Registration Dossier

Diss Factsheets

Administrative data

Key value for chemical safety assessment

Effects on fertility

Description of key information

Grouping of soluble inorganic Se compounds

Inorganic selenium compounds all release Se ions upon dissolution in aqueous media. These dissolved Se ions can be considered a common transformation product across all inorganic Se substances (whether the transformation occurs via hydrolysis, dissociation or (bio)degradation), which will be responsible for any potential effect, rather than the initial non-transformed substance.

The released selenium ion can be present under various oxidation states and forms under normal physiological and environmental conditions. The speciation is determined by relevant physicochemical conditions of the media, and not by the source Se substance. Indeed, for assessment of toxicity to human health, studies on Na2SeO3and Na2SeO4indicate no consistent difference in toxicity among these soluble tetravalent or hexavalent inorganic Se substances. Because all selenium compounds do share the very same metabolic fate in the human body, in that after their resorption, reduction to the selenide moiety [Se2-] takes place, all data for soluble inorganic Se compounds are grouped for the assessment of reproductive and developmental effects of inorganic Se substances. Consequently, studies on different inorganic Se compounds are assessed in the following paragraphs, and the conclusions drawn are valid for all inorganic Se compounds.

A detailed rationale for the read-across hypothesis has been outlined in the read-across report that was generated according to the principles laid out in the Read-Across Assessment Framework (RAAF) and can be found in section 13.


Weight-of-evidence evaluation of reproductive toxicity studies on Se compounds

Highest weight must be given to a reproductive and developmental screening study which has been conducted under the National Toxicology Program (NTP) according to GLP and similar to OECD TG 421 (Wolfe, 1996). In this study, 10 male and 10 female Sprague Dawley rats received 0, 7.5, 15, or 30 mg sodium selenate/L drinking water (males: 0, 0.4, 0.7 or 1 mg sodium selenate/kg bw/d; corresponds to 0, 0.18, 0.3 or 0.46 mg selenium/kg bw/d; female animals: 0, 0.5, 0.8 or 1.1 mg sodium selenate/kg bw/d; corresponds to 0, 0.23, 0.37 or 0.5 mg selenium/kg bw/d). Male animals were treated for a total of 26 days, starting 6 days before mating. Group A females were treated before, during, and after mating to test the ability of the exposure to interfere with ovulation, mating, fertilisation, and/or implantation, Group B females were exposed only during gestational development and not mated Group C females were treated over 30 days, providing data on the ability of selenate to alter oestrous cycle parameters (female vaginal cytology). In this study, 15 ppm and 30 ppm of sodium selenate in the drinking water induced systemic toxicity by reducing feed and water consumption in male and female Sprague-Dawley rats, resulting in body weight loss and reduced body weight gain. With regard to reproduction, sodium selenate did not appear to adversely affect male reproductive function and no statistically significant effect on sperm parameters was observed. In females, the 30 ppm rats were drinking only approximately 20% of control intake, had lost weight and had clinical signs of dehydration. Only at this concentration, reproductive toxicity was noted in both Group A and B animals based on increased gestation length and decreased number of live pups, live total pup weight, proportion of pups born alive, number of implants and corpora lutea per litter, and pup survival. In Group C females, an increase in oestrous cycle length was observed at 30 ppm only. The study authors concluded that the observed effects might be secondary to dehydration. In addition, survival was adversely affected at this test concentration of 30 ppm; 1/10 males, 1/10 Group A females, 7/12 Group B females, and 2/10 Group C females were either found dead or killed moribund during the study. Mortality was primarily observed in the 30 ppm Group B females just prior to or during littering. In conclusion, this well-conducted and highly reliable short-term reproductive study of the effects of sodium selenate in drinking water on rats reported some female reproductive toxicity but only at a concentration that produced signs of severe maternal toxicity, including a large reduction in water consumption, dehydration, reduced body weight and even mortality. Doses that produced signs of systemic toxicity did not cause any increase in sperm abnormalities or lesions of the testis or epididymis. In this study, sodium selenate was not a selective reproductive toxicant, as it caused severe maternal toxicity at the test concentration at which some reproductive effects were observed, and induced body weight decreased at doses even below those that affected reproduction.

These findings are supported by a repeated dose toxicity study in rats and mice (Abdo, 1994). In more detail, the sub-chronic oral toxicity of sodium selenate and sodium selenite was assessed in mice and rats in a GLP study similar to OECD Test Guideline 408 by continuous treatment of test animals via drinking water for 13 weeks. Mice were continuously treated with 2, 4, 8, 16 and 32 ppm sodium selenite in drinking water (corresponding to 0.14, 0.3, 0.5, 0.9 or 1.6 mg Se/kg bw/d), or with 3.75, 7.5, 15, 30 and 60 ppm sodium selenate in drinking water (corresponding to 0.3, 0.5, 0.8, 1.5, and 2.6 mg Se/kg bw/d). Neither sodium selenite nor sodium selenate affected spermatid or epididymal spermatozoal measurements in treated males. Sodium selenite caused significant increases in oestrus cycle lengths in female mice of the highest treatment group (32 ppm; 1.6 mg Se/kg bw/d). This test concentration also impacted on body weight, however body weight was depressed by less than 10%. In contrast, sodium selenate did not cause significant effects on oestrus cycle in female mice upon exposure up to 60 ppm (2.6 mg Se/kg bw/d). Rats were continuously treated with concentrations of 2, 4, 8, 16 and 32 ppm sodium selenite in drinking water (corresponding to 0.08, 0.13, 0.2, 0.4 or 0.8/0.9 mg Se/kg bw/d), or with 3.75, 7.5, 15, 30, and 60 ppm sodium selenate in drinking water (corresponding to 0.1, 0.2, 0.4, 0.6, 1.1 (males), and 0.8 (females) mg Se/kg bw/d). In male rats, treatment with sodium selenite caused a significant decrease in epididymal sperm concentrations in all groups evaluated (4, 8 and 16 ppm), however without dose-relation. No effects were observed on other male spermatozoal parameters in rats. Because there were also no effects on spermatid and epididymal spermatozoal measurements in male mice, this finding was considered not to be treatment related. Treatment with sodium selenate led to a significant decrease in epididymal spermatozoal motility in the 30 ppm group. In addition, significant effects were observed on the number of testicular spermatid heads and the spermatid count per gram of testis in the 3.75 and 15 ppm groups, but not in the 30 ppm group, indicating no dose-dependence of this effect as well. In females, increases in oestrus cycle lengths were observed in rats treated with 16 ppm sodium selenite (32 ppm not evaluated), and in all evaluated females treated with sodium selenate (3.75, 15 and 30 ppm). Similar to the findings in mice, this test concentration also impacted on body weight, however body weight was depressed by less than 10%. In female rats, other findings at high exposure concentrations were regarded as confounded by dehydrated and unthrifty conditions of the animals. In sum, no significant treatment-related effect on sperm parameters was found in male rats and mice, and increases in oestrus cycle length were found in mice exposed to sodium selenite but not selenate, and in rats exposed to sodium selenite and sodium selenate. These increases in oestrus cycle length were accompanied by slightly depressed body weights of females.

Absence of specific adverse effects of Se on sperm parameters is further supported by a study in male rats (Zhou, 2017). Here, weanling male rats were fed a selenium-deficient basal diet (<0.01 mg Se/kg) for 5 weeks to deplete their selenium stores and to adjust their baseline selenium status. Groups of 12 rats were treated with sodium selenite at concentrations of 0.25 (adequate), 3 or 5 (excess) mg Se/kg in diet (corresponding to 0.03, 0.35 and 0.6 mg Se/kg bw/d), while one group was continuously fed basal diet (selenium-deficient) for 4 weeks. In rats fed 5 mg Se/kg the rate of body weight gain tended to be lower compared with those fed adequate Se level. There were no differences in testes weights or testis index scores between the 4 groups. At 3 mg Se/kg no significant changes were observed. Sodium selenite caused increased sperm deformity and reduced sperm density in rats fed 5 mg/kg compared to rats fed adequate Se and 3 mg Se/kg. No further information on the overall condition of rats fed the higher excess level of 5 mg Se/kg are provided besides decreased body weight gain.

Two studies have been identified which report the effect of Se on male reproductive organs at excess levels of Se.

In a study published by Kaur (2000), groups of 12 adult male albino rats were fed 6 and 8 ppm sodium selenite in diet for 6 and 9 weeks (corresponding to 0.2 and 0.28 mg Se/kg bw/d). Each male consuming 6 ppm sodium selenite was mated with two untreated females; their offspring were allowed to mature up to 12 weeks of age. Both test concentrations of sodium selenite caused a striking body weight loss, reaching a loss of body weight in the range of ca. -20 to -30 g after 9 weeks compared to day 0, whereas control animals gained ca. 60 g of weight throughout the study period. Furthermore, the rats receiving diets containing 6 and 8 ppm sodium selenite showed clear signs of systemic toxicity as evident by loss of body fur at certain points and reddening of claws starting after 6 weeks of dietary Se treatment which became more prominent as the duration of feeding was prolonged to 9 weeks. In addition, loss of vision in one animal was reported without specifying at which test concentration. Under these conditions, the relative testicular weight was concentration-dependently reduced after 9 weeks but not 6 weeks, whereas the relative weight of cauda epididymis was reduced at both time points. Furthermore,pathomorphological changes in testis and an increase inmorphologically abnormal spermatozoaare reported. In addition, the offspring of treated males showed lower body weight gain as compared to offspring from control animals. The results provide an indication that the general health of the treated rats was crucially impaired at the selected test concentrations of Se. Thus, the reported effects on male reproductive parameters in this study are clearly observed in the context of severe generalised toxicity caused by Se concentrations above systemically tolerable dose levels.

Another study addressed the influence of selenium nanoparticles (SeNPs) on the reproductive performance of male Sprague-Dawley rats (Liu, 2017). A suspension of SeNPs was administered by oral gavage for 2 weeks at 0, 0.2, 0.4, 0.8, 2, 4, or 8 mg Se/kg bw/d to groups of 10 male Sprague-Dawley rats. During the study none of the rats died, however the animals treated with 2 mg Se/kg bw/d and above showed symptoms of poisoning including mental fatigue, decreased mobility and body weight. No further parameters related to systemic toxicity were evaluated. In the 4 and 8 mg Se/kg bw/d groups, reduced testes wet weight, testes atrophy and histopathological changes in testes and epididymis were observed, as well as lower sperm concentration, motility and movement parameters. In contrast, animals treated with 2 mg Se/kg bw/d had no significant effect on sperm concentration and sperm vitality, while some sperm movement parameters were significantly increased. In addition, the level of serum testosterone was slightly but statistically significantly increased at 8.0 mg Se/kg bw only. In sum, lower doses (0.2, 0.4 and 0.8 mg Se/kg bw/d) were found to have a beneficial effect on male reproductive parameters, while dose levels higher than 4 mg Se/kg bw/d showed a negative effect. As signs of generalised poisoning are shown at concentrations higher than 2 mg Se/kg bw/d, these effects are only observed in the context of systemic toxicity. Considering the multitude of physiological roles of Se and the variety of potential targets for toxicity at excess dose levels, this study also does not provide evidence for a selective adverse effect of Se on sperm parameters or male fertility.

A potential effect of Se on oestrous cyclicity and ovulation in rats was also studied by Parshad (1999). Normal cycling female Wistar rats were exposed to sodium selenite, via intraperitoneal injected at a concentration of 2 or 4 mg/kg bw daily for 30 days (corresponding to 0.9 and 1.8 mg Se/kg bw), by single injection either during proestrous or oestrous, or by four intraperitoneal doses of sodium selenite at metoestrous, dioestrus, proestrous and oestrous stages. Oestrus cyclicity, ovarian follicles, ovulation, implantation and pregnancy outcome were determined on day 14 of gestation. Unfortunately, both selected dose levels caused severe systemic toxicity, as administration of 2 or 4 mg/kg bw sodium selenite for 30 days caused increased mortality of animals of 13.6% and 40%, respectively, occurring mainly after day 21 of treatment. Moreover, treatment with 4 doses of sodium selenite during oestrous cycle of rats evaluated on day 14 of gestation showed a mortality after mating of 12 % (low dose) and 28 % (high dose).

Continuous treatment with sodium selenite had no effect on the duration of first two oestrous cycles but thereafter the rats remained at the dioestrus stage. Examination of the ovaries from treated rats on day 31 showed cystic follicles in 21 % (low dose) and 60 % (high dose). The ovaries of rats, where cysts were absent, showed no signs of corpora lutea. Single selenite treatment during the oestrous cycle preceding mating affected the implantation and pregnancy outcome in a dose-related manner. A single dose of 2 mg/kg bw administered either at proestrous or oestrous had no effect on different reproductive parameters investigated in this study, however the 4 doses during the oestrous cycle stages reduced the number of corpora lutea and implantations as compared to saline injected control female rats. Similar effects of a single dose of sodium selenite (4 mg/kg body weight) when injected at proestrous were recorded. The high dose of sodium selenite administered at oestrous or throughout the cycle decreased the number of implantations, but in addition, also increased the resorption rate/litter on day 14 of gestation. However, these changes occurred in the context of significant and very severe systemic toxicity evident by clearly increased mortality up to 40%, so these observations do not provide a robust indication for a specific effect of Se on female reproductive parameters, especially considering the fact that the oestrous cycle is tightly controlled by hormones which would be relevantly affected at much more subtle toxicity.

As mentioned above, changes in oestrus cycle length may in general be indicative for effects on the hormonal system. However, a recent epidemiological study found no significant association between maternal and peripubertal Se exposure and reproductive hormones nor sexual maturation and progression in girls (Ashrap et al., 2019).

The results of the above-mentioned NTP study are further supported by an earlier 4-generation study published by Schroeder, 1971. In this study, five pairs of weanling Long- Evans mice were exposed to sodium selenate via drinking water at a concentration of 3 ppm (corresponding to ca. 0.1 mg Se/kg bw/d) and allowed to breed over a period of 6 months. At this dose level, no significant difference in number of litters, the average pair age at first litter, average litter size and average interval between litters was observed between exposed mice and control animals in the F1 and F2 generation. In addition, effects were noted with regard to the survival and presence of runts (i.e. animals with large heads and small bodies) in the litters starting in the F1 generation. Also, in the F1 generation one dead litter was noted for the exposed animals and 13 young deaths were observed. In the F2 generation, there were two failures to breed in exposed mice but none in the control group. In the fourth generation (F3) there were only three litters (23 in control group) showing reduced litter size with a total of 23 mice, of which 16 were runts.

Unfortunately, only one dose level was tested in this study and no information regarding (confounding) adverse effects to the maternal animals is given. However, in a following study from the same working group (Schroeder, 1972), effects of selenium after life-span exposure of mice are reported. Although mice continuously exposed to the same test concentration of 3 ppm selenate via drinking water (corresponding to ca. 0.1 mg Se/kg bw/d) showed no significant change in body weight (except for an increase in male body weight at the age of 360 days), selenium-fed mice were reported to be less active than their controls, did not appear healthy and had poor coats. In addition 40 % of the selenate treated mice showed oedema of the extremities shortly before or at death. Based on these systemic effects described in adult animals it can be concluded that the effects observed during the treatment of 4 successive generations were confounded by parental/systemic toxicity.

While these studies demonstrate effects of Se on reproductive parameters in the presence of systemic/maternal toxicity, another study conducted by Nobunaga (1979) found no significant adverse effect of oral exposure to 3 and 6 ppm sodium selenite in drinking water (corresponding to 0.34 and 0.68 mg Se/kg bw/d) sodium selenite for 30 days before gestation and following mating until day 18 of gestation, neither on parental animals nor on reproduction or embryo/foetal development. In more detail, no significant effect was found on the number of implants, dead embryos or foetuses, resorptions and implantation sites as well as surviving foetuses and litter size. The mean body weight of surviving foetuses on Day 18 of gestation was slightly, but statistically significantly, reduced in the 6 ppm group. However, no increased incidence in gross malformations or skeletal anomalies indicating retarded foetal development could be identified. This further supports the conclusion that adverse effects observed on reproductive parameters or prenatal development are linked to significant parental systemic toxicity.

Beside these, further animal studies addressing some reproductive parameters were found in the open literature. However, these studies are considered as not adequate for the evaluation of reproductive toxicity of Se due to several methodological deficiencies as described in detail in a summarising robust study summary in section 7.8.1, or due to the complete lack of reporting of any parental observations which hampers the evaluation of the significance and relevance of reported findings. Thus, these studies are not included in this weight-of-evidence evaluation.

Effect on fertility: via oral route
Endpoint conclusion:
no adverse effect observed
Effect on fertility: via inhalation route
Endpoint conclusion:
no adverse effect observed
Effect on fertility: via dermal route
Endpoint conclusion:
no adverse effect observed

Effects on developmental toxicity

Description of key information

Weight-of-evidence evaluation of developmental toxicity studies on Se compounds

Beside some of the above-mentioned studies, further studies in rats and mice are available which focussed on the potential developmental toxicity of sodium selenite or sodium selenate.

In a prenatal developmental toxicity study published by Hardin (1987), pregnant CD-1 mice were treated with sodium selenite at a concentration of 3.5, 5, 7 or 14 mg/kg bw/d (corresponding to 1.6, 2.3, 3.2 or 6.4 mg Se/kg bw/d) via oral gavage on gestation day (GD) 6-13 and were then allowed to deliver litters. Litter size, birth weight, and neonatal growth and survival to postnatal day 3 were recorded as indices of potential developmental toxicity. Sodium selenite had no apparent adverse effect on the maternal or developmental endpoints at doses of 7 mg/kg bw/d or less. At a dose of 14 mg/kg bw/d, sodium selenite caused 44% maternal mortality and significantly reduced maternal body weight gain of surviving animals. This demonstrates the very narrow range between tolerable and severely adverse dose levels. At the significantly toxic dose level of 14 mg/kg bw/d, sodium selenite also adversely affected the number of viable litters, the survival of the pups and the birth weight and postnatal weight gain. In conclusion, this study again shows the absence of embryo-/ foetotoxicity of sodium selenite at maternal nontoxic dose levels.

This conclusion is further supported by another prenatal developmental toxicity study conducted in rats (Juszkiewic, 1993). In this study, pregnant rats were exposed to sodium selenite at single doses of 0.5, 1 or 2 mg Se/kg bw via subcutaneous injection on gestation day (GD) 9 or daily doses of 0.5 mg Se/kg bw/d on GD 6-15. Dams and foetuses were examined before term (on GD 21). No developmental delayed foetuses and no evidence for teratogenic malformations were found in any of the treatment groups, furthermore Se did not cause signs of significant maternal toxicity in any treatment group. This provides further evidence that sodium selenite has no embryo-/foetotoxic properties in the absence of significant generalised toxicity.

Prenatal developmental toxicity of Se was also studied by Helal (2010). In this study, pregnant rats were exposed to sodium selenite at dose levels of 0, 5 or 10 µg Se/kg bw/d via oral gavage from gestation day (GD) 7 to 19. Animals were euthanised on GD 20 and parameters of maternal toxicity and foetal development were evaluated. Pregnant rats orally treated with 5 and 10 μg/kg bw/d revealed signs of toxicity as illness, lethargy, weakness, and diarrhoea during the exposure period; furthermore, the maternal body weight was significantly decreased at the low and high dose of sodium selenite. Maternal mortality was observed in the high dose group. In addition, sodium selenite treatment revealed significant reduction of placental and liver weights in treated dams. This was accompanied by further changes in the liver (increase in DNA fragmentation, marked reduction of hepatic DNA content, and several histopathological changes). At both exposure levels, mean foetal weights and lengths were significantly lower on GD 20, moreover the foetal skeleton showed signs of developmental delay in skull and limbs of the treated groups, clearly indicating delayed foetal developmental at these dose levels. Furthermore, foetal deaths were observed at the high dose of selenite. At both dose levels, the mean number of observed foetuses was significantly decreased and the number of early complete resorbed foetuses as well as the mean number of post implantation loss was increased. In line with the above-mentioned studies, the results demonstrate that the dosages chosen led to abnormal foetal development in the presence of maternal toxicity.

The effect of Se on prenatal development was also studied by Yonemoto, 1984. Pregnant CD-1 mice received a single injection i.v. of sodium selenite or selenodiglutathione on day 12 of gestation at a concentration of 16.4, 20.5, 25.6 or 32.0 µmol/kg (corresponding to 1.3, 1.6, 2.0 or 2.5 mg Se/kg bw), while the control group received saline only. Unfortunately, the selected test concentrations showed significant generalised toxicity, as a dose-dependent increase of maternal mortality was observed starting already at the second-lowest test concentration of 20.5 µmol/kg (12.5% and 22.2 % mortality after exposure to sodium selenite and selenodiglutathione), yielding 80 % and 100 % dead dams in the highest dose group after sodium selenite and selenodiglutathione treatment, respectively. Maternal deaths occurred rapidly after treatment, during GD 13 to 16. All surviving dams littered on GD 19 and no cases of abortion were observed. However, three surviving dams gave stillbirths (1/5 surviving dams treated with 25.6 µmol/kg sodium selenite, 1/7 and 1/1 surviving dams treated with 20.5 or 25.6 µmol/kg selenodiglutathione, respectively), all of which did not show an increase of body weight after injection of treatment solution indicating a lethal effect of these highly maternally toxic concentrations on the foetuses as well. The number of implantations and number of live births per dam were not different between the treated groups, while the body weight of offspring was slightly but statistically significantly lower in all treated groups, albeit the effect was not dose-dependent. Conclusively, in this study developmental toxicity was only observed at highly toxic doses causing high incidences of maternal mortality.

The impact of Se on early postnatal development with focus of cataract effects was studied in rats (Ostadalova and Babicky, 1980). On postnatal day 10 groups of 30-32 juvenile male Wistar rats were treated with a single subcutaneous injection of sodium selenate at a concentration of 0, 10, 20, 40, 60, 80, or 100 µmol/kg bw (corresponding to approx. 0.8, 1.6, 3.2, 4.7, 6.3 or 7.9 mg Se/kg bw). Mortality was monitored, and the occurrence of cataract was registered daily after the eyes opened (postnatal day 14-16). Animals were weaned at the age of 30 days and henceforth fed with a standard laboratory diet and water ad libitum. The experiment was terminated on postnatal day 60. At 60 µmol/kg bw, spontaneous death of treated rats was first observed, and the percentage of mortality increased with higher dose levels. A dose on 100 µmol/kg bw had a 100% lethal effect within 24 h after treatment. Other selenium compounds were also tested and also showed a dose-dependent increase of mortality. Sodium selenate induced cataract starting at doses of 20 µmol/kg bw. Similarly, cataract was also induced by the administration of selenomethionine and selenocystine at doses lower than the lethal doses, while no cataract effects were noted for dimethyl selenide and trimethylselenonium ion. Beside mortality, no other observations are reported. However, considering the differences in pre-/postnatal eye development between rodents and humans as well as considering the much higher internal exposure level after postnatal injection compared to indirect interuterine exposure, the results may not per se be indicative for comparable effects in humans. 

A series of further animal studies addressing developmental toxicity of selenium compounds have been published. However, due to various methodological and/or reporting deficiencies (e.g. only one dose level tested, no data on maternal findings provided, etc.), these do not provide sufficient information for a robust evaluation of the results and are thus not considered in this weight-of-evidence evaluation. However, these are mentioned in a summarising robust study summary in section 7.8.2 for completeness.

Of course, much less information on developmental effects in humans is available. A recent epidemiological study with 888 pregnant women reports an association between increased maternal Se level and congenital heart defects in offspring, however differences in maternal Se level between groups were marginal with even identical mean hair Se levels in both case and control group (Guo et al., 2019). In contrast, a case report on a woman who accidentally received a very high level of 200 mg Se/day from gestation week 7 to 12 experienced clear clinical signs of selenosis such as nausea, vomiting, hand and foot paresthesia, fatigue as well as loss of fingernails and hair; however no complications during pregnancy occurred and her child was born at term with no congenital defects (D’Oria et al., 2018). Several studies addressed the question of potential neurodevelopmental toxicity of excess levels of Se in humans, however inconsistent results are reported in the open literature. For example, a U-shaped association between maternal Se status and neuropsychological development in children (infants and preschool children) was identified after evaluation of 651 or 490 mother-child pairs, respectively (Amoros et al., 2018 a, b), whereas Oken et al. (2016) found no association between maternal Se uptake during pregnancy and mid-childhood cognition at the age of age 7.9 in 872 mother-child pairs. A recent study in 484 newborns reports an association between placental Se and muscle tone in newborns (Tian et al., 2020). With regard to potential developmental immunotoxicity, a multicentre prospective cohort study with 410 mother-child pairs found no significant association between maternal Se status and atopic dermatitis, food allergy or asthma in 1st or 2nd year of life (Podlecka et al., 2017). These human studies indicate that variations in internal Se level can affect developmental parameters which may not be unexpected considering the multiple physiological roles of Se compounds; however no clear indications for a significant adverse effect could be identified even after unintended repeated exposure to an Se level well above the tolerable dose level.

Effect on developmental toxicity: via oral route
Endpoint conclusion:
no adverse effect observed
Effect on developmental toxicity: via inhalation route
Endpoint conclusion:
no adverse effect observed
Effect on developmental toxicity: via dermal route
Endpoint conclusion:
adverse effect observed

Toxicity to reproduction: other studies

Description of key information

Please refer to description of key information above.

Justification for classification or non-classification

General introduction and objective of this assessment

Selenium is an essential micro-nutrient that is used in the body in enzymes that protect the body from oxidative damage, or in enzymes related to growth and metabolism.

The majority of the daily intake of selenium occurs via oral uptake (food, and to a lesser extent, via drinking water), but can also be taken up via inhalation. Organic selenium compounds in food (e.g. selenomethionine and selencysteine in grains, cereals, and forage crops) are easily absorbed by the human body, and selenium is made available where needed. Inorganic selenium compounds such as sodium selenate and sodium selenite are the main speciation forms that are found in drinking water, and are also absorbed from the digestive tract. These inorganic forms are then transformed into (organic) forms that can be used by the human body. Upon absorption, selenomethionine and selenocysteine play multiple roles in physiological processes both at cellular and organismal level, and undergo different metabolic pathways in the body. The biological effects are mainly carried out by selenoproteins. For example, various selenoproteins are known in humans with functional relevance in skeletal and cardiac muscle metabolism, T-cell mediated immunity, selenium homeostasis and transport, and many other processes. Being an essential element, the human body has several mechanisms that regulate the internal Se-concentration levels (uptake and elimination mechanisms). However, if exposure levels are very high, or if exposure occurs over a long period of time, selenium build-up in the body can take place (in liver, kidney, blood, lungs, heart, testes, nails and hair), resulting in adverse dermal and neurological effects referred to as selenosis. Considering the multiple physiological roles of selenoproteins, it is highly plausible that Se deficiency also results in adverse health effects (muscle pain, heart problems, lung effects in premature babies). Two endemic diseases in selenium-poor regions of China are related to Se deficiency: Keshan Disease and Kashin-Beck Disease.

In light of the toxicity of Se compounds at excess levels and deficiency, this weight-of-evidence evaluation assessed whether Se compounds specifically cause adverse effects on reproductive or developmental parameters and whether classification and labelling of Se compounds as reproductive toxicant according to Regulation (EC) No 1272/2008 (CLP Regulation) is warranted.

General toxicity of Se compounds at excess levels and deficiency

Evidence from animal studies as well as human data clearly indicate adverse systemic toxicity upon exposure to excess Se levels, the clinical signs being well-described as generalised effects in humans including symptoms such as weight loss, fatigue, anaemia, irritated skin, mucous membranes and eyes, as well as irritation in the pharyngeal, bronchial, and gastrointestinal tracts with a garlic odour on breath and in sweat. Furthermore, human selenosis includes loss of hair and nails, skin lesions, and nervous system abnormalities such as polyneuritis. In animals, excess levels of Se cause effects such as rough hair coat/loss of hair, malformed hooves, nervous system abnormalities (impaired vision and paralysis), lesions in liver, heart and kidneys, up to severe systemic toxicity resulting in mortality.

Considering the adverse effects occurring at deficiency as well as excess levels of Se, the dose-response relationship for Se compounds can be best described as a U-shaped curve. The broad database of available studies on Se compounds reveals a very narrow dose range for an adequate plateau of no adverse effect, indicating optimal homeostasis. However, this optimal dose range clearly varies not only between species but also depending on further conditions such as e.g. background plasma selenium levels influenced by dietary exposure among others (ATDSR, 2003; MAK, 2011). Nevertheless, it is clear that only adverse effects caused by exposure to Se exceeding tolerable levels are relevant for classification and labelling according to CLP Regulation.

With regard to reproductive and developmental toxicity, it is of utmost importance to carefully analyse the reported findings at excess Se levels considering its potential to cause severe (parental) systemic toxicity. In general, it is well-known that maternal systemic toxicity impacts on pre- and postnatal development. For instance, a decrease in body weight gain, or even weight loss in dams during gestation is linked to delayed foetal development leading to decreased foetal body weights, retarded foetal growth and delayed ossification, presumably caused via malnutrition (for further reading please be referred to Tyl, 2012; Fleeman et al., 2005; Chernoff et al., 2008).

This well-known link between parental systemic toxicity and reproductive capacity / development is relevant mainly for the decision on classification based on data from animal studies and is also taken into account in the CLP criteria for classification and labelling of a substance as reproductive toxicant.


Criteria for classification as reproductive toxicant according to CLP Regulation

The criteria for classification as Repr. Category 1B and Category 2 read as follows:

Cat. 1B: “The classification of a substance in Category 1B is largely based on data from animal studies. Such data shall provide clear evidence of an adverse effect on sexual function and fertility or on development in the absence of other toxic effects, or if occurring together with other toxic effects the adverse effect on reproduction is considered not to be a secondary non-specific consequence of other toxic effects. However, when there is mechanistic information that raises doubt about the relevance of the effect for humans, classification in Category 2 may be more appropriate.”

Cat. 2: “Substances are classified in Category 2 for reproductive toxicity when there is some evidence from humans or experimental animals, possibly supplemented with other information, of an adverse effect on sexual function and fertility, or on development, and where the evidence is not sufficiently convincing to place the substance in Category 1. If deficiencies in the study make the quality of evidence less convincing, Category 2 could be the more appropriate classification.Such effects shall have been observed in the absence of other toxic effects, or if occurring together with other toxic effects the adverse effect on reproduction is considered not to be a secondary non-specific consequence of the other toxic effects.”


Beside data gained from reproductive or developmental toxicity studies, adverse effects observed in repeated dose toxicity studies relevant for the reproductive function can also contribute to the classification of a substance as reproductive toxicant. However, also in these cases the CLP Regulation specifies that these effects are only relevant if observed “in the absence of significant generalised toxicity” (“ Adverse effects or changes, seen in short- or long-term repeated dose toxicity studies, which are judged likely to impair reproductive function and which occur in the absence of significant generalised toxicity, may be used as a basis for classification, e.g. histopathological changes in the gonads.”).


Set of studies on Se compounds considered for assessment of reproductive toxicity of Se

In line with CLP criteria as described above, all relevant findings from the reproductive, developmental and repeated dose toxicity studies are evaluated in the following using a weight-of-evidence approach, also considering the presence or absence of systemic (parental) toxicity. This is of utmost importance especially for substances with the potential for the induction of severe systemic toxicity such as Se compounds. In line with this, the Guidance on the Application of the CLP criteria specifies that “Adverse effects on fertility and reproductive performance seen only at dose levels causing marked systemic toxicity (e.g. lethality, dramatic reduction in absolute body weight, coma) are not relevant for classification purposes.” (ECHA, 2017).

A very broad range of studies on Se have been conducted demonstrating the absence of adversity at low supplementary levels, or even beneficial effects of such low levels on parental animals or offspring (e.g. Laureano-Melo, 2015). Furthermore, a series of in vitro studies have been conducted to address mechanistic aspects underlying the multiple physiological roles of Se in various processes such as spermatogenesis (e.g. Shi et al., 2017). As these do not provide relevant information regarding adversity needed for deciding on classification and labelling of Se according to CLP criteria, this kind of studies was not considered in this evaluation. Nevertheless, a series of animal studies addressing potential reproductive toxicity and/or pre-/postnatal developmental toxicity of Se at levels above the tolerable dose range is available. These studies have been assessed regarding their reliability and the usefulness of the results for the evaluation of these endpoints. The relevant findings from all studies considered as sufficiently reliable are evaluated within a comprehensive weight-of-evidence approach (see above in "description of key information" for fertility and developmental toxicity, respectively).

Overall conclusion

Reliable reproductive toxicity studies as described above in detail demonstrate effects of Se on reproductive parameters in the presence of systemic/maternal toxicity, whereas no relevant adverse effects on reproductive parameters were identified at lower dose levels in the absence of general toxicity. Thus, the available studies clearly indicate that adverse effects observed on reproductive parameters or prenatal development are linked to significant parental systemic toxicity. In line with this evaluation of the available reproductive toxicity studies, the German MAK Commission, after detailed assessment (MAK, 2011), summarised the available data set as follows (MAK, 2014): “In various studies there was no significant influence of selenium on various sperm parameters, such as sperm concentration, number and motility, sperm volume and sperm morphology or on the activities of the reproductive hormones testosterone, follicle-stimulating hormone, luteinising hormone, prolactin, oestradiol and progesterone.”

With regard to potential developmental toxicity of Se compounds, results of reliable studies as described above in detail demonstrate that developmental effects observed in animal studies consistently occurred at dose levels causing severe maternal toxicity, and human data do not indicate an increased sensitivity of humans compared to rodents with regard to developmental toxicity. This conclusion is clearly supported by the robust evaluation of a comprehensive set of animal studies on potential developmental toxicity of selenium compounds by the German MAK Commission (MAK, 2011). In brief, they concluded that “only maternally toxic dose levels showed teratogenic effects” (MAK, 2011).

Taken together, these data clearly indicate that Se is not a selective reproductive nor developmental toxicant. Thus, no classification as toxic to reproduction according to CLP Regulation is warranted.


ATSDR (Agency for Toxic Substances and Disease Registry). Toxicological Profile on Selenium. 2003.

Chernoff N, Rogers EH, Gage MI, Francis BM. The relationship of maternal and fetal toxicity in developmental toxicology bioassays with notes on the biological significance of the "no observed adverse effect level". Reprod Toxicol. 2008;25(2):192-202. doi: 10.1016/j.reprotox.2007.12.001.

ECHA. Guidance on the Application of the CLP Criteria. Guidance to Regulation (EC) No 1272/2008 on classification, labelling and packaging (CLP) of substances and mixtures. Version 5.0, 2017.

Fleeman TL, Cappon GD, Chapin RE, Hurtt ME. The effects of feed restriction during organogenesis on embryo-fetal development in the rat. Birth Defects Res B Dev Reprod Toxicol. 2005;74(5):442-9. doi: 10.1002/bdrb.20056.

MAK (German Commission on Occupational Exposure Levels). Selenium. 2011.

MAK (German Commission on Occupational Exposure Levels). Selenium. 2014.

Tyl RW. Commentary on the role of maternal toxicity on developmental toxicity. Birth Defects Res B Dev Reprod Toxicol. 2012;95(3):262-6. doi: 10.1002/bdrb.21015.

Additional information