Registration Dossier

Diss Factsheets

Administrative data

Link to relevant study record(s)

Description of key information

In this dossier, the endpoint toxicokinetics is not addressed by substance-specific information, but instead by a weight of evidence approach based on collected information for all substances of the CoRC cobalt category. The category consists of inorganic cobalt substances for which the toxicity is governed by the cobalt cation. The bioavailability of the cation varies in a predictable manner and is assumed to be dependent on the bioaccessibility of the respective cobalt substance. The assumption that in vivo bioavailability correlates with the in vitro bioaccessibility was verified in an in vivo blood kinetics study in rats, using three representative substances. Consequently, the in vitro and in vivo toxicokinetic behaviour of the cobalt category substances form the basis for the oral-systemic read-across. For this reason and for the ease of reference all available toxicokinetic data for the cobalt category substances is discussed below.

Further details on the read-across approach are given in the reports attached in IUCLID section 13.2.


In vitro Studies

In vitro bioaccessibility testing:

The bioaccessibility of 26 different cobalt substances was investigated by measuring their solubility in simulated physiological fluids, performed in adaptation according to ASTM D 5517 (”A Standard method for Determining the Solubility of metals in Art materials”). The test items covered the full range of commercially relevant cobalt substances, such as cobalt metal, cobalt salts, cobalt oxides and hydroxides, cobalt salts with organic anions ("carboxylates"). The results of the bioaccessibility testing can be expressed in different units, common units being “percentage release” (“% release”), “release concentration” or “release rate”. Percentage release is the percentage of available Co in the compound being tested. Release concentration is the absolute value of the Co concentration observed in the receptor fluid after the standardised incubation time. In the case of the Co substances, 26 very different substances are being compared, due to the large variety of substances on the market. In order to adjust for the differences in individual particle size, the surface area of the substances is introduced in the unit, by expressing the release rate. The bioaccessibility results are given over three time points, i. e. 2 or 5, 24 and 72 hours, in the following test media:


1. Artificial interstitial fluid, pH: 7.4

2. Artificial alveolar fluid, pH: 7.4

3. Artificial lysosomal fluid, pH: 4.5-5.0

4. Serum, pH. 7.2 -7.4

5. Artificial synovial fluid, pH: 7.4

6. Artificial gastric juice, pH: 1.5

7. Artificial intestinal juice, pH: 7.4

8. Artificial perspiration (by EN 1811 method), pH 6.4-6.6


Four measures were used to evaluate the bioaccessibility of cobalt species if inhaled, using synthetic equivalents of alveolar fluid, interstitial fluid, lysosomal fluid, and human serum. In these fluids, the selected cobalt species showed variations in bioaccessibility. Cobalt substances showing a high solubility in neutral lung fluid can be assumed to be readily dissolved upon sedimentation. Cobalt salts and cobalt metal powder are readily soluble in a gastric juice surrogate. Ingestion of dusts or liquids containing any of the cobalt compounds would however, be associated with exposure to ionised cobalt that would be available for absorption. Employing gastric and intestinal fluid surrogates for the assessment of bioaccessibility of cobalt species is useful in predicting the bioavailability of cobalt compounds and determining doses related to health risks caused by their ingestion. For the sake of brevity, detailed data on the cobalt release following incubation are given in the IUCLID study record.


In vitro dermal absorption data:

For the assessment of the percutaneous absorption potential of cobalt substances, the absorption of Cobalt following topical application of four cobalt substances to human skin in vitro was investigated. Matar (2020) investigated the dermal absorption of cobalt metal, cobalt bis(2-ethylhexanoate), cobalt dichloride hexahydrate and tricobalt tetraoxide. The dermal absorption of the cobalt compounds was investigated after applying 100 mg/cm² to human skin in the presence of artifical sweat. Each compound was either applied to the skin for an 8-hour or 24-hour exposure period before being washed from the skin surface. Each piece of skin was sectioned using a cryostat to determine the depth of penetration of cobalt. The following results were recorded for the four cobalt compounds:

Generally, the largest proportion of the applied dose was consistently recovered from the wash from both 8- and 24-hour exposures:

Cobalt metal: 93.26 % and 93.45 % after 8 hour and 24-hour exposure, respectively

Cobalt bis(2-ethylhexanoate): 94.87 % and 94.35 % after 8 hour and 24-hour exposure, respectively

Cobalt dichloride hexahydrate: 96.88 % and 98.24 % after 8 hour and 24-hour exposure, respectively

Tricobalt tetraoxide: 94.27 % and 93.73 % after 8 hour and 24-hour exposure, respectively

The potentially absorbed dose calculated as the sum of the total amount cobalt recovered with the receptor fluid and skin available for absorption (first 20 µm of skin omitted) was for the four cobalt compounds as follows:

Cobalt metal:

8 hour exposure: 6 µg cobalt/cm² (equivalent to 0.006% of the applied dose)

24 hour exposure: 19 µg cobalt/cm² (equivalent to 0.02% of the applied dose)


Cobalt bis(2-ethylhexanoate):

8 hour exposure: 16 µg cobalt/cm² (equivalent to 0.093% of the applied dose)

24 hour exposure: 62 µg cobalt/cm² (equivalent to 0.36% of the applied dose)


Cobalt dichloride hexahydrate:

8 hour exposure: 8 µg cobalt/cm² (equivalent to 0.033% of the applied dose)

24 hour exposure: 31 µg cobalt/cm² (equivalent to 0.12% of the applied dose)


Tricobalt tetraoxide:

8 hour exposure: 1 µg cobalt/cm² (equivalent to 0.0014% of the applied dose)

24 hour exposure: 23 µg cobalt/cm² (equivalent to 0.031% of the applied dose)


The rationale for subtracting the first 20 μm was to remove any compound that was potentially retained within the stratum corneum and therefore only quantifying the amounts within the skin. However, it is not evident from the data if the compound had diffused through the stratum corneum into the underlying skin or had deposited deep within the hair follicles.

In the study by Roper, 2010 cobalt dichloride was applied to human skin at two different application rates: ca 100μg/cm² and ca 1000μg/cm² (equivalent to cobalt surface doses of ca 31.9μg/cm² and 319μg/cm², respectively). The exposure duration was 8 hours after which the solutions were removed by washing, followed by a ca 64 hours post exposure monitoring period. For the two exposure concentrations, the absorbed dose of cobalt was 0.15 and 0.87%, respectively. The dermal delivery of cobalt was 0.29% and 0.99%, respectively. The majority of the applied dose was removed during the washing procedure with 99.51% and 98.83% at 8 hours. For the purposes of risk assessment, the potentially absorbable dose (sum of absorbed dose, skin and stratum corneum tape strips 6-20) is used, corresponding to 0.38% for the low exposure scenario and 1.08% for the high exposure scenario. In conclusion, the dermal absorption of cobalt has been shown to be very low, thus rendering percutaneous uptake a negligible route of entry into the body in comparison to oral/inhalation routes. Despite being a guideline study conducted under GLP, the results of this study were disregarded for the hazard and risk assessment of the cobalt category substances, as the loading used in this study of 31.9 and 319 µg/cm² was not sufficiently high, compared with the loading used in the study by Matar (2020) of 100 mg/cm². Consequently, the results obtained in the Matar (2020) study will be used in the hazard and risk assessment of the cobalt category substances.

Lastly, further references on the in vitro and in vivo dermal absorption were identified which do not fulfil the relevance, reliability and adequacy criteria as foreseen by the OECD manual for investigation of HPV chemicals, chapter 2. These studies are included in the IUCLID as summary entry (Larese, 2007; Larese-Filon; 2004, Lacy, 1996).


In vitro inhalation deposition/absorption model predictions:

The fate and uptake of deposited particles depends on the clearance mechanisms present in the different parts of the airway. In the head region, most material will be cleared rapidly, either by expulsion or by translocation to the gastrointestinal tract. A small fraction will be subjected to more prolonged retention, which can result in direct local absorption. More or less the same is true for the tracheobronchial region, where the largest part of the deposited material will be cleared to the pharynx (mainly by mucociliary clearance) followed by clearance to the gastrointestinal tract, and only a small fraction will be retained (ICRP, 1994). Once translocated to the gastrointestinal tract, the uptake will be in accordance with oral uptake kinetics.

In consequence, the material deposited in the head and tracheobronchial regions would be translocated to the gastrointestinal tract without any relevant dissolution, where it would be subject to gastrointestinal uptake, depending on the oral absorption rates determined in in vivo experiments. The material that is deposited in the pulmonary region is conservatively assumed by default to be absorbed to 100% (recognised as over-conservative for very poorly soluble cobalt substances such as tricobalt tetraoxide, based on extensive lung clearance studies). This absorption value is chosen in the absence of relevant scientific data regarding alveolar absorption although knowing that this is a conservative choice. Thus, the following predicted inhalation absorption factors can be derived for cobalt compounds.



Studies in Animals


Oral route:


A relative bioavailability study involving serum kinetics over a period of 72 hours p. a. involving an i. v. dosing of a soluble cobalt reference substance (cobalt dichloride) compared to single oral doses of cobalt dichloride, tricobalt tetraoxide, cobalt distearate, cobalt sulfide and cobalt lithium dioxide. For details, please refer to the corresponding robust study summaries. In brief, 20 animals (10m/10f) per group received single doses of (1) 0.1 mg/kg cobalt dichloride intravenously, (2) 10 mg/kg cobalt dichloride via oral gavage, (3) 300 mg/kg tricobalt tetraoxide, (4) 300 mg/kg cobalt lithium dioxide, 300 mg/kg cobalt sulfide via oral gavage and (4) 5 mg/kg cobalt stearate via oral gavage. Blood samples were taken at 0, 0.5, 1, 2, 4, 8, 12, 24 and 72 hours post exposure and blood plasma samples were prepared and analysed for cobalt. To reduce the stress for the animals, not every animal was sampled at every time-point. Instead, 10 animals (5m/5f) from each group were sampled at every second time-point. For comparison, the average (n=40) concentration of cobalt in plasma taken before exposure at t=0 h was (0.0082 ± 0.0077) µg/L (min=0.004; max=0.041). The plasma concentrations for the orally administered animals declined post dosing with an elimination half-life ranging from 10.1 to 18.2 hours. A relative bioavailability of 9.3% was calculated for cobalt dichloride (6.8 % in males, 11.7% in females) following oral administration compared to intravenous administration, 0.1 % for tricobalt tetraoxide (0.1 % in males, 0.1 % in females), 13 % for cobalt stearate (9.4 % in males, 16.6 % in females), 0.3% for cobalt lithium dioxide (0.29 % in males, 0.25 % in females) and 0.09% for cobalt sulfide (0.09 % in males, 0.09 % in females).


Two reasonably well described studies (Ayala-Fierro et al. 1999, Firriolo et al. 1999) assessed the oral absorption/excretion in rats following single exposure to cobalt dichloride (single dose: 33.3 mg Co/kg bw) and cobalt naphthenate (dose groups: 0.33, 3.33, 33.3 mg Co/kg bw). Eight dose groups of three rats each were held in metabolic cages for up to 36 hours after oral exposure. Total cobalt content was measured via GF-AAS in urine and faeces 8, 12, 18, 24, and 36 hours post-dosing. Tissues were collected 8, 12, 18, 24, and 36 hours after exposure. These studies indicate that cobalt absorbed in the gastrointestinal tract is primarily retained in the liver, but was also found in the kidneys, heart, stomach, and intestine. However, the overall body burden 36 hours post-exposure was low since a total of 98.4% was excreted.

The major route of excretion was via faeces with an elimination rate of 74.5% for cobalt dichloride and 73.1% for cobalt naphthenate (high dose group), 36 hours after exposure. Excretion via urine was 23.9% and 26.3% (high-dose group), respectively. The absorption/excretion of both compounds indicates that the organic anion of cobalt naphthenate does not have an impact on the toxicokinetics, since both cobalt substances show similar excretion rates.

In another series of experiments, an interspecies comparison was conducted to determine the 57Co excretion pattern following ingestion of tricobalt tetraoxide particles. Excretion of cobalt was followed for up to eight days after ingestion of monodisperse 0.8 and 1.7µm diameter tricobalt tetraoxide particles by baboons, guinea-pigs, rats (HMT and Fischer-344), mice and hamsters. (Andre et al. 1989, Bailey et al. 1989, Collier et al. 1989, Patrick et al. 1989, Talbot et al. 1989). An overview of all species is given below in tabulated format.

The test item was given to the animals via gavage in a single application. Two batches with different particle sizes, 1.7 and 0.8µm were used, suspended in a suitable vehicle. After administration, the animals were placed in metabolic cages for separate collection of urine and faeces. The 57Co content of urine and faeces was usually determined daily. After the observation period of up to 8 days, animals were sacrificed and the 57Co content of selected tissues was determined. Excretion and retention rates were calculated for both tricobalt tetraoxide batches.

Table. Summary of measurements of retention and excretion of 57Co following intragastric administration of Co3O4 particles. Mean percentage of recovered activity at seven days after administration.


Cumulative faecal excretion [%]

Cumulative faecal excretion [%]

Whole body retention [%]

Whole body retention [%]

Cumulative urinary excretion [%]

Cumulative urinary excretion [%]

Absorption [%]

Absorption [%]

Particle size



























Rat (HMT)









Rat (F-344)




























Overall, it can be stated that faecal elimination of the poorly soluble tricobalt tetraoxide is the primary route of excretion in animals following oral exposure. Faecal clearance has been noted to decrease as cobalt particle solubility increases. In several species, oral exposure to tricobalt tetraoxide (with 57Co tracer) resulted in little gastrointestinal absorption and a rapid elimination in faeces (>96%). No significant differences in tricobalt tetraoxide elimination were observed among species. Cobalt dichloride, which is more soluble, was excreted primarily via faeces (70–83% of the administered dose) in rats, with urinary excretion accounting for the remainder of the dose.


Inhalation route:

Inhalation absorption rates have not been reported for any cobalt substance, which is why orientating model predictions were performed (see in vitro section below). However, detailed lung clearance investigations have been conducted, which provide a basis for the assessment of the fate of inhaled cobalt particles, and in particular document substantial differences between soluble cobalt substances (i.e., cobalt chloride and nitrate) and poorly soluble cobalt substances (represented by tricobalt tetraoxide). The most relevant studies are summarized briefly below:

The clearance of soluble cobalt (as radioisotopic cobalt chloride) deposited intratracheally in the lungs of various species was measured directly over 100 days (>1000 days in the case of dogs). Despite significant interspecies differences in whole-body retention and urinary/faecal excretion, all species studied had a high concentration of 57Co in the trachea and in the lung tissue relative to the concentration in the whole body. The mean fraction of 57Co retained in lungs for >100 days ranged from 0.13-0.58% of dose. By comparison with previous studies on tricobalt tetraoxide (see below), it was concluded that the interspecies differences in absorption rates for inhaled tricobalt tetraoxide particles were not a result of differences in the fraction of dissolved cobalt retained in lung tissue (Patrick et al., 1994).

Menzel et al. (1989) similarly studied respiratory tract burdens following inhalation of soluble aerosols in Sprague-Dawley rats, involving exposed to cobalt dichloride aerosols for 2 h in a head only apparatus either as a single exposure or 7d repeated exposure. Co aerosols (0.80–0.93 MMAD, Δg 1.32–1.41) were deposited in the nasopharyngeal region (NP) and lung with measured efficiencies of 2.3% and 5.4%. The kinetics of cobalt removal were first-order, with estimated clearance rate of 0.4 h−1 or a t½ of 1.8 h. Both the acute and repeated exposure clearance rates were faster for soluble cobalt than those for insoluble particles removed by mucociliary clearance, allowing the conclusion that Co removal from the lung is due to removal of the Co ion.

In an extensive collaborative project, an interspecies comparison of the lung clearance of inhaled monodisperse cobalt oxide particles was conducted. Animals were exposed nose-only to aerosols of 0.8 and 1.7 µm tricobalt tetraoxide (geometric diameter). Whole body retention and fractional lung retention were measured, as well as urinary/faecal excretion in some cases.

- Baboons were exposed nose-only, and for comparative purposes by i.v. injection and gavage administration. Average biological retention half-lives for 0.8 and 1.7 µm particles were 90 and 150 days, respectively. Faecal excretion decreased continuously with time, on average 0.28-0.04% of the lung burden per day. Urinary excretion remained fairly constant for the larger particles (0.11-0.17% per day), whereas this was somewhat higher for the small particles (0.24-0.54% per day). 6 months after cessation of inhalation exposure, cobalt activity was still associated with particulate material in alveolar macrophages, with 99% of the activity bound to cell pellets and only 1% in the supernatant. The authors conclude that the amount of cobalt activity translocated from the lungs is roughly equivalent to the amount excreted via urine, and that the long-term clearance pattern emphasises the importance of intra-cellular dissolution (Andre et al., 1989, Bailey et al. 1989).

- In a follow-up publication, an interspecies comparison of lung clearance in baboons, rats and dogs is reported. Lung retention at 6 months after inhalation ranged from 60% of the initial lung deposit in baboons to 5%, in rats and in all three species clearance of 1.7 µm particles was considerably slower than that of porous 0.8 μm tricobalt tetraoxide particles. Lung clearance rates due to translocation of dissociated 57Co to the blood, S(t), and due to particle transport to the GI tract M(t) were calculated. Initially, S(t) ranged from 0.1% of the contemporary lung content day−1 in baboons to 0.7% in rats. For each species, a correlation was found between the initial translocation rates and the specific surface area of the three particles used (Kreyling et. al., 1991; Kreyling et. al., 1993).

- In a series of in total eight publications, this group of authors conducted an in-depth inter-species (humans, baboons, dogs, guinea-pigs, rats (three strains), mice and hamsters) comparison of the lung clearance of the poorly soluble 57Co3O4, plus supplementary experiments involving gavage administration (57Co3O4) or i.v. injection (57cobalt nitrate). At six months after inhalation of the 0.8 μm particles, lung retention ranged from 1% of the initial lung deposit (ILD) in HMT and Sprague-Dawley rats to 45% ILD in man; and for the 1.7 μm particles from 8% ILD in HMT rats to 56% in man. Supplementary experiments were conducted to determine 57Co excretion patterns following injection of Co(NO3)2 into the blood and following ingestion of cobalt oxide particles, in order to calculate lung clearance rates due to translocation of dissociated 57Co to the blood, S(t), and due to particle transport to the GI tract, M(t). Initially, S(t) for 0.8 μm particles ranged from 0.4% of the contemporary lung content day−1 in baboons to 1.6% day−1 in HMT rats. Initial values for 1.7 μm particles were lower in all species, and ranged from 0.2% in baboons to 0.6% day−1 in HMT rats (Bailey et al., 1989). From the above, the authors concluded that (i) particle dissolution was the predominant clearance pathway, and (ii) broncho-alveolar lavages revealed phagocytosed particles in alveolar macrophages from epithelial surfaces up to 500 days after inhalation (Andre et al. 1989, Bailey et al. 1989, Collier et al. 1989, Drosselmeyer et al., 1989; Foster et al., 1989; Kreyling et al., 1989; Patrick et al. 1989, Talbot et al. 1989).


As shown above for the in vitro inhalation deposition/absorption model predictions, the default pulmonary absorption factor of 100% was used due to an absence of animal or human data.


Dermal route:

There are no reliable in-vivo dermal absorption data in animals.



Cobalt is not subject to any metabolism in its true sense: regardless of its original chemical speciation. Cobalt transforms rather quickly to divalent cobalt cations upon dissolution. In this form, it is available via diet or drinking water, and represents the physiologically relevant cobalt species. Once systemically available, cobalt is stable in the cobalt cation and not subject to any changes in speciation or valence.



Since cobalt is an essential metal and a component of vitamin B12, it has been found in most tissues, such as muscle, lung, lymph nodes, heart, skin, bone, hair, stomach, brain, pancreatic juice, kidneys, plasma, urinary bladder, and liver (highest levels).

“Dietary vitamin B12 is absorbed through a receptor mediated mechanism in the ileum. Food-bound vitamin B12 has first to be liberated through peptic digestion and gastric acid secretion in the stomach. The ‘free’ vitamin B12 becomes then bound to haptocorrins (or R-proteins) secreted by the salivary glands and the gastric mucosa. In the small intestine the R-binders are degraded by pancreatic protease action and the cobalamins are subsequently bound to the Intrinsic Factor (IF), a glycoprotein secreted by the parietal cells of the stomach. Uptake in the ileum is specific for the IF-cobalamin complex. Fractional absorption decreases as the oral dose is increased. Ileal receptors are saturated with dosages between approximately 1.5 and 2.5 μg of vitamin B12 per meal. At intakes around 1 μg about 50% is absorbed, at dosages around 25 μg only 5% is absorbed. Very small amounts (ca 1%) can be absorbed by passive diffusion, in the absence of IF. Vitamin B12 is an exceptional B-vitamin as it can be stored in significant amounts, especially in the liver and the kidney. The average concentration in human liver is between 0.5-1 μg/g; the total body pool size is estimated between 2-3 mg. The main excretion is through the bile, but there is a considerable reabsorption of these biliary cobalamin losses in the ileum (enterohepatic circulation). Average daily losses via the stool are estimated at ca 0.5 μg.” (EFSA, 2006)

Laboratory animal studies in various species indicate that cobalt absorbed via the gastrointestinal tract is primarily retained in the liver (Ayala-Fierro et al., 1999). Cobalt was also found in the kidneys, heart, stomach, and intestines (Ayala-Fierro et al., 1999).


The study conducted by Hansen (2015) investigated the tissue distribution of cobalt in male and female CD rats. Cobalt dichloride in water (30 mg/kg bw/day) was given to a group of 10 male and 10 female rats via oral administration (gavage) daily for a duration of 90 days. A vehicle control group was run concurrently. In addition, the animals were placed individually in metabolic cages one day before scheduled sacrifice for 24 hours to collect urine and faeces for cobalt analysis. The analysis of the urine and faeces samples confirms the findings of the mass balance study after bolus oral exposure, that after 90-days, cobalt is mostly excreted via faeces and only a small portion of the cobalt is excreted via urine. Samples of organs (ovary, prostate, uterus, testis, brain, kidney, bone marrow, adrenal medulla, pancreas, liver, intestine, lungs) were obtained after 13 weeks of administration and analysed for cobalt content expressed as µg cobalt per g wet tissue using ICP-MS. In general all organ of the cobalt dichloride treated animals showed elevated cobalt levels compared to the concurrent control animals. However, higher concentrations were found in organs relevant for the elimination, such as kidney and liver of male and female rats. Elevated cobalt concentrations were also measured in the adrenal gland and the pancreas. Cobalt was also found in the ovaries of the treated females, but to a lesser extend as in the aforementioned organs.




Following ingestion, the excretion is primarily via the faeces: (i) ca. 70-83% for soluble cobalt substances with a correspondingly high oral absorption rate (e.g. cobalt dichloride), and (ii) ca. 95% for substances with a low oral absorption rate (e.g. tricobalt tetraoxide), with urinary excretion accounting for the remainder of the dose.

The overall elimination after systemic uptake is very rapid, with whole body retention rates of ca. 1.5% 36 hours after administration of cobalt dichloride and <1% 6-8 days after administration of tricobalt tetraoxide.

The biliary excretion of 58Co after i.v.-administration of two different doses (177 and 1770 µg Co/kg bw) was studied in rats. 24h p.a., the cumulative biliary excretion reached 2.67 ± 1.98% (low dose) and 7.33% (4.6-10.9, high dose) of dose, respectively. The highest rate of excretion was between 10 and 30 minutes after administration, and at the high dose, the urinary excretion was reduced compared to that of the low dose (Cikrt & Tichy, 1981). In another study (Gregus & Klaassen, 1986), tissue distribution (2h p.a.) as well as faecal (0-4 days), urinary (0-4 days) and biliary (0-2 hr) excretion were investigated following single i.v. administration of cobalt chloride (dose 0.3 mg Co/kg bw) to Sprague-Dawley rats (m, 200-300g). Total (faecal + urinary) excretion was relatively rapid (87.7% of dose in 4d); biliary excretion in the first 2h p.a. ranged from 2.3-4.7% of dose and increasing with dose (0.03-1.0 mg Co/mg bw). The highest tissues levels (2 hrs p.a. in % of dose) were in livers (5.1-7.0), kidney (14.5-19.6) and pancreas (1.8-4.7), whereas all other tissues were at or below 1%; levels in testes were between 0.5-0.7%.


The mass balance study by Leuschner (2018), conducted according to OECD 417 and under GLP, cobalt excretion was investigated in groups of five male and five female Crl: CD(SD) rats after a single oral application via gavage of cobalt dichloride (10 mg/kg bw), tricobalt tetraoxide (300 mg/kg bw) or cobalt stearate (5 mg/kg bw). The results showed that the excretion of cobalt appears to be rapid and occurs mostly within the first 24 hours after the bolus dose of cobalt dichloride, tricobalt tetraoxide or cobalt stearate. In the case of tricobalt tetraoxide, 100 % of the bolus dose is excreted within the first 24 hours, whereas cobalt dichloride and cobalt stearate show a slightly lower total excretion after 24 hours of 95 % and 85 % of the total cobalt dose administered. This shows that all three cobalt substances are excreted rapidly and quantitatively, with no indication for bioaccumulation. For all three substances, most of the cobalt is excreted via faeces (Co3O4: 99.87%, CoCl2: 90.28%, CoStearate: 83.22%, mean for males and females). Some urinary excretion of cobalt is seen from cobalt dichloride (9.72%) and cobalt stearate (16.78%). For cobalt dichloride there appears to be a sex difference in the urinary excretion (males: 16.94%, females: 2.50%), whereas for cobalt stearate the urinary excretion is almost equally distributed (males: 15.31%, females: 18.26%). The urinary excretion of cobalt from tricobalt tetraoxide is negligible (males: 0.12 %; females: 0.13 %).



Studies in Humans

Only two substantial toxicokinetic studies in human volunteers have been published to date:

In an older study (Smith et al., 1972), 60Co chloride was administered either i.v. or orally together with varying amounts of unlabelled cobalt chloride to 24 subjects, followed by whole body monitoring as well as blood, urine and faeces sampling. Cobalt was eliminated from blood rapidly (30% of dose within 24h p.a.), and the liver was thought to initially retain ca. 20% of the dose. However, due to the employed imprecise methodology and corresponding large variances in results, definitive conclusion cannot be drawn from this study.

A study comparing the oral absorption of cobalt chloride and tricobalt tetraoxide was conducted in 23 human volunteers, with a reference group of 118 non-exposed subjects (Christensen et al., 1993). Daily dosing with either substance was for 10 consecutive days, with blood and urine sampling. Due to limited sampling time points and the lack of a mass balance, only the following semi-quantitative observations can be made: blood/urine concentrations in non-exposed reference subjects were about 2-3 fold higher in females than in males - this was mirrored to some extent in cobalt-treated subjects. Median blood and urine values from males dosed with cobalt oxide did not differ significantly from the reference values and females showed only a minor elevation of cobalt concentrations in comparison to control, indicating that the oral absorption of tricobalt tetraoxide was low to negligible. In contrast, oral dosing with soluble cobalt chloride increased blood and urine concentrations in comparison to the reference by 14-20-fold and 16-59-fold, respectively.

Four studies have documented cobalt tissue concentrations in human autopsy specimens from non-exposed individuals, which however either suffer from a lack of statistical power or from methodological shortcomings; in one case, the results are also contradictory and implausible. Overall, these data are not considered to represent a data base that would allow a meaningful conclusion on “background” tissue concentrations in humans (Yamagata, 1962; Tipton & Cook, 1963; Schroeder et al. 1967; Takemoto et al, 1991).

A group of seven studies report on follow-up investigations of radioactive cobalt levels in humans contaminated accidentally either in nuclear power plants or other, laboratory-type facilities. Unfortunately, neither chemical composition nor the nature of the contaminants are described to any extent that would allow conclusions concerning the toxicokinetic behaviour of individual cobalt substances (Sedlet et al., 1959, Morsey & El-Assaly, 1970; Newton & Rundo, 1971; Hedge et al, 1978; Beleznay & Osvay, 1994; Davis et al., 2007).

Six investigations from the past have been published on aspects of diagnosing physiological disorders and dietary (iron) status by administering 57cobalt or 60cobalt (as chloride) to patients, with subsequent elimination monitoring. However, in most cases the degree of clinical or nutritional disorder in these studies renders them of minor relevance for human health risk assessment which conventionally considers normal, healthy individuals (Harp & Scoular, 1952; Paley & Sussman, 1963; Valberg et al., 1969; Sorbie et al. 1971; Valberg et al., 1972; Ishihara et al., 1987).

Urinary biomonitoring has been used in several cobalt producing or processing industries. However, the correlation between inhalation exposure and urinary excretion is not always straightforward. This is likely due to several factors, among them differences in bioavailability of the particular cobalt substance handled, as well as variance in personal hygiene management between individual workers, for example involving inadvertent oral exposure from hand-to-mouth transfer. Nevertheless, based on the toxicokinetic profile of cobalt, urinary biomonitoring appears to be a valuable tool in the assessment of systemic exposure (Teraoka, 1981; Hartung et al., 1983; Gerhardsson, 1984; Christensen & Mikkelsen, 1985; Hewitt, 1988; Apostoli et al., 1994; Scansetti et al., 1994; Linnainmaa & Kiilunen, 1996).

All available, published human investigations associated with toxicokinetic issues are summarised in tabular format. Reliability scores were not assigned, since these are not considered applicable to such human information.



Based on the in vivo test results for oral absorption, two groups of substances with different oral absorption factors can be identified: the in vivo oral absorption varies from (i) 0.1% for tricobalt tetraoxide to (ii) 9.3 and 13% for cobalt dichloride and cobalt distearate, respectively.

Particle-size dependant fractional deposition modelling coupled with gastric bioaccessibility data indicate that inhalation absorption may be expected to be in the range 2-25%. However, lung clearance data for very poorly soluble substances suggest that translocation to the gut with subsequent dissolution is the major elimination pathway, and in view of the poor bioavailability, these above predictions may be over-conservative.

Reliable in vivo dermal absorption data for cobalt substances do not exist. However, the dermal absorption of soluble substances can be assessed based on a guideline-conform in vitro percutaneous absorption study on cobalt dichloride conducted under GLP. The observed dermal absorption was found to be 0.38% for the low exposure scenarios (ca 31.9μg Co/cm²) and 1.08% for the high exposure scenarios (ca 319μg Co/cm²). These values also account for part of the material associated with the stratum corneum and the test was conducted with a highly water soluble form of Cobalt in an aqueous solution.

Cobalt is not subject to any metabolism or change in valence, but instead when becoming systemically available is in the form of divalent cobalt cations.

Ingested bioavailable cobalt is widely distributed throughout the body, and laboratory animal studies in various species indicate that cobalt is primarily retained in the kidney and the liver, and to a lesser extent in the adrenal gland, pancreas and bone marrow.

Following ingestion, the excretion is primarily via faeces: (i) ca. 83-90% for soluble cobalt substances with a correspondingly high oral absorption rate (e.g. cobalt dichloride and cobalt stearate), and (ii) >99% for substances with a low oral absorption rate (e.g. tricobalt tetraoxide), with urinary excretion accounting for the remainder of the dose. The overall elimination after systemic uptake is very rapid and almost quantitatively within 72 hours after single oral administration. Biliary excretion in rats has been shown to represent only a minor fraction of the dose, ranging from approx. 2-7% of dose. There is no data suggesting that cobalt has any bioaccumulation potential.

Key value for chemical safety assessment

Bioaccumulation potential:
no bioaccumulation potential

Additional information