Registration Dossier

Data platform availability banner - registered substances factsheets

Please be aware that this old REACH registration data factsheet is no longer maintained; it remains frozen as of 19th May 2023.

The new ECHA CHEM database has been released by ECHA, and it now contains all REACH registration data. There are more details on the transition of ECHA's published data to ECHA CHEM here.

Diss Factsheets

Administrative data

Description of key information



A read-across category-approach is used for the assessment of the toxicological properties of selenium and selenium compounds. The following Se-substance are included in the category:

  • Se-metal (massive, powder)
  • Disodium selenate
  • Disodium selenite
  • Selenium dioxide / selenious acid
  • Zinc selenite
  • Barium selenite

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). In summary, the physico-chemical behavior of elemental selenium (once it has formed an ion-from its metal state), disodium selenite, disodium selenate and selenium dioxide/selenious acid is the same with regard to their metabolic fate. All selenium compounds (organic and inorganic, including elemental selenium), do share the very same metabolic fate in that after their resorption, reduction to the selenide moiety [Se2-], which is the single common precursor for its further metabolic conversion, takes place.

Therefore, there seems to be good evidence that different selenium moieties will behave very similar also for their ability to form reactive species which may play a decisive role in the generation of cytotoxicity followed likewise by unspecific and secondary clastogenicity and read-across can be made from the available data for disodium selenite. It is concluded that additional testing for each individual member of the proposed Se-category is not necessary and scientifically not meaningful.

In the case of inorganic salts like barium selenite and zinc selenite, uptake is always associated with a dissolution of the substance, i.e. dissociation into the metal cation (Zn2+, Ba2+) and the selenite anion (SeO32-). It can safely be assumed that the selenium/selenite moiety of barium/zinc selenite is generally of higher toxicological relevance than the zinc/barium cations. Therefore, the subsequent assessment of the toxicity of barium/zinc selenite focuses on the selenium moiety. As no in vivo toxicokinetic data or in vitro bioaccessibility data are available for a comparative assessment of relative bioavailability of various selenite substances, water solubility is adopted as a surrogate for bioavailability. Disodium selenite is readily soluble, with a water solubility of 800-900 g/L at 20°C. Barium selenite and zinc selenite, on the other hand, are poorly soluble salts (water solubility at 20°C of 66.7 mg/L and 16 mg/L, respectively, i.e. a difference of four/five orders of magnitude). Based on that, an intrinsically very conservative read-across from highly soluble forms to the poorly soluble barium/zinc selenite is proposed as the latter are assumed to have a lower solubility. It should also be noted that selenite anions in the tests with disodium selenite are formed under most physiological relevant conditions (i.e. neutral pH), thus facilitating unrestricted read-across between the various substances. In slightly acid conditions (pKa:8.32) the hydrogen selenite ion (HSeO3-) is formed whereas in more acidic conditions (pKa:2.62) the formation of selenious acid is observed (H2SeO3). Based on such existing equilibrium conditions, read-across between selenites, hydrogen selenites and selenious acid (solubility of 1670 g/L at 20°C) is justified.


Read-across from sodium selenite and selenious acid to barium/zinc selenite

Based on a comparison between toxicity reference values of zinc compounds and selenium compounds, it can safely be assumed that the selenium/selenite moiety of zinc selenite is generally of higher toxicological relevance than the zinc cations. Comparing the DNELs for the zinc/barium ion itself with the zinc/barium levels that are associated with the DNELs for barium/zinc selenite (based on selenite-data) indicated significantly higher values (in the range of factor 10 to 20) for the DNELs derived for the barium/zinc ion itself. Therefore, the subsequent assessment of the toxicity of barium/zinc selenite focuses on the selenium moiety.

Several reliable short-term repeated dose and sub-chronic studies are availble for the oral route:

  • Abdo (1994), NOAEL rat: 0.4 mg Se/kg bw/d (sub-chronic test with sodium selenite), NOAEL 0.4 mg Se/kg bw/d (sub-chronic test with sodium selenate)
  • Abdo (1994), NOAEL mouse: 0.9 mg Se/kg bw/d (sub-chronic test with sodium selenite), NOAEL 0.8 mg Se/kg bw/d (sub-chronic test with sodium selenate)
  • Bioulac (1992), NOAEL rat: 0.12 mg Se/kg bw/d (test with sodium selenite)
  • Johnson (2000), NOAEL mouse: 0.36 mg Se/kg bw/d (test with sodium selenite)

Based on these data, a NOAEL for rats of 0.4 mg Se/ kg bw/day has been selected as the key value for repeated dose toxicity via the oral route for the different Se-compounds within the current category. No studies on repeated dose toxicity via inhalation or dermal route are available.

Yang et al. (1989): NOAEL man: Se-intake of 850 µg Se/day per person; this figure is used as starting point for DNEL derivation.

It has to be emphasized, that the NOAEL according to Yang et al. (1989), which is used as starting point for DNEL derivation is based on human data. The existing studies on humans are considering a wealth of toxicological endpoints and overrule the available animal-based data by far.

Key value for chemical safety assessment

Additional information


The ATSDR Toxicological Profile on Selenium (2003), which is currently the most comprehensive review, was used as key source of relevant data on selenium compounds because it contains a detailed evaluation of toxicity data, performed by a renowned scientific body. More recent reviews, e. g. the work conducted for the Canadian Soil Quality Guidelines / Sudbury Soil Study, were also been screened for additional data. The underlying assumption is that the key literature considered by renowned international organisations such as ATSDR has usually already been subjected to a reliability assessment.


Nevertheless, all key references identified by ATSDR for Se-compounds were re-evaluated and re-assessed for use in the REACH dossiers according to Klimisch and with respect to the requirements for risk assessment. Studies which were assessed as not adequate, not relevant or unreliable by expert judgement during the screening procedure were assigned to "disregarded study", and rated as "not reliable" (RL=3), therefore data are not included as an endpoint study record.

In addition, after the evaluation of the ATSDR review and further secondary literature (as documented in IUCLID section 12), a follow-up endpoint-specific literature search was conducted in bibliographic databases restricted to references on repeated dose toxicity by inhalation or the dermal route, published since the beginning of 2000 (as documented in IUCLID section 12). The results of this search were also been screened for use in the REACH dossiers.


Evaluation of relevant references


Repeated dose toxicity, oral:

According to the evaluation criteria used by the experts of ATSDR (2003), a selection of studies on repeated dose oral toxicity with Se-compounds providing reliable, quantitative estimates of No-Observed-Adverse-Effect Levels (NOAELs) or Lowest-Observed-Adverse-Effect Levels (LOAELs) are available (reported in ATSDR, 2003). All these references were obtained and re-evaluated according to the criteria of REACH.


From the most reliable (RL=1) 90-day toxicity studies (Abdo, 1994: NTP technical report 38) with oral administration of sodium selenite via drinking water conducted in Fischer rats and B6C3F1 mice it can be concluded that the rat is more sensitive than the mouse. Based on mortality, body weight depression, decreased water consumption and renal papillary lesions in rats, the lowest estimated no-observed-adverse-effect level (NOAEL) was 0.4 mg selenium/kg body weight per day. For mice the NOAEL based on body weight depression and decreased water consumption can be established at 0.9 mg selenium/kg body weight per day.


Two additional subchronic/subacute toxicity studies are available with a special focus on hepatotoxicity in Sprague-Dawley rats (Bioulac 1992) and immunotoxicity in Balb/c mice (Johnson 2000). In the rat hepatotoxicity study, a somewhat lower NOAEL of 0.12 mg Se/kg bw/d was determined based on specific histopathological examinations of livers. However, this study was a dietary study and the difference of the resulting NOAEL to the key study NOAEL might result from the difference in routes of exposure and default calculation of dose levels based on amounts of test item in the diet. The NOAEL of 0.36 mg Se/kg bw/d obtained in the immunotoxicity study in mice with administration via drinking water is in the same range as the NOAEL for rats from the NTP study and somewhat lower than the NOAEL for mice from the NTP study (0.9 mg Se/kg bw/d).


For selenium a lot of human data is available. As laid down in the most recent publications of the German MAK commission (see chapter 5.11. of the CSR) the study of Yang et. al. (1989) revealed the most prominent effect of increased selenium levels in human serum. The results show that clinical signs (increased prothrombin time) may occur at a marginal blood-Se level of approximately ≥ 1 µg/mL or a corresponding marginal daily Se-intake of 850 µg/day.


Repeated dose toxicity, inhalation:

No data available.

For “Repeated dose toxicity, inhalation”, an endpoint-specific literature search was performed (see IUCLID section 12) since in the ATSDR (2003) publication only studies on elemental selenium, selenium dioxide, and various other Se-compounds which are not considered adequate for risk assessment of Se-compounds that are covered by the current category (e.g. selenium oxychloride, hydrogen selenide, dimethyl selenide). However, also from the literature search, no reliable studies on repeated dose toxicity of Se-compounds via the inhalation route were identified. Therefore, this endpoint needs to be covered by route-to-route extrapolation (e.g. using multiple path deposition modelling).



Repeated dose toxicity, dermal:

No data available.

For “Repeated dose toxicity, dermal” an endpoint specific literature search has been performed (see IUCLID section 12) since according to the ATSDR (2003) publication, no studies were available for evaluation that provide reliable, quantitative estimates of No-Observed-Adverse-Effect Levels (NOAELs) or Lowest-Observed-Adverse-Effect Levels (LOAELs).However, no reliable studies on repeated dose toxicity of Se-compounds via the dermal route were identified by this literature search.


According to the data requirements as outlined in section 8.6, column 2, Annexes VIII-IX, of Regulation (EC) 1907/2006 a repeated dose toxicity study shall be performed via the most appropriate route of administration, having regard to the likely route of human exposure. The dermal route is not considered as the most appropriate route of exposure for the various Se-compounds. Furthermore, the repeated dose dermal toxicity study shall be performed only if inhalation of the substance is unlikely, skin contact in production and/use is likely and the phys. -chem. and toxicological properties suggest potential for a significant rate of absorption through the skin.

Justification for classification or non-classification

In the current version of Regulation (EC) No 1272/2008 (3. ATP) selenium is classified as STOT RE 2; H373**: May cause damage to organs through prolonged or repeated exposure (labelling: GHS08: Health Hazard depending on endpoint). This classification is obviously based on the previously existing classification R33 (based on DSD), even though there is no directly corresponding classification to R33 in Regulation (EC) No 1272/2008.

Sodium selenite also has a harmonised classification in Annex VI of Regulation (EC) No 1272/2008 (CLP Regulation), but is not classified for this endpoint. 

No substance-specific harmonised classification exists for any of the other Se-compounds that are covered by the assessed Se-category. These Se-substances however, are formally intrinsically included in a group entry named “selenium compounds with the exception of cadmium sulphoselenide and those specified elsewhere in this Annex“ and is therefore subject to harmonized classification according to Regulation (EC) No 1272/2008. In the current version of Regulation (EC) No 1272/2008 (as amended by 4th ATP) this substance group is classified as STOT RE 2; H373**: May cause damage to organs through prolonged or repeated exposure (labelling: GHS08: Health Hazard depending on endpoint). This classification was determined to be of equivalent concern, but not directly equivalent to the classification end-point R33 as indicated in Directive 67/548/EEC. In a weight of evidence approach the following aspects should be considered:


(i) According to Regulation (EC) No 1272/2008 classification in Category 2 is applicable, when significant toxic effects observed in a 90-day repeated-dose study conducted in experimental animals are seen to occur within the guidance value ranges of 10 < effect level < 100 mg/kg bw/d. However, although for sodium selenite some single incidences of liver toxicity or renal findings were observed in studies with laboratory animals, these cannot be regarded as significant health effects for humans, because they are not consistently observed in various studies.


(ii) In addition, information and results of data presented in the peer-reviewed ATSDR (2003) publication is included in the weight of evidence approach. It is stated that in many aspects, similar patterns of toxicity have been reported for selenium in human and animal systems. “However, species-specific differences in toxicity are present (e.g., the main effect of selenium toxicity in rodents is damage to the liver, which is not observed in humans) and this may represent evidence of underlying differences in how selenium is metabolized. [...]The lack of evidence of liver damage in humans due to selenosis, despite the animal data to the contrary, suggests a problem with the animal models of the disease.“


According to the weight-of-evidence approach above, no classification would be appropriate for specific target organ toxicity of any of the Se-compounds within the assessed Se-category. As the harmonised classification of sodium selenite does not classify this substance for the Repeated Dose endpoint, no no classification is applied for this endpoint..