitro mammalian chromosomal aberration test with dose levels ranging from 1 to 3 µg/mL (18-hour harvest) and at 3 µg/mL (28-hour harvest). The most frequently observed type of chromosome damage was exchange figures. However, no signs of structural chromosomal aberrations were observed in the in vivo bone marrow chromosomal aberration test. Triclosan was also negative in an unscheduled DNA synthesis assay in rat primary hepatocytes at the concentrations tested (US EPA 2008b).

3.1.8 Carcinogenicity Potential in Humans

The US EPA's Cancer Assessment Review Committee of the Office of Pesticide Programs reviewed the carcinogenic potential of triclosan based on a chronic toxicity/carcinogenicity study in hamsters, carcinogenicity studies in mice and rats, metabolism and mutagenicity studies, as well as additional documentation regarding the significance of the mouse study results for human health. The Cancer Assessment Review Committee determined that there was sufficient evidence supporting activation of peroxisome proliferator-activated receptor alpha (PPARα) as the primary mode of action (MOA) for triclosan-induced hepatocarcinogenesis in the mouse. Mutagenic and cytotoxic MOAs were ruled out based on the overall negative in vivo genotoxicity database for triclosan and the lack of evidence supporting a sustained regenerative cellular proliferative response, respectively.

The proposed MOA for liver tumours in mice was found to be theoretically plausible in humans. Although human cells contain PPARα, its activity is approximately 10 times lower than that of mouse hepatocytes. Thus, the human liver would be less susceptible to peroxisome proliferation than the mouse liver. Further, peroxisome proliferators (including hypolipidemic drugs) that are known carcinogens in rodents have not been shown to be carcinogenic in other species, including humans. Consequently, based on quantitative species differences in PPARα activation and differences in toxicokinetics, triclosan-induced carcinogenicity by the proposed MOA was considered by the US EPA to be quantitatively implausible and unlikely to take place in humans. In accordance with the US EPA Final Guidance for Carcinogen Risk Assessment, the US EPA's Cancer Assessment Review Committee classified triclosan as "Not likely to be carcinogenic to humans" (US EPA 2008c).

According to the European Union and Australian classification systems, triclosan is not considered classifiable as a carcinogen (SCCP 2009, NICNAS 2009).

3.1.9 Neurotoxicity

In a 14-day neurotoxicity study in rats exposed to triclosan at a dose level of 0, 100, 300, 1000 or 2000 mg/kg bw per day, a slight inhibition of movement, decreased muscular tone, polydypsia (excessive thirst) and polyuria (increased urination) were observed at 300 mg/kg bw per day, with more pronounced signs at 1000 mg/kg bw per day. No changes in brain weights or histopathology and no changes in peripheral nerves were observed at any dose level tested (US EPA 2008b).

3.1.10 Thyroid Effects

In a published short-term (4-day) study by Crofton et al. (2007), weanling female Long-Evans rats (27-29 days old) were exposed via oral gavage to triclosan at a dose of 0, 10, 30, 100, 300 or 1000 mg/kg bw per day. Decreased serum total thyroxine (T₄) concentrations and increased liver weights were reported in exposed animals. Serum T₄ concentrations were reduced in a dose-dependent manner by 28%, 34% and 53% at 100, 300 and 1000 mg/kg bw per day, respectively. No significant changes were seen at 10 or 30 mg/kg bw per day. The study authors did not report thyroid-stimulating hormone (TSH) levels. The study NOEL was 30 mg/kg bw per day, and the lower 95% confidence limit on the benchmark dose (BMDL) (calculated by the study authors) for a 20% reduction in T₄ was 35.6 mg/kg bw per day.

In a published study by Zorrilla et al. (2009), the effect of triclosan on the thyroid was investigated using the pubertal assay. Weanling male rats were dosed via oral gavage for 30 days starting on PND 23. Animals were exposed to 0, 3, 30, 100, 200 or 300 mg/kg bw per day. Mean serum T₄ concentrations were decreased in a dose-dependent manner by 47%, 50%, 80% and 81% at 30, 100, 200 and 300 mg/kg bw per day, respectively. Triiodothyronine (T₃) was affected only at 200 mg/kg bw per day, while TSH was not affected statistically significantly at any dose. Mean liver weight in male rats was increased significantly at 100 mg/kg bw per day and above, suggestive of hepatic enzyme induction and increased clearance of thyroid hormones. However, the study noted no induction of liver uridine diphosphate-glucuronyl transferase at 3 or 30 mg/kg bw per day. In the same study, decreased serum testosterone was observed at 200 mg/kg bw per day only, although the onset of puberty (balano-preputial separation) and growth of androgen-dependent reproductive tissues (including epididymides and testis) were not altered. At the highest dose, a few animals showed testicular degeneration (multinucleated giant cells within the seminiferous tubule epithelium); however, this change was minimal and not correlated with decreased testosterone or testis weight in the individual animals. The study NOEL was 3 mg/kg bw per day, and the BMDL (calculated by the study authors) for a 20% reduction in T₄ was 7.23 mg/kg bw per day.

In a published study by Paul et al. (2010a), exposure of weanling female Long-Evans rats by oral gavage to triclosan at a dose of 10, 30, 100, 300 or 1000 mg/kg bw per day for 4 days starting on PND 27 resulted in dose-dependent decreases in thyroid hormones, more pronounced for serum T₄ than for T₃. Total T₄ decreased to 43% of control at 1000 mg/kg bw per day, and total T₃ decreased to 89% and 75% of control at 300 and 1000 mg/kg bw per day, respectively, while TSH levels remained unchanged. The study authors speculated that triclosan-induced hypothyroxinemia was likely due to the observed upregulation of hepatic enzymes (i.e., induction of cytochrome P450 2B1/2 [CYP2B1/2] and pentoxyresorufin O-depentylase activity) and increased glucuronidation and sulfation of thyroid hormones. In contrast, the lack of CYP1A1 (ethoxyresorufin O-deethylase) induction indicated that the minor dioxin contaminants found in the triclosan sample used in this study (2,8-dichlorodibenzo-p-dioxin [2,8-DCDD] and 2,4,8-trichlorodibenzo-p-dioxin [2,4,8-TriCDD]) did not induce aryl hydrocarbon receptor-mediated effects on phase I and phase II hepatic enzymes. The NOEL was 30 mg/kg bw per day, and the BMDL (calculated by the study authors) for a 20% reduction in T₄ was 65.6 mg/kg bw per day.

Three additional studies investigated the effects of triclosan on thyroid hormone levels in pubertal and maternal animals, as well as offspring.

In a published study by Stoker et al. (2010), the effects of triclosan on thyroid hormones were investigated in a 21-day female pubertal assay and an immature rat uterotrophic assay (3-day exposure). Wistar rats were dosed orally by gavage after weaning with triclosan doses up to 300 mg/kg bw per day (PNDs 22-42 in the pubertal assay; for 3 days in the uterotrophic assay, either alone or co-treated with ethinylestradiol at 3 mg/kg bw per day). A dose-dependent decrease in thyroid hormone levels was observed at doses of 37.5-150 mg/kg bw per day following the 21-day exposure, and free serum T₄ was decreased at 75 and 150 mg/kg bw per day. There was no significant difference in the mean serum TSH concentration following a 21-day exposure. The NOEL for the decrease in total serum T₄ level was 9.4 mg/kg bw per day; the lowest-observed-effect level (LOEL) was 18.75 mg/kg bw per day in this study (no BMDL was calculated). In the pubertal exposure study, the highest dose of triclosan (150 mg/kg bw per day) resulted in a significant earlier age of onset of vaginal opening and increased uterine weight, which, according to the authors, was indicative of an estrogenic effect. There was also a non-significant decrease in age of first estrus at the highest dose. In the uterotrophic assay measuring the estrogenicity of the compound, triclosan enhanced the uterine response to ethinylestradiol, but did not alter uterine weight or histopathology when tested alone at doses as high as 300 mg/kg bw per day.

In a published study by Paul et al. (2010b), pregnant Long-Evans rats were exposed to triclosan at a dose of 0, 30, 100 or 300 mg/kg bw per day by oral gavage from gestational day 6 through PND 22. Perinatal maternal exposure to triclosan resulted in hypothyroxinemia in dams and young neonates and a 31% and 27% decrease in serum T₄ levels in dams (PND 22) and pups (PND 4) at 300 mg/kg bw per day, respectively. No changes in serum T₄ levels were reported in pups on PND 14 or PND 21 at any dose level. TSH levels were not reported by the study authors. The NOEL was 100 mg/kg bw per day for both dams and pups. The BMDLs calculated by the study authors for a 20% reduction in T₄ were 104 mg/kg bw per day and 58 mg/kg bw per day for dams and pups, respectively.

In a subsequent study by Paul et al. (2012), pregnant Long-Evans rats were exposed to triclosan at a dose of 0, 10, 30, 100 or 300 mg/kg bw per day by oral gavage from gestational day 6 through PND 21. At 300 mg/kg bw per day serum T₄ decreased approximately 30% in GD20 dams and fetuses, PND4 pups and PND22 dams. The NOEL for a decrease in serum T₄ was 100 mg/kg bw per day for GD20 dams and 30 mg/kg bw per day for PND22 dams. For offspring, serum T₄ was decreased by 28% in GD20 fetuses and by 26% in PND4 neonates at 300 mg/kg bw per day. The computed BMDLs for a 20% reduction in serum T₄ were 33 and 61.8 mg/kg bw per day for GD20 fetuses and PND4 neonates, respectively. There was no effect on serum T₄ for PND14 or PND21 neonates from any treatment group. There was no effect on T₃ or TSH in samples tested. Triclosan concentrations in fetal and neonatal serum as well as liver were observed to decrease with animal age from PND 4 to PND 21, suggesting that the lack of effect on T₄ at PND 14 and PND 21 is due to lower exposures at these ages. According to the authors, the obtained data demonstrate that fetal or neonatal rats do not experience more exposure or greater effects with triclosan compared to the perinatally exposed dam. Further, the authors conclude that in the rat, triclosan is a low-potency and low-efficacy thyroid hormone disruptor.

In a published report by Axelstad et al. (2013), two studies investigating the effects of triclosan on T₄ levels in rats were presented. In a first short-term (10-day) study, Wistar rat dams (10 per group) were exposed via oral gavage to triclosan (99%) in corn oil at a dose of 0, 75, 150, or 300 mg/kg-bw per day on GD 7-16. Significantly decreased T₄ levels were observed in dams on GD 15 and PND 16, but no significant effects on T₄ levels were observed in offspring at the end of lactation. Similar to a previous study by Paul et al (2010), the study authors suggested that the lack of effect on T₄ may have been caused by lack of triclosan entering the maternal milk. T₄ levels were decreased by 59%, 72% and 72% in pregnant dams (GD15) and by 38%, 55% and 58% during lactation (PND16), respectively. Decreased body weight gain during pregnancy was observed from GD 7-21 at 300 mg/kg bw per day. Unaffected measures included gestation length, gender distribution, post-implantation loss and litter size, neonatal deaths, and offspring body weights. No effects were observed on male or female anogenital distance or nipple retention. Absolute and relative thyroid gland weights were unaffected by triclosan exposure in both dams and offspring and no histopathological effects were observed in offspring thyroids at the highest dose tested. The LOEL was 75 mg/kg bw per day for dams. The results of this study showed that exposure to triclosan at all doses tested significantly lowered T₄ serum levels in dams but did not significantly affect T₄ levels in the offspring at the end of the lactation period.

In a second study, male and female pups, but not dams, were dosed directly via gavage daily between PND 3-16 to 50 and 150 mg/kg bw per day of triclosan in corn oil (Axelstad et al. 2013). It should be noted that all of the control pups were from the same litter and T₄ levels trended higher than those in the first study control group. A significant, dose-responsive and comparable decrease in T₄ levels in both male and female offspring was observed at PND 16 at 50 (16%) and 150 mg/kg bw per day (39%). There were no signs of general toxicity or significant effects on pup body weights or weight gain. The absence of an effect in offspring exposed to triclosan indirectly via nursing compared with the presence of an effect of triclosan via direct oral exposure in the Axelstad studies supports the concept that triclosan exposure through lactation is inadequate to disrupt the thyroid system in offspring (Witorsh, 2014).

The proposed adverse outcome pathway for the effects of triclosan on the thyroid hormone system includes the activation of the pregnane X receptor (PXR) and/or the constitutive androstane receptor (CAR) in rat liver by triclosan as an initiating event, leading to the effect on the circulating free T₄. The activation of these receptors was shown to result in upregulation of hepatic phase I and phase II enzymes and hepatic transporters, leading to an increased catabolism of thyroid hormones in rats (US EPA 2011a). To compensate for the movement of free T₄ into the liver, a compensatory mechanism is activated, and T₄ moves from the protein-bound state into the free pool. Due to the constant removal of T₄ from the free fraction into the liver, free T₄ concentrations remain decreased, and T₄ storage in the serum (i.e., protein-bound T₄) decreases, as manifested by a decrease in total T₄, with a subsequent potential impact on neurological development (Figure 3-1). At this time, evidence supporting an alternative mode of action for triclosan-induced decrease in the T₄ level, i.e., disruption of thyroid hormone synthesis as a result of f triclosan-induced thyroperoxidase (TPO) inhibition remains elusive (Paul et al. 2013a; Paul et al. 2014).

Figure 3-1: Proposed adverse outcome pathway for the effects of triclosan on the thyroid hormone system [TR = thyroid receptor] (See long description below) — **Figure 3-1. Proposed adverse outcome pathway for the effects of triclosan on the thyroid hormone system [TR = thyroid receptor]**

Recently published study investigating the effect of triclosan on human, rat and mouse PXR and CAR activity in vitro showed that triclosan acts as an agonist for human PXR but not rat or mouse PXR. The authors concluded that failure to measure activation of rodent PXR in vitro may accurately reflect the in vivo biological response, or may be inherent to the model used or to the concentration range tested, i.e., activation of rodent PXR may require a higher concentration of triclosan than activation of human PXR. The study showed that triclosan can act as an inverse agonist for both human CAR1 and rodent CAR and as a weak agonist for human CAR3 (Paul et al. 2013b). The opposite effects of triclosan on human PXR and human CAR1 receptors are not unexpected given that similar opposite effects of xenobiotics on these orphan nuclear receptors have been previously reported (Moore et al. 2000).

Interestingly, when the study authors compared the potential human oral exposure dose estimated by them (0.13 mg/kg bw per day) to the approximate concentration required to activate human PXR in vivo (15 mg/kg bw per day), they concluded that it would be insufficient to activate human PXR and human CAR (Paul et al. 2013b).

There are also uncertainties as to whether the magnitude of the observed thyroid hormone alteration is sufficient to affect brain development in rats. In the existing animal database for triclosan, no neurodevelopmental effects were reported following triclosan exposure. However, these in vivo screens and tests were originally designed to evaluate effects of the test material on reproduction and development, and not alterations in cognitive or behavioural function. Further, a developmental neurotoxicity study with triclosan is not available. Thus, there is uncertainty associated with whether triclosan-induced alterations in T₄ levels may have an effect on brain development or cognition in rats.

In general, triclosan-induced hypothyroxinemia would be expected to manifest itself in several systemic effects. One of the early indications of a reduction of T₄ in the rat is an increase in serum cholesterol. In the rodent database with triclosan, animals were shown to demonstrate decreases in cholesterol level. Hypothyroxinemia would also have an effect on the reproduction system. In human and rodent males, thyroid hormones regulate testis development through promotion of Sertoli cell differentiation. The effect is proposed to occur through activation of thyroid receptor alpha 1 (TRα1) in both species. In general, hypothyroxinemia-induced alterations in the reproductive system, such as decreased sperm count and decreased libido, are observed in adult male laboratory animals and humans (Bourget et al. 1987; Jannini et al. 1995). Prepubertal hypothyroxinemia is associated with precocious sexual development (enlargement of the testes without virilization) and absence of libido and ejaculate in rats (Jannini et al. 1995; Longcope 2000). In adult female rats, hypothyroxinemia is generally associated with altered menstrual and estrous cycles (Fisher and Brown 2000; Krassas 2000). Fetal hypothyroxinemia in female rats alters reproductive tract development, but a similar effect is not seen in human females. Hypothyroxinemia in the prepubertal period is associated with delayed sexual maturity in female rats and humans. However, in the rodent database with triclosan, alterations in the reproductive system either were not noted or were only observed at high doses of triclosan (e.g., chronic toxicity study with hamsters, the Stoker et al. 2010 study with rats). Thus, the paucity of clear indicators of hypothyroidism and associated clinical or histopathological indices in the rat with triclosan exposure suggests that decreases in T₄ may not be sufficient to cause overt hypothyroxinemia in the animal model.

Extrapolation of thyroid hormone data obtained in rats to human risk should be tempered by toxicodynamic and toxicokinetic differences in thyroid hormone homeostasis between humans and rats. In general, humans are considered less sensitive than rats to chemical-induced perturbation in thyroid hormone homeostasis due to the presence of high-affinity binding proteins (thyroxine-binding globulin) in human serum, which results in a longer serum T₄ half-life in humans (5-9 days in humans compared with 0.5-1 day in rats) (Glinoer 1997; Choksi et al. 2003). In rats, most T₄ in serum is bound to transthyretin, which has a lower binding affinity for T₄, resulting in a higher rate of T₄ clearance in adult rats compared with humans (Savu et al. 1987; Rouaze-Romet et al. 1992; US EPA 2011a). The increased clearance of thyroid hormones results in a higher rate of production of T₄ per unit of body weight in rats to maintain normal concentrations of T₄ (US EPA 2011a). These differences have been linked to increased susceptibility of rats to thyroid follicular tumours compared with humans (US EPA 2011a). Thus, it is likely that humans will be less responsive to any triclosan-induced changes in serum T₄ levels. As well, less than 1% of T₄ in humans is freely circulating and therefore available for destruction by liver enzymes, resulting in humans having a greater resistance than the rat model to thyroid toxicity, which occur secondary to liver enzyme activation.

Though triclosan can activate both human PXR and CAR3 in vitro, there is no evidence available supporting the up-regulation of Phase I and Phase II enzymes or triclosan-induced hypothyroxinemia following human exposure to triclosan (Paul et al. 2013b). The available literature reports no significant effect of triclosan on thyroid hormone homeostasis in humans.

In a published short-term (14-day) study by Allmyr et al. (2009), the effect of triclosan on thyroid hormone status was measured in 12 adult humans following exposure to triclosan-containing toothpaste. The plasma triclosan concentrations increased from 0.009-0.81 to 26-296 ng/g upon exposure. The highest serum concentration was determined to be equivalent to a triclosan dose of 0.1 mg/kg bw per day. Despite this, there were no significant changes in plasma levels of either 4β-hydroxycholesterol (indicative of CYP3A4 induction) or thyroid hormones during the exposure (Allmyr et al. 2009), demonstrating that triclosan-induced alterations in T₄ levels are unlikely to occur in healthy adult humans.

More recently, the effect of triclosan on thyroid hormone status was measured in 132 human subjects (predominantly male) with coronary heart disease (~ 61 years of age) in Brisbane, Australia, 64 of whom were exposed to triclosan-containing (0.3%) toothpaste and 68 of whom were exposed to placebo toothpaste for over four years (Cullinan et al. 2012). Serum measures of TSH, free T₄, free T₃, antithyroglobulin antibody and antithyroid peroxidase antibody were made at year 1 and year 5 of the study. Serum concentrations of triclosan were not directly measured in this study but were based on results of a previous study (Allmyr et al. 2008). No significant changes in thyroid function as indicated by changes plasma levels of thyroid hormones or antibodies were observed, except for a significantly higher level in free T₄ in the triclosan group compared to the placebo group in year 5. The study authors indicate that this result was due to a reduction in free T₄ level in the placebo group rather than a treatment related increase in T₄ in the triclosan group. Authors also evaluated haematological and clinical chemistry parameters and indicated no evidence of liver function changes, suggesting that the liver would not be a target organ in humans (Cullinan, personal communication 2014; unreferenced). Overall, triclosan exposures in this study appeared to have no adverse effect on thyroid measures. As well, the authors have reported that there were no adverse effects of triclosan exposure on human hematology, clinical chemistry, or measurements of liver function (Cullinan, personal communication 2014; unreferenced) at these exposure levels.

In another recent epidemiological study by Koeppe et al. (2013), results from 1831 subjects (less than or equal to 12 years of age) were studied for an association between urinary biomarkers of triclosan and serum thyroid measures from 2007-2008 National Health and Nutrition Examination Surveys (NHANES) data conducted in the United States. Study participants were stratified by age (i.e., adolescents: ages 12-19; adults: ages 20+) for regression modelling. Single samples of urine and serum were obtained from each individual. Further analyses were performed with sex as a variable. Urinary biomarker concentrations of triclosan were significantly elevated in females compared with males and age was also positively associated with triclosan concentrations. The only positive association between triclosan and any thyroid measure was an interquartile range (IQR) increase in urinary triclosan associated with a 3.8% increase in total serum triiodothyronine (T₃) concentrations in the smaller sized age group of adolescents. No associations between triclosan and T₄, free T₃ or TSH levels were observed and this age group showed consistently lower urinary concentrations of triclosan compared to adults. The authors suggested that while differences in distribution kinetics or metabolism between adolescents and adults might account for small changes in total T₃, it is likely that this result might simply be the result of residual confounding or chance.

In 2011, both the EU SCCS and the US FIFRA Scientific Advisory Panel considered the effects of triclosan on thyroid hormone homeostasis in rats and their relevance to humans. In light of the evidence demonstrating that the rat as being more sensitive to chemically induced alterations in thyroid hormone levels, the SCCS regarded a decrease in rat T₄ levels following exposure to triclosan as a biochemical marker that is not linked to an adverse effect (SCCS 2011). In consideration of the fact that the observed triclosan toxicity does not fit the typical pattern expected from perturbations of thyroid homeostasis, the FIFRA Scientific Advisory Panel recommended further revisions and refinements to the proposed adverse outcome pathway for triclosan before it could be used predictively. Although subtle perturbations of the T₄ level may have little or no effect due to the operation of homeostatic processes, the FIFRA Scientific Advisory Panel noted that additional data are needed "to determine the magnitude of perturbation of T₄ alone or in combination with other thyroid hormones that would lead to adverse neurodevelopmental effects" (US EPA 2011a).

In summary, with respect to the observed effects of triclosan on thyroid hormone system in the rat and their relevance to humans, the following can be observed:

a decrease in T₄ levels in rats is caused by disruption in the target organ (liver), based on rodent-specific metabolism of triclosan,
a decrease in T₄ levels in rats is likely to occur via up-regulation of hepatic catabolism and elimination of T₄ following exposure to triclosan,
there are no indications of adverse effects on thyroid function in the animal database,
the available human data show no changes in thyroid hormone levels or liver function after chronic, low dose exposure from toothpaste use,
humans have a much greater capacity to adapt to deviations in T₄ levels.

Based on the above, the overall database does not currently support effects of triclosan on thyroid function as a critical effect for risk characterization in humans. This is supported by a recent critical review of the endocrine activity of triclosan and its relevance to human exposure by Witorsch (2014). According to this review, there is little evidence that triclosan exposure, specifically through personal care product use, presents a risk of adverse health effects in humans via an endocrine mode of action.

3.1.11 Immunotoxicity

An analysis of the available information from subchronic and chronic mouse, rat, dog, and hamster studies focused on hematology, serum chemistry profiles, routine histopathology, and weight changes in specific organs in evaluating the immunotoxic potential of triclosan. No statistically significant, permanent, dose- or treatment-related findings were observed. More specifically, there were no indications of changes in white blood cell count, serum protein in combination with abnormal albumin:globulin ratios, gross findings during histological evaluations of lymphoid organs (spleen, lymph nodes, thymus, or bone marrow) or organ weights in subchronic mouse and dog as well as chronic rat and hamster studies (US EPA 2008b).

An in vitro study in rats monitored the level of degranulation in a mast cell model of rat basophilic leukemia cells in order to investigate the potential anti-inflammatory effect of triclosan. In response to various stimuli, mast cells degranulate, releasing allergic mediators such as histamine. The authors found that triclosan strongly dampened the release of granules from activated rat mast cells starting at 2 µM and above in a dose-responsive manner and further postulated that triclosan could be used for topical treatment for allergic skin disease (Palmer et al 2012). The overall interpretation of this study is limited.

A study by Udoji et al. (2010) examined the ability of triclosan to suppress human natural killer cell function in vitro. Triclosan was able to inhibit natural killer cell lytic function by 87% within 24 hours. These negative effects persisted following a brief (1-hour) exposure, indicating that the impairment of function cannot be eliminated by removal of triclosan under in vitro conditions. Clayton et al. (2011) investigated the association of triclosan with markers of immune function using 2003-2006 NHANES data by comparing triclosan levels with serum cytomegalovirus antibody levels and diagnosis of allergies or hay fever in US adults and children 6 years of age and older. Triclosan showed a positive association with hay fever diagnosis in the less than 18 year age group, although triclosan levels were not associated with cytomegalovirus antibody levels.

Savage et al. (2012) compared urinary levels of triclosan with IgE levels in 860 children (6-18 years of age) from the 2005-2006 NHANES data. A statistically significant increase in odds of aeroallergen sensitization with level of triclosan was observed in male subjects only, however the interaction between triclosan level and sex was not statistically significant. Also, a statistically significant increase in odds of aeroallergen and food sensitization with level of triclosan was observed when analyzed with both sexes combined. It should be noted that the allergen sensitization as an outcome was limited by lack of clinical correlation of allergic disease.

In summary, as with many epidemiological studies, it is difficult to determine a direct causal or even a reverse causal relationship between an environmental exposure and an adverse health outcome and these studies inherently have multiple limitations, such as use of general public questionnaires in lieu of medically diagnosed outcomes, cross-sectional versus prospective analysis, etc. The potential of triclosan to affect the immune system may warrant further investigation, but based on the lack of significant immune response in subchronic and chronic animal studies, triclosan-induced immunotoxicity does not appear to be demonstrated in multiple mammalian species.

3.2 Toxicological Endpoints for the Human Health Risk Assessment

3.2.1 Completeness of the Database

There is high confidence in the health effects database. The database for triclosan consists of the full array of toxicity studies currently required for hazard assessment purposes and is therefore adequate to define the majority of the toxic effects that may result from exposure to triclosan.

In examination of the database as a whole, the principal toxicity in rodents and dogs following ingestion of triclosan is mainly hepatic in nature, as demonstrated by hepatocellular necrosis, vacuolization, inflammation and other morphological changes in the liver, with the mouse being the most sensitive species. Triclosan produced hepatic effects and hepatic tumours in mice, but only limited hepatic effects and no tumours in rats. There is evidence that liver effects observed in mice were typical of a PPAR agonist.

A FIFRA Scientific Advisory Panel convened in 2003 reviewed the issue of PPARα agonist-mediated hepatocarcinogenesis in rodents and its relevance to human health risk assessment (SAP 2004). Overall, the majority of the Panel felt that there was adequate evidence in support of the proposed MOA for PPARα agonist-induced rodent hepatocarcinogenesis and that there are relevant data indicating that humans are less sensitive than rodents to the hepatic effects of PPARα agonists, although the opinions of the experts ranged from full agreement to complete disagreement. The basis for the disagreement was the lack of human data and the evidence that would be necessary to fully support the proposed MOA and its relevance to humans.

More recently, two different transgenic PPARα-humanized mouse models have been generated, demonstrating that while peroxisome proliferators can activate human PPARα expression, the mitogenic and hepatocarcinogenic effects do not occur (Cheung et al. 2004; Morimura et al. 2006). It was suggested that the difference in species response may be due to species-specific regulation of a micro-ribonucleic acid (RNA) (Shah et al. 2007; Peters 2008).

Although it is generally accepted that hepatocarcinogenesis in rodents by a PPAR agonist is irrelevant to humans, the same cannot be concluded for activation of PPARα, which alters the expression of genes involved in lipid metabolism that induce hypolipidemia (SAP 2004). Further, it cannot be excluded that non-cancer liver effects observed in rodent studies may also be a result of other modes of triclosan toxicity, such as CAR and PXR activation.

Toxicity in hamsters and baboons was different from that observed in rodents and dogs. Hamsters showed no increased liver toxicity and no tumours following chronic exposure (US EPA 2008b), which is consistent with the apparent differences in triclosan toxicokinetics metabolite profile in this species. Chronic toxicity was characterized by urinary and stomach lesions, which is consistent with the rapid conjugation and urinary excretion of triclosan. Chronic oral administration of triclosan via capsule to baboons did not lead to systemic toxicity, with the exception of clinical signs of vomiting and diarrhea occurring 4-6 hours after dosing, consistent with stomach irritation (US EPA 2008b). Similar to hamsters, liver toxicity was absent. Limited subchronic studies in rabbits also showed no clinical signs of toxicity from triclosan exposure (SCCP 2009).

Minor changes in hematology were considered adaptive, and alterations in biochemical parameters observed following short-term (mice), subchronic and chronic oral exposures to triclosan in rats and mice were considered secondary to liver toxicity in these species.

The data from the reproductive study in rats provide evidence of reduced viability of the offspring in the early postnatal days and a reduced weaning index in both generations. In a developmental toxicity study in mice, an irregular ossification was reported in fetuses (US EPA 2008b). These effects in rodents were observed at doses that also caused maternal toxicity. Increased liver weights in adult mice and increased incidence of liver discoloration in adult rats were observed in these studies; however, no histopathological assessment was undertaken (US EPA 2008b). The data from studies examining triclosan effects on male reproduction parameters in rats and hamsters provide conflicting evidence with regards to the potential testicular toxicity following exposure to triclosan. No association between exposure to triclosan and infertility was found in rats.

Triclosan exposure results in a modest decrease in serum T₄, but not T₃ or TSH levels in rat. However, there is uncertainty as to whether the observed magnitude of triclosan-induced maternal or early neonatal hypothyroxinemia is sufficient to affect brain development in rats.

In summary, with respect to the observed effects of triclosan on thyroid hormone system in the rat and their relevance to humans, the following can be observed:

a decrease in T₄ levels in rats is caused by disruption in the target organ (liver), based on rodent-specific metabolism of triclosan;
a decrease in T₄ levels in rats is likely to occur via up-regulation of hepatic catabolism and elimination of T₄ following exposure to triclosan;
there are no indications of adverse effects on thyroid function in the animal database;
the available human data show no changes in thyroid hormone levels or liver function after chronic, low dose exposure from toothpaste use; and
humans have a much greater capacity to adapt to deviations in T₄ levels.

Consequently, the overall database does not currently support effects of triclosan on thyroid function as a critical effect for risk characterization in humans.

Even though the level of concern for developmental neurotoxicity is low, an additional 3-fold uncertainty factor for database deficiency is being applied by Health Canada to all exposure scenarios to account for the lack of a confirmatory neurodevelopmental study in the rat.

3.2.2 PCPA Hazard Characterization

For assessing risks from exposure to chemicals in products used in or around homes or schools, the PCPA requires the application of an additional 10-fold factor to threshold effects to take into account completeness of the data with respect to the exposure of and toxicity to infants and children and potential prenatal and postnatal toxicity. A different factor may be determined to be appropriate based on reliable scientific data.

With respect to the completeness of the toxicity database as it pertains to the toxicity to infants and children, the database for triclosan contains the full complement of required studies, including developmental toxicity studies in rats, mice and rabbits and a two-generation reproductive toxicity study in rats. The lack of a developmental neurotoxicity study was accounted for through the use of an uncertainty factor for database deficiency.

With respect to identified concerns relevant to the assessment of risk to infants and children, in the developmental toxicity study in mice, a decrease in fetal weight was observed at a dose that also caused maternal toxicity. No treatment-related developmental effects were observed in developmental toxicity studies in rats and rabbits (US EPA 2008b). No evidence of increased susceptibility was observed in offspring in the available two-generation reproductive toxicity study conducted with rats. Effects in offspring, including reduced pup weight and viability in both generations, were observed following in utero and/or lactational exposure at a dose that was also associated with maternal toxicity (NOAEL of 50 mg/kg bw per day, LOAEL of 150 mg/kg bw per day; US EPA 2008b).

Reduced pup viability is considered a serious endpoint and, if selected for risk assessment purposes, would be subject to the application of the PCPA factor. As concern for this endpoint is tempered by the occurrence of maternal toxicity at the same dose level, the PCPA factor would be reduced from 10-fold to 3-fold for scenarios involving in utero and lactational exposure; however, when a point of departure less than or equal to the NOAEL of 50 mg/kg bw per day is utilized for risk assessment, the concerns identified under the PCPA 3-fold factor are considered to be subsumed by the 3-fold uncertainty factor for database deficiency, to temper compounding conservatism. Accordingly, the PCPA factor was reduced to 1-fold, since uncertainties with respect to the completeness of the data were accounted for through application of the database deficiency factor, and there was a low level of concern for prenatal and postnatal toxicity, given the endpoints and uncertainty factors selected for risk assessment.

It should be noted that the submission of a developmental neurotoxicity study could result in the potential removal of the uncertainty factor for database deficiency, pending the results of the study. However, reference doses would need to be reconsidered in totality to determine whether they remain protective of all vulnerable populations.

3.2.3 Acceptable Daily Intake (All Populations)

A number of studies were considered in the selection of the acceptable daily intake (ADI), an estimate of a daily intake of a substance over a lifetime that is considered to be without appreciable health risk, for the general population. Subchronic oral studies in the dog were not considered suitable for endpoint selection due to a number of factors, including study deficiencies, limited reporting, the age of the studies and the inconsistent results obtained (i.e., capsule studies demonstrated a LOAEL of 25 mg/kg bw per day, whereas a dietary study demonstrated no effects at this same level; US EPA 2008b). The results of the 1-year baboon study (NOAEL of 30 mg/kg bw per day, LOAEL of 100 mg/kg bw per day) were similarly disregarded, as the effects observed (i.e., diarrhoea and vomiting) following administration by capsule were thought to reflect the irritant properties of triclosan rather than systemic toxicity (US EPA 2008b).

In the remaining species tested, the mouse exhibited a NOAEL of 25 mg/kg bw per day (LOAEL of 75 mg/kg bw per day) in the 90-day and developmental toxicity studies for non-cancer effects (liver effects and decreased fetal body weight), compared with NOAELs of approximately 50 mg/kg bw per day in the rat (reduced pup weights and reduced pup viability in a reproductive toxicity study and liver effects in a 2-year oral toxicity study) and 75 mg/kg bw per day in the hamster (kidney effects in a 90-week study) (US EPA 2008b). Liver effects observed at the NOAEL in the mouse studies (e.g., increased liver weights, hypertrophy) were typical of a PPAR agonist. However, it cannot be excluded that the observed liver effects may also be the result of other triclosan modes of toxicity, such as PXR and CAR activation. Additional effects on hematology (mild decreases in erythrocyte parameters in the 90-day study), clinical chemistry parameters (decreased cholesterol) and liver pathology (vacuolization) were observed at the NOAEL that progressed to adversity at higher dose levels. It is well recognized that humans are generally less sensitive to PPARα agonist-induced hepatocarcinogenesis, primarily due to a reduced quantity of functional receptors in the human liver (compared with the mouse). That said, humans are at least as sensitive to activation of PPARα, which alters the expression of genes involved in lipid metabolism that induce hypolipidemia (SAP 2004).

Considering the current available information on the adverse effects of triclosan, a database NOAEL of 25 mg/kg bw per day was identified from a 90 day oral toxicity study in mice and was conservatively selected to be protective of a number of effects observed in multiple species with LOAELs ranging from 50 to 75 mg/kg bw per day. This NOAEL was considered protective for potential liver effects, if any, that could occur in humans as well as effects in other organs and systems. Standard uncertainty factors of 10-fold for interspecies extrapolation and 10-fold for intraspecies variability have been applied. An additional uncertainty factor of 3-fold has been applied to account for database deficiency (i.e., lack of a developmental neurotoxicity study). For the reasons outlined in Section 3.2.2, the PCPA factor was reduced to 1-fold for risk assessment purposes. This results in a composite assessment factor (CAF) (or target margin of exposure [MOE]) of 300.

The ADI for all populations is calculated according to the following formula:

AD1 = NOAEL/CAF = 25 mg/kg bw per day/ 300 = 0.08 mg/kg bw per day

This ADI provides a margin of greater than 600 to the NOAEL for reduced pup viability (50 mg/kg bw per day) and is considered protective for pregnant women and their fetuses as well as nursing infants.

3.2.4 Toxicological endpoints for residential and occupational risk assessment

3.2.4.1 Incidental oral exposure (directly exposed children)

For short-term incidental oral exposure (object-to-mouth and hand-to-mouth scenarios) of all children, the database NOAEL of 25 mg/kg bw per day was considered the most appropriate endpoint (as per the ADI). Standard uncertainty factors of 10-fold for interspecies extrapolation and 10-fold for intraspecies variability have been applied. An additional uncertainty factor of 3-fold has been applied to account for database deficiency (i.e., lack of a developmental neurotoxicity study). For the reasons outlined in Section 3.2.2, the PCPA factor was reduced to 1-fold for risk assessment purposes. This results in a target MOE (or CAF) of 300.

3.2.4.2 Dermal Exposure

For dermal exposure of all durations for all populations, the NOAEL of 40 mg/kg bw per day from a 90-day dermal toxicity study in rats was considered the most appropriate endpoint. Treatment-related effects at the LOAEL of 80 mg/kg bw per day included minor hematological changes (males), reduced triglyceride (males) and cholesterol levels (males and females), occult blood in urine and a slight focal degeneration of cortical tubules (males) (US EPA 2008b). Standard uncertainty factors of 10-fold for interspecies extrapolation and 10-fold for intraspecies variability have been applied. An additional uncertainty factor of 3-fold has been applied to account for database deficiency (i.e., lack of a developmental neurotoxicity study). For the reasons outlined in Section 3.2.2, the PCPA factor was reduced to 1-fold for risk assessments pertaining to residential scenarios. This results in a target MOE (or CAF) of 300 for the general population.

3.2.4.3 Inhalation Exposure

For inhalation exposure assessments, the NOAEL of 3.21 mg/kg bw per day from a 21-day inhalation toxicity study in rats was considered the most appropriate endpoint for all populations. Effects at the LOAEL of 7.97 mg/kg bw per day included changes in body weight, hematology and clinical chemistry and a slight increase in respiratory irritation (US EPA 2008b). The selected NOAEL is considered protective of effects observed in other species. Standard uncertainty factors of 10-fold for interspecies extrapolation and 10-fold for intraspecies variability have been applied. An additional uncertainty factor of 3-fold has been applied to account for database deficiency (i.e., lack of a developmental neurotoxicity study). For the reasons outlined in Section 3.2.2, the PCPA factor was reduced to 1-fold for risk assessments pertaining to residential scenarios. This results in a target MOE of 300 for the general population. The target MOE (or CAF) for all inhalation scenarios and populations is therefore 300.

3.2.5 Aggregate Exposure Scenarios

Aggregate exposures of adults and children to triclosan in products used by consumers (e.g., treated clothing, cosmetics, toothpaste and toys) are expected in residential settings. Exposures are expected to occur via the oral and dermal routes; inhalation exposure to triclosan is expected to be a negligible contributor to the aggregate exposure due to its low volatility.

For assessing aggregate exposure of the general population, the assessment can be performed using the endpoints and assessment factors selected for the ADI for the general population. Both oral and dermal studies have shown minor, but consistent, effects on hematology parameters at the LOAEL, as well as effects on cholesterol. Consequently, the database NOAEL of 25 mg/kg bw per day, was considered the most appropriate endpoint for assessing aggregate risks for all populations (US EPA 2008b). Standard uncertainty factors of 10-fold for interspecies extrapolation and 10-fold for intraspecies variability have been applied. An additional uncertainty factor of 3-fold has been applied to account for database deficiency (i.e., lack of a developmental neurotoxicity study). For the reasons outlined in Section 3.2.2, the PCPA factor was reduced to 1-fold for risk assessments pertaining to residential scenarios. This results in a target MOE (or CAF) of 300 for the general population.

3.2.6 Cancer Risk Assessment

Hepatic adenomas and carcinomas were observed in both sexes of mice in an 18-month dietary study; however, there was no evidence of carcinogenicity in long-term dietary studies in rats or hamsters (US EPA 2008b). Based on the available data, triclosan was not considered genotoxic, suggesting that the mouse tumours occurred as a result of a non-genotoxic MOA. It was determined that the hepatic tumours in mice were the consequence of a species-specific response to the peroxisome-proliferating properties of triclosan. This specificity has been demonstrated both morphologically and biochemically. Notably, mouse livers have shown dose-dependent increases in the numbers of peroxisomes and sensitivity to biochemical indicators of peroxisome proliferation, such as peroxisomal fatty acid β-oxidation, 11- and 12-hydroxylation of lauric acid and levels of CYP4A proteins. In comparison, effects in rats and hamsters are less pronounced (i.e., no increases in numbers of peroxisomes and biochemical indicators either unaffected or affected at high doses only) (Klaunig et al. 2003). It is generally accepted in the scientific community that mouse liver tumours induced through the MOA of peroxisome proliferation are of little relevance to humans (Section 3.1.8). While PPAR can be activated in humans following exposure to known agonists with resulting hypolipidemia, there is little evidence to indicate that hepatocellular proliferation and clonal expansion of initiated hepatocytes (required for tumour development) occur in humans. Accordingly, no quantitative cancer risk assessment is warranted for triclosan.

Endpoints of toxicological concern selected for use in the human health risk assessment are summarized in Appendix A.

3.3 Human Health Exposure and Risk

The approach taken in the health portion of this assessment report is to examine various lines of technical information and develop conclusions based on a weight of evidence approach and applying precaution as required under CEPA. The assessment of general population exposure to triclosan is based on several Canadian biomonitoring studies including the Canadian Health Measures Survey (CHMS), the Plastics and Personal-Care Product Use in Pregnancy (or P4) Study, and the Maternal-Infant Research on Environmental Chemicals (or MIREC) Study. These data encompass exposures to triclosan from all potential sources and routes, and are considered the most accurate estimates of total exposure of the general population in Canada to triclosan. Additional exposure characterization was undertaken as appropriate.

3.3.1 General population exposure and risk assessment

The potential sources of exposure to triclosan for Canadians include products used by consumers which are treated with or containing triclosan (including, but not limited to, drugs, cosmetics, and natural health products), drinking water, breast milk and household dust. Triclosan has also been measured in biosolids/wastewater sludge in Canada (Lee et al. 2013; Lee and Peart 2002; Chu and Metcalfe 2007; CCME 2010a; Sabourin et al. 2012), and, in some cases, has been taken up by plants such as soybeans, carrots, lettuce and radishes (Wu et al. 2010a; Macherius et al. 2012; Pannu et al. 2012a). However the overall exposure to the general population via food is expected to be minimal. Domestic-class pest control products containing triclosan are not registered in Canada. The biomonitoring data for triclosan provide actual internal measures of exposure, not only because they include specific measurements of triclosan in urine, but also because they reflect the integrated exposure to triclosan from all sources and pathways including use of products used by consumers which contain triclosan.

In April 2013, Health Canada released the second cycle of biomonitoring data collected as part of the Canadian Health Measures Survey (CHMS), an ongoing nationally representative survey that collects important health and wellness data as well as biological samples from individuals across the country (Health Canada 2013). Total triclosan (conjugated and free forms) was measured in spot urine samples for approximately 2500 individuals aged 3 to 79 years at 18 sites across Canada from 2009 to 2011. According to Statistic Canada (2013a) and Health Canada (2013), triclosan was detected in urine in approximately 72% of the population indicating that the majority of the Canadian population was exposed to this chemical. The CHMS does not include individuals living on reserves or in other Aboriginal settlements in the provinces, residents of institutions, full-time members of the Canadian Forces, persons living in certain remote areas, and persons living in areas with a low population density (Health Canada 2013).

Another study initiated by Health Canada in 2008, the Plastics and Personal-Care Product Use in Pregnancy (referred to as the P4 Study), recruited 80 pregnant women from the Ottawa, Ontario, area from December 2009 to December 2010, in order to collect multiple maternal urine samples, detailed products used by consumers/food packaging diaries, infant urine and meconium samples, breast milk and infant formula. Total triclosan (conjugated and free forms) was detected in more than 80% of the maternal urine samples (Arbuckle et al. 2015b).

The Maternal-Infant Research on Environmental Chemicals (MIREC) Study also measured various substances in Canadian pregnant women. The MIREC Study recruited approximately 2000 women in their first trimester of pregnancy from 10 cities across Canada between 2008 and 2011 (Arbuckle et al. 2013). Health Canada in collaboration with the US National Institute of Child Health and Human Development and the National Children's Study analysed total urinary triclosan (conjugated and free forms) from stored urine samples from this study. Total triclosan was detected in over 99% of the maternal urine samples; however, a more sensitive method was employed than in the P4 Study (Arbuckle et al. 2015b). The median maternal urinary triclosan concentration in the P4 Study was 25.3 µg/L (based on 1247 urine samples from 80 women) and was 8.74 µg/L in the MIREC Study (based on one urine sample each from 1861 women).

A follow-up study was initiated, MIREC-Child Development Plus (MIREC-CD Plus), which measured urinary triclosan concentrations in children from the original MIREC Study on a subsample of 200 children aged 23 to 36 months (unpublished data, personal communication September 2014 from Environmental Health Science and Research Bureau, Health Canada to Existing Substances Risk Assessment Bureau, Health Canada; unreferenced).

General population (3 to 79 years of age) daily dose estimates derived from the CHMS, P4 and MIREC studies will be used to assess risks to the Canadian population (3 to 79 years of age) in section 3.3.2. For children less than 3 years of age, daily dose estimates are derived using a combination of biomonitoring data from the P4 Study, preliminary results from MIREC CD Plus, as well as deterministic estimates to account for potential exposures via breast milk, household dust and mouthing of triclosan-treated plastic products (refer to section 3.3.4 for details).

3.3.2 Estimation of the daily exposure dose based on the urinary triclosan concentration

Given that the health effect levels are expressed in mass of the substance (e.g., in milligrams) per kilogram body weight per day, it is necessary to convert the triclosan concentration in spot urine samples to estimates of daily exposure.

In order to interpret urinary concentrations for any chemical, it is important to adjust the urine concentrations to account for hydration status (Haddow et al. 1994; Miller et al. 2004). There are various methods available to adjust for urine dilution including normalization of urine concentration by urinary creatinine concentration, osmolality, specific gravity, and/or estimation of total urine output using urine flow-rate. Selecting a particular method depends on availability of relevant data, and the substance being measured. Considering wide variations in urine dilution and creatinine excretion due to wide fluctuations in fluid intake and differences in physiology in the general population, the preferred option would be to use the 24-hour urine samples. However, these data are not available in the CHMS, P4 or MIREC Studies. The CHMS, P4 and MIREC Studies measured specific gravity and/or creatinine thus urinary triclosan concentrations have been adjusted using creatinine and/or specific gravity adjustments. Creatinine is commonly used to correct urine spot samples in occupational and environmental monitoring studies; however, it varies greatly with age, time of day, season, as well as exercise and consumption of red meat (Barr et al. 2005, Pearson et al. 2009), and therefore, may be problematic for populations experiencing rapid physiological changes such as pregnant women (Abduljalil et al. 2012), newborns and infants (Matos et al. 1999, Quigley 2012). For this reason, it was not used to adjust urinary triclosan concentrations in the P4 and MIREC Studies. Although specific gravity is used less often in biomonitoring studies, it is considered less variable than creatinine (Pearson et al. 2009) and was considered to be slightly more correlated with the urinary excretion rate (often considered the “true” dilution status of the sample) in a recent study conducted by Koch et al. (2014). Given the potential issues related to creatinine adjustment for infants and pregnant women, and that specific gravity was considered slightly more correlated to true excretion, only the specific gravity adjusted concentrations and daily intake estimates will be presented in the assessment report. The unadjusted and creatinine adjusted methods are presented in the appendices.

The approach used to estimate daily intakes from the CHMS data included the use of individual adjusted body weight urinary concentrations (derived by Statistics Canada (2013a), see Appendix B) along with a range of typical urine volumes (L/day) reported in the literature (see Appendix C), as well as daily creatinine excretion derived from the Mage equations described in Huber et al. (2011) (Appendix D). A similar approach was used to estimate daily intakes from the unadjusted and specific gravity-adjusted urinary concentrations from the P4 and MIREC data; however, these values were derived by Health Canada (January 2014 personal communication from the Environmental Health Science and Research Bureau, Health Canada to Existing Substances Risk Assessment Bureau, Health Canada; unreferenced) (see Appendix B).

In addition to adjusting for urine dilution, urinary triclosan concentrations were adjusted for incomplete excretion of triclosan in urine while computing exposure estimates. Based on pharmacokinetic studies (Table 3-1) investigating the absorption, metabolism and excretion of triclosan in humans with several different routes of administration, including oral exposure to triclosan-containing products (e.g., toothpaste), oral ingestion of capsules, aqueous solutions and dental slurries (i.e., following brushing with triclosan-containing toothpaste) and percutaneous exposure (in vivo and in vitro), the SCCP (2009) concluded that ingested triclosan is almost completely absorbed, whereas oral cavity and percutaneous exposure to triclosan-containing products (e.g., toothpaste, soap, cream) results in limited absorption. The SCCP (2009) also concluded that following all routes of administration, absorbed triclosan is nearly totally converted to glucuronic and sulfuric acid conjugates (varied relative proportions), with only trace amounts of the parent compound detected in the plasma, and the predominant route of excretion was the urine, with the majority of the compound appearing as the glucuronide conjugate.

Table 3-1. Summary of triclosan excretion data in humans
Type of administered dose	% of dose excreted in urine	% of dose excreted in feces	References
Single or multiple oral doses, capsule	57-87	10-33	Stierlin 1972; Ciba-Geigy 1976b; Lucker et al. 1990
Single oral dose, aqueous	24-83	Not measured	Sandborgh-Englund et al. 2006
Dermal dose	2-14	0.5-2	Stierlin 1972; Caudal et al. 1974; Thompson et al. 1975; Queckenberg et al. 2010
Intravenous dose	65	21	Maibach 1969

Following single and multiple oral doses of triclosan, 57-87% of the administered dose was excreted in urine, with much smaller amounts appearing in the feces (10-33% of the administered dose), based on studies by Lucker et al. (1990), Stierlin (1972) and Ciba-Geigy (1976b). In a study using single doses of aqueous solutions containing triclosan, the major fraction was excreted within 24 hours of exposure, with between 24% and 83% (median 54%) of the oral dose excreted within the first 4 days after dosing (Sandborgh-Englund et al. 2006). For dermal dosing, the excretion profile was similar, with the predominant route of excretion in the urine (2-14%) based on studies by Stierlin (1972), Caudal et al. (1974) and Thompson et al. (1975), with much smaller amounts appearing in the feces (0.5-2% of the applied dose) (SCCP 2009). The SCCP (2009) also concluded that excretion data obtained from an intravenous study were consistent with those obtained from the oral studies, with the majority of the dose (approximately 65%) excreted in the urine, while approximately 21% was excreted in the feces (Maibach 1969).

To account for variability in urinary excretion of triclosan between individuals, a conservative median urinary excretion of 54%, as reported in the Sandborgh-Englund et al. (2006) oral study, was assumed for all individuals (i.e., 54% of triclosan is excreted in the urine). This value is considered appropriate given the high absorption of triclosan via the oral route and limited absorption via the dermal route combined with similar excretion noted via intravenous administration (65%). Consequently, all exposure estimates were adjusted by a factor of 0.54 to account for incomplete urinary excretion following exposure via multiple routes. For children, although there are limited pharmacokinetic data, the SCCP (2009) concluded that the rate of elimination is comparable to that of adults; therefore, the same correction factor was applied to the assessment for children under the age of 3 years.

Estimated daily doses of triclosan for the general population of Canada were derived using urinary concentrations (adjusted and unadjusted) per kg body weight and the median urinary excretion fraction of 0.54 (see Appendix D). The estimated daily doses derived using the specific gravity-adjusted urinary triclosan concentrations resulted in the highest estimated doses and are presented below in Table 3-2. The estimated daily doses derived using the unadjusted and creatinine-adjusted urinary triclosan concentrations are presented in Appendix D.

3.3.2.1 CHMS cycle 2

The second biomonitoring report published by Health Canada in 2013 contains summary statistics for the unadjusted and creatinine-adjusted urinary triclosan concentrations (Health Canada 2013). Statistics Canada (2013a) provided additional analyses of this data: specifically, urine concentrations were adjusted by specific gravity and all urinary concentrations were divided by each individual's body weight (µg/L/kg or µg/g/kg) for use in the estimation of daily doses. In order to perform these analyses, the CHMS Data Users Guide was used (Statistics Canada 2013b). Additional details on these analyses, including the methods used for specific gravity and creatinine adjustments, can be found in Appendices B-D.

The geometric mean and 95^th percentile unadjusted urinary triclosan concentrations for males and females aged 3-79 year olds are 16 µg/L and 710 µg/L, respectively (Health Canada 2013). When the data were adjusted using specific gravity, the geometric mean and 95th percentile urinary triclosan concentrations are 22 µg/L and 990 µg/L, respectively (Statistics Canada 2013a). The geometric mean and 95^th percentile creatinine adjusted urinary triclosan concentrations are 15 µg/g, and 620 µg/g, respectively (Statistics Canada 2013a). Based on the 95% confidence intervals, there is no apparent difference in triclosan urine concentrations (both unadjusted and specific gravity adjusted) between males and females; however, urinary triclosan concentrations were significantly lower in children 3 to 11 year olds compared to 12 to 59 year olds. Based on the 95% confidence intervals for the creatinine adjusted values, there appears to be no significant difference between males and females or between age groups. These unadjusted urinary concentrations are in similar range to recent levels reported in CHMS cycle 3 (Health Canada 2015) as well as those used in the preliminary assessment from NHANES 2007-2008 (geometric mean of 15.3 µg/L for 6 years of age and older) and to more recent data reported from 2009-2010 and 2011-2012 (geometric mean of 14.5 µg/L and 11.8 µg/L for 6 years of age and older, respectively) (CDC 2015). The Canadian urinary triclosan concentrations are somewhat higher than those reported in the Korean National Human Biomonitoring Survey (Kim et al. 2011) (geometric mean of 1.68 µg/L), as well as several other smaller studies from Belgium (Pirard et al. 2012; Den Hond et al. 2013), Denmark (Frederiksen et al. 2013a, 2013b), Greece (Asimakopoulos et al. 2014) and China (Li et al. 2013; Chen et al. 2012, 2013; Engel et al. 2014). Given the distribution of the dataset from the CHMS Cycle 2, the geometric means for different age groups were used for estimating the mean daily doses as described in Section 3.3.3.

3.3.2.2 Urinary triclosan concentrations from the P4 study

In the P4 Study, triclosan was detected in more than 80% of the maternal urine samples (multiple samples throughout pregnancy from each participant) (Arbuckle et al. 2015b). The unadjusted and specific gravity adjusted urinary triclosan concentrations for pregnant women are shown in Appendix E, Table E-4. The geometric means and 95^th percentile unadjusted and specific gravity adjusted urinary concentrations are 21.61 µg/L and 833.4 µg/L, and 22.9 µg/L and 774.9 µg/L, respectively (Arbuckle et al. 2015b). Temporal variability of urinary triclosan concentrations has been reported in a recent publication of the P4 Study (Weiss et al. 2015). In addition, this publication showed that the ability of a single spot urine sample collected at any time during or post-pregnancy to predict an individual's geometric mean urinary triclosan levels corresponding to low, medium or high exposure was 86.7%.The authors noted that since the data reflect a small subset of the Canadian population, the study results may not be generalizable to other populations. (Weiss et al. 2015). Information from this study on the presence of triclosan in infant urine, meconium, breast milk and infant formula will be examined in section 3.3.4.

3.3.2.3 Urinary triclosan concentrations from the MIREC study

Almost all of the women in the MIREC Study had detectable levels of triclosan in their urine (one sample from each participant) and the results are shown in Appendix E, Table D. The geometric mean and 95th percentile unadjusted and specific gravity adjusted urinary concentrations are 12.64 µg/L and 697.58 µg/L, and 14.36 µg/L and 571.10 µg/L, respectively (Arbuckle et al. 2015a).

The urinary triclosan concentrations from both the P4 and MIREC Studies are similar to those reported in CHMS including females of child-bearing age (13 to 49 years of age), and were similar or slightly lower than those identified from several other studies that measured triclosan in urine of pregnant women including the United States (Wolff et al. 2008, Biomonitoring California 2013, Philippat et al. 2013, Mortensen et al. 2014), Puerto Rico (Meeker et al. 2013), Denmark (Tefre de Ranzy-Martin et al. 2014) and Norway (Bertelsen et al. 2013).

3.3.2.4 Uncertainties associated with dose conversion

There are several uncertainties associated with using triclosan concentrations in spot urine samples to estimate human exposures to triclosan. Spot urine samples (Appendix E) were used as a surrogate for 24-hour urine samples. In order to estimate daily doses from these spot samples, a range of typical daily urine volumes specific to a given subpopulation were used (Appendix C). There is high variability with daily urine volumes both between and within individuals therefore the range of typical urine volumes identified from various sources was selected to account for this variability. The 95^th percentile urinary triclosan concentration from a spot urine sample will likely overestimate the 95^th percentile from a 24-hour urine sample (Summit Toxicology 2013); therefore, the 95th percentile urinary triclosan concentration adjusted for body weight and the high end of the mean urine volumes were used to calculate an upper-bounding estimate of exposure for the general population of Canada. Since it has been shown that there is a statistically significant inverse relationship between body-weight adjusted urine triclosan concentrations and urinary flow rate in all age groups (Summit Toxicology 2013), the full range of urine volumes was not used as this would result in an overestimate of actual upper bound daily intakes given the use of 95th percentile body weight adjusted concentrations.

Another uncertainty in the dose conversion of spot urine samples for all age groups is the assumption that absorption, distribution, metabolism and elimination parameters are the same for all Canadians and remain constant within individuals over time. There is uncertainty associated with the use of the median value of 54% to account for urinary excretion of triclosan for all individuals, as the values were highly variable (24-83%) and were based on oral dosing (Sandborgh-Englund et al. 2006). However, according to Krishnan et al. (2010), the data from the Sandborgh-Englund et al. (2006) study were considered to be fairly robust. In addition, the SCCP (2009) concluded that, although there are limited pharmacokinetic data for children and no direct comparisons with adults were possible given differences in doses and dosing formulations in various studies, elimination was determined to be essentially the same for children and adults based on an oral dosing study with toothpaste and dental slurry. Given the number of potential sources of exposure via the dermal route, there is uncertainty in correcting spot urine samples for incomplete excretion using an oral dosing study. However, given the high absorption of triclosan via the oral route and limited absorption via the dermal route combined with similar excretion noted via intravenous administration (65%), correction using a median of 54% via oral dosing is considered appropriate.

There is also some uncertainty with converting spot urine samples to a daily dose, as the routes of exposure and timing of exposure in relation to the timing of sampling are unknown. However, given the short half-life of triclosan in urine of 11 hours (Sandborgh-Englund et al. 2006), and the widespread daily use of triclosan-containing products, the spot urine samples for triclosan represent a range of short- and long-term measurements of exposure. Since the dose estimation likely represents a range of exposure durations, the high percentage (72%) of individuals with detectable levels of triclosan in urine (Statistics Canada 2013), and that triclosan is found in a number of products used by consumers that could be used more than once a day, it is reasonable to assume that the elimination of triclosan in urine of individuals in the CHMS data is at steady state.

3.3.3 Aggregate risk assessment for the general population (3-79 years of age)

The CHMS data provide information on the total exposure to triclosan of individuals 3-79 years of age. As such, exposure and risk for children less than 3 years of age were assessed separately (see section 3.3.4). The unadjusted and adjusted urinary concentrations from all Canadian biomonitoring studies used in this assessment are presented in Appendix E; however, only the concentrations, and estimated daily intakes, adjusted for specific gravity are shown in the text. The risk for the Canadian population (≥ 3 years of age) was characterized by comparing the estimated daily dose for each population subgroup with the relevant health effect endpoint identified by Health Canada (Appendices A, D).

The methods used to estimate daily doses from the spot urinary triclosan concentrations are described in Appendix D. The mean daily dose estimates were derived based on the geometric means of data from the CHMS Cycle 2 (2009-2011) (Health Canada 2013, Statistics Canada 2013a), the P4 (Arbuckle 2015b) and the MIREC Studies (Arbuckle 2015a), as summarized in Table 3-2.

Table 3-2. General population risk based daily dose estimates derived from geometric mean and 95th percentile specific gravity adjusted urinary concentrations and a range of typical urine volumes
Study^a	Group	Mean estimated daily dose^b (µg/kg bw per day)	MOE range^c (means)	95th percentile estimated daily dose^d (µg/kg bw per day)	MOE range^c (95^th percentile)
CHMS Cycle 2	Children 3-5 years of age	0.47-0.74	33 784-53 191	5.99-9.33	2680-4174
CHMS Cycle 2	Children 6-11 years of age	0.15-0.61	40 983-166 667	4.87-20.27	1233-5133
CHMS Cycle 2	Adolescents 12-19 years of age	0.31-0.99	25 252-80 645	9.80-31.11	803-2551
CHMS Cycle 2	Adults 20-59 years of age	0.41-1.39	17 986-60 976	16.67-56.39	443-1500
CHMS Cycle 2	Adults greater than or equal to 60 years of age	Not shown- greater than 40% (less than LOD)	NA	4.63-44.44	563-5400
CHMS Cycle 2	Females: 13-49 years of age	0.48-1.62	15 432-52 083	17.78- 60.15	416-1406
P4 and MIREC	Females: pregnant	0.31-1.60	15 625-80 645	12.59-57.10	438-1986

Table Notes

Abbreviations: NA, not applicable

^a CHMS Cycle 2 (2009-2011), P4 (2008-2011), MIREC (2008-2011).

^b Estimated daily dose using the geometric mean urinary concentrations per kg body weight (Appendix 4) and a range of mean urine volumes (Appendix 5).

^c MOE (Margin of Exposure) = NOAEL (µg/kg bw per day) / exposure dose (µg/kg bw per day), where the NOAEL of 25 000 µg/kg bw per day, with a target MOE of 300, was selected for all populations.

^d Estimated daily dose using the 95^th percentile urinary concentrations per kg body weight (Appendix 4) and a range of typical urine volumes (Appendix 5).

To account for uncertainties with respect to the dose estimation (e.g., high variability between individuals’ pharmacokinetic data for triclosan) and potentially higher exposure of some individuals due to high use of products used by consumers which contain triclosan or a single event such as swallowing toothpaste prior to sampling, exposure estimates were also determined based on the 95th percentile urine concentrations and a range of typical urine volumes (Table 3-2).

Based on an analysis on the relationship between spot urine concentrations, 24-hour composite average, and longer-term averages it was found that single spot urine samples of triclosan are reliable for measuring individual’s longer-term exposures (Summit Toxicology 2013). The 95th percentile spot urine samples were also found to overestimate the 95th percentile of the 24-hour composite urine samples (for substances with shorter half-lives); however, there is some uncertainty in the percentile estimates for the 24-hour composite samples due to the small number of data points (n=8) (Summit Toxicology 2013).

Based on the results of the aggregate risk assessment, it can be concluded that exposure of adults (including pregnant females) and children over the age of 3 years to triclosan residues is below the level of concern.

3.3.4 Aggregate risk assessment for children younger than 3 years of age

Although CHMS did not sample children younger than 3 years of age, triclosan has been measured in the urine of infants and children younger than 3 years of age, as reported in other Canadian studies. Results of the spot urine triclosan concentration are available from the Canadian P4 Study for infants under 1 month of age and 2-3 months of age (or 0 to 3 months old). Triclosan was measured in 61% of the urine samples from infants aged 0 to 3 months old (some infants were measured at both less than 1 month and at 2-3 months of age) (Arbuckle et al. 2015b). The MIREC-CD Plus Study also measured urinary triclosan concentrations in children aged 23 to 36 months. Triclosan was detected in 58% of the 200 urine samples (Personal communication Sept 2014 from Environmental Health Science and Research Bureau, Health Canada to Existing Substances Risk Assessment Bureau, Health Canada; unreferenced).

Other biomonitoring data for children in the under 3 years of age group were identified, including a study that collected urine from 42 premature infants in Boston, Massachusetts (Calafat et al. 2009), a study from Belgium that included urine data for children 0-6 years old (Pirard et al. 2012), and urine samples from 56 children 3-6 years of age collected in Guangzhou, China (Li et al. 2013). The results of these three studies are shown in Table 3-3.

Table 3-3. Unadjusted concentrations of total triclosan in urine of children less than 6 years of age
Location	Age	Number of Samples	Triclosan concentration (µg/L)	Limit of detection (µg/L)	Reference
Canada	Infants (0-3 months)^b	100^c	2.8^d (GM) (95P = 52.0)	3.0	Arbuckle et al. 2015b
Canada^a	23-36 months	200	3.64 (GM) (95P = 140.65)	3.0	Personal Communication (Health Canada 2014)
Boston, MA	Premature infants	42	greater than 2.3-16.7	2.3	Calafat et al. 2009
Belgium	0-6 years	21	1.70 (GM)	0.33	Pirard et al. 2012
China	3-6 years	56	3.77 (GM)	0.0009	Li et al. 2013

Table Notes

Abbreviations: GM, geometric mean; 95P, 95^th percentile.

^a Preliminary data from MIREC-CD Plus Study.

^b Formula-fed Infants (n=6) were noted to have higher triclosan concentrations compared to nursing or combination of the two (breast fed and formula) (n=47) (Arbuckle et al. 2015b).

^c For P4 Study, it refers to number of urine samples.

^d Authors used results that were less than the limits of detection in their calculations.

Among young children, infants 6-12 months of age are likely to have the highest exposure to triclosan, given that children in this age group display a number of additional behavioural activities that are not captured by the 6-11 years age category. These behaviours include nursing, "object-to-mouth" (e.g., mouthing plastic toy), "hand-to-mouth" (e.g., touching triclosan-impregnated products or crawling) and inhalation of contaminated dust (created as a result of children's activities on the floor/carpet). Younger age groups (i.e., birth to less than1 month, 1 to less than 3 months and 3 to less than 6 months) are considered to have lower exposures relative to body weight due to less frequent contact with treated objects (i.e., hand-to-mouth and object-to-mouth activities). Older age groups (i.e., 1 to less than 2 years, 2 to less than 3 years, 3 to less than 6 years of age) are expected to have lower exposures due to the cessation of nursing and a reduction in hand-to-mouth activities (US EPA 2011b).

Although NHANES did not sample children younger than 6 years of age, triclosan has been measured in the urine of infants and children younger than 6 years of age, as reported in the other studies. Preliminary results of the spot urine triclosan concentration are available from the Canadian P4 study for infants under 1 week of age and infants 2-3 months of age. Other biomonitoring data for children in the under 6 years of age group were identified, including a study that collected urine from 42 premature infants in Boston, Massachusetts (Calafat et al. 2009) and urine samples from 56 children 3-6 years of age collected in Guangzhou, China (Li et al. 2011). The results of these three studies are shown in Table 5.

Among young children, infants 6-12 months of age are likely to have the highest exposure to triclosan, given that children in this age group display a number of additional behavioural activities that may not be captured by the 3-5 year age category. These behaviours include nursing, "object-to-mouth" (e.g., mouthing plastic toy), "hand-to-mouth" (e.g., touching triclosan-impregnated products or crawling) and inhalation of contaminated dust (created as a result of children's activities on the floor/carpet). Younger age groups (i.e., birth to less than 1 month, 1 to less than 3 months and 3 to less than 6 months) are considered to have lower exposures relative to body weight due to less frequent contact with treated objects (i.e., hand-to-mouth and object-to-mouth activities). Older age groups (i.e., 1 to less than 2 years, 2 to less than 3 years, 3 to less than 6 years of age) are expected to have lower exposures than infants due to the cessation of nursing and a reduction in hand-to-mouth activities (US EPA 2011b).

Using the same method as was used for individuals 3 years of age and older to convert spot urine samples to dose, the estimated daily dose for infants ranged from 0.018 to 13.07 µg/kg bw per day based on the mean and 95^th percentile specific gravity adjusted urinary concentrations from the P4 Study (Arbuckle et al. 2015b), range of typical urine volumes (Appendix C), and a factor of 54% to account for urinary excretion. The estimated daily dose for children aged 23 to 36 months ranged from 0.22 to 10.67 µg/kg bw per day based on the mean and 95^th percentile specific gravity adjusted urinary concentrations from the MIREC-CD Plus Study (Personal communication Sept 2014 from Environmental Health Science and Research Bureau, Health Canada to Existing Substances Risk Assessment Bureau, Health Canada; unreferenced. Although there are limited pharmacokinetic data for children and no direct comparisons with adults, elimination of triclosan was determined to be essentially the same for children and adults based on an oral dosing study with toothpaste and dental slurry (SCCP 2009). Using the database NOAEL of 25 mg/kg bw per day and estimated daily doses, the resulting MOEs are greater than 2300 (target MOE of 300).

3.3.4.1 Meconium and amniotic fluid

The presence of triclosan in both meconium and amniotic fluid provides some evidence of transplacental exposure in utero (over a period of time during gestation). Meconium is the fecal material that is passed during the first few days of birth (Abudu 2011). It is considered to be a repository for substances that the fetus has been exposed to throughout pregnancy from approximately the 12th week of pregnancy (Ostrea et al. 2006). Amniotic fluid is the liquid that surrounds an unborn baby during pregnancy and is considered to be a potential matrix for measuring fetal exposure (NLM 2014, Cooke 2014).

One of the components of the P4 Study was to measure triclosan in meconium as a potential matrix for measuring in utero exposure. Triclosan was detected in approximately 81% of the meconium samples ranging from below the limit of detection (0.49 ng/g) to 77.0 ng/g with a geometric mean of 2.24 ng/g and a 95th percentile of 68.8 ng/g (Arbuckle et al. 2015b). According to the authors, triclosan concentrations in meconium were significantly correlated with both maternal (during pregnancy) and infant urinary concentrations shortly after birth. The authors also reported that triclosan concentrations in meconium from female infants were significantly higher than those measured in males (Arbuckle 2015b).

Philippat et al. (2013) assessed the relationship between maternal urine and amniotic fluid concentrations of nine environmental phenols, including triclosan, among pregnant women. Triclosan was measured in 69 samples of amniotic fluid but was only detected in 6% of the samples (limit of detection = 2.3 µg/L) with a median of less than the LOD and a 95th percentile of 19.4 µg/L (Philippat et al. 2013). The authors concluded that amniotic fluid may not be a suitable matrix for assessing fetal exposure to certain phenols given the infrequent detection and the lower concentrations measured in amniotic fluid compared to maternal urine.

Based on the presence of triclosan in meconium measured by Arbuckle et al. 2015b, there is evidence of potential foetal exposure to triclosan in utero; however, there is uncertainty in deriving daily exposure estimates and characterizing risk from this matrix for instance, the potential contamination from infant urine could not be ruled out.

3.3.4.2 Infant-specific exposure scenarios

3.3.4.2.1 Nursing

Triclosan has been measured in human breast milk in Canada, the United States, Australia, Europe and China (Arbuckle et al. 2015b; Adolfsson-Erici et al. 2002; Allmyr et al. 2006; Dayan 2007; Ye et al. 2008; Toms et al. 2011; Azzouz et al. 2011; Wang et al. 2011). A summary of the results of these studies is shown in Table 3-4.

Table 3-4. Concentration of total triclosan in human breast milk
Location	Number of Samples	Mean (µg/kg lipid)	Minimum (µg/kg lipid)	Maximum (µg/kg lipid)	LOD (µg/kg lipid)	Reference
Canada	52	2.50^a (geomean)	greater than LOD	2287.0	14.1^b (0.58 µg/L)	Arbuckle et al. 2015b
United States	62	Not specified	greater than LOD	2100	0.150	Dayan 2007
United States	4	Not specified	greater than LOD	353	24.3	Ye et al. 2008
Sweden	36	8.3-13.5 (median)	greater than LOQ	23.8	0.45	Allmyr et al. 2006
Sweden	5	Not specified	greater than 20	300	Not specified	Adolfsson-Erici et al. 2002
Spain and Morocco	3	Not specified	greater than LOD	6.3	0.015	Azzouz et al. 2011
China	10	Not specified	greater than MQL	309	3.5	Wang et al. 2011
Australia	151	32.5	greater than LOQ	475	0.39-0.46	Toms et al. 2011

Table Notes

Abbreviations: LOD, limit of detection; LOQ, limit of quantification; MQL, method quantification limit.

^a Authors used results that were less than the limits of detection in their calculations.

^b The LOD was calculated using the following assumptions: density of human breast milk of 1.03 g/mL and fat content in human breast milk of 4% (US EPA 2011c). To convert µg/kg lipid concentration to µg/kg whole milk concentration simply multiply the µg/kg lipid x 4% lipid content in milk (or used measured lipid content from each sample). To convert ng/mL milk to µg/kg lipid: ng/mL milk / density of milk g/mL / 4%.

Daily exposure of infants to triclosan in breast milk was estimated by Health Canada (Table 3-5), assuming the geometric mean and maximum concentration of triclosan in breast milk of 0.051 µg/kg fresh weight and 73.18 µg/kg fresh weight, respectively (Arbuckle et al. 2015b) from Canadian mothers. Additional assumptions included mean breast milk intakes of 770 mL/day and 620 mL/day for infants under 6 months and 6-12 months of age, respectively, the density of human milk of 1.03 g/mL, and the body weights of 6 kg and 9.2 kg for infants less than 6 months of age and 6-12 months of age, respectively (US EPA 2011c).

Table 3-5. Exposure of infants to triclosan in breast milk
Exposure scenario	Triclosan concentration in milk (mg/kg)	Daily milk intake (mL/day)	Milk density (g/mL)	Body weight (kg)	Mean Estimated daily dose^a (mg/kg bw per day)	Maximum Estimated daily dose^a (mg/kg bw per day)
Birth to 6 months	0.000051-0.073	770	1.03	6	6.7 × 10^-6	0.010
6-12 months	0.000051-0.073	620	1.03	9.2	3.5 × 10^-6	0.005

Table Notes

^a Estimated daily dose (mg/kg bw per day) = triclosan concentration in milk (mg/kg) × daily intake (mL/day) × milk density (g/mL) × conversion factor (0.001 kg/g) / body weight (kg).

For infants under 6 months of age and 6-12 months of age, maximum daily exposures to triclosan in breast milk were estimated to be 0.010 mg/kg bw per day and 0.005 mg/kg bw per day, respectively. Using these maximum estimated daily exposures and the database NOAEL of 25 mg/kg bw per day, the resulting MOEs are 2500 and 5000 (target MOE of 300) for infants under 6 months and 6-12 month of age, respectively.

3.3.4.2.2 Object-to-mouth activity

Incidental oral exposure of children to triclosan resulting from object-to-mouth behaviours was assessed for 6- to 12-month-old infants mouthing a plastic toy. The following assumptions were used in the assessment of a plastic toy being mouthed: maximum surface area of 50 cm² that can be mouthed, plastic weight of 5 g, application rate of 0.5% a.i., 0.5% a.i. available on the surface of the toy, a saliva extraction efficiency of 50% (US EPA 2011b) and an average infant body weight of 9.2 kg (US EPA 2011c). The exposure dose for children mouthing a plastic toy was estimated to be 0.0068 mg/kg bw per day (Table 3-6).

Table 3-6. Incidental oral exposure of a 6- to 12-month-old infant mouthing a toy made from plastic treated with triclosan
Scenario	Surface area mouthed (cm²)	Plastic weight (g)	Amount available on plastic surface (% a.i.)	Maximum application rate (% a.i.)	Surface residue^a (mg a.i./cm²)	Saliva extraction efficiency (%)	Estimated daily dose^b (mg/kg bw per day)
Child mouthing a plastic toy	50	5	0.5	0.5	0.0025	50	0.0068

Table Notes

^a Surface residue (mg a.i./cm²) = toy weight/toy surface (g/cm²) × % a.i./100 × % a.i. available on surface/100 x conversion factor (1000 mg/g) = 0.0025.

^b Estimated daily dose (mg/kg bw per day) = surface residue (mg a.i./cm²) × saliva extraction efficiency (%)/100 x surface area (cm²) / body weight (kg).

Using this estimated daily exposure and the database NOAEL of 25 mg/kg bw per day, the resulting MOE is 3676 (target MOE of 300).

3.3.4.2.3 Dust ingestion

Triclosan has been measured in indoor dust in Canada, Belgium and Spain (Canosa et al. 2007a, 2007b; Geens et al. 2009; Fan et al. 2010). A summary of the results of each study is shown in Table 3-7.

Table 3-7. Triclosan in household dust
Location	Number of Samples	Mean (ng/g)	Minimum (ng/g)	Maximum (ng/g)	Limit of detection (ng/g)	Reference
Canada	63 homes	Median = 571 (fresh sample)	87	3040	8.7	Fan et al. 2010
Canada	63 homes	Median = 378 (composite sample)	82	4090	8.7	Fan et al. 2010
Canada	261 homes	733 (Median = 415)	32	7849	8.7	Unpublished data^a
Belgium	18 homes	484	25	1828	0.5	Geens et al. 2009
Spain	10 homes	702	240	2200	Not specified	Canosa et al. 2007a
Spain	8 homes	1134	90	2444	1.2	Canosa et al. 2007b

Table Notes

^a Unpublished Canadian House Dust Study data: 2011 e-mail from Environmental Health Science and Research Bureau, Health Canada, to Existing Substances Risk Assessment Bureau, Health Canada; unreferenced.

Incidental oral exposure of children resulting from hand-to-mouth activities was assessed based on a representative scenario of a 6- to 12-month-old infant crawling on a floor or carpet and ingesting triclosan-contaminated dust stuck to his or her hands. The assessment of potential oral exposure of children from hand-to-mouth activities was based on modelled estimates using a range (20 to 74 mg/day) of dust ingestion rates reported by Özkaynak et al. (2011) and Wilson et al. (2013), and the mean and maximum dust concentrations from the unpublished Canadian study.

Estimated daily doses for 6- to 12-month-old infants with a body weight of 9.2 kg ranged from 1.59 × 10^-6 mg/kg bw per day to 6.31 × 10^-5 mg/kg bw per day.

For 6- to 12-month-old infants, the daily exposure resulting from ingestion of triclosan-contaminated dust was estimated to be 3.27 x 10^-6 mg/kg bw per day (mean dust concentration from the unpublished study and mean dust ingestion rate for toddlers from Wilson et al. 2013). Using this estimated daily exposure and the NOAEL of 25 mg/kg bw per day, the resulting MOE is greater than 7 000 000 (target MOE of 300).

3.3.4.2.4 Inhalation of triclosan-contaminated dust

Inhalation exposure of infants to triclosan in household dust was estimated using available dust standards (US EPA 2008d) and a maximum triclosan concentration in dust of 7849 ng/g from the unpublished Canadian House Dust study (see Table 3-7). Additional assumptions included an inhalation rate for a child less than 1 year of age of 5.4 m3/day (US EPA 2011c) and the average body weight of a 6- to 12-month-old infant of 9.2 kg. The estimated daily doses ranged from 6.45 × 10^-8 to 6.45 × 10^-6 mg/kg bw per day.

For 6- to 12-month-old infants, the maximum daily exposure resulting from inhalation of triclosan-contaminated dust was estimated to be 6.91 x 10^-5 mg/kg bw per day. Using this estimated daily exposure and the NOAEL of 3.21 mg/kg bw per day from the inhalation study in rats, the resulting MOE is greater than 46 450 (target MOE of 300). This is considered a conservative estimate and is based on the assumption that all triclosan-contaminated dust is bioaccessible and thus readily absorbed.

3.3.4.2 Aggregate exposure of children less than 3 years of age

The Canadian biomonitoring data available for children less than 3 years of age are limited to infants 0 to 3 months old and children aged 23 up to 36 months of age. No Canadian biomonitoring data are available for 4- to less than 23 months.

The aggregate risk for children was estimated by combining estimated daily doses from infant-specific scenarios with estimated daily doses derived from biomonitoring data for either children 23 to 36 months (MIREC-CD Plus; unpublished) or infants (0-3 months old) (Arbuckle et al. 2015b) (Table 3-8). A combined MOE approach is used to aggregate estimated daily exposures from scenarios with the same target MOE. According to information obtained for infants in the P4 Study, the majority of the urine concentrations are from breast fed or breastfed and formula-fed infants (Arbuckle et al. 2015b); therefore, inclusion of exposure (and MOE) for 0- to 3-month old infants is expected to account for all potential routes of exposure, including nursing. The following aggregation equation was used to aggregate "unitless" MOEs into a total MOE (MOE_T):

MOE_T = 1 / [1/MOE₁ + 1/MOE₂+… + 1/MOE_n]

where MOE₁, MOE₂, …, MOE_nrepresent route-specific scenarios (i.e., object-to-mouth, hand-to-mouth and P4 biomonitoring data for a 0- to 3-month-old infants or MIREC-CD Plus biomonitoring data for children 23 to 36 months. A total MOE greater than the target MOE of 300 indicates that risk is not of concern.

Table 3-8. Aggregate risk estimates for children less than 3 years of age
Scenario	Estimated daily dose (mg/kg bw per day)	MOE	Details
Urine data (MIREC-CD Plus)	3.37 × 10^-4	74250	Preliminary MIREC-CD Plus urine data for children 23-36 months (mean concentration and high end mean urine volume)
Urine data (P4 Study)	6.90 × 10^-4	36 232	P4 urine data for infants 0-3 months of age (mean exposure dose)
Nursing	3.50 × 10^-6	7 142 857	Infants 6-12 months of age (geomean concentration)
Dust ingestion	3.27 × 10^-6	7 645 260	Infants 6-12 months of age (mean concentration and mean ingestion rate from Wilson et al. 2013)
Object-to-mouth	6.8 × 10^-3	3676	Infants 6-12 months of age
Combined MOE approach^a	NA	3500	Preliminary MIREC-CD Plus (children 23-36 months) + dust ingestion + object-to-mouth + nursing
Combined MOE approach^a	NA	3336	P4 (infants 0-3 months old) + dust ingestion + object-to-mouth

Table Notes

Abbreviations: NA, not applicable.

^a Combined MOE = 1 /(1/MOE₁ + 1/MOE₂ + … + 1/MOE_n), where MOE₁, MOE₂, …, MOE_n represent route-specific scenarios.

The inhalation estimate was not included in the aggregate exposure assessment, since the contribution of inhalation exposure was considered negligible when compared with other potential routes of exposure (see above).

Using the combined MOE approach, aggregate exposure of 6- to 12-month-old infants resulted in combined MOEs ranging from 3336 to 3500 (target MOE of 300). The results of this highly conservative risk assessment indicate that the aggregate risk for children less than 3 years of age, including breastfed infants, is below the level of concern.

3.3.4.3 Uncertainties associated with aggregate risk assessment for children

There are uncertainties and conservative assumptions in conducting an aggregate exposure and risk assessment for children, due to a lack of adequate data to fully characterize the exposure of young children to triclosan. These uncertainties are highlighted below.

Similar to the uncertainties identified in section 3.3.2.5 for individuals aged 3 and older, there are generally recognized uncertainties associated with using spot urine samples to estimate human exposures to triclosan. To account for this uncertainty, a range of mean infant daily urine volumes were used (Appendix C). Another uncertainty in the dose conversion of infant spot urine samples is the use of the median value of 54% to account for urinary excretion of triclosan in infants and therefore assuming that absorption, distribution, metabolism and elimination parameters are the same for all individuals and remain constant within individuals over time. The SCCP (2009) concluded that, although there are limited pharmacokinetic data for children and no direct comparisons with adults were possible given differences in doses and dosing formulations in various studies, elimination was determined to be essentially the same for children and adults, rapid, based on an oral dosing study with toothpaste and dental slurry. No pharmacokinetic data related to triclosan for infants was identified. However, it is known that infants in their first year of life, do not have the fully mature metabolic capacity that adults do, and that certain renal clearance mechanisms are also not fully developed in infants up to 6 months of age (Alcorn and McNamara 2002).

There is an uncertainty with respect to the dose estimation for breastfed infants due to the high variability of triclosan measurements in breast milk, possibly related to its short half-life. It is unknown if high levels of triclosan in some breast milk samples were the result of abundant use of products used by consumers or the result of an isolated contamination of sample. For that reason, an assumption of the maximum triclosan concentration detected in breast milk for the nursing exposure scenario is considered highly conservative.

There is uncertainty regarding the potential co-occurrence of all identified scenarios in practice. An assumption that a child will be exposed daily to high triclosan residues as identified for each scenario is considered conservative. The assumption that all potential exposure scenarios will co-occur also represents conservatism in the aggregate assessment for infants 6-12 months of age. Further, assumptions used in incidental oral exposure assessments (i.e., hand-to-mouth and object-to-mouth) are considered conservative, since it is unlikely that all plastic toys will be made with material treated with triclosan.

There is also uncertainty regarding the inclusion of the MIREC-CD Plus estimate for children 23 to 36 months in the aggregate risk assessment for 6- to 12-month-old infants. The inclusion of the MIREC CD Plus estimate is expected to err on the side of overestimating the potential aggregate dose, since additional sources of exposure that are not relevant to the infant scenario are also captured (e.g., washing hand with antimicrobial soap).

3.3.5 Human health risk assessment for workers exposed to pest control products containing triclosan

Workers can be exposed to triclosan via inhalation and dermal contact with this active ingredient while handling the chemical during the manufacturing process or when handling the manufactured goods.

3.3.5.1 Handler Exposure and Risk

There were no chemical-specific exposure studies available for triclosan. Health Canada's PMRA assessed occupational exposure in industrial settings using exposure data from the Chemical Manufacturers' Association (CMA) Antimicrobial Exposure Assessment Study (CMA 1990). The objective of the CMA Study was to measure occupational exposure of industrial workers during mixing or transfer of antimicrobials to industrial systems. The study monitored workers' exposure to chemicals used as preservatives in metal working fluids, paints and coatings, in wood, pulp and paper facilities and in cooling towers. Worker exposure was measured for different application methods, including a liquid pour (open mixing/transfer) and liquid pump (closed mixing/transfer).

Dermal and inhalation exposures of individuals involved in the transfer of the antimicrobial (as many transfers as are normally conducted in a workday) from the container to the production batch were monitored in the study. Dermal exposure was assessed by inside and outside gauze patch dosimeters through one layer of clothing. Exposure of the hands was measured using cotton fabric gloves. Inhalation exposure was measured by using a personal sampling pump. Due to the diversity of the products used, there was significant variability in the types of protective clothing worn. Most individuals wore long-sleeved shirts and long pants. Each replicate was representative of the time spent performing the antimicrobial-related task in 1 day; therefore, the data were not normalized. Laboratory and field recoveries were measured; however, recoveries were highly variable due to an insufficient number of spiked samples, poor collection efficiency of sample media, difficulty in the analysis for the active ingredient and poor storage stability. These are considered limitations of the CMA exposure study.

Monitoring times and the amount of active ingredient handled daily in plants manufacturing paints and coatings, in plants using metal working fluids and in cooling towers ranged from 2 to 285 minutes and from 0.006 to 265 kg, respectively. In all scenarios, exposure was primarily dermal. Total exposure for each replicate was calculated by summing the total dermal and inhalation doses for each replicate. Since applications of biocides in industrial processes are similar regardless of the use site (e.g., cooling towers, pulp and paper), it was considered appropriate to combine replicates based on the application method. Thus, the replicates with liquid pour and pump application in material preservatives, cooling towers and pulp and paper scenarios were combined to generate exposure estimates. Given the limitations of the exposure study (low and variable laboratory and field recoveries), the 90th percentiles generated from the input CMA data were used by Health Canada's PMRA to estimate potential risks to operators handling industrial products containing triclosan. Dermal and inhalation exposure estimates represent the 90th percentile of exposure dose normalized to a 70 kg body weight (Table 3-9). Since most individuals in the CMA Study wore long sleeves, long pants and cotton gloves, these data are considered representative of an individual wearing a single layer and gloves.

Table 3-9. Occupational risk assessment for industrial handler
Application method	Dermal exposure^a (mg/kg bw per day)	Inhalation exposure^a (mg/kg bw per day)	MOE^b Dermal	MOE^b Inhalation
Liquid, pour	0.1034	0.0010	387	3210
Liquid, pump	0.0268	0.0032	1493	1003

Table Notes

^a 90^th percentile of the exposure dose normalized to a 70 kg body weight (CMA 1990).

^b MOE = NOAEL (mg/kg bw per day) /daily exposure dose (mg/kg bw per day), where the NOAEL of 40 mg/kg bw per day with a target MOE of 300 was selected for the dermal scenarios, while the NOAEL of 3.21 mg/kg bw per day with a target MOE of 300 was selected for the inhalation scenarios.

The results of the occupational risk assessment for workers applying triclosan in industrial settings via the closed delivery system or an open pour method indicate that risks are below the level of concern.

3.3.5.2 Occupational Post-application Exposure and Risk

Occupational post-application exposure of workers handling manufactured products is not expected to be of concern based on the registered use pattern, since triclosan is applied at low application rates during the manufacturing process and is expected to be embedded in the finished product.

3.3.5.3 Uncertainties in Worker Exposure Estimation

There are uncertainties and conservatisms in conducting occupational risk assessments due to a lack of adequate tools and data to fully characterize exposure from all possible routes. Some of these uncertainties are highlighted below.

Occupational exposure estimates are based on data from the CMA Antimicrobial Exposure Assessment Study. Even though there are a number of limitations associated with the study, it is currently the only occupational study available with which to assess potential exposure from antimicrobial uses of pest control products. Low and variable laboratory and field recoveries were obtained in this study, which may affect the validity of the reported exposure estimates. However, since the 90th percentile estimates from this study were used for risk assessment purposes, exposure estimates are not expected to be underestimated.

Because of the limitations described above, the exposure estimates from the CMA Antimicrobial Exposure Assessment Study were not normalized to the amount of active ingredient handled per day. The activities monitored in the study were considered representative of a typical workday; thus, no normalization was conducted. In addition, many of the activities do not involve direct handling of the biocide, but rather a change in coupling or hose from the biocide container. It is uncertain whether the amount of triclosan handled per day by workers is within the range of kilograms of active ingredient handled in the CMA Antimicrobial Exposure Assessment Study.

3.4 Cumulative Effects

Health Canada's Science Policy Notice SPN2001-01, Guidance for Identifying Pesticides that have a Common Mechanism of Toxicity for Human Health Risk Assessment, describes the steps for identifying mechanisms of toxicity of pesticides that cause a common toxic effect, the types of data needed and their sources, how these data are to be used in reaching conclusions regarding commonality of mechanisms of toxicity, and the criteria Health Canada applies for categorizing pesticides for the purpose of cumulative risk assessments. No relevant evidence indicating that triclosan shares a common mechanism of toxicity with other pesticides or shares a toxic metabolite produces by other pesticides has been identified (US EPA 2008a, US EPA 2014.

3.5 Transformation Products

There are a number of potential environmental transformation products of triclosan to which the general population may be exposed, including methyl-triclosan, 2,4-dichlorophenol (2,4-DCP) and PCDDs (Section 4.2).

Methyl-triclosan is a major environmental transformation product formed as a result of biomethylation in soil and water systems (see Sections 4.1.2.2 and 4.2.5.2). It is also formed during the aerobic treatment of wastewater and is discharged in effluents from WWTPs with triclosan. While there is limited monitoring information for methyl-triclosan in the environment and there is uncertainty regarding the observed half-lives and bioaccumulation estimates for this compound, the available laboratory and aquatic field evidence indicates that methyl-triclosan is likely to be both more persistent and more bioaccumulative than triclosan.

2,4-DCP and the lower chlorinated dioxins 2,7/2,8-DCDD are major photoproducts of triclosan (see Section 4.2.3). In addition, 2,4-DCP as well as PCDDs (1,2,8-trichlorodibenzo-p-dioxin [1,2,8-TriCDD], 2,3,7-TriCDD and 1,2,3,8-TCDD) can form in natural water as a result of further phototransformation of chlorinated triclosan derivatives (formed during the disinfection of wastewater). A SIDS Initial Report for 2,4-DCP (under the OECD High Production Volume [HPV] Chemicals Programme) indicated that human exposure to this chemical from the use of products containing 2,4-DCP and from the environment is expected to be low (OECD 2007). Dioxins usually enter and are present in the environment as complex mixtures. The toxicity of different dioxins is expressed on a common basis using the international toxicity equivalency factors that recognize and compare the similarities and differences between the toxic actions of the dioxins. The lower chlorinated dibenzodioxins (2,7/2,8-DCDD, 1,2,8-TriCDD, 2,3,7-TriCDD and 1,2,3,8-TCDD) are not listed on the list of 17 dioxins and furans that are of the greatest concern to human health based on international toxicity equivalency factors (NATO 1988), which means that they will be expected to contribute comparatively little to the toxicity of a complex mixture. On this basis, the potential for general population risk from these dioxins is expected to be low.

Triclosan was also shown to react with chlorine ion in tap water to form chloroform (Rule et al. 2005). The 2001 Government of Canada Priority Substances List Assessment Report for Chloroform (Canada 2001) indicated that human exposure to chloroform from all potential routes and sources of exposure is expected to be considerably less than the level to which a person may be exposed daily over a lifetime without harmful effect.

3.6 Antimicrobial Resistance

The potential of triclosan to induce antimicrobial resistance (AMR) was reviewed in assessments published by the Australian Department of Health and Ageing (NICNAS 2009) and the European Commission (SCENIHR 2009, 2010; SCCS 2010).

In 2009, NICNAS concluded, based on a comprehensive review of literature published in scientific journals between 2002 and 2005 and the 2002 European Commission Scientific Steering Committee review of triclosan AMR (European Commission 2002), that there was "no evidence that the use of triclosan is leading to an increase in triclosan-resistant bacterial populations or that there is any increased risk to humans regarding antibiotic resistance" (EU 2002; NICNAS 2009).

In 2009 and 2010, the European Commission also published comprehensive reviews of available scientific data on the antibiotic resistance effects of triclosan. The studies reviewed by the EU's Scientific Committee on Emerging and Newly Identified Health Risks (SCENIHR) indicated that triclosan-resistant bacteria can be found in health care settings and in products used by consumers. Although laboratory studies showed that it is possible to develop bacterial mutants with reduced susceptibility to both triclosan and antibiotics, no notable selection of antibiotic resistance in bacteria exposed to triclosan was observed in the environmental studies. In addition, the lack of data on other biocide compounds prevented the SCENIHR from reaching a conclusion regarding the potential for triclosan to induce bacterial antibiotic resistance under field use conditions (SCENIHR 2009). The EU's Scientific Committee on Consumer Safety (SCCS) concluded that, based on the available scientific information, it was not possible to quantify the risk of development of AMR induced by triclosan applications, including its use in cosmetics (SCCS 2010). This position was confirmed in a parliamentary response in May 2013 (European Parliament 2014).

A more recent review of all available information on the potential for triclosan to induce AMR was conducted on behalf of Health Canada via an external expert review (Tetra Tech 2014). This review, spanning publicly-available literature between 2009 to 2013 with judicious review of the older literature, addressed the following: development of triclosan resistance, bacterial cross resistance arising from triclosan exposure, uses of triclosan as it relates to clinical versus environmental and household settings, as well as fate and environmental occurrence of triclosan.

Consistent with the previous conclusions stated by SCENIHR (2009), this review identified studies that indicated that there is the potential for triclosan-resistant bacteria to exist in clinical settings (environments not representative of general population exposure), but this has not been documented outside of clinical use (e.g. household settings) (Larson et al 2003; Lanini et al 2011, Skovgaard et al., 2013; Guiliano et al 2015).

More recently, a large-scale study examining 3319 clinical isolates from three different locations (research lab, hospital, and university) did not find any significant evidence of triclosan resistance^Footnote2 (Morrissey et al 2014). It should also be noted that uses of triclosan-containing soaps in clinical settings differ from consumer (household) settings in: the formulations used, the concentration of triclosan in the soap, the duration of scrubbing and the frequency of scrubbing and these factors are important in interpretation of relevance for the general population.

There were no studies demonstrating the development of AMR after repeated sublethal exposures of triclosan to bacteria found in household settings through use of antimicrobial soaps compared to plain soaps (Cole et al 2003), triclosan-containing flooring (Møretrø et al., 2011), dental samples (McBain et al., 2004), triclosan-containing boxes (Braid and Wale, 2002), or in sinks and drains of users and non-users of biocidal cleaning agents, including triclosan (McBain et al., 2003; Marshall et al 2012).

A more recent study by Cullinan et al (2013) investigated whether long-term continuous use of triclosan-containing toothpaste (0.3% w/w) selected for triclosan-resistant bacteria commonly found in the mouth. Common species between both the placebo and triclosan user dental plaque isolates showed similar Minimum Inhibitory Concentrations (MICs) for a range of concentrations (125-1000 µg/ml) of triclosan leading the authors to conclude that continuous use of triclosan-containing toothpaste over 5 years did not result in the development of triclosan-resistant bacteria in the mouth. This is further supported by annual results from a 19- year-long evaluation (1991-2010) of dental plaques of 58 subjects who used triclosan dentifrice for at least 5 years and showed no changes in oral microbial susceptibility to triclosan over this long period of time (Haraszthy et al. 2014).

Previous studies have reported the potential for the generation of triclosan-resistant bacterial strains in the laboratory (Heath et al., 1998; Heath et al., 1999; McMurray, 1999). However, in the few cases where resistant organisms have been isolated, there is little data to suggest that this resistance was the result of triclosan. There have been cases where cross-resistance was claimed in laboratory and clinical triclosan resistant strains in vitro (Aiello et al., 2007), but there have also been several papers that were unable to find evidence of cross-resistance (Suller and Russell 2000; Wingnal et al. 2008; Cottell et al. 2009; Saleh et al. 2010; Skovgaard et al. 2013). Furthermore, differences in formulations of triclosan used in clinical and laboratory settings compared to commercial applications are unknown.

Concentrations of triclosan observed in Canadian surface waters and wastewaters are well below those required to inhibit bacterial growth (Koburger et al. 2010, Latimer et al. 2012, Blair et al. 2013). Resistant phenotypes develop over a wide range of concentrations, depending on the organism, but even the lowest triclosan concentration (0.23 µg/mL; Latimer et al. 2012) at which resistant phenotypes have been observed is at least an order of magnitude above the highest observed surface water concentration (Table 4-3). In addition, correlation between the presence of triclosan in surface water or waste water and the presence of bacteria resistant to other antimicrobials and antibiotics is not consistently observed (Novo et al., 2013; Carey and McNamara 2015).

Overall, although there is the potential for triclosan-resistant bacteria to exist in clinical and laboratory settings, this has not been well documented outside of clinical use (e.g. households, toothpaste use, and environmental waters), and, due to the unavoidable limitations in clinical and laboratory studies as they relate to the potential for triclosan to induce AMR outside of these settings, interpretation for relevance to the general population is also limited. Therefore, based on available information, induction of AMR from current levels of triclosan has not been identified as a concern for human health.

Top of Page

Page 4

Date modified:: 2016-11-25

Final Assessment

3. Human Health

3.1 Toxicology Profile of Triclosan

3.1.1 Metabolism and Toxicokinetics

3.1.2 Acute Toxicity

3.1.3 Subchronic Toxicity

3.1.4 Reproductive Toxicity

3.1.5 Developmental Toxicity

3.1.6 Chronic Toxicity

3.1.7 Genotoxicity

3.1.8 Carcinogenicity Potential in Humans

3.1.9 Neurotoxicity

3.1.10 Thyroid Effects

3.1.11 Immunotoxicity

3.2 Toxicological Endpoints for the Human Health Risk Assessment

3.2.1 Completeness of the Database

3.2.2 PCPA Hazard Characterization

3.2.3 Acceptable Daily Intake (All Populations)

3.2.4 Toxicological endpoints for residential and occupational risk assessment

3.2.4.1 Incidental oral exposure (directly exposed children)

3.2.4.2 Dermal Exposure

3.2.4.3 Inhalation Exposure

3.2.5 Aggregate Exposure Scenarios

3.2.6 Cancer Risk Assessment

3.3 Human Health Exposure and Risk

3.3.1 General population exposure and risk assessment

3.3.2 Estimation of the daily exposure dose based on the urinary triclosan concentration

3.3.2.1 CHMS cycle 2

3.3.2.2 Urinary triclosan concentrations from the P4 study

3.3.2.3 Urinary triclosan concentrations from the MIREC study

3.3.2.4 Uncertainties associated with dose conversion

3.3.3 Aggregate risk assessment for the general population (3-79 years of age)

Table Notes

3.3.4 Aggregate risk assessment for children younger than 3 years of age

Table Notes

3.3.4.1 Meconium and amniotic fluid

3.3.4.2 Infant-specific exposure scenarios

3.3.4.2.1 Nursing

Table Notes

Table Notes

3.3.4.2.2 Object-to-mouth activity

Table Notes

3.3.4.2.3 Dust ingestion

Table Notes

3.3.4.2.4 Inhalation of triclosan-contaminated dust

3.3.4.2 Aggregate exposure of children less than 3 years of age

Table Notes

3.3.4.3 Uncertainties associated with aggregate risk assessment for children

3.3.5 Human health risk assessment for workers exposed to pest control products containing triclosan

3.3.5.1 Handler Exposure and Risk

Table Notes

3.3.5.2 Occupational Post-application Exposure and Risk

3.3.5.3 Uncertainties in Worker Exposure Estimation

3.4 Cumulative Effects

3.5 Transformation Products

3.6 Antimicrobial Resistance

Page details