Registration Dossier

Toxicological information


Currently viewing:

Administrative data

Description of key information

No relevant studies in humans on the carcinogenicity of DEHP is available. The results show that DEHP is carcinogenic in rats and mice. A statistically significant increase in the incidence of hepatocellular tumours with a dose-response relationship was observed in rats and mice of both sexes and a significant dose-related increase in the incidence of Leydig cell tumours was observed in male rats. It was also noted that low doses did not cause hepatocellular tumours, which suggests a threshold for this effect.

However, there is a plausible mechanism for the PPs-induced hepatocarcinogenicity in rodents (activation of PPARα) and there is evidence showing that humans are less sensitive to the hepatotoxic effects of PPs by the suggested mechanism. Therefore, the relevance for humans of the liver tumours in rodents induced by DEHP, a weak PPs, is regarded to be negligible.

Leydig cell tumours have been reported in only one study in Sprague–Dawley rats at doses that have been used in two other studies (using F-344 rats). It has been argued that the differing results are the consequence of the high spontaneous incidence of Leydig cell tumors in Fischer 344 rats compared to Sprague–Dawley rats.

In a 104-week rat study, an increased incidence of mononuclear cell leukaemia (MCL) was also noted. The relevance of MCL is unknown, but it was only seen in one of two rat studies and in neither of the two mouse studies. Moreover, this tumour type is well known to occur spontaneously with high incidence in the F344 rat strain used in the study.

The LOAEL and the NOAEL for tumour induction in rats (both liver tumours and MCL) were established as 2500 ppm (146.6 mg/kg bw/d for males) and 500 ppm (28.9 mg/kg bw/d for males) respectively (David et al., 1999, 2000). In mice, the LOAEL and the NOAEL for induction of liver tumour were 1500 ppm (292 mg/kg bw/d for males) and 500 ppm (98 mg/kg bw/d for males) respectively (David et al., 1999, 2000).

Key value for chemical safety assessment

Carcinogenicity: via oral route

Endpoint conclusion
Endpoint conclusion:
adverse effect observed

Carcinogenicity: via inhalation route

Endpoint conclusion
Endpoint conclusion:
no study available

Carcinogenicity: via dermal route

Endpoint conclusion
Endpoint conclusion:
no study available

Additional information

Recently, the International Agency for Research on Cancer (IARC) re-evaluated the carcinogenicity of DEHP. Based on mechanistic information the Agency concluded that activation of peroxisome proliferator-activated receptorαrepresents an important mechanism in rodents, but that “multiple molecular signals and pathways in several cell types in the liver, rather than a single molecular event, contribute to the induction of cancer in rats and mice.” The Agency concluded on a classification in Group 2B (possibly carcinogenic to humans) (Rusyn and Corton, 2012).

It should be emphasised that no convincing data are available pointing to involvement of genotoxic modes of action. Furthermore, humans do not appear susceptible to peroxisomal proliferation (PP) as exemplified by clinical studies of populations exposed for long periods to hypolipidemic drugs (rodent hepatocarcinogens and strong rodent PPs). Liver tumours and the modes of action (including alternative modes of action as discussed by IARC) by which they form are likely not relevant to humans; however, a data gap for demonstrating the lack of cell proliferative events (key event in the rodent MOA) in humans was acknowledged.

In conclusion, these new data and evaluation by the IARC Working Group don’t warrant a modification of the actual non-classification of DEHP as carcinogenic according to the criteria of Regulation (EC) N° 1272-2008 (CLP) and will not impact the risk assessment performed under Regulation (EC) N° 1907-2006 (REACH), as DNELs for DEHP have been derived from the reproductive toxicity data, which is the most sensitive end-point.

Studies in humans

No relevant studies in humans on the carcinogenicity of DEHP is available.

Studies in animals

Long-term carcinogenicity studies in rats and mice have been published. The studies are comparable to guideline studies and performed according to GLP.


Rat studies

In a study compliant to OECD guideline 453,F-344 rats (70/sex/group) received DEHP in the diet at doses of 0, 100, 500, 2500, or 12500 ppm (M/F: 0/0, 5.8/7.3, 28.9/36.1, 146.6/181.7, or 789/938.5 mg/kg bw/d) for 104 weeks(Davidet al., 1999, 2000). In an additional recovery group, rats (55/sex/group) were administered 12500 ppm DEHP for 78 weeks, followed by a 26-week recovery period. Increases in hepatocellular adenomas and mononuclear cell leukaemia (MCL) in males at 2500 ppm and above and hepatocellular carcinomas in males and females at 12500 ppm, were observed. However, the incidence of hepatocellular adenomas/carcinomas was decreased in recovery animals at 12500 ppm (2-week recovery period), compared with the same dose group at the end of the dosing period. Peroxisome proliferation was induced from 2500 ppm. Effects on the liver, kidney and testis induced at 2500 ppm and above are described in Section “repeated dose toxicity”.The LOAEL for tumour induction (hepatocellular neoplasms and MCL in male rats) was 2500 ppm (147 mg/kg b.w. per day for males). The NOAEL was 500 ppm (28.9 mg/kg bw/d, males).

In a study performed following a protocol comparable to OECD guideline 451, groups of 50 male and 50 female Fischer 344 rats, five to six weeks of age, were fed diets containing 6000 or 12 000 mg/kg diet (ppm) di(2-ethylhexyl) phthalate (> 99% pure) for 103 weeks(NTP, 1982; Kluwe et al., 1982 and Kluwe et al., 1983). All surviving rats were killed at 104–105 weeks. Other groups of 50 males and 50 females served as controls. There was a dose-related decrease in body weight gain in both sexes but no effect on survival. More than 60% of the animals survived to the end of the study. High-dose male rats had significant increases (p= 0.01, Fisher’s exact test) in the combined incidence of hepatocellular carcinomas and neoplastic nodules (control, 3/50; low-dose, 6/49; high-dose, 12/49). The Cochran–Armitage test also indicated a significant trend (p= 0.007). [The IARC Working Group (2000) noted that the term neoplastic nodule is now generally assumed to represent hepatocellular adenomas.] The incidence of hepatocellular carcinomas alone or neoplastic nodules alone was not significantly increased. In female rats, the incidence of hepatocellular carcinomas was increased in high-dose rats (8/50;p= 0.003, Fisher’s exact test) compared with controls (0/50) and that of neoplastic nodules was also increased in high-dose females (5/50;p< 0.028) compared with controls (0/50). The incidence of hepatocellular carcinomas and neoplastic nodules combined was also increased in low-dose (6/49;p= 0.012) and high-dose (13/50;p< 0.001) females compared with controls (0/50).The LOAEL for tumour induction in rat was 6/000 ppm DEHP in the diet (320 mg/kg/day for male rats).

DEHP was administered in the diet at 0, 600, 1897, and 6000 mg/kg to male Sprague-Dawley rats beginning at an age of 90–110 days and continuing for the remaining lifetime of the animals (up to 159 weeks) (Voss et al., 2005). DEHP dose levels were 0, 30, 95, and 300 mg/kg bw/d. Significantly increased incidence of hepatocellular adenomas and carcinomas were observed at the highest dose. The percentage of benign Leydig cell tumors in the highest dose group was almost twice as high as the percentage in the control group (28.3% versus 16.4%). There was a significant dose-related trend in incidence of hepatic neoplasms and Leydig cell tumours. Leydig cell tumours have not been reported in previous studies in Sprague-Dawley rats, most likely due to late appearances outside the normal observation ranges of carcinogenicity studies.

Male F-344 rats were fed a diet containing 2% di-(2-ethylhexyl)phthalate (DEHP) for 95 weeks (Rao et al., 1987). Liver nodules and/or hepatocellular carcinomas (HCC) developed in 6/10 rats fed DEHP and none were found in controls (P less than 0.005 by chi 2 test). All the nodules and HCC were negative for gamma-glutamyl transpeptidase. In the non-tumorous portions of liver, the hepatocytes contained an increased number of peroxisomes and extensive accumulation of lipofuscin. By immunocytochemical analysis, the liver peroxisomes in rats treated chronically with DEHP had visually detectable decrease in the H2O2-degrading catalase and increase in H2O2-producing fatty acyl-CoA oxidase. These results show that higher dietary level of DEHP, which causes substantially greater degree of peroxisome proliferation than the 1.2% dietary level used in the National Toxicology Program bioassay (1982, Publication no. NTP-80-37, Tech. Report Series No. 217), can induce liver tumors in male rats.

F-344 male rats were given a diet containing 2% DEHP ad libitum for 108 wk (Rao et al., 1990). At necropsy livers were quantitatively analyzed for total tumor incidence and the number of lesions per liver after slicing the entire organ at 1- to 2-mm intervals. Neoplastic nodules and/or hepatocellular carcinomas were observed in 11 of 14 rats (78.5%). When evaluated according to the size, 57, 16, and 36% rats contained nodules ranging from 1 to 3, 3 to 5, and greater than 5 mm in size, respectively. The number of nodules per liver ranged from zero to four. These results indicate that DEHP induces tumors in a large number of animals at 2% dose levels.


In a study compliant to OECD guideline 453,B6C3F1 mice (70/sex/group) received DEHP in the diet at concentrations of 0, 100, 500, 1500, or 6000 ppm (M/F: 0/0, 19.2/23.8, 98.5/116.8, 292.2/354.2, or 1266.1/1458.2 mg/kg/d) for 104 weeks(Davidet al., 1999, 2000). In an additional recovery group, mice were dosed with 6000 ppm of DEHP for 78 weeks, followed by a 26-week recovery period. Significantly increased incidences of hepatocellular adenomas and carcinomas were observed at 1500 ppm and 6000 ppm in male mice. In these two high dose groups, induction of peroxisome proliferation but not hepatocellular proliferation was more pronounced in both sexes. In the 6000 ppm recovery group, the incidence of hepatocellular adenomas, but not carcinomas, was less than in the 6000 ppm group. Non-tumour endpoints are described in Section “repeated dose toxicity”. The LOAEL for tumour induction (hepatocellular neoplasms in male mice) in this study was 1500 ppm (292 mg/kg bw/d). The NOAEL was 500 ppm (98 mg/kg bw/d).

In a study performed following a protocol comparable to OECD guideline 451, groups of 50 male and 50 female B6C3F1mice, six weeks of age, were fed diets containing 3000 or 6000 mg/kg diet (ppm) di(2-ethylhexyl) phthalate (> 99% pure) for 103 weeks weeks(NTP, 1982; Kluwe et al., 1982 and Kluwe et al., 1983). All surviving mice were killed at 104–105 weeks. There was a clear dose-related decrease in body weight gain in females. Survival at the end of the study was more than 60% in males and more than 50% in females. High-dose males had a slightly decreased body weight gain. In male mice, significant increases in the incidence of hepatocellular carcinomas were observed (control, 9/50; low-dose, 14/48; high-dose, 19/50;p= 0.022, Fisher’s exact test). The Cochran-Armitage test also indicated a significant trend (p= 0.018). The incidence of hepatocellular adenomas and carcinomas combined was also increased in males (control, 14/50; low-dose, 25/48,p= 0.013; highdose, 29/50,p= 0.002, Fisher’s exact test). In females, significant increases in the incidence of hepatocellular carcinomas were seen (control, 0/50; low-dose, 7/50,p= 0.006; high-dose, 17/50,p< 0.001, Fisher’s exact test) and of hepatocellular adenoma and carcinoma combined (control, 1/50; low-dose, 12/50; high-dose, 18/50,p< 0.001, trend and Fisher’s exact tests).The LOAEL for tumour induction in mice was 3,000 ppm DEHP in the diet (670 mg/kg/day for male mice).


No reliable studies have been reported.


No studies have been reported.

Studies on tumour initiating and/or promoting activity

Since DEHP is considered to be a non-genotoxic substance, it has been suggested that the carcinogenic effect is exerted during the promotion phase of hepatocarcinogenicity. DEHP has therefore been tested in several initiation/promotion experiments in rats and mice where the end-point has been the number and/or volume of altered liver cell foci.

In conclusion, DEHP have no tumour initiating activity (Garvey et al., 1987; Williams et al., 1987; Ward et al., 1983), a positive promoting activity in mice liver (Ward et al, 1983, 1984; Schuller and Ward, 1984; Hagiwara et al., 1986; Ward et al, 1990; Weghorst et al, 1994), a weak or no promoting activity in rat liver (Popp et al., 1985; Williams et al, 1987; Oesterle et al., 1988; Ito et al., 1988; Gerbracht et al., 1990) and a promoting activity in rat kidneys (Kurokawa et al., 1988).


Concerning the hepatocarcinogenicity of DEHP, previously, two different modes of action have been suggested for DEHP and other Peroxisome Proliferators (PPs):

-induction of peroxisome proliferation leading to oxidative stress and generation of electrophilic free radicals and/or

-increased hepatocyte proliferation/ suppression of hepatocellular apoptosis which could lead to fixation of a previously existing DNA-damage; enhancing the conversion rate of initiated cells to tumor cells; as well as increasing the susceptiblity of hepatocytes to replication and a subsequent neoplastic transformation.

None of these mechanistic premises provides a wholly satisfactory explanation of the mechanism of cancer induction caused by DEHP. However, in view of the available evidence, a mechanism due to oxidative stress seems to be the least likely to play a major role (Cattley et al., 1998; Youssef and Badr, 1998; IARC, 1995).

A third and a more feasible mechanistic basis for hepatocarcinogenicity through activation of Peroxisome Proliferator activated Receptor alpha (PPARα) has been accepted by most of the experts in this field. Activation of PPARαis also required for the induction of peroxisome proliferation, cell proliferation, and most probably also of several other aspects of the multifaceted effects brought about by the PPs (Peters et al., 1997, Ward et al., 1998, Cattley et al., 1998). The role of PPARαin the toxicity of DEHP has been investigated, recently, in a subchronic study in PPARα-null and wild-type male Sv/129 mice (Ward et al., 1998). Whereas the wild-type mouse fed DEHP exhibited typical lesions in the liver (such as increase in the number of peroxisomes, induction of replicative DNA-synthesis, and hepatomegaly), kidney, and testes, no signs of liver toxicity was detected in the PPARα-null mice. On the other hand, evidence of lesions in kidneys and testes were found also in the PPARα-null mice, indicating a PPARα-independent pathway for induction of toxicity in these organs. It has been demonstrated in another study with knockout mice fed the potent peroxisome proliferator Wy-14,643 (Peters et al., 1997) that PPARαis required for the hepatocarcinogenicity of this substance. However, there is still no clear evidence showing that the carcinogenicity of DEHP in rodent is mediated through activation of PPARα.

Species differences are evident regarding the response to the different effects of the PPs on the liver. Rats and mice are very sensitive, Syrian hamsters appear to exhibit an intermediate response, whereas guinea pigs and monkeys appear to be relatively insensitive. The potential human response to PPs has been examined in liver biopsies obtained from patients treated with hypolipidemic drugs with no evidence of peroxisome proliferation. The low sensitivity of human liver to the effects of PPs could be explained by the low level of PPARαfound in human liver (1-10% of the level found in rat and mouse liver) and genetic variations that render the human PPARαless active as compared to PPARαexpressed in rodent liver (Palmer et al., 1998; Tugwood et. al, 1996; Woodyatt et al., 1999). The potential carcinogenic risk of hypolipidemic therapy with fibrates, potent PPs, has been evaluated in two limited clinical trials with no evidence for carcinogenesis obtained. No relevant data are available on humans exposed to DEHP.

It has been suggested that the hepatocarcinogenic effects of PPs, such as DEHP, in experimental animals are rodent-specific and irrelevant for human. This position is held by a number of experts and is a defensible conclusion based on the available mechanistic data. However, the following arguments still indicate that a certain human cancer risk cannot, with certainty, be excluded:

1. The arguments for rodent-specificity of the liver tumours and the irrelevance of the experimental data for humans are based on the overall evidence available for all the PPs together. The weight of evidence available for each of the PPs, for example DEHP, is weaker.

2. The available data indicate a quantitative but not a qualitative, species variation in the expression of PPARα. Humans express PPARαin liver, albeit in levels lower than those found in rodents (Tugwood et al., 1996; Palmer et al., 1998). Therefor, a certain human cancer risk may still exist for some of the highly potent peroxisome proliferators. Also inter-individual differences in expression of human PPARαhave been demonstrated (Tugwood et al., 1996). This evidence supports the conclusion reported by Vanden Heuvel (1999) “Therefore, although PPs may pose little risk to the population as a whole, the potential human carcinogenicity of these chemicals cannot be summarily ignored. ”

3. DEHP has shown positive activity in several cell transformation assays and this effect is correlated with inhibition of gap junctional intercellular communication. It may be argued that these effects on cell transformation and intercellular communication by DEHP points at a different mechanism of carcinogenicity independent of PPARα(Dybing and Sanner, 1997; Mikalsen and Sanner, 1993; Tsutsui et al., 1993).

4. An association between non-peroxisomal effects of PPs and the carcinogenic process could exist. Possible changes in non-peroxisomal parameters (such as mitochondrial effects; regulation of cytochrome P-452, hormonal disturbances; and effects on cellular biology and ion homeostasis) in experimental animals and the relevance of such effects to humans have not been well studied (Youssef and Badr, 1998; Eagon et al., 1996).

Most recently, a Working Group of the “International Agency for Research on Cancer”(IARC) have concluded that the mechanism by which DEHP increases the incidence of liver tumours in rodents (activation of PPAR-á) is not relevant to humans. Therefore, and based on the overall evaluation of the available data, the DEHP-induced liver tumours in rats and mice will not be considered in the present Risk Assessment Report on DEHP.

Leydig cell (LC) tumours in rats

An increase in the incidence of testicular interstitial cell tumours (LC tumours) was observed in rats exposed to DEHP in a long-term study (Voss et al., 20005). In this study, 2,170 male Sprague-Dawley rats were exposed, lifelong, for the three non-genotoxic liver carcinogens DEHP, phenobarbital-sodium (PHB), and carbon tetrachloride (CCl4) to assess their hepatocarcinogenic effects alone or in combination. DEHP at dose levels of 30, 95 and 300 mg/kg, in the diet, did not cause liver tumours but induce LC tumours.

In a two-generation reproduction toxicity range-finding study in Wistar rats (Schilling et al., 1999), DEHP (approximately 0, 110, 339 and 1,060 mg/kg/d) was administered, in the diet, to groups of 10 males and 10 females sexually immature animals (F0 parental generation). Males and females from the same dose group were mated 70 days after the start of treatment. The females were allowed to litter and rear their pups (F1generation pups) until day 21 post partum. All male and female F1generation pups with the exception of one male and one female pup/litter (each first surviving pup/sex) were sacrificed on day 21 post partum.

The selected pups (F1generation pups) were reared for at least 10 weeks to become the F1generation parental animals. The male animals of the F1generation parental animals were killed after the mating period. In the 3-month-old F1males, a slight (grade 2) diffuse LC hyperplasia were observed in all six high dose animals. The authors considered the lesion to be treatment-related.

This DEHP-related effect is critical as the prenatal exposure period was relatively short, the onset of the lesions is early and the development of LC hyperplasia to LC tumours is possible if the 3-month-old animals had been allowed to age. This report support the results of Voss et al., (2005) showing induction of LC tumours in Sprague-Dawley rats exposed for DEHP.

The relevance for humans of rodent LC tumours has recently been evaluated in an international workshop (summarised in Clegg ED, Cook JC, Chapin RE, Foster PM and Daston GP (1997) Leydig cell hyperplasia and adenoma formation: mechanisms and relevance to humans. Reproductive Toxicology;11(1), 107-121) as well as in a published review (Cook JC, Klinefelter GR, Hardisty JF, Sharpe RM and Foster PMD (1999) Rodent Leydig cell tumorigenesis: a review of the physiology, pathology, mechanisms, and relevance to humans. Crit. Rev. Toxicol.29, 169-261). It was concluded that the pathways for regulation of the Hypothalamo-Pituitary-Testis (HPT) -axis in rats and humans are similar and hence, compounds that induce LCTs in rats by disruption of the HPT-axis pose a risk to human health with exception of two classes of compounds GnRH and dopamine agonists. Since it has been demonstrated that DEHP and other phthalates has a direct effect on the foetal testes the two latter mechanisms are not relevant for phthalates, and the induction of LC tumours in rats exposed for pthalates should be regarded as relevant to humans taking into consideration the species differences in sensitivity (Jones et al., 1993; Mylchreest et al., 1999; Foster, 1999).


The observed increase in the incidence of MCL in F344 rats is within the range of NTP’s historical control data. However, the concurrent study control groups remains most appropriate for comparisons, and the historical control data, if considered, must be from the test laboratory itself. Therefore, the increase in the incidence of MCL in male rats (Davis et al., 2000) may be DEHP-related, as the incidence were significantly increased compared to the study control and in addition to the historical control data from the same laboratory.

Additionally, it should be noted that increases in the incidence of MCL in F344 rats exposed to other phthalates, for example, diisononyl phthalate, diallyl phthalate, and butylbenzyl phthalate have been reported.

Whereas Ward and Reynolds (1983, Large granular lymphocyte leukemia, a heterogeneous lymphocyte leukemia in F-344 rats. Am. J. Pathol.111, 1-10) consider MCL in F344 rats as having similar pathology to an uncommon human tumour (large granular lymphocytic leukemia) and representing a unique model for study of natural tumour immunity, other experts regard MCL as F344 rats-specific, with little relevance for humans (Caldwell DJ (1999). Review of mononuclear cell leukemia (MNCL) in F-344 rat bioassays and its significance to human cancer risk: a case study using alkyl phthalates. Regul. Toxicol. Pharmacol.30(1), 45-53). Based on the available data the relevance for humans of the DEHP-induced MCL in F344 rats is not clear.

Justification for classification or non-classification

Based on the overall evaluation of the available data, no classification for carcinogenicity is warranted.