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Please be aware that this old REACH registration data factsheet is no longer maintained; it remains frozen as of 19th May 2023.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

Ecotoxicological information

Endpoint summary

Currently viewing:

Administrative data

Description of key information

Additional information


1)     Chronic data – establishing the dataset

A significant amount of data is available on Cd toxicity to soil or litter microflora, soil fauna and higher plants in the EU risk assessment (RA; ECB 2007). The quality and relevancy of those data have been reviewed in detail during the EU risk assessment process. Reliability indices 1, 2 and 3 (RI 1, RI 2 and RI 3) data were used in the PNEC derivation, while reliability 4 data were excluded. Because the RI 1 and 2 data group has limited number of species, the RA has proposed to include the reliability 3 data too as basis for deriving the PNEC. This is mainly due toplant data that are excluded from the group RI 1-2, whereas plants seem to be the most sensitive group.For the present analysis, the same approach was followed, to be conform with the EU RA. Moreover, studies assigned RI 3 are still quite well documented and therefore be considered reliable.

In this assessment, an update of the literature was made and new toxicity data for Cd in soil were found useful to be added to the dataset. Reliability was reviewed based on the same reliability indices as those in the RA and using the same criteria.For each test, a RI was given according to the following criteria (RA, ECB 2007):

RI 1: standard test, including the OECD 207 acute toxicity test withEisenia fetidain OECD-soil and the ISO 1994: soil quality effects of soil pollutants on Collembolla (Folsomia candida): method for the determination of effects on reproduction.

RI 2: no standard test but complete background information is given, i.e. the following information is present:

a)     soil pH

b)     soil organic matter or carbon content

c)      texture (class or texture fractions)

d)     total Cd content of the soil at zero Cd application if the NOEC or LOEC value is below 2mg/g

e)     equilibration time after soil contamination and prior to the test

f)      statistical analysis of the dose-response relationship

g)     no varying metal contamination along with increasing Cd application

h)     the control soil must be tested along with at least two Cd concentrations above the background concentration

i)       the soil must be homogeneously mixed with the metal prior to the test


RI 3: no standard test and one or more of the following information from the above-mentioned list is missing as background information: b), c), e) or f). All other information from that list is present.


RI 4: no standard test and one or more of the following information from the above-mentioned list is missing as background information: a), d), g), h) or i). The requirement d) is critical since some tests reporting LOEC values < 2mg/g are considered unreliable. Background Cd concentrations in soil typically range between 0.1 and 0.5mg/g and the lack of reporting the background concentration may underestimate the total Cd concentration in soil at which the first toxic effects are found. Unbounded NOECs were not used. Tests performed in substrates that were judged as not representative for soils (e.g. pure quartz sand or farmyard manure) were not included in this effects assessment.

Additional toxicity literature for cadmium was also checked according to the general criteria for data quality:

·        Toxicological endpoints, which may affect the species at the population level, are taken into account. In general, these endpoints are survival, growth and reproduction.

·        If for one species, several NOEC values on the same endpoint are available, the geometric mean of the NOEC values was first calculated and the most sensitive endpoint was taken forward in the SSD for PNEC derivation.

·        Only the results of tests in which the organisms were exposed to cadmium alone were used, thus excluding tests with metal mixtures.

·        Like in the RA, unbounded NOEC values were not used in the assessment.

·        Like in the RA, only the results of tests with soluble Cd2+ salts were used.

·        The NOECs used are reported as nominal values and were taken as such for the PNEC derivation. No correction for natural background was thus applied.


From the present revision of the terrestrial dataset, four new species of macro-organisms were found to respond to the criteria. Among them, three additional arthropod species and one plant species were included, allowing for a revision of the invertebrates + plants SSD. New data on species already figuring in the RA – database were also considered and species geometric mean NOEC values were recalculated based on new information. A new HC5 “plants and invertebrates” could subsequently be calculated (see Section 3).


2)     Single-species data for Cd toxicity in soil

The available database of chronic terrestrial toxicity tests for single-species with cadmium provides information on several species of soil micro-organisms, invertebrates and plants. These species are routinely utilized for assessing the toxicity of substances in spiked soils and standard test protocols exist.


Microflora dataset

The microflora dataset of the RA contains 21 entries (12 tests on respiration, 4 tests on N-cycle, 4 tests on soil enzymes and one test on N2fixation). The individual NOEC values varied from 3.6 mg/kg for the N2fixation endpoint up to 3000 mg/kg for respiration. No new data on microflora was found in this update and the microflora dataset from the RA remains thus unchanged. The microflora entries are summarized in the table 1.


Invertebrates dataset

The invertebrates dataset now contains 9 species among which: four species of annelids, one species of nematod and four species of arthropods. The available data on macroinvertebrate organisms include only long-term tests, from 21 to 294 days, which cover growth and reproduction effects. Three new species of arthropods (Onychiurus yodai, Sinella umesaoiandParonychiurus kimi) were added to the RA-dataset with NOEC values for reproduction of 50, 25 and 25 mg Cd/kg, respectively. A new NOEC value for growth of 12.5 mg Cd/kg was also found for the annelidL. rubellusand two new NOEC values of 25 and 80 mg Cd/kg were found for the reproduction endpoint the collembolFolsomia candida.Considering those new values, the endpoint reproduction becomes the most sensitive endpoint forF. candida. Overall, the individual NOEC values varied from 5 mg/kg for the annelidEisenia foetidaup to 320 mg/kg forF. candida. The invertebrates entries are summarized in the table 1.


Plants dataset

The updated plants dataset provides information on 15 species, including exposure times from 14 to 100 days and covering growth (length, weight, biomass) and germination effects. As compared to the RA, one new species was added to the dataset, i.e.Brassica campestrisL. cv.Chinensiswith a NOEC for growth (biomass) and NOEC for germination of 25 and 100 mg Cd/kg, respectively. New NOECs were found for the wheat seedling (Triticum aestivum; NOEC root elongation of 20 mg Cd/kg), for the oat (Avena sativa; NOEC biomass of 6.3 mg Cd/kg and NOEC germination of 25 mg Cd/kg) and the lettuce (Lactuca sativa; NOEC biomass of 3.1 mg Cd/kg and NOEC germination of 12.5 mg Cd/kg). The individual NOEC values varied from 1.8 mg/kg for the speciesPicea sitchensisup to 100 mg/kg forB. campestris var. Chinensis. The plants entries are summarized in the table 1.

The geometric mean NOEC values calculated for invertebrates and plants on the most sensitive endpoint are reported in Table 1. NOECs of soil microbial assays have not been averaged across soils because of the intrinsic variability of the microbial population between soils.


Table 1: Summary table of species geometric mean NOECs for the most sensitive endpoints of plants and invertebrates used in the SSD. New species to the ones mentioned in the RA or species for which new information was found are highlighted in bold. The newly added individual NOECs are underlined in the last column.

organismphylum/classOrderfamilyendpointSpecies geometric mean NOEC(µg g-1) 
Dendrobaena rubidaAnnelidaHaplotaxidaLumbricidaeReproduction10Geometric mean of 10, 10
Eisenia fetidaAnnelidaHaplotaxidaLumbricidaeReproduction5 
Lumbricus rubellusAnnelidaHaplotaxidaLumbricidaeGrowth43.3geometric mean of 150,12.5
Eisenia andreiAnnelidaHaplotaxidaLumbricidaeGrowth13.4Geometric mean of 10, 18
Onychiurus yodaiArthropodaIsotonidaOnychiuridaeReproduction50 
Sinella umesaoiArthropodaIsotonidaOnychiuridaeReproduction25 
Paronychiurus kimiArthropodaIsotonidaOnychiuridaeReproduction25 
Folsomia candidaArthropodaCollembolaIsotomidaeReproduction50.5geometric mean of25, 80
Plectus acuminatusNematodaAraeolaimidaPlectidaeGrowth32 
Avena sativaAvena sativaCyperalePoaceaeGrowth8.6geometric mean of6.3, 10, 10
Picea sitchensisPinopsidaPinalesPinaceaeGrowth1.8 
Triticum aestivumLiliopsidaCyperalesPoaceaeGrowth16.9geometric mean of 7.1, 20,20, 29
Glycine maxMagnoliopsidaFabalesFabaceaeGrowth6.6geometric mean of 2.5, 5, 10, 10, 10
Raphanus sativusMagnoliopsidaCapparalesBrassicaceaeGrowth20geometric mean of 10 and 40
Lactuca sativaMagnoliopsidaAsteralesAsteraceaeGrowth9geometric mean of 2, 2.5,3.1, 3.2, 5, 5, 10, 10, 20, 20, 32, 40, 40
Lycosperisicon esculentumMagnoliopsidaSolanalesSolanaceaeGrowth50.6geometric mean of 32 and 80
Phaseolus vulgarisZea maysCucurbita pepoLepidium sativumBrassica rapaDaucus carotaMagnoliopsidaLiliopsidaMagnoliopsidaMagnoliopsidaMagnoliopsidaMagnoliopsidaFabalesCyperalesViolalesCapparalesCapparalesApialesFabaceaePoaceaeCucurbitaceaeBrassicaceaeBrassicaceaeApiaceaeGrowth20108051010 
Brassica campestris var. chinensisMagnoliopsidaCapparales Brassicaceae Growth25 
Beta vulgarisMagnoliopsidaCaryophyllalesChenopodiaceaeGrowth34geometric mean of 20, 20, 20, 40,‘40, 40, 40, 80



3)     PNEC derivation

a.      Statistics on the species sensitivity distribution (SSD)

As in the RA, the statistical extrapolation approach is proposed in the PNEC derivation. We tested the lognormal distribution in the statistical approach as default option using the RIVM program ETX version 2.0. As in the RA (ECB, 2007), the HC5 is calculated for three different scenarios of data selection. The first scenario is by using microflora NOEC values only. The values were taken as such from the Cd RA, as no new data were added to the dataset. The second scenario is based on the use of the revised database for invertebrates and plants, (going from 20 to 24 species). The third scenario is making use of the whole terrestrial toxicity database, i.e. using data on microflora, plants and invertebrate organisms, as applied to other RA on metals (Cu, Ni) and recommended by the Scientific Committee for Health and Environmental Risks (SCHER) for the zinc RA.  The statistics of the curve-fitting on the chronic NOEC data are summarized in Table 02.


Table 02. Summary statistics for the SSD on chronic NOEC values for cadmium in soil

ScenarioNHC5 at 50% (Lower estimate on HC5) mg Cd/kgA-D test and significance levelK-S test and significance levelStatistical acceptance
Microflora212.3 (0.7)0.50(P>0.1)0.74(P>0.1)Accepted
Plants+Invertebrates243.6 (2.0)0.29(P>0.1)0.63(P>0.1)Accepted
Microflora+plants+invertebrates452.4 (1.0)0.46(P>0.1)0.69(P>0.1)Accepted


Using both the Anderson-Darling (A-D) and the Kolmogorov-Smirnov (K-S) tests for normality, the lognormal distribution fits significantly at a level of 1% for all tested scenarios.

The terrestrial data set is split in two groups: microbial processes and soil invertebrates + higher plants. The endpoints for microbial processes are relevant at the ecosystem functioning level, while the endpoints for soil fauna and plants are relevant at the species level. The principle of splitting the terrestrial data in two groups is open to criticism: there is no scientific argument (e.g. field validation) for either option. However, this approach was taken forward in the Cd RA (ECB, 2007) and is therefore proposed in the present assessment. As in the RA, the lowest NOEC selection approach was not performed because such a selection would not yield a representative data set for the terrestrial ecosystem (e.g. all clay soils would be excluded). The HC5 for the microflora is lower than the HC5 for soil fauna and plants. In conclusion, we propose to use the HC5 based on the microflora data set for the PNEC derivation i.e.HC5microflora= 2.3 μg Cd/g d.w. This approach is in accordance with the Cd RA (ECB, 2007) and results in the lowest of the three HC5 values following from the three tested scenarios.

For the sake of comparison, if the assessment factor approach would be applied, using the lowest NOEC divided by an assessment factor (AF) 10, this would yield a PNEC soil of 1.8 µg g-1/AF10 or 0.18 µg/g. This value is within the range of cadmium background concentrations in soils which typically range between 0.1 and 0.5 µg/g (Cd RA, 2007).



b.      Setting the PNEC soil

The PNEC soil is set based on the lowest observed HC5 derived by statistical extrapolation from the microflora data, i.e.2.3 µg Cd/kg d.w. In the Cd RA, an AF 1 or 2 was considered. The current analysis rather suggests using an AF1 on the HC5 to derive the PNEC. It is noted that thePNECsoilbased on secondary poisoning is of 0.9 µg Cd/g d;w., which is below the proposed value. The latter value is therefore proposed and used for PNECsoil in this analysis. This is in accordance with the Cd RA.