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Please be aware that this old REACH registration data factsheet is no longer maintained; it remains frozen as of 19th May 2023.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

Link to relevant study record(s)

Description of key information

In rats and guinea pigs ca. 40% of the applied formaldehyde is absorbed via the skin. In in vitro experiments using guinea pig skin the percutaneous absorption rate was ca. 30% after 1 h of exposure.
Dermal absorption (%): 4% for penetration through the skin possibly leading to systemic effects; 15% for penetration through and into the skin possibly leading to local effects.

Key value for chemical safety assessment

Additional information

Endpoint summary basic toxicokinetics including summary of reviews

Endogenous occurrence

In humans as well as in animals formaldehyde is an essential metabolic intermediate. It is endogenously formed from serine, glycine, methionine and choline and is produced in the demethylation of N-, O-, and S-methyl compounds. Formaldehyde is an essential intermediate in the biosynthesis of purines, thymidine and several amino acids (IARC, 1995; summary reviews). The mean endogenous concentration of free and reversible bound formaldehyde in blood of unexposed humans was 2.61 µg/g blood (range 2.05- 3.09 µg/g), in rats 2.24 µg/g and in monkeys 2.42 µg/g (Heck et al., 1985; Casanova et al., 1988;), i.e. about 0.1 mmol/L. EFSA (2014) estimated that endogenous turnover of formaldehyde was estimated to be approximately 0.61-0.91 mg/kg bw per minute and 878-1310 mg/kg bw per day (corresponding to about 60-90 g/adult person per day) assuming a half life of 1-1.5 min.In rats the concentrations of free and acid labile formaldehyde were 12.6, 6.03, 8.4, and 2.9 µg/g wet tissue weight in the nasal mucosa, liver, testes, and brain, respectively (Heck et al., 1982).



Consistent with the high water solubility of formaldehyde and its reactivity with macromolecules the substance is deposited and absorbed after inhalation in the upper respiratory tract, the site of first contact.

The amount is directly proportional to the concentration. (WHO, 2002; summary reviews; Heck et al., 1983; for detailed documentation). Differences between species were found in the actual sites of uptake; in obligate nose breathers like the rat absorption occur in the nasal passages and in oronasal breathers like humans and monkeys in nasal passages but also in oral passages, nasopharynx, trachea, and proximal bronchi (WHO, 2002; IARC, 1995; summary reviews). Generally, the localisation of uptake in each species is determined by nasal anatomy, mucus coating and clearance mechanisms. It could be demonstrated for the rat model, that the main part of flow intake at the nostrils passes into the middle and lateral meatuses with less flow to the dorsal and ventro-medial pathways (BfR, 2006; summary reviews). These results are confirmed by experimental data demonstrating histopathological damage (Section Repeated dose toxicicity: dermal) and DNA-protein cross-links (DPC) formation (Section Genetic toxicity in vivo) at these sites. Similar results were presented for the monkey. Formaldehyde induced nasal lesions (see Casanova et al., 1991; Section Genetic toxicity in vivo or DPC as well as chapter Distribution) correlated (with some exceptions) well with regions of high bulk flow, secondary flow and turbulence. Using different models it could be calculated that in humans the highest surface fluxes were predicted in the nasal airways. Oronasal breathing which occurs in humans at higher activity resulted in higher tracheobronchial flow compared to nasal breathing (BfR, 2006; summary reviews).

The overall uptake by the nasal passages at resting airflow rates has been predicted to be 90% in rats, 67% in monkeys, and 76% in humans (Kimbell et al., 2001) (BfR, 2006; summary reviews).

As reported in an abstract (Patterson et al., 1986), the retention by the nasal passages in rats was > 93% of the dose (2, 6, 15, or 50 ppm; exposure duration 30 minutes) of inhaled radio-labelled formaldehyde, regardless of the airborne concentration.

Based on previous studies performed at CIIT a „computational fluid dynamics“ (CFD) model has been developed for rats, monkeys and humans to localize the impact of inhaled formaldehyde in the nose and different parts of the total respiratory tract. Overton et al. (2001) extended the CFD model of the human nasal passages to include the total human respiratory tract, divided into the upper and lower part. The upper respiratory tract was partitioned into the “proximal” and “distal” segments. The proximal segment represented either the mouth cavity or the nasal airway passage, but not both, to take account of oronasal breathing at high physical activities; for example at a minute volume of 50l/min only 46 % of the air passes through the nasal passages. The distal upper respiratory tract segment comprised the air passage distal to the nasal airway and the mouth cavity up to the proximal end of the trachea. The lower respiratory tract was divided into the tracheobronchial region starting from the trachea down to the terminal bronchioles (generation 0-16) and the pulmonary region starting from the first respiratory bronchioles to the alveolar sacs (generation 17-23). Formaldehyde surface flux (pmol/mm²/h/ppm) was calculated for the different regions of the respiratory tract for four activity states from sleeping with nasal breathing and a minute volume of 7.5 l/min up to heavy exercise with oronasal breathing and a minute volume of 50 l/min. For nasal breathing the highest surface fluxes were predicted for the nasal airways. As minute volume increased, formaldehyde penetrated further into the respiratory tract even for nasal breathing. During oronasal breathing (50 l/min) only 46 % of the inhaled air flowed through the nasal airways. The peak fluxes for oronasal breathing occurred at generation 3 of the tracheobronchial region. The major conclusions obtained by this modeling were the following:

– for each activity state more than 95 % of the inhaled formaldehyde is retained by the total respiratory tract,

– in the lower respiratory tract flux will increase for the first generations and then rapidly decreases,

– compared to the pulmonary region fluxes, the tracheal bronchial fluxes are over 1000 times larger and there is essentially no flux in the alveolar sacs.

Franks (2005) developed a mathematical model for the absorption and removal of inhaled formaldehyde in the nasal tissue of humans to predict the proportion of formaldehyde entering the blood. The spatial distributions of the formaldehyde concentration and of the metabolic capacity within the mucosa were used to determine the concentration in the mucus, epithelium, and blood. Modeling a worker at rest with nasal breathing it was found that a steady-state profile was attained within a few seconds of exposure. When an exposure to 1.9 ppm over 8 h was simulated, the increase of the formaldehyde concentration in blood was predicted to be insignificant compared with the preexisting levels in the body. Formaldehyde was efficiently removed by the nasal tissue.

In a recent study Garcia et al. (2009,) studied the interhuman variability in dosimetry of nasal uptake and gas absorption in humans via computational models. Under resting conditions water soluble reactive gases like formaldehyde are mainly absorbed in the anterior portion of the nasal passages and interhuman variability was predicted to be 1.6-fold among the 7 individuals studied. Children and adults displayed very similar patterns of nasal gas uptake and no significant differences were noted between the two age groups.

In summary, models were developed to simulate airflow in the respiratory tract and transfer of formaldehyde to the linings of the airway passages of rats, the rhesus monkey, and humans. These models were based on

- mathematical simulation of regional airflow in the nasal passages by “computational fluid dynamics (CFD)”- partitioning of the nasal surface into “bins” defined by the flux of formaldehyde to these areas

- inclusion of the total human respiratory tract down to the alveolar sacs

- simulation of airflow patters at different inspiratory air flows in humans.

The CFD model reflected the airflow visualized experimentally in nasal molds of the rat and rhesus monkey. Hot-spots for the impact of formaldehyde on the nasal linings could be identified.

The combined airflow modeling showed

- a good match to the high- and low-tumor regions in the rat

- a good match to the distribution of the formaldehyde induced lesions and DPC formation in the rat and monkey in the mucus covered respiratory epithelium

- a high impact of formaldehyde in the nasal vestibule covered with squamous epithelium, but as no tumors are found in rats at this location this tissue seems to be resistant to formaldehyde injury

- a similar wall mass flux of formaldehyde in the rat and monkey at sites of cell proliferation

- an increase of the maximum flux under resting conditions in the order of humans < rats < monkey

- an increase of formaldehyde impact on regions distal from the nose in the order of rats < monkey < humans

- attainment of steady state of formaldehyde tissue disposition within minutes as compared to the hours-long exposure durations

- delivery of the largest portions of air in humans under nasal breathing conditions to the middle and ventral nasal airways

- delivery under oronasal breathing to the nasopharynx of humans as a potential hot-spot by the airflow pattern

- with increasing respiratory airflow in humans a shift of the nasal surface flux distally and an increase of the maximum flux

- for oronasal breathing peak fluxes in humans occur at the 3rd generation of the tracheobronchial region with essentially no flux in the alveolar sacs.

- interhuman variability of nasal uptake was predicted to be 1.6-fold and no significant differences were noted between children and adults.

- when modeling exposure of humans to 1.9 ppm over 8 h the increase of the formaldehyde in blood was insignificant and formaldehyde was efficiently removed by the nasal tissue.

A literature search after the last IUCLID update was carried out up to April 20, 2015 and provided the following new information:

In a supporting study, Conolly et al. (2004) calculated the maximum likelihood estimate for human respiratory tract cancer risk based on 1) the use of computational fluid dynamics to model local impact of formaldehyde, 2) the association of the local impact with DPC formation and cytolethality leading to regenerative cell proliferation, and 3) a two-stage clonal growth model to link DPC and cell proliferation with tumor formation. The model incorporated a hockey stick shaped and a J-shaped dose response relationship for cell proliferation, the latter because this was indicated by the data available at that time. Maximum likelihood estimates of additional risks were calculated for different exposure levels and physical workloads. As in the most recent study of Andersen et al. (2010) no clear indication for a J-shaped dose response was obtained, only risks based on the hockey shape will be given here. For example, for a non-smoking worker with “light work” occupational exposure (80 year lifetime with an environmental exposure of 4 ppb and 40 years of work at 0.3 ppm, 8 h/d, 5 d/week) the additional risk was 1.79x10-7 and for a smoker 4.14x10-6. The additional risks for lifetime exposure at 0.1 ppm are for non-smokers 3.33x10-7 and for a smoker 5.34x10-6. Although the robustness of this model was challenged by Subramaniam et al. (2007, 2008, supporting) and Crump et al. (2008, supporting), it has to be acknowledged that this model was the only one trying to include the wealth of data available for a risk assessment of formaldehyde.

Some refinement of this basic approach have been published without leading to substantial changes of the risk estimates of Conolly et al. (2004). Miller et al. (2011, supporting) took into account the number of cells at risk for becoming cancerous, i.e. the fraction of progenitor cells. They considered the basal and non-ciliated cells in the transitional and respiratory epithelium in the nose and the type II cells in the alveoli to be such cells at risk. By taking into consideration cell cycle times of these cells they concluded that the carcinogenic risks calculated by Conolly et al. (2004) under the assuption the all all cells are at risk would lead to an overprediction of 35% between exposures of 2.5 and 5 ppm formaldehyde. Asgharian et al. (2012, supporting) modeled dosimetry of vapor uptake and tissue disposition of different aldehydes specifically for the lung. They concluded that formaldehyde, the most water soluble of the series under consideration, will be taken up by the upper tracheobronchial airways with only shallow penetration into the lung. Schroeter et al. (2014, supporting) modeled nasal uptake of formaldehyde by taking into consideration its endogenous concentration using the anatomically accurate computational fluid dynamics models of rat, monkey and humans. Flux and tissue uptake of external formaldehyde were not significantly affected by internal levels at exposure concentrations >500 ppb, however at lower concentrations (<10 ppb) predicted uptake was clearly reduced and at <1 ppb even desorption from nasal tussue was predicted for humans.

Two different approaches have recently been described. In the key study by Andersen et al. (2010)  carried out a benchmark calculation for a compartmental pharmacokinetic model that accounted for formaldehyde production, formaldehyde interaction with GSH, and saturable metabolism of the formaldehyde thioacetal to formic acid. Non-linear effects in nasal tissue occur due to saturable metabolism and the finite concentration of GSH. Taking also into account changes in gene expression (see below) they concluded that concentrations of 1 or 2 ppm would not increase the risk of cancer in the nose or any other tissue or affect formaldehyde homeostasis within the epithelial cells.

Starr and Swenberg (2013, key) proposed a bottom-up approach for assessment of low dose human cancer risk from exposure to chemicals that produce the same specific DNA adducts from endogenous and exogenous sources. Taking into account background (endogenous) exposure the approach is consistent with the “additivity to background concept” and provides central and upper bound risk estimates that are linear at all doses. The endogenous and exogenous dG adducts of FA measured in cynomolgus macaques (Moeller at al., 2011) at 2 ppm after two 6 h exposures were taken as a surrogate for humans for continuous life-time exposure. The build-up of adducts was estimated by kinetic modelling of the Starr and Swenberg(2013) rat data yielding an elimination half-life of 63 h. Thereby they arrived at an upper bound life time risk of 3.8x10-4 for continuous exposure at 1 ppm. This risk estimate is nearly 29-fold lower than that calculated by the approach of the US EPA (1.1x10-2). For exposure at the workplace a simple linear modelling would then result in an upper bound risk at 0.3 ppm of 1.6x10-5 (exposure of 5 d/week, 8 h/d over 45 years). The authors noted several reasons why their model should be considered conservative, because for example all background risks for NPC are only ascribed to dG adducts (and not also to the endogenous dA adducts not formed by exogenous FA) or linearity is assumed for all exposure levels without taking into consideration cytotoxicity or cell proliferation enhancing mutations. On the other hand, the half-life of dG adducts has recently been shown to be longer, i.e. 7.1 d about 2.7-fold higher than the half-life used in this extrapolation (Yu et al., 2015).


Male Sprague-Dawley rats were gavaged with 29 mg/kg bw 14C-formaldehyde (concentration ca. 1%) and radioactivity measured in urine and feces for 168 h and in organs and carcass at sacrifice; in additional experiments the excretion via urine, faeces and exhalation was measured within 24 h after gavage. Fasted rats excreted after 24h about 6.7% of applied radioactivity via urine and 1.7% via faeces. After 168h, excretion was 9.4% via urine and 3.2% via faeces. In the carcass 10% were found after 168h and about 2% in organs. In non-fasted rats 8.2% of applied radioactivity was excreted via urine after 24h and 16.6% via faeces. After 168h excretion was 10.8% via urine and 18.3% via faeces. In the carcass 7.3% were found after 168h and 2% in organs. In urine feces and carcass only radioactivity was determined without identification of the chemical entities. Due to the rapid metabolism of formaldehyde the radioactivity cannot be ascribed to formaldehyde per se because of the metabolic incorporation via the C1-poole (see below). Exhalation of formaldehyde was investigated in detail in one non-fasted animal. Until 12h after application 57% of applied radioactivity was exhaled as 14C-CO2. After 24h 63% was exhaled. In this animal excretion via urine and faeces were about 6-7% after 24h. In conclusion, more than 90% of applied radioactivity was absorbed in rats after oral application via gavage; the main route of excretion was exhalation of CO2 (BASF AG, 1983).


See dermal absorption below.



Chang et al. (1983) have analyzed distribution of radioactivity by autoradiography in mice and rats exposed for 6 h to 15 ppm 14C-formaldehyde and sacrificed immediately after the exposure period. Beside widespread distribution high amounts of radioactivity were found in nasal cavity, trachea, lung and gastrointestinal tract. Radioactivity in gastro-intestinal tract is considered to be related to grooming and mucociliary clearance.

The aim of a GLP inhalation study performed by TNO Triskelion (2012) was to investigate whether formaldehyde (FA) can enter the blood stream after single exposure by inhalation and can thus reach the bone marrow or other tissues far from the port of entry. Ten male rats were exposed for six hours and three minutes at a target concentration of 10 ppm in a nose-only exposure chamber. The additional 3 minutes of exposure were due to the time it took to collect a blood sample. To be able to differentiate between exogenous FA and endogenous FA after inhalation exposure, stable isotope labeled FA (13C-FA) was administered to the rats during the inhalation study. After the rats were exposed to the 13C-FA their blood was analyzed to determine total FA, as well as the ratio between labeled and endogenous FA in blood. It was concluded that the inhalation of 13C-FA at 9.65 ppm (target concentration 10 ppm) had no significant effect on the total FA concentration in blood. Exogenous 13CFA was not detected in the blood stream of exposed rats during or after (up to 30 minutes) inhalation exposure (6 hours and 3 minutes) using the experimental conditions of this study. The applied method would have allowed the detection of exogenous 13CFA in blood at a concentration approximately 1.5% of the endogenous FA blood concentration.

In further experiments in rats uptake and disappearance of radioactivity in blood was measured during and after inhalation exposure (8 ppm, 6 h). Plasma concentration of radioactivity increased during the exposure period (peak reached at the end of the exposure period) and slowly declined during a period of several days.

The estimated terminal half-life was 55 h. These observations suggested metabolic incorporation of radioactivity into serum proteins (half-life of these proteins also 2.9 days in rats). The radioactivity in the packed cell fraction of the blood exhibited a multi-phase kinetic profile with a maximum 35 h post exposure. In the terminal phase a very slow decline took place which was consistent with metabolic incorporation of radioactive formaldehyde into the erythrocytes, since the half-life of these cells in rats is about 17 d. After intravenous injection of either radiolabeled formaldehyde or formate quantitatively similar profiles of radioactivity with time were obtained for plasma and cell fraction. These findings show that oxidation of formaldehyde to formate and incorporation of the latter via the one-carbon metabolism determines long-term plasma pharmacokinetics of radioactivity following exposure to radioactive formaldehyde. Therefore it is not possible to draw any conclusion about metabolism or distribution of the formaldehyde moiety after application of radioactive formaldehyde without chemical analysis specifically for this chemical (Heck et al., 1983).

Because of the rapid oxidation to formic acid (see section Metabolism), the concentration of formaldehyde in the blood of humans (exposed to 2 ppm for 40 minutes) and rats (14.4 ppm for 2 h) (Heck et al., 1985) as well as in monkeys (6 ppm, 6 h/day, 5 days/week for 4 weeks; measured after the last exposure; Casanova et al., 1988) was not increased immediately after the exposure period. This might be the cause for the lack of systemic effects after inhalation exposure in chronic studies due to high first-pass effects in the upper respiratory tract.

DPC formation is a good parameter for the local distribution of formaldehyde. Heck et al. (1989) compared DPC formation after a single 6 h exposure in rats at 0.3, 0.7, 2, 6, or 10 ppm and rhesus monkeys at 0.7, 2 or 6 ppm with 14C-formaldehyde. In both species DPC formation increased non-linearly with the airborne concentration. Concentrations of DPC in the turbinates and anterior nasal mucosa were significantly lower in monkeys than in rats by a factor of about 10. In contrast to rats, DPC were also formed in the nasopharynx and trachea of monkeys, clearly at least at 6 ppm while at 2 ppm radioactivity indicative of DPC formation were less than twice the background. No cross-links were found in these tissues in monkeys exposed to 0.7 ppm.

According to Casanova et al. (1991) DPC found at 2 ppm were low but statistically significant in the larynx/trachea/carina and in the major intrapulmonary airways. There was no indication for DPC formation in the sinus, proximal lung, or bone marrow of monkeys. Formation of DPC in nasopharynx and trachea of monkeys is consistent with histopathological findings of the respiratory tract of monkeys showing that formaldehyde can penetrate to deeper regions of the respiratory tract of monkeys than of rats. In contrast to rats breathing only by nose, monkeys may breathe by nose as well as by mouth. Similarly, there was no evidence of tissue damage by histopathology in the sinus of monkeys after exposure to formaldehyde (Monticello et al., 1989). In parallel to the severity of histopathological lesions the concentrations of DPC in the more distal regions of the monkey respiratory tract were generally much less than in the nasal turbinates.

Casanova et al. (1991) developed a pharmacokinetic model for rat to monkey interspecies scaling of DPC formation and extrapolation to humans. The model included saturable and non-saturable elimination pathways (elimination of formaldehyde by metabolic detoxification, formation of other products, and DNA binding) and described regional differences in DNA binding as having an anatomical rather than a biochemical basis. The major predictors of DPC formation were minute volume and quantity of nasal mucosal DNA. By this model, DPC formation in nasal mucosa of adult humans was predicted based on the experimental data in rats and monkeys. The result suggested that formaldehyde would generate lower concentrations of DPC in the nasal mucosa of humans than in monkeys and much lower concentrations than in rats. There was a 2- to 3-fold difference between the fitted concentration-response curve for monkeys and the predicted curve for humans, but that was not statistically significant at concentrations below 1 ppm. The authors conclude that if the concentration of DPC differs between monkeys and humans, it is more likely to be higher in monkeys than in humans. Therefore, risk estimates for humans based on monkey DPC data are probably on the conservative (health-protective) side.

These studies related to DPC formation after a single exposure of naive (previously unexposed) rats and monkeys. Casanova et al. (1994) investigated whether the quantity of DPC may differ in subchronically exposed animals in comparison to naive ones. In addition, the dissection procedure was refined in order to determine DPC yields in different regions of the respiratory epithelium. Male rats were pre-exposed to 0.7, 2, 6, or 15 ppm (6h/d, 5d/week, 11 weeks plus 4 days). Naive rats were exposed to room air. On the 5th day of the 12th week pre-exposed and naive rats inhaled 14C-formaldehyde at the same concentrations as used for pre-exposure over 3 h. DPC yields were measured in the mucosal lining of the nasal lateral meatus (LM) (site of high tumor incidence in the bioassay) and in the medial and posterior meatus (M:PM) (low tumor site). DPC yields in the LM were approximately 6-fold higher than in the M:PM. At 0.7 and 2 ppm there was no difference between the pre-exposed and naive rats in either tissue. At 6 and 15 ppm the DPC yields in the pre-exposed rats were approximately half of those of naive rats in the LM, but no differences were detected in the M:PM. These data were referred to “acute DPC yield” as they related only to DPC formation caused by the last 14C-formaldehyde exposure in naive and pre-exposed rats at the end of the experiment. In addition “cumulative DPC yield” was determined by measuring interfacial DNA after exposure to 6 and 10 ppm. Thereby the total DPC accumulating over the whole exposure period could be obtained, but with a lower sensitivity. “Cumulative DPC yields” increased in a concentration-dependent manner in both naive and pre-exposed rats, but the yields were smaller in the pre-exposed than in the naive rats. Thereby it was demonstrated that no accumulation of DPC occurred by pre-exposure over 12 weeks and it implies that cumulative and acute DPC yields of rats exposed to formaldehyde are essentially identical. DPC produced in any single day’s exposure must be completely or almost completely removed by the time of the next day’s exposure, i.e. within at most 18 h.

Site specificity of DPC yield and its dependence on pre-exposure history were modeled by Heck, Casanova (1995) using the approach of Casanova et al. (1991). Basically, the model describes the absorption of formaldehyde into tissue and its elimination as metabolites and other products, including DPC, as a function of kinetic parameters. The most important compartment (site) specific parameters are the quantity of DNA in and the formaldehyde flow into the compartments. Pre-exposure is not assumed to affect the kinetics of elimination (Casanova-Schmitz et al., 1984) nor the flow to different compartments. On the other hand, the amount of DNA is increased at the site of high tumor yield (lateral meatus) at pre-exposures of =/>6 ppm by induction of hyperplasia and squamous metaplasia. This increase of DNA does not occur at low exposure concentrations (=/<2 ppm) nor at low-frequency tumor sites (medial and posterior meatus). This modeling confirmed that in naive rats DPC yield at high tumor sites is by a factor of 5-6 higher than that at low tumor sites and that DPC formation is a non linear function of the airborne concentration. In pre-exposed rats at up to 2 ppm DPC yields at different sites correspond to those to those in naive rats. But at pre-exposure concentrations of 6 ppm and above DPC yields were smaller than those of naive rats at high tumor sites because of the larger quantity of DNA. In conclusion, this model shows that site specificity of DPC formation is driven by local airflow characteristics and not by specific differences in formaldehyde metabolism.

Casanova, Heck (1987) investigated the effects of glutathione depletion on metabolic incorporation and DPC formation by the 3H/14C ratio. Male rats were exposed for 3 h to 3H and 14C-formaldehyde at concentrations of 0.9, 2, 4, 6, or 10 ppm one day after a single 3 h pre-exposure to the same concentrations of unlabelled formaldehyde. Two hours prior to the second exposure the animals were injected with phorone or corn oil. Phorone treatment decreased the concentration of non-protein sulfhydryl in the nasal respiratory mucosa down to 10% of baseline control corn oil treatment. This decrease of glutathione should decrease formaldehyde metabolic detoxification by formaldehyde dehydrogenase. In accordance with this assumption metabolic incorporation of radioactive formaldehyde into DNA, RNA and proteins in the respiratory and olfactory mucosa and in bone marrow was significantly decreased. Correspondingly, DPC formation was significantly increased in the respiratory mucosa of glutathione depleted rats at all formaldehyde exposure concentrations. On the other hand, covalent binding of formaldehyde to macromolecules in the bone marrow was not detected, neither in DNA, RNA, or protein. The results show that glutathione-dependent oxidation of formaldehyde catalyzed by formaldehyde dehydrogenase is an important defense mechanism against DPC formation and that even after glutathione depletion there is no covalent binding of formaldehyde in the bone marrow.

In summary, in rats DPC formation only occurs in the nasal respiratory mucosa, but not in the olfactory mucosa or bone marrow. In monkeys DPC concentrations in the nasal respiratory mucosa are by a factor of about 10 less than in rats. Pharmacokinetic modeling for interspecies scaling indicates that DPC concentrations in humans may even be less as compared to monkeys. While DPC formation in rats is confined to the respiratory epithelium of the nose, in monkeys it descends down to the nasopharynx and trachea but in much smaller quantities. DPC were not found in the sinus, proximal lung, or bone marrow of monkeys. A NOEC could not be determined down to the lowest exposure concentrations investigated, i.e. 0.3 ppm in rats and 0.7 ppm in monkeys. The exposure response curve is highly non-linear and in rats the slope below 2 ppm is much smaller than at higher concentrations. There is no accumulation of DPC by prolonged exposure and DPC are substantially removed by one day after exposure. In regions of high tumor incidence DPC yield is decreased after prolonged pre-exposure to high formaldehyde concentrations in comparison to that of naïve rats. DPC formation is driven by local airflow characteristics and not by specific differences in formaldehyde metabolism.

A literature search after the last IUCLID update was carried out up to April 20, 2015 and provided the following new information:

Kleinnijenhuis et al (2013, key) performed an inhalation experiment with FA in rats in order to study whether FA can enter the blood and thus cause systemic toxicity in remote tissues. To differentiate between exogenous and endogenous FA, the rats were exposed (10 ppm for 6 hours) to stable isotope 13C-labelled FA by inhalation. During and after exposure, blood was analysed to determine the ratio between labelled and endogenous FA in blood and the total blood concentration of FA. With the method applied, exogenous 13C-FA could have been detected in blood at a concentration approximately 1.5 % of the endogenous FA blood concentration. However, exogenous 13C-FA was not detectable in the blood of rats, neither during nor up to 30 min after the exposure. It was concluded that the inhalation of FA, even at 10 ppm for 6 hours, did not result in an increase of the total FA concentration in blood.


In rats and guinea pigs significant proportions of the dose (ca. 15% at 0.1 or ca. 4% at 11.2 mg/animal) are found after 72 h at the site of application and in the remaining carcass (approx 25-30%, independent of dose). 6-8% of the applied radioactivity was excreted via urine and faeces (mainly urine). The large majority of the applied dose is lost to evaporation. During the collection period fairly constant but low blood concentrations were measured (ca. 0.1% in total blood of applied dose). There were no significant differences between the high and low dose level except the amount of radioactivity at the site of application (much lower at the high dose level). There is good evidence that the skin of the monkey is less permeable to formaldehyde than that of rodents using a similar experimental design (Jeffcoat et al., 1983).



The enzymatic oxidation of formaldehyde results in detoxification and protects from elevated endogenous and exogenous formaldehyde concentrations. There exist several enzyme systems able for oxidation of formaldehyde (BfR, 2006; summary reviews). Formaldehyde reacts spontaneously and non-enzymatically with glutathione, the product of this reaction is S-hydroxymethylglutathione. This adduct is predominant in vivo, since circulating glutathione concentrations are reported to be 50 times those of formaldehyde in humans (Sanghani et al., 2000). On the basis of the glutathione content in animal cells and the dissociation constant, Uotila, Koivusalo (1989) calculated that the majority (50-80%) of formaldehyde in animal cells will be reversibly bound to glutathione. Furthermore, it was shown, that about 90% of glutathione must be depleted before the activity of FAD becomes impaired (Dicker, Cederbaum, 1986).

The importance of the glutathione status in rat nasal mucosa was described by Casanova, Heck (1987). Male rats were exposed to 3H-and 14C-formaldehyde at 0.9, 2, 4, 6, or 10 ppm for 3 h. Two h prior to exposure, the animals were injected either with phorone or with corn oil. Phorone injection led to a decrease of non-protein sulfydryls in the nasal respiratory mucosa to 10 % of that of the corn oil injected rats. Thereby a marked decrease in formaldehyde metabolism occurred, leading to the following effects: The metabolic incorporation of radioactive formaldehyde into DNA, RNA, and proteins in the respiratory and olfactory mucosa and in the bone marrow (femur) was significantly decreased. In contrast, the concentration of DNA protein cross links was significantly greater in glutathione-depleted than in glutathione-normal rats at all airborne concentrations. By the method employed no DNA-protein cross links were detected in undepleted rats at 0.9 ppm, but cross linking did occur at this concentration in depleted rats. In addition the amount of radioactive formaldehyde covalently bound to protein was calculated. Depletion of glutathione had no measurable effect on the quantity of formaldehyde covalently bound to proteins.

The glutathione status has a marked effect on formaldehyde detoxification and in consequence on formaldehyde toxicity. On the other hand, formaldehyde at concentrations under discussion here does not lead to a reduction of glutathione in the respiratory tract. Repeated exposure of rats to formaldehyde (15 ppm, 6h/d, 10 d) did not affect the concentration of non-protein sulfydryls in respiratory mucosal homogenates of animals directly sacrificed after exposure. Furthermore, the specific activity of FAD was not affected by repeated formaldehyde exposure (Casanova-Schmitz et al., 1984).

Cassee, Feron (1994) exposed rats during 6 consecutive periods of 12 h for 8 h to 3.6 ppm formaldehyde followed by a 4 h period of non exposure. This treatment regime had no effect on aldehyde dehydrogenase with formaldehyde as substrate, on formaldehyde dehydrogenase (FAD) using formaldehyde and glutathione as substrates, or on cytosolic glutathione levels all measured in the nasal respiratory epithelium. In summary, glutathione depletion markedly decreases formaldehyde metabolism and increases formaldehyde toxicity. On the other hand, formaldehyde inhalation even at high toxic concentrations does not lead to a reduction of glutathione or enzymatic activity in the respiratory tract.

S-hydroxymethylglutathione is oxidized by formaldehyde dehydrogenase (FAD) to form S-formylglutathione with concomitant reduction of NAD+ (Uotila, Koivusalo, 1996).

The cytosolic formaldehyde dehydrogenase (FAD) is present in all animal tissues tested (OECD, 2002; summary reviews). Also in humans it has been identified in different organs and tissues, e.g. liver, brain, oral mucosa, or erythrocytes (BfR, 2006; summary reviews).

As the glutathione-dependent FAD plays a key role in the detoxification of endogenous formaldehyde, it is highly conserved in many tissues of different animals (Uotila, Koivusalo, 1989). Uotila & Koivusalo (1996) found FAD and S-formyl glutathione hydrolase (for the subsequent hydrolysis, see below) in all the 16 rat tissues they investigated and concluded that these enzymes may be ubiquitous in animal cells. The specific activities varied 12- to 30-fold in the tissues investigated. Liver, kidney, stomach, colon, and small intestine were among the richest sources for these enzymes (the nasal epithelium was not investigated).

In the respiratory tract of dogs FAD activity appeared to be highest in the nose and lowest in distal bronchi, lung, and liver parenchyma (Maier et al., 1999) indicating to a very efficient defense in nasal epithelium of this species.

According to Höög et al. (1994) the abundant occurrence of FAD probably reflects the need for scavenging of formaldehyde in cytoprotection. Estonius et al., (1996) assign a house keeping character to this enzyme due to its ubiquitous expression and its important role in formaldehyde scavenging.

In contrast to rats, humans are not obligatory nose breathers and especially under physical activity the oral mucosa may be exposed to inhaled formaldehyde. Therefore, Hedberg et al. (2000) investigated the glutathione-dependent FAD in oral tissue specimens and cell lines. The mRNA of FAD was expressed in basal and parabasal cell layers of oral epithelium, whereas the protein was detected throughout the cell layers. m-RNA and protein were further detected in homogenates of oral tissues and various oral cell cultures, including normal, immortalized, and tumor keratinocyte cell lines. The mRNA half life was 7 h, but a decay of the protein was not observed throughout a 4-day period in normal keratinocytes. The protein content correlated to the oxidizing activity for the S-hydroxymethyl glutathione substrate. A substantial capacity for formaldehyde detoxification was shown, indicating that FAD is the major enzyme involved in formaldehyde oxidation in oral mucosa

Keller et al. (1990) localized histochemically formaldehyde dehydrogenase in the rat. Beside the expected localizations in liver, kidney, or brain this enzyme was detected in the respiratory and the olfactory epithelium of the nasal cavity. But the qualitative study data are not sufficient for explaining the differences in the affected regions in the nose after formaldehyde inhalation (lesions predominantly in the respiratory epithelium and less in the olfactory epithelium of the nasal cavity; differences might be related to the different local concentrations of formaldehyde in the nasal passages).

Rats were exposed to 15 ppm for 6 h per day for 10 days and thereafter nasal mucosa samples collected. The rate of oxidation of formaldehyde in the presence and the absence of glutathione (GSH) was determined in these tissues. In the absence of glutathione formaldehyde oxidation in tissues of unexposed rats was catalysed by a single enzyme, considered to be an isoenzyme of aldehyde dehydrogenase. In the presence of GSH a second enzyme catalysed the oxidation of formaldehyde considered to be formaldehyde dehydrogenase. The specific activity of both enzymes was similar in tissues of unexposed and exposed rats. Glutathione (GSH) is a cofactor for formaldehyde dehydrogenase. GSH and formaldehyde react reversibly and form S-hydroxymethylglutathione and this adduct is the true substrate for formaldehyde dehydrogenase. Formaldehyde dehydrogenase in the presence of GSH is more effective in oxidation of formaldehyde than aldehyde dehydrogenase. GSH dependent oxidation of formaldehyde catalysed by formaldehyde dehydrogenase is an important defence mechanism against the formation of covalent binding of formaldehyde to DNA. Repeated exposure of rats to 15 ppm formaldehyde over 10 days did not change the specific activities of FAD, aldehyde dehydrogenase or the nonprotein sulfhydryls in the respiratory mucosa and thus the detoxification of formaldehyde is not affected by high formaldehyde exposures (Casanova-Schmitz et al., 1984; Casanova & Heck, 1987).

Data on repeated dose toxicity have shown a sharp increase in toxicity with increasing exposure. Lesions of the epithelium in the nasal cavity can be observed at formaldehyde concentration >= 6 ppm (Section Repeated dose toxicity: dermal). This is in accordance with pharmacokinetic data presented by Casanova et al. (1989). The detoxification pathway (mainly via formaldehyde dehydrogenase) in rats is half saturated at an airborne formaldehyde concentration of 2.6 ppm (Casanova et al., 1989).

Formaldehyde dehydrogenase catalyses the formation of S-formyl-glutathione (see above). In a further step the enzyme S-formyl-glutathione hydrolase catalyses the hydrolysis of S-formylglutathione to formic acid and glutathione (Uotila & Koivusalo, 1996). S-formyl-glutathione hydrolase is suggested to have an ubiquitous tissue distribution and a higher activity than formaldehyde dehydrogenase (16 different rat tissues examined, all showed a high activity; Uotila & Koivusalo, 1996). When comparing the activities of FAD and S-formylglutathione hydrolase, S-formylglutathione hydrolase was by a factor of 600 – 2000 more active than FAD. The hydrolysis of S-formylglutathione to formic acid and glutathione is also catalysed by glyoxalase II (ubiquitous tissue distribution is suggested) but the activity of this enzyme is lower than that of S-formylglutathione hydrolase (10-30 fold; Uotila & Koivusalo, 1996). Formic acid can be excreted via urine as its sodium salt, or oxidized to CO2, which is exhaled. As formate, an uptake into the carbon one metabolic pathway is possible (BfR, 2006; see also Figure documented in attachment of “summary reviews”).

As mentioned above formaldehyde dehydrogenase in the presence of GSH is more effective in oxidation of formaldehyde than aldehyde dehydrogenase (Casanova-Schmitz et al., 1984; Casanova & Heck, 1987).

In the study of Uotila & Koivusalo (1997) it has been also shown that the specific activity of formaldehyde dehydrogenase is in most tissues higher than that of aldehyde dehydrogenase.

The oxidation of formaldehyde by catalase is a possible pathway which becomes important after depletion of glutathione (BfR, 2006; summary reviews).

Uotila & Koivusalo (1989) reviewed the different enzyme systems for the metabolism of formaldehyde. The following enzymes of animal tissues are known to catalyze in vitro the removal of formaldehyde: The glutathione-dependent cytosolic FAD described above, non-specific aldehyde dehydrogenases, catalase, and tetrahydrofolate-dependent reactions. The latter two have probably only minor quantitative significance for the removal of formaldehyde. Besides the cytosolic FAD, the mitochondrial “low-Km” isoenzyme of non-specific aldehyde dehydrogenase may be of some importance. Overall, Uotila & Koivusalo (1989) conclude, that the glutathione-dependent formaldehyde dehydrogenase (FAD) is the major pathway of formaldehyde utilization in animal tissues in vivo.

Further biological pathways (BfR, 2006; summary reviews)

1) Reversible binding of formaldehyde to cysteine resulting in the formation of thiazolidine-4-carboxylate.

2) Formaldehyde can be reversible bound to urea to form hydroxymethyl adducts.

3) Reversible binding to a protein like albumin or mucus proteins in the nasal or oral mucosa.

4) Irreversible reaction with two proteins resulting in protein-protein-cross-links.

5) Irreversible reaction with one proteins and DNA resulting in DNA-protein-cross-links (DPC).

6) Non-enzymatic binding to tetrahydrofolic acid followed by an uptake into the carbon one metabolic pathway and incorporation into biological macromolecules (synthesis of purine, thymidine and certain amino acids).

Formaldehyde bound to tetrahydrofolate was estimated not to be more than 10 % of the formaldehyde analyzed in liver samples (Heck et al., 1982). Therefore, tetrahydrofolate adducts were not considered to be a major source of endogenous formaldehyde.

These multiple metabolic pathways lead to a rapid oxidative detoxification of formaldehyde that is extremely important for cell survival as formaldehyde is a highly reactive and toxic normal constituent in all living tissues. Rietbrock (1965; 1969) reported a detoxification of formaldehyde to formic acid with a half-life of 1 min in the dog, cat, rabbit, guinea pig, and rat. This was later confirmed by McMartin et al. (1979) in monkeys after intravenous infusion of radioactive formaldehyde and determination of the unmetabolized formaldehyde by the dimedon method. Formaldehyde was rapidly eliminated from the blood with a half-life of about 1.5 min.

For the assessments of health effects caused by exposure to formaldehyde the question of polymorphism of the detoxifying enzyme systems must be addressed. Polymorphism of the glutathione-dependent FAD being a pivotal protection against endogenous formaldehyde in all living systems seems a priori relatively improbable. This detoxifying pathway is highly conserved not only in animals but also in plants and yeast.

Castle & Board (1982) analyzed formaldehyde dehydrogenase by electrophoresis and did not find any genetic variants in 249 human liver biopsy samples from Australian Caucasians, Chinese, and Indians. Uotila & Koivusalo (1987) analyzed red cell hemolysates from non-related Finns by electro focusing on polyacrylamide gel and located formaldehyde dehydrogenase by an activity-staining method. Three forms of the enzyme were constantly found for all 217 individuals and no variants were observed in this population. From a number of individuals diluted human liver cytosole was investigated in addition to the hemolysates. The liver samples also gave three formaldehyde dehydrogenase activity bands with identical locations. In the largest study blood samples were analyzed by iso-electric focusing on polyacrylamide gels and stained for FAD activity. Three activity bands were observed with blood samples from 200 Koreans, 69 Singapore Chinese, 160 Hungarians, and 171 Germans. No distinct intra- and interpopulation differences were found in the intensity of the 3 bands (Benkmann et al., 1991). 

The importance of an efficient enzymatic detoxification system becomes obvious when considering the large amounts of formaldehyde produced endogenously throughout the whole body. Cascieri and Clary (1992) calculated the endogenous production of formaldehyde to 2450 mg/h for an adult human. They assumed a half-life of formaldehyde in blood of about 1.5 min, an equal distribution in aqueous body fluids, A body weight of 70 kg for an adult corresponding to 49 kg of water, and an equilibrium level of formaldehyde in the aqueous system of the body of 2.5 mg/l. The hepatic detoxification capacity of adult humans was calculated to be about 22 mg of formaldehyde converted to CO2 in a minute (i.e. 1320 mg/h) (Owen et al., 1990).

In summary, glutathione-dependent cytosolic formaldehyde dehydrogenase (FAD) is probably the most efficient detoxifying enzyme system. FAD is highly conserved in yeasts, plants and animals. It was found in all tissues investigated, including the respiratory tract and nasal respiratory and olfactory mucosa. There are other non-specific aldehyde dehydrogenases that may contribute to the metabolic detoxification of formaldehyde, the most important one being probably the “low-Km-“mitochondrial aldehyde dehydrogenase. These enzyme systems lead to a rapid metabolism of formaldehyde with a biological half life of 1-1.5 min. The specific activity of FAD is not affected by pre-exposure to high concentrations of formaldehyde. The nuclear membrane is the last barrier to a genotoxic/mutagenic action of formaldehyde. In a large number of individuals with different ethnic origin no indication for polymorphism of the glutathione-dependent FAD protein was found.

A literature search after the last IUCLID update was carried out up to May 2022 and provided the following new information:

Further studies are available on the variability and polymorphism of formaldehyde dehydrogenase (FAD = ADH5 according to the new nomenclature) and on polymorphism of glutathione transferases.

The former studies mentioned above included in total more than 1000 samples from human donors without an indication for polymorphism of ADH5 on the protein level. A recent study with DNA analyses from Wu et al., (2007, supporting) identified two single nucleotide polymorphisms that, according to the authors, might possibly be association with childhood asthma. And Hedberg et al. (2001, supporting) found a polymorphism in promoter region with reduced transcriptional activity in vitro. But as the biological meaning of these polymorphisms for formaldehyde related toxicity and genotoxicity remained unclear, further studies were carried out by the group of Speit.

Just et al. (2011, supporting) investigated 3 polymorphisms of ADH5 in the blood of healthy German volunteers. The polymorphism of Hedberg et al. (2001) was not detected in 150 subjects and another polymorphism described in literature was not detected in 70 subjects. A third polymorphism was identified in 105 subjects: 43 were heterozygous, 46 homozygous for one allel and 16 homozygous for the other allel. As the comet assay with blood samples of homozygous subjects showed no difference in strand breaks or DPX formation, i.e. no influence on in vitro genotoxicity of formaldehyde, no biologically relevant polymorphisms of the ADH5 gene could be identified. Formaldehyde exposure did not lead to alterations of ADH5 expression in human volunteers at concentrations up to 0.7 ppm or 0.4 ppm plus peaks of 0.8 ppm (Zeller et al., 2011a; details of exposure see Mueller et al., (2013, key).

No differences for inter-individual susceptibility could be identified with 30 male smokers, 30 female non-smokers and 30 school children when leukocytes were incubated with formaldehyde. The endpoints studied included in vitro formation and removal of DPC by the Comet assay, in vitro induction and persistence of SCE and expression of mRNA levels of the ADH5 gene by real-time PT-PCR. In addition there was no association of GSTM1 and GSST1 polymorphism with in vitro genotoxicity (Zeller et al., 2012, supporting). When the leukocytes of the volunteers of the Mueller et al. (2013) study were subjected to the same in vitro battery of tests no differences were identified for the subgroups hyper- and hyposensitive persons to CO2 induced nasal irritation (Zeller et al., 2011b, supporting).




Rats were exposed for 6 h to 0.63 or 13.1 ppm 14C-labelled formaldehyde. Immediately after exposure the rats were transferred in metabolism cages (collection of expired air, urine, faeces) for 70 h. Most of radioactivity was excreted via exhalation (ca. 40%). Nearly the same amount remained in the tissues and carcass but this cannot be ascribed to the formaldehyde moiety. 17% were excreted via urine and 5% via faeces. The exposure concentration had no influence on the relative amount in the different fractions. The excretion via expired air was multiphasic with an initial high rate of exhalation, which declined rapidly over a period of 12 h and followed by a much slower phase (Heck et al., 1983)

Riess et al. (2010, key) developed a highly specific method based on the acetylacetone derivatisation. The exhaled air of 10 volunteers was analysed. Prior to measurement the subjects exercised on an ergometer for 35 min at a resistance up to 50% of their maximal perceived exertion. In addition they were asked to eat methanol rich fruits and to drink methanol rich alcoholic beverages. Formaldehyde concentrations in exhaled air did not exceed 2 ppb. Fuchs et al. (2010, key) determined formaldehyde by an independent approach by GC-MS after derivatisation to the stable oxime. Lung cancer patients, smokers, and healthy subjects (each 12) were studied. The mean expired formaldehyde concentrations after correction for its concentration in the inspired air was 1,582 ppb, similar to the results obtained by Riess et al. (2010, key).


In gavage studies (BASF, 1983) fasted rats excreted after 168h about 9.4% of applied radioactivity via urine and 3.2% via faeces. In the carcass 10% were found after 168h and about 2% in organs.

In non-fasted rats 10.8% of applied radioactivity was excreted via urine after 168h and 18.3% via faeces, presumably due to binding to corresponding components of the diet. In the carcass 7.3% were found after 168h and 2% in organs. Exhalation of formaldehyde was investigated in detail in one non-fasted animal. After 24h 63% was exhaled. In this animal excretion via urine and faeces were about 6-7% after 24h.


In humans as well as in animals formaldehyde is an essential metabolic intermediate and found in all tissues of animals and humans. The concentrations of formaldehyde stemming from endogenous production is exhaled air is in the range of 1 ppb.

Formaldehyde is absorbed and deposited after inhalation in the upper respiratory tract, the site of first contact. The localisation of uptake in each species is determined by nasal anatomy, mucus coating and clearance mechanisms. The overall uptake by the nasal passages at resting respiratory rates has been predicted to be 90% in rats, 67% in monkeys, and 76% in humans. For the total respiratory tract the uptake in humans was estimated to be more than 95%. The physiological level of formaldehyde in the blood of humans and experimental animals is not increased after inhalation exposure due to the rapid metabolism (t ½ ~ 1 min) and reactivity at the site of first entry. This might be the cause for the lack of systemic effects after inhalation exposure due to high first-pass effects.

Formaldehyde is rapidly and nearly completely absorbed from the intestinal tract after oral exposure. After dermal exposure in rats and guinea pigs ca. 40% of the applied formaldehyde is absorbed via the skin; the skin of monkeys is less permeable.

The data available on metabolism and biological pathways of formaldehyde are summarized in the Figure presented in an attachment (BfR, 2006; summary reviews). The oxidation of formaldehyde catalysed by formaldehyde dehydrogenase (ADH5) (see Figure) is considered to be the main defense mechanism against the formation of covalent binding of formaldehyde to macromolecules like proteins or DNA. After depletion of glutathione also other oxidation pathways, catalysed by aldehyde dehydrogenase and catalase, may become important (not shown in the Figure). But glutathione depletion is not to be expected even at high inhalative exposures to formaldehyde.

Formaldehyde reacts spontaneously and non-enzymatically with glutathione to form S-hydroxymethylglutathione (1; see Figure). In the presence of NAD+S-hydroxymethylglutathione can be converted to formylglutathione (2) catalysed by formaldehyde dehydrogenase (FAD). FAD is highly conserved in all species, is found in all tissues and may be regarded as a house-keeping enzyme necessary for protection against the highly toxic endogenous formaldehyde. The production of endogenous formaldehyde was estimated to be 2450 mg/h for an adult human. EFSA (2014) estimated that endogenous turnover of formaldehyde was estimated to be approximately 0.61-0.91 mg/kg bw per minute and 878-1310 mg/kg bw per day (corresponding to about 60-90 g/adult person per day) assuming a half life of 1-1.5 min. In the presence of water formylglutathione is cleaved by S-formylglutathione hydrolase to glutathione and formic acid (3). Formic acid can be excreted as its sodium salt via urine or oxidized to CO2and exhaled. As formate an uptake into the carbon-1-metabolic pathway is also possible [see also (9)].

The most important enzyme for detoxification of formaldehyde is ADH5 (FAD) found in all tissues of all species. No polymorphism has been identified on the protein level. Although some polymorphisms have been described on the DNA level, it has been shown that these do not affect the efficiency of formaldehyde detoxification or protection against DNA damage, apart from patients with Fanconi Anemia (see below).

Other biological pathways: Formaldehyde can be reversibly bound to cysteine to form thiazolidine-4-carboxylate (4). Formaldehyde reacts reversible with urea or protein to form hydroxymethyladducts (5) or protein-adducts (6), respectively. Irreversible reaction with two proteins results in protein-protein-cross-links (7) and with DNA and protein in DNA-protein-cross-links (8) (see Section Genotoxicity). The uptake into the one-carbon metabolic pathway is possible after non-enzymatical binding of formaldehyde to tetrahydrofolic acid (9). In this activated form formaldehyde is an essential intermediate for the synthesis of purine, thymidine and certain amino acids which are incorporated in proteins and DNA.

PBPK modelling taking account of nasal airflow pattern, local DPC formation, and a clonal 2-stage growth model for tumor formation calculated a maximum likelihood cancer risk for the workplace at 0.3 ppm and for environmental exposure at 0.1 ppm in the range of 10-6– 10-7. Penetration of inhaled formaldehyde into the lung is negligible for humans. Similar results were obtained in more recent assessments based on formaldehyde detoxification by GSH interaction and changes of gene expression or the relationship of DNA adducts formed by endogenous vs. exogenous formaldehyde.

Toxicokinetics, metabolism and distribution

Formaldehyde is an essential metabolic intermediate in humans as well as in animals. After inhalation formaldehyde is absorbed and deposited in the upper respiratory tract, the site of first contact. The localisation of uptake in each species is determined by nasal anatomy, mucus coating and clearance mechanisms. The overall uptake by the nasal passages at resting airflow rates has been predicted to be 90% in rats, 67% in monkeys, and 76% in humans. The physiological level of formaldehyde in the blood of humans and experimental animals is not increased after inhalation exposure due to its rapid oxidation to formic acid and reactivity at the site of first contact.

After oral exposure formaldehyde is rapidly and nearly completely absorbed from the intestinal tract of rats and mice.

After dermal application to rats and guinea pigs ca. 40% of the applied formaldehyde is absorbed via the skin and in monkeys 15%.

Formaldehyde reacts spontaneously and non-enzymatically with glutathione to form S-hydroxymethylglutathione.

In the presence of NAD+, S-hydroxymethylglutathione can be converted to formylglutathione catalysed by formaldehyde dehydrogenase (FAD). In the presence of water, formylglutathione can be cleaved by S-formylglutathione hydrolase to glutathione and formic acid. Formic acid can be excreted as its sodium salt via urine or oxidized to CO2 and exhaled. As formate an uptake into the carbon-1-metabolic pathway is also possible.

In inhalation studies in rats using 14C-labelled formaldehyde 40% of applied radioactivity was excreted within the following 70 hours via exhalation, 17% via urine and 5% via faeces. Oral studies have shown that ca. 60% of the applied radioactivity were exhaled as CO2 within 12 h after gavage and minor amounts via urine and faeces.

Data on toxicokinetics were also evaluated in a recent review by IARC (2006); a summary is presented in IUCLID Section 7.1.1.

Dermal absorption

Jeffcoat et al. (1983) investigated dermal absorption of 14C formaldehyde in rats, guinea pigs and monkeys by open application. This mode of exposure corresponds to the requirement of the OECD Guidance Document for the conduct of skin absorption studies (OECD Series on Testing and Assessment, No. 28, March 25; 2005) that the experimental procedure should mimic normal exposure conditions as for example in production and downstream use of formaldehyde. For volatile chemicals evaporation after skin contact must be taken into account. Basically the test design followed OECD TG 427 (adopted April 13: 2004) (e.g. use radioactive test compound, determination of radioactivity in urine, feces, expired air, carcass and application site in metabolism cages up to 24 h) with some major deviations:

- the specific application device proposed by OECD was not used that would have allowed trapping and determination of evaporated test substance by charcoal

- for rats and guinea pigs only total radioactivity and CO2 could be determined in air traps, but evaporated formaldehyde could not be determined analytically

- for monkeys only exhaled CO2 could be determined in the air traps. Evaporated formaldehyde was not measured

- no determination of radioactivity in the carcass of monkeys.

- total recovery of 100 +/-10 % was not reached.

But in spite of these deviations, the amount of formaldehyde systemically absorbed and that remaining at the application site in rats and guinea pigs can reliably be estimated from this investigation. As for monkeys radioactivity was not determined in the carcass, an estimate of the absorption is only possible by comparing the amounts excreted in the 3 species (see below). Overall, this investigation enables a determination of dermal absorption of formaldehyde.

Rats and guinea pigs were kept in individual metabolism chambers for collection of urine, feces and a combination of expired air and evaporated products. Aqueous solutions were applied to a 2-cm2 shaved area of the lower back at dose levels of 0.1 mg formaldehyde in 10 µl (1%) or 11.2 mg in 40 µl (28%) containing approximately 30 µCi 14C formaldehyde. Blood samples were collected via a catheter from the carotid artery at 1, 2, 3, 4, 7, and 24 h after dosing. Air was pulled through the chambers and passed through traps containing 0.5 N NaOH. Traps were changed each time a blood sample was taken and at 48 and 72 h after dosing. After 72 h the animals were sacrificed and selected tissues were removed. Radioactivity was measured in blood, excreta, tissues, air traps and carcass. Air trap contents was analyzed for formaldehyde by reaction with 2,4-dinitrophenylhydrazine and for CO2.

Monkeys were placed in restraining chairs with a Plexiglass hood around the head fitted snugly to the neck. Air was drawn from the hood and trapped with 0.5 N NaOH for collection of CO2. An aqueous solution containing 2 mg radioactive formaldehyde in 200 µl (1%) was applied to an 18-cm² area on the lower portion of the monkeys posterior. Samples were collected and analyzed as described for rats and guinea pigs.

The mean values (% of applied dose) for the distribution of 14C during 72 h after application are given in the table. Radioactivity in blood and selected tissues is not listed; at maximum it only slightly exceeded 0.1 % of the dose.



Dose (mg)

Air traps



Skin (appl. site)

Remaining carcass

Total recovery









Guinea pig
















Guinea pig














Not analyzed



Total recovery only averaged to 67 % for rats and guinea pigs and this was ascribed by the investigators to evaporation of formaldehyde during the time between application of the test material and placement of animals into the metabolism chamber. For monkeys total recovery was even lower with only 10 %. According to the authors this very low recovery was mainly due to the method for trapping volatiles (only CO2 trapped) and because radioactivity was not determined in the remaining carcass; the experimental design used for monkeys only enabled trapping of radioactive CO2, but not of evaporated formaldehyde. But this low recovery does not invalidate the dermal absorption rate calculated by radioactivity found in the major determinants for dermal absorption, namely in urine, feces, carcass, exhaled air (as CO2), and application site for rats and guinea pigs. For monkeys a rough estimate for the remaining carcass can be made by comparing the amounts excreted by the different species. Thereby the major determinants for calculation of absorption are also covered for monkeys enabling a semi-quantitative estimate of the total absorption in this species, too.

For guinea pigs about 21-24% of the applied dose was collected in the air traps, most in the first 2 h. Only <3% of the trapped radioactivity was present as CO2, the rest would have been evaporated test material but analysis specifically for formaldehyde was not possible due to its instability in NaOH. About 8% was excreted in urine and feces. No large accumulation of radiolabel occurred in any internal organ and in blood the average concentration was about 0.1% of applied dose. At the application site about 16% (0.1 mg dose) and 4% (11.2 mg dose) were recovered, and in the remaining carcass 27-28%. Generally there was not much of a difference between the high and low doses apart from the application site, where the radioactivity recovered was clearly lower for the high dose. Total absorption including application site was about 52% (low dose) and 43% (high dose), and excluding application site 36% (low dose) and 39% (high dose) if approximately 3% for CO2 in expired air are added. Total recovery was in the range of 64-70%.

A similar pattern was observed with rats administered the same amounts of formaldehyde. Total absorption including application site was about 53% (low dose) and 41% (high dose), and excluding application site 32% (low dose) and 38% (high dose) if approximately 3% for CO2 in expired air are added. Total recovery was in the range of 60-73%.

The skin of the monkey is less permeable than that of rats or guinea pigs. Monkeys showed a much higher animal to animal variation that, according to the authors, might be expected from the greater heterogeneity of their breeding. Therefore, the quantitative data for monkeys may not be as reliable as those for the other two species. But nevertheless the data show qualitatively less penetration through monkey skin.

The major findings in monkeys were:

- the sum of excreta (air, urine, feces) were less than 1% of the applied dose

- radioactivity in all organs examined totaled less than 0.05% of applied dose and was fairly evenly distributed

- less than 2% (mean about 0.4%) of applied dose was trapped as CO2, i.e. about 10 % of the amount collected from rodents (but in rodents the trapped radioactivity also contained evaporated formaldehyde)

- excretion of radioactivity in urine and feces was about 10% of that excreted by rodents

- the dose remaining at the site of application was 9.5%, intermediate between that of rodents receiving 0.1 and 11.2 mg formaldehyde

- the fractional dose in the blood circulation was also only about 10% of that observed in rodents.

In monkeys total absorption could not be determined because the total remaining carcass was not analyzed. For the remaining carcass a valid comparison to rodents can be made if only the amounts of radioactivity in excreta (urine, feces, air as CO2) with and without the application site are used:

- rats, 1 mg; excreta (including ~3% CO2 in expired air): ~10%; application site: ~16%

- rats, 11.2 mg; excreta (including ~3% CO2 in expired air): ~12%; application site: ~3.4%

- monkeys, 2 mg; excreta (including 0.4 % CO2 in expired air): ~0.8%; application site: ~9.4%

This analysis shows that in monkeys as compared to rodents only about 10% of formaldehyde applied dermally appears as radiolabel in excreta. It is justified to conclude that the amount of radioactivity in the remaining carcass of monkeys would also be in the range of 10% of that in rats and guinea pigs, and that systemic absorption in monkeys only is in total 10% of that of the other species.. The amount remaining at the application site after 72 h is unlikely to be free formaldehyde; the radioactivity is either bound to or metabolically incorporated into macromolecules.

It is proposed that for systemic effects the absorption rate without the amount at the application site is used while for local effects the application site is taken into account as well.

A decision has to be made whether risk assessment should be based on rat or monkey data. It is well established that skin absorption is much higher in rodents than in humans. On the other hand several investigators have shown that the monkey is a better model for humans (e.g. Franklin et al., 1986; Bronaugh et al., 1990; Scott et al., 1991 cited in ECETOC Monograph 20, 1993). After systematic review of in vitro and in vivo data for various substances, ECETOC (1993, Monograph 20) concluded that the permeability of monkey skin is generally lower than that of other species and more like that of man, while mouse, rabbit, rat and guinea pig skin is more permeable. In addition it is concluded that the skin of monkey (and pig) is the most promising for use as an in vitro model for human skin. Therefore, it is proposed to use the monkey data for risk assessment.

For systemic effects dermal penetration in monkeys cannot be directly derived from the data as radioactivity in total carcass was not determined. In excreta (air as CO2, urine, feces) of monkeys only about 10% of the amount recovered from rodents was found. It is proposed to use a similar relationship for the remaining carcass excluding the site of application leading to a total systemic absorption for monkeys of 4% (i.e. systemic absorption in rodents 30-40%).

For local effects the absorption at the site of application has to be taken into consideration as well, i.e. 9.5% for a formaldehyde concentration of 1% in monkeys. In rodents about 15% and 4% were recovered from the site of application for formaldehyde concentrations of 1 and 28%, respectively. Thus, local penetration decreases as concentration increases. Therefore, for local effects a dermal penetration of 15% will be used (4% systemic plus 10% local penetration) as found in monkeys for a 1% solution. It is admitted that the monkey data show the highest variability. But nevertheless the monkey data should be the basis, as skin permeability in rodents is not comparable to that humans.

Bartnik et al. (1985) studied the percutaneous absorption of 14C-labelled formaldehyde in rats. According to the authors the labeled sample of formaldehyde may have contained up to 2% methanol and 3% formic acid which may have contributed to the dermal penetration data that was determined by liquid szintillation counting.

Formaldehyde was incorporated into a typical O/W-cream as used in cosmetic products (77% water, 9% fatty alcohols, 6% cosmetic oils, 3% fatty acid glycerine ester, 3% polyols, 0.5% perfume, 0.5% PHP-ester, 0.5% polyacrylates, 0.5% neutralising agents) at a concentration of 0.1%. Stability of formaldehyde in this formulation was not analytically verified. The test material was applied under occlusive and non-occlusive conditions using non-perforated and perforated glass capsules. The results obtained are presented in the table giving the % of recovered radioactivity over 48 h related to the amount applied:



Non-occlusive; 8 males

Non-occlusive; 4 females

Occlusive; 2 males









CO2 (expired)








Total systemic absorption




Glass capsule




Volatile formaldehyde




Treated skin









In comparison to the results obtained by Jeffcoat et al (1983) for rats the study of Bartnik et al. (1985)

- showed a higher overall recovery

- slightly lower amounts excreted in urine, feces and expired air (as CO2)

- clearly lower amounts in the remaining carcass

- much higher amounts at the application site.

The total systemic absorption was between 3.4 and 9.2 %. There was no clear difference in the results after occlusive and non-occlusive application. Autoradiography of the treated skin at the end of study showed a very sharp localization of the remaining radioactivity in the uppermost skin layers.

This study gives an indication of the absorption rate of low concentrations of formaldehyde, as e.g. used in cosmetics. This data cannot be extrapolated to aqueous solutions of formaldehyde that are relevant for production or downstream uses because

- the stability of formaldehyde in the formulation used is unknown

- it is unclear whether the dermal penetration of formaldehyde is modified by the formulation used.

In summary, a dermal absorption for formaldehyde

- of 4% is taken for systemic effects

- and of 15% for local effects

based on the monkey data of Jeffcoat et al. (1983).