8-hour TWA: 0.3 ppm (0.369 mg/m3)
STEL: 0.6 ppm (0.738 mg/m3)
BLV: -
Additional categorisation: SCOEL carcinogen group C
(genotoxic carcinogen with a mode-of action based threshold)
Notation: Sensitisation (Dermal)
Recommendation Executive Summary
When reviewing the scientific data available for formaldehyde (FA), SCOEL recognised that FA is a very well investigated substance, for which a high number of reliable high-quality studies relevant for the occupational situation are available. This includes a variety of epidemiological studies on exposed workers, studies on human volunteers for sensory irritation and a broad database on experimental animal studies.
SCOEL has assessed all available information. FA has a potential to cause adverse health effects and is therefore a hazardous chemical agent. FA also is a genotoxic carcinogen, for which a mode-of-action based limit value can be derived. For FA the available information is adequate for deriving a health-based OEL (8-hour TWA and STEL).
Analytical measurement systems exist to determine the recommended levels of formaldehyde with an appropriate level of precision and accuracy.
Due to the high water solubility and the high reactivity of FA, it shows intrinsic hazardous properties predominantly with respect to local effects. In addition, directly induced systemic effects of inhalation at concentrations relevant for the workplace are considered unlikely. The following key effects were considered as being relevant for the protection of workers and in particular the OEL derivation:
(a) the potential of the substance to produce respiratory irritation and chemosensory effects, both in humans and animals, and
(b) the local carcinogenicity in studies with experimental animals exposed by inhalation.
Ad (a): Sensory irritation has been investigated in experimental animals, in exposed workers, and most importantly also under controlled exposures in volunteers.
Ad (b): Tumour induction of the upper respiratory tract has been studied in experimental animals including mechanistic investigations on events that will trigger carcinogenesis, like DNA-protein crosslinks (DPX), DNA-adducts and sustained cytotoxicity leading to cell proliferation. In addition, several high quality epidemiological studies are available on exposed workers. A review by RAC (ECHA 2012) concluded that these data would not provide sufficient evidence to classify FA as a human carcinogen but a classification as Cat. 1B carcinogen (H350 “May cause cancer”; based on CLP criteria) would be appropriate.
Mechanistic studies have provided strong evidence that tumour induction in the nasal mucosa of rats and mice is the result of chronic proliferative processes caused by the cytotoxic effects of the substance in combination with DNA alterations by endogenous and exogenous FA. The dose-response relationships for all parameters investigated, such as damage to the nasal epithelium, cell proliferation, tumour incidence, the formation of DPX and DNA-adducts, is very flat for low level exposures and becomes much steeper at higher concentrations. For these endpoints no-effect concentrations were demonstrated with the exception of the formation of DPX and DNA-adducts. However, at the lowest concentrations investigated so far (0.7 ppm), adducts caused by the endogenous, physiological FA by far exceeded the amounts caused by exogenous FA. The background incidence of nasal tumours in rodents and of nasopharyngeal tumours in humans is very low in spite of the appreciable amount of endogenous DNA adducts. One of the reasons may be the low physiological proliferation rate of the respiratory epithelium, and as long as this is not increased (which requires exposure to concentrations of more than 2 ppm), the probability of tumour formation also is low. At prolonged exposure at 2 ppm in rats, the half-life of the most sensitive biomarker of DNA-adducts, N2-hydroxymenthy-dG, was 7 days. At 2 days of exposure in monkeys, the biomarker was estimated to be by a factor of 5-11 lower for the exogenous adduct than that of the endogenous adduct in the nasal epithelium. Comparing short term exposures, the relationship of exogenous/endogenous DNA–adducts was by a factor of about 5-fold lower for monkeys than for rats, suggesting monkeys being a less sensitive species than rats. Taking into consideration the strong non-linearity of the dose response curve after a single exposure at lower exposure concentrations, the ratio between exogenous/endogenous adducts will at low exposures be dominated by the endogenous adducts, but the ration will increase disproportionately with increasing FA concentrations. Also in the low dose range, cell proliferation is not increased. It has therefore been considered that the genotoxicity of FA plays no or at most a minor role in a potential carcinogenic effect at this exposure-range.
Therefore SCOEL considers FA as a group C carcinogen (genotoxic carcinogens for which a limit value derived from mode-of-action based threshold is supported) (SCOEL, 2008).
Experimental studies support that the local carcinogenesis at the portal-of-entry is pivotal. In the sensitive rat species, the apparent LOAEC was 6 ppm, and the apparent NOAEC was 2 ppm for nasal cancer. Experimentally, the histopathological NOAEC for nasal effects of FA in rats and monkeys is 1 ppm and the NOAEC for regenerative cell replication 2 ppm. At these NOAECs, the FA-DNA adducts were less in monkeys than in rats as was the relationship of exogenous/endogenous DNA adducts, which is in line with the assumption that humans should be a less sensitive species. The new studies confirm that local FA-DNA adducts show a highly non-linear relationship with external FA exposures. At ≤ 2 ppm FA, the FA DNA-adducts induced by external exposures comprise a minor portion of the total FA-DNA adducts, which were driven mainly by internal (naturally generated) FA. This is supported by considerations on toxicokinetics, concluding that the intracellular FA concentration increases only slightly, and the intracellular glutathione concentration decreases only slightly in this range and that the homeostasis within the epithelial cells would not be affected. Therefore, the apparent NOAEC of 1 ppm can be considered a mode-of-action based NOAEC for carcinogenic effects at the portal-of-entry.
Ad (a): Preventing histopathological effects, like irritation, inflammation and regenerative cell replication caused by cytotoxic irritation, will also prevent nasal cancer as at such low exposure concentrations (< 1 ppm) the total intracellular FA concentration is dominated by the internal (natural) FA. This experimentally derived paradigm, namely the avoidance of cell proliferation in the upper respiratory tract being critical to prevent local carcinogenicity, also holds valid for humans. Ideally the lower sensitivity against cytotoxic irritation of humans as compared to rats should be taken into consideration. While cytotoxic irritation cannot be investigated in humans, mainly for ethical reasons, there is a broad database available for sensory irritation from volunteer studies under controlled exposure conditions. By derivation of limit values for sensory irritation of eye and upper respiratory tract in humans also the critical effects of irritation-induced local cell proliferation and subsequent possible carcinogenesis shall be covered (Brüning et al 2014).
In this respect, numerous studies, comprising in total more than 400 volunteers, have addressed human sensory irritation effects of FA. The Paustenbach et al (1997) review [and two similar reviews of Bender (2002) and Arts et al. (2006)], concluded that sensory irritation would seldom be observed at 0.5 ppm FA and extrapolated these results to suggest that a limit of 0.3 ppm would prevent sensory irritation in nearly all occupational exposed individuals. Two recent chamber studies (Lang et al. 2008; Mueller at al. 2013) found no pure sensory irritation, as measured by objective parameters, in the concentration range from 0.5 to 0.7 ppm at a constant exposure to FA during a 4-hour period. Both studies applied slightly different concentration regimes. Exposures with 4 superimposed peaks being most relevant for derivation of an OEL with STEL were 0.3 ppm + peaks of 0.6 ppm and 0.5 ppm + peaks of 1 ppm in the Lang study, and in that of Mueller 0.3 ppm + peaks of 0.6 ppm and 0.4 ppm + peaks of 0.8 ppm. Objective signs of irritation were only observed at 0.5 ppm + peaks of 1 ppm. Because 0.3 ppm + peaks of 0.6 ppm was a consistent NOAEC in both of these investigations this exposure regime is taken forward for derivation of the OEL, TWA with STEL. The recent study (Mueller et al. 2013) was conducted with hypo- and hyper-sensitive individuals, who showed no difference in sensory irritation sensitivity to FA, but the hypersensitive individuals reported significantly higher effects for olfactory induced symptoms as ”perception of impure air”.
Based on these experimental studies in human volunteers SCOEL derives an OEL of 0.3 ppm (8 h TWA) with a STEL of 0.6 ppm. As sensory irritation is a concentration rather than a cumulative dose-driven effect, a STEL value is appropriate. This OEL based on sensory irritation will also protect workers from undue annoyance and discomfort at the workplace.
Ad (b): The OEL of 0.3 ppm derived from human volunteer studies is supported by data in experimental animals. The histopathological NOAEC for nasal effects of FA in rats and monkeys is 1 ppm and for regenerative cell proliferation in rats is 2 ppm. Preventing these effects will also prevent nasal cancer. As a strong support, toxicokinetic studies suggest that at an exposure level of 1 ppm the local intracellular concentration of formaldehyde is dominated by the internal (naturally produced) FA. Backed by this finding, SCOEL considers an uncertainty factor of 3 to be sufficiently protective. This supports the proposed OEL (8h-TWA) of 0.3 ppm.
As a result of the predominantly local effects of FA, a “skin” notation is not required. FA is a well-known contact allergen to the skin (skin sensitizer). A notation sensitisation (Dermal) is therefore added. Against the background of a widespread use, respiratory sensitization has been reported only occasionally, and therefore the designation as respiratory sensitizer is not warranted.
A biological limit value (BLV) or biological guidance value (BGV) is not proposed.
For additional details, which were considered in the OEL derivation by SCOEL, see chapter 7.11.