Registration Dossier

Diss Factsheets

Administrative data

Description of key information

Data gap filling for the acute inhalation toxicity endpoint is achieved using the category approach according to ECHA guidance on read-across (ECHA, 2017c). For this endpoint, all effects are consistent with the hypothesized MoA and direct electrophilic reactions of NCO with biological nucleophiles. Modified MDI substances contain different higher molecular weight constituents, and all have in common a high content of bioaccessible low molecular weight MDI constituents responsible for presenting NCO reactivity, scenario 4 or 6 according to the RAAF considered as most appropriate due to a common mechanism. Selection between scenario 4 and 6 depends essentially upon the presence of variation in the properties i.e. magnitude of effect. Since it has been demonstrated that the bioaccessible low molecular weight MDI constituents are responsible for presenting NCO reactivity, and the higher molecular weight MDI constituents do not contribute to the observed toxicity it is reasonable to assume that their presence in these mixtures attenuates toxicity. Further, as a worst-case approach is adopted in which 4,4’-MDI isomer is used for read-across to all substances of the MDI category, then use of RAAF Scenario 4 (variations in the properties observed among source substances) is justified over scenario 6.


Acute oral toxicity

In the case of oral exposure, before the reactive NCO groups present on the substances of the MDI category have opportunity to react locally, or be absorbed, they polymerize in the acid environment of the stomach to form solid polyureas that are excreted via the feces without being absorbed. Consequently, if exposure were to occur by the oral route this would not lead to local or systemic effects.

For MDI Mixed Isomers, the key study (Reliability 1) did not record mortality up to the limit dose of 2,000 mg/kg bw . A supporting study describing the acute oral toxicity of pMDI conducted similar to OECD 401 guideline (Reliability 2) also did not find any mortality up to the maximum dose tested, hence the LD50 is greater than 10,000 mg/kg bw . Other studies on MDI substances are consistent with this, albeit with lower reliability. 

Four additional acute oral toxicity studies (Reliability 1) have been conducted on other representative substances of the category subgroups (i.e. ‘MDI, its condensation products and the reaction products with glycols’ and ‘MDI and its reaction products with glycols’). In all cases, there was no mortality up to the limit dose (5,000 mg/kg). The lack of mortality in the available acute oral toxicity across the available studies, alongside the lack of gross lesions in distal tissues (e.g. liver, kidney etc.) supports the lack of systemic bioavailability. The NCO groups present on MDI substances react with acids within the stomach leading to formation of an insoluble polymerized mass that is excreted in the feces without being absorbed (see Toxicokinetics). 

Supporting evidence comes in the form of several accidental ingestion reports in dogs where ingestion of MDI based glues produced no intrinsic toxic effects other than the formation of a solid polyurea mass that may lead to gastric obstruction. When the mass was removed by surgery, rapid and complete recovery was achieved (Horstman et al., 2003; Ohngren, 2007). 


 Acute dermal toxicity

In the case of dermal exposure, before the reactive NCO groups present on MDI substances have opportunity to be absorbed to any significant extent through the stratum corneum they react with proteins and moisture at the skin surface leading to the formation of an insoluble polymerized mass thereby limiting dermal absorption and systemic availability (Leibold 1999). Modified MDI substances, having a higher molecular weight than mMDI isomers and due to their higher molecular volume, increased octanol-water partition coefficient and decreased water solubility will in any event not be able to penetrate the stratum corneum (Bartels 2021).

The available acute dermal toxicity studies indicate that all substances of the MDI category have low acute dermal toxicity. The key study describing the acute dermal toxicity of pMDI in rabbits did not find lethality up to the maximum dose tested, and the LD50 was greater than 9,400 mg/kg bw (Wazeter et al., 1964a). Other less reliable studies on pMDI or 4,4’-MDI are consistent with this.  Observed differences in LD50 values between pMDI and 4,4'-MDI/TPG are not considered to be significant or represent a trend since they are significantly higher than the limit for classification and are indicative of a lack of systemic exposure. The available data for the substances of the MDI category is consistent with the hypothesis that NCO groups present on MDI substances react with proteins and moisture at the skin surface leading to the formation of an insoluble polymerized mass resulting in limited dermal absorption and systemic availability. The available data and hypothesis is supported by the key dermal absorption study that shows that MDI substances have very low systemic bioavailability (<1 %) (Leibold et al., 1999). By comparison, modified MDI substances, with molecular weight greater than mMDI, will demonstrate even further reduced dermal absorption based upon physico-chemical properties (i.e. increased octanol-water partition coefficient, decreased water solubility, and increased molecular weight), and this has been confirmed with GastroPlus™ modeling (Bartels, 2021). Although a reliable, acute dermal in vivo toxicity data is only available on two category substances, it is considered sufficient for assessment of this endpoint for the category. Due to the low predicted dermal bioavailability of all category substances, and the lack of systemic toxicity demonstrated in the oral acute toxicity studies, additional testing is not justified as all substances of the MDI category would be predicted to have comparable or reduced acute dermal toxicity potential to tested substances.


Acute Inhalation Toxicity

Following inhalation exposure the initiating event in hypothesised MoA for acute toxicity in the lung is the reaction of the MDI substance with GSH in the airway lining fluid (adduct formation). Subsequent development of toxic effects is driven by the rate of depletion of GSH. This depletion begins with the reduction in extracellular GSH, which leads to a reduction in intracellular GSH disturbing the redox balance in the cell. With increasing amounts of NCO exposure (e.g. via exposure concentration or bioaccessibility), the protective GSH system gradually becomes overwhelmed, and toxicity evolves along the path: (1) no cytotoxicity; (2) cytotoxic effects; (3) reduced cell viability; and (4) cell death. This is accompanied by increasing extravasation because of increased junction permeability and epithelial damage ultimately causing edema. 


The rate of nucleophile depletion by MDI-based substances is driven by the availability of the NCO-group, which itself is a function of (1) the NCO value of the substance and (2) the molecular weight of its constituents (driving its reactive dissolution). Monomeric MDI isomers have been shown to become available at a similar rate in toxicokinetic studies (Wisnewski, 2018; Wisnewski et al., 2019a) which is consistent with the generally comparable LC50 values for all of the isomers. Conversely, higher molecular weight constituents have both a reduced NCO value and exhibit reduced water solubility, making them less accessible to react with GSH. Therefore, the substances with the highest available NCO value and bioaccessibility (mMDI and three-ring oligomers) are the most toxic, while those with increasing amounts constituents less able to react with GHS demonstrate reduced toxicity.


Tests also show that toxicity is limited to portal-of-entry effects. The absence of systemic toxicity is due to the extracellular reactions described above, combined with transcarbamoylation to proteins described in more detail in the Chapter (Toxicokinetics), constitute a detoxification mechanism. Acute toxicity is only observed when this protective mechanism becomes overwhelmed and is limited to the lung.

This mode of action is supported with high confidence by reliable acute inhalation data available for multiple MDI isomers and modified MDI substances (described in more detail below). The toxicity of MDI substances will decrease with increasing average molecular weight as these substances will have constituents that are less bioaccessible and with a lower NCO value. For these substances, higher exposure concentration is required to induce toxic effects, which is consistent with the observed results from the available acute inhalation toxicity tests. 

Testing proposal:

While testing is available on 8 MDI category memeber (including all sub-groups) acute toxicity testing (OECD 403) will be performed on an additional 4 MDI substances.  This information will further support the category hypothesis as well as help to define substance selection and study design for repeat-dose bridging studies. 


Available data:

The ‘Monomeric MDI’ subgroup

Reliable acute inhalation studies are available for all three isomers of mMDI (2,2’-; 2,4’-; and 4,4’-MDI) in accordance with OECD Guideline 403 in a series of studies by (Pauluhn, 2008d; Pauluhn, 2008e; Pauluhn, 2008f). All three substances consist of more than 98 % pure mMDI, corresponding to NCO value of 33%. In all cases, mortality was linked to local effects of the respiratory system that included severe irritation, pulmonary edema, and ultimately death occurring within one to two days following exposure consistent with the hypothesized MoA. LC50s were comparable and ranged from 368 to 598 mg/m3 for males and from 559 to 686 mg/m3 for females. For all studies, exposure parameters met internationally recognized recommendations for MMAD and GSD and were similar for all three isomers.

The ‘Oligomeric MDI’ subgroup

Polymeric MDI (approximately 40 % mMDI; 33 % NCO value, with viscosity of approximately 200 mPas) was tested in an acute inhalation toxicity study according to OECD 403 (Pauluhn, 2008c). Mortality was linked to local effects of the respiratory system that included severe irritation, pulmonary edema, and ultimately death occurring within one to two days following exposure consistent with the hypothesized MoA. An LC50 (95 % confidence interval) of 310.2 (266.4-361.3) mg/m3 was determined for pMDI. 

Polymeric MDI was also tested following depletion of monomeric MDI resulting in a mixture of 1.2 % mMDI and 98.8 % of higher (> two-ring) oligomers according to OECD 403 (Pauluhn, 2011a). The combined LC50, for male and female rats, for ‘monomer-depleted pMDI’ was greater than 2,188 mg/m3. Average mean mass aerodynamic diameter (MMAD) and geometric standard deviation (GSD) was generally comparable to that of the pMDI containing mMDI (85-87 %).

The ‘MDI and its condensation products’ subgroup

The acute inhalation toxicity of MDI Mixed isomers/PIR (60 % mMDI and NCO value of 26 %) was tested in an acute inhalation toxicity study according to OECD 403 (Pauluhn, 2012). The combined LC50 for male and female rats was 1,088 mg/m3, mortality was linked to portal of entry effects of the respiratory system, including severe irritation and pulmonary edema. Mortality occurred up to two days post-exposure and was causally related to an acute pulmonary edema.

The ‘MDI and its reaction products with glycols’ subgroup

Two reliable acute inhalation studies are available for substances of the ‘MDI and its reaction products with glycols’ subgroup.

The acute inhalation toxicity of 4,4'-MDI/1,3-BD/TPG/PG (60 % mMDI; 23 % NCO) was conducted according with OECD 403 (Kopf, 2016).  The LC50 was calculated to be 518 mg/m3, and mortality was linked to local effects of the respiratory system that included severe irritation, pulmonary edema, and ultimately death occurring within one to two days following exposure consistent with the hypothesized MoA. 

The acute inhalation study of 4,4'-MDI/DPG/HMWP (50 % mMDI; 25 % NCO) was tested according to OECD 403 (Hotchkiss and Weidemoyer, 2020). The LC50 is 1,110 mg/m3 for male rats and 1,250 mg/m3 for female rats. The four-hour LC50 is 1.15 mg/L for male and female rats combined. Similar to the other LC50 studies, mortality was linked to local effects of the respiratory system that included severe irritation, pulmonary edema, and ultimately death occurring within one to two days following exposure consistent with the hypothesized MoA.


An acute inhalation study was performed in rats at only one concentration level of 2.24 mg/L/1h (Pauluhn 2003, 2004). This study was specifically designed to comply with NFPA 704, and also complied with the limit test of the OECD guideline 403 with deviations (only 1 hr exposure, concentration lower than limit test concentration) and is therefore reliable with restrictions. Exposure of 4,4’-MDI for 1 hr resulted in mortality shortly after exposure of one out of ten rats. Clinical signs were characterised by typical signs of respiratory tract irritation. Necropsy findings were unremarkable in surviving rats, whilst the rat that succumbed displayed signs of lung oedema which was considered to be the cause of death. The LC50 >2.24 mg/L/1h (analytical) in both males and females was determined. 



Adequacy of the available data for risk assessment and classification purposes

Using the strict GHS LC50 cut-off for classification, the LC50 values obtained for the mMDI would trigger a Category 2 (or Category 3) according to GHS CLP. However, classification for these substances according to ECHA CLP Guidance (2017) text allows for the application of scientific judgement. It must be considered that the LC50 cut-off of 500 mg/m3 (approximately 50 ppm for pMDI), is over 2,500-fold above the saturated vapor concentration for pMDI. This difference is even further exacerbated in the pre-polymer mixtures where the presence of the higher molecular weight fraction even further reduces the vapor pressure making exposure less likely. 


Furthermore, the aerosols were generated using sophisticated techniques in the laboratory, whereby extremely small particles are generated in order to meet international guidelines for testing. This size and concentration of aerosol is not generated in the workplace even under foreseeable worst-case conditions (Ehnes et al., 2019). The particle size distribution of aerosols formed during actual spraying applications has virtually no overlap with that of the highly respirable aerosol generated in inhalation studies (see EC (2005)). Due to a very low vapor pressure (<0.01 Pa) MDI substances are not inherently toxic by inhalation since the saturated vapor concentration would be orders of magnitude below toxic concentration. It is only with modification and input (in terms of heat, cooling and size screening) that MDI substances become toxic after inhalation. In the EU risk assessment report (EU 2005) MDI is classified as  harmful by inhalation.


The acute inhalation data of pMDI and 4,4’-MDI data were considered by EU experts, and their conclusion that MDI be classified as “Harmful” and  reported in the 25th Adaptation to Technical Progress (ATP) to the Dangerous Substances Directive (67/548/EEC). This was endorsed in the 28th ATP and both MDI substances remain as “Harmful” in the 30th ATP (adopted by Member States on 16 February 2007 and published 15th September 2008). The original decision was upheld in the EU Risk Assessment of MDI (Directive 793/93/EEC, 3rd Priority List) published in 2005, noting that considering “the exposure assessment, it is reasonable to consider MDI as harmful only and to apply the risk management phrase ‘harmful by inhalation’. This classification was also endorsed by the Scientific Committee on Toxicity, Ecotoxicity and the Environment (CSTEE, now SCHER) in giving their opinion on the Risk Assessment (EC, 2008). With the enforcement of the CLP regulation (Regulation (EC) No 1272/2008) in 2009, the Dangerous Substance/Preparation Directive (DSD) was repealed and harmonized classifications were formally transferred to the CLP regulation; MDI is classified with Acute Tox. 4 H332 (Annex VI Regulation (EC) No 1272/2008 (CLP regu lation).

Given the mechanism of action of the MDI substances and the changes in physical chemical properties imparted by the modifications in the modified MDI substances, the entire category is consistent with this guidance and classification, and the classification should not be changed.


The classification as “Harmful”, is equivalent to GHS Category 4. For these reasons, the GHS proposal follows the EU Regulatory lead accepting that the animal data are inappropriate and classified pMDI as GHS acute toxicity category 4 (ISOPA 2007).




 Assessment of the available acute toxicity data indicates that inhalation exposure to the aerosols of MDI results in toxicity confined predominantly to the respiratory tract. In terms of hazard characterization, MDI is harmful by inhalation according to EU (H332) and GHS (Cat. 4) classification. MDI is non-toxic after single oral and dermal exposure.



Justification for classification or non classification:


 EU classification according to CLP: H332 


 GHS classification (GHS UN rev.2, 2007): Inhalation route (vapour): Acute Category 4. 


 Not toxic by the dermal or oral routes.

Key value for chemical safety assessment

Acute toxicity: via oral route

Endpoint conclusion
Endpoint conclusion:
no adverse effect observed
Dose descriptor:
> 2 000 mg/kg bw
Quality of whole database:
Results are consistent within the key study Bomhard (1990) (reliability1)and data from a supporting study (Wazeter et al. 1964) (reliability 2).

Acute toxicity: via inhalation route

Endpoint conclusion
Endpoint conclusion:
adverse effect observed
Dose descriptor:
431 mg/m³

Acute toxicity: via dermal route

Endpoint conclusion
Endpoint conclusion:
no adverse effect observed

Additional information

Acute inhalation toxicity: Sub-lethal irritation studies 


Numerous acute studies are also available describing the sub lethal and are described in detail according to the. (Respiratory Irritation) the MDI category justification document.  These results are summarized below.

Like the acute mortality studies, in sub-lethal studies, the MDI substances enters the lung where the NCO reacts with nucleophilic biomolecules at the MDI/lung fluid interface to form conjugates (primarily MDI-GSH adducts). Formation of MDI-adducts depletes the electrophile scavenging capacity of the lung resulting in pulmonary irritation and inflammatory cell influx which is consistent with the available in vivo inhalation data. As shown by acute inhalation toxicity studies, the magnitude of effects is dependent upon NCO value as attenuated by solubility (i.e. octanol-water partition coefficient) to dictate its bioaccessibility. Low molecular weight substances (e.g. mMDI, pMDI) not only have high NCO value but, due to their higher solubility, are also more available for reacting with nucleophiles in the lung fluid. Conversely, the higher molecular weight constituents of modified MDI substances have lower NCO value and are less available for reacting with biomolecules due to reduced solubility. As all substances of the MDI category have a significant quantity of bioaccessible low molecular weight constituents to induce irritation, the levels of toxicity are expected to be comparable for all category substances.


  • Acute inhalation studies

Several acute inhalation studies are available on MDI substances. Three repeated exposure studies are also included as they inform on respiratory irritation. While most of the data is on a limited sample of category substances (4,4’-MDI, pMDI, and 4,4’-MDI/DPG/HMWP), the effects noted are highly consistent and support the proposed mechanism of respiratory irritation. These studies are outlined in Table 55 of the category document.

The ‘Monomeric MDI’ subgroup

In an acute inhalation study with mice, pulmonary irritation of 4,4’-MDI was assessed by recording respiratory rates (Weyel and Schaffer, 1985). Groups of four male Swiss-Webster mice were exposed for 240 minutes to 4,4’-MDI at 6.7, 10.2, 19.6, 25.8, 40.3, 58.5 mg/m³. The average respiratory rates were monitored during exposure, and all animals were killed 24 hours post exposure. Upon sacrifice, the lung weights were measured. RD50 (concentration required to reduce the respiratory rate by 50%) were determined. To confirm that 4,4’-MDI acted as a pulmonary and not solely as a sensory irritant, mice were also exposed via tracheal cannulation to MDI aerosol at a concentration of 23.6 mg/m3 with the aim of eliminating the influence of sensory irritation by bypassing the trigeminal nerve. The exposure concentration was that expected to be the RD50 in control mice. Except for the cannulation, all other experimental conditions were identical to those in the inhalation exposure groups. At the lowest concentrations (6.7 and 10.2 mg/m³), an increase in respiratory rate above the control was observed for almost three hours of the exposure, followed by a gradual decline in respiratory rate during the last hour. For the highest concentration (58.5 mg/m³) there was only a slight increase in respiratory rate, of short duration, followed by a rapid decline during the last three hours. For the intermediate concentrations, the respiratory rate was initially elevated above the control for approximately one hour and gradually declined for the last three hours of exposure. A plateau in response was reached during the last 30 minutes of exposure. Little or no recovery was observed during 20 minutes following each exposure. The RD50 was estimated to be 32 mg/m3. Dose-dependent increases in lung weight were found for all MDI concentrations, which the authors concluded were the result of pulmonary edema. In the tracheal cannulation test, a decrease in respiratory frequency was observed, which was interpreted by the authors as a clear pulmonary irritant response. Based on the results, the authors concluded that 4,4’-MDI acted primarily as a pulmonary irritant, evoking little or no sensory irritation.

Further, two studies which did not measure respiratory parameters but focused on the dose-response relationship of bronchoalveolar lavage (BAL) parameters in 4,4’-MDI-exposed rats are available:

A study by Hotchkiss et al. (2017) was designed to provide data on concentration- and time-dependent effects on BAL cells in the lung of Wistar rats after a single, acute 6-hour inhalation exposure to 4,4'-MDI aerosol and included evaluation of a) macrophage activation, b) oxidative stress, and c) cytotoxicity. Groups of twelve male Wistar rats were subjected to a single six hours aerosol exposure of 4,4’-MDI at 4, 12 or 27 mg/m3. Half of the rats in each exposure group were euthanized immediately after exposure with the remaining rats in each group euthanized approximately 18 hours later. At necropsy, bronchoalveolar lavage was performed on all experimental animals. The lavage fluid supernatant was analyzed for biomarkers of injury and inflammation (total protein, LDH, alkaline phosphatase, β-glucuronidase) and oxidative stress (GSH and GSSG levels). The lavage cells were analyzed for total and differential cell counts, targeted gene expression (including signals for inflammation, macrophage activation, apoptosis, and oxidative stress), and markers of apoptosis (Annexin V and Caspase-3). All animals survived to the scheduled euthanasia and there were no test substance-related clinical observations or effects on body weight.

Results included:

  • Dose-dependent increases in total protein and β-glucuronidase were noted at the end of the exposure and 18 hours post-exposure at ≥4 mg/m3(statistically significant in 27 mg/m3 exposure group).

  • Small increases in LDH activity were detected at the end of exposure and 18 hours post-exposure (statistically significant in 4 and 27 mg/m3 exposure groups).

  • The ratio of GSH/GSSG was reduced (indicative of oxidative stress) compared to controls at the end of exposure at ≥4 mg/m3. GSH/GSSG ratio remained lower than control rats in 12 and 27 mg/m3 exposed rats at 18 hours but was slightly elevated in rats exposed to 4 mg/m3 due to a significant increase in GSH levels

  • Neutrophilic inflammation was noted at >12 and 27 mg/m3 immediately and was still evident at 18 hours post-exposure.

  • A concentration-dependent increase in % Annexin (+) / PI (-) cells in nucleated BAL cells was observed in 4,4’-MDI exposed animals immediately after exposure and 18 hours post-exposure

In summary, a single six-hour exposure to respirable aerosols of 4,4’-MDI to 4, 12, and 27 mg/m3 resulted in concentration- and time dependent increases in oxidative stress, inflammation, and markers of apoptosis. There was some evidence of inflammation, oxidative stress, and/or apoptosis at every dose, although the clearest effects were observed in the 12 and 27 mg/m3 exposures and especially by 18 hours post exposure. Based on the increases in total protein in the BALF (which has been demonstrated to be a good marker for respiratory irritation of pMDI, see Pauluhn (2002b), it can be concluded that the LOAEC regarding respiratory irritation is 4 mg/m3 in that study.

This study was followed up with a second acute inhalation toxicity in a combination with an in vivo genotoxicity toxicity study according to OECD TG 489 (genotoxicity assessed by Comet assay) (Randazzo, 2017). Wistar rats were exposed 2.5, 4.9 or 12 mg/m3 (measured concentrations) of 4,4’-MDI administered via a single nose-only inhalation exposure for six hours. A concurrent control group received filtered air on a comparable regimen. Six rats/group were sacrificed approximately one hour post-exposure (ca. one hour after termination of the six-hours exposure) and the other six/group approximately 18 hours post-exposure. All animals survived to the scheduled euthanasia and there were no test substance-related clinical observations or effects on body weight. Bronchoalveolar lavage (BAL) was performed in all animals at the scheduled necropsies, and the BAL fluid (BALF) and cells (BALC) was assessed for biomarkers of cytotoxicity and inflammation. The endpoints alkaline phosphatase (type II alveolar epithelial cell cytotoxicity), lactate dehydrogenase (tissue damage/cytotoxicity), Annexin + flow cytometry (apoptosis and necrosis), and total protein (cytotoxicity, blood/air barrier dysfunction) were determined to assess the cytotoxicity. β-glucuronidase (indicator for macrophage activation: activated macrophages secrete various inflammatory mediators such as cytokines/chemokines) and cell differential (with particular focus on the percent neutrophils, the influx of which is the hallmark of the typical acute inflammatory response in the rat lung) were determined to assess the inflammatory potential of the test substance.

The following results were found:

  • Dose-dependent induction of β-glucuronidase was observed one hour post-exposure at ≥4.9 mg/m3 and ≥2.5 mg/m3 at 18 hour post-exposure, reaching statistical significance for the high dose-group.

  • Dose-dependent increases in LDH at ≥2.5 mg/m3 at one-hour timepoint. At the 18 hour timepoint, LDH was only increased in BALF at 12 mg/m3.

  • Clear dose-dependent increases in total protein was observed ≥2.5 mg/m3 at both time points.

  • At both 1 and 18 hours after cessation of exposure, the percent neutrophils in BALF was increased in in the 4,4’-MDI-exposed groups at 12 mg/m3 indicating an acute inflammatory response.

  • An increased of late apoptosis and necrotic cells was observed one-hour post-exposure at ≥2.5 mg/m3 as identified by Annexin V expression which returned to baseline by 18 hours. Similarly, the number of early apoptotic cells increased one hour after exposure at ≥4.9 mg/m3. However, by 18 hours, an increase was only observed at 12 mg/m3.

In summary, local cellular toxicity, characterized by an increased concentration of total protein and the macrophage activation marker β-glucuronidase in BALF, an increase in apoptosis/necrosis, were observed in animals exposed to ≥2.5 mg 4,4’-MDI/m3. Except for β-glucuronidase and total protein, all other parameter returned to control levels by 18 hour post-exposure at the low- and mid-dose groups. The influx of neutrophils (a hallmark of an acute inflammatory response) was induced at 12 mg/m3 at both sacrifice days confirming MDI induces a persistent inflammatory response at this dose level. Based on the increases in total protein in the BALF (which has been demonstrated to be a good marker for respiratory irritation of pMDI, see Pauluhn (2002b), it can be concluded that the LOAEC regarding respiratory irritation is 2.5 mg/m3 in that study.

The ‘Oligomeric MDI’ subgroup

Pauluhn (2000a) examined the time course of the relationship between acute pulmonary irritation and acute pulmonary response of Wistar rats exposed to respirable polymeric MDI (pMDI) aerosol of 0, 0.7, 2.4, 8, or 20 mg pMDI/m³ for 6 hours. The time-response relationship of MDI-induced acute lung injury was examined at 0 hours (directly after cessation of exposure), 3 hours, 1 day, 3 days, and 7 days after exposure. Bronchoalveolar lavage (BAL) fluid was analyzed for markers indicative of injury of the bronchoalveolar region. Results suggested that respirable pMDI aerosol interacts directly with the air/blood barrier causing increased extravasation of plasma constituents because of increased permeability of capillary endothelial cells. A transient dysfunction of the pulmonary epithelial barrier occurred at a level as low as 0.7 mg/m³ and was interpreted as a dysfunction of pulmonary surfactant. Such dysfunction is thought to correspond to a physiological response. These results show that single or repeat (sub-acute) exposure to highly respirable pMDI aerosols at high concentrations results in lung effects consistent with exposure to an irritant particulate but that recovery occurs upon cessation of exposure. As such, following a six-hours exposure to pMDI aerosol, the NOAEC is 0.7 mg/m³ air whereas the 2.4 mg/m³ air, caused borderline biochemical effects (= LOAEC). This study, among others, was used by Pauluhn (2002b) to estimate an acute irritant threshold concentration of 0.5 mg/m3 for pMDI.

In another study in rats with a similar design, single six-hours exposures to pMDI at 10, 30, or 100 mg/m³ resulted in signs of respiratory tract irritation (abnormal respiratory noise, breathing rate reduced and depth increased, mucous secretions from the nose) and a pattern of lung responses that was consistent with exposure to irritant aerosols (Kilgour et al., 2002). These effects were observed in all exposure groups. An exposure concentration related body weight loss and increase in lung weight were seen post-exposure, with complete recovery by day ten. Analysis of lung lavage fluid revealed irritation-related changes in the lung over the initial days following exposure. These consisted of a pattern of initial toxicity, rapid and heavy influx of inflammatory cells (alveolar macrophages) and soluble markers of inflammation and cell damage, increased lung surfactant, with a subsequent recovery and epithelial proliferative phase (e.g. bronchiolar and type II cell hyperplasia). Finally, by day 30 post-exposure, a return to the normal status quo of the lung was observed.

  • Short-term Repeated exposure studies


The ‘Oligomeric MDI’ subgroup

In a short-term inhalation toxicity study of polymeric MDI in rats, changes in respiratory parameters in a single-exposure scenario were correlated to the alteration of surfactant activity in a two-week repeated-dose scenario (Pauluhn et al., 1999b). For the single-exposure part of the study, rats were exposed to 0, 2.4, 6.7, 15.8 or 38.7 mg pMDI/m3 air for 150 minutes. Respiratory rate and tidal volume were examined before and during exposure using nose-only exposure restrainers modified to function as flow plethysmographs. Based on concentration-dependent changes in tidal volume, it was concluded that pMDI caused acute pulmonary irritation starting at 2.4 mg/m3 (minimal extent).


In the second part of the study, rats were exposed to 0, 1.1, 3.3 or 13.7 mg/m³ (original target concentration of 10 mg/m3 was raised to 16 mg/m3 during week two due to absence of marked effects) for 14 days (6 h/d, 5 d/week during first week 1, 7 d/week during second week). The rats of each group were allocated to three different examination groups: bromodeoxyuridine (BrdU) labeling, histopathology and organ weight determinations were performed in the first fraction of the animals per group, arterial blood-gas determinations and lung lavage in a second fraction, and lung function measurements, including two female rats/group for electron microscopy, in a third fraction. Rats assigned to lung lavage, blood gas measurements, histopathology, BrdU labeling, and organ weight determinations were sacrificed one day after the last exposure to pMDI. Rats assigned to ultrastructural examinations were sacrificed three days after the last exposure. Lung function measurements were made two to four days after the last exposure. All animals survived until the scheduled necropsies. At 3.3 mg/m3 and 13.7 mg/m3, clinical signs and changes in breathing patterns indicative of respiratory irritation (irregular and labored breathing, rales, dyspnea, tachypnea, bradypnea, at 13.7 mg/m3 additionally nasal discharge) were observed. These changes were partly transient at 3.3 mg/m3. Lung weights were statistically significantly increased in female rats at 13.7 mg/m3. Determination of arterial blood gases, lung function, and carbon monoxide diffusing capacity did not demonstrate specific effects. BAL revealed changes indicative of marked inflammatory response and/or cytotoxicity in rats exposed to 13.7 mg/m3 (statistically significantly increased activities of LDH, beta-NAG, and protein). In terms of non-inflammatory parameters, phospholipid concentrations and gamma-GT were already increased in rats exposed to 1.1 mg/m3 and above. Light and transmission electron microscopy revealed focal inflammatory lesions and an accumulation of refractile, yellowish-brownish material in alveolar macrophages with concomitant activation of type II pneumocytes at 3.3 and 13.7 mg/m3. A concentration-dependent increase of BrdU-labeled epithelial cells in the terminal bronchioles was observed in all exposure groups. In summary, the results show that rats experienced mild signs of respiratory tract irritation which appeared to exacerbate during the study. The LOAEC for respiratory irritation is considered 3.3 mg/m3 in this study. Furthermore, the findings obtained suggest that the interaction of pMDI with surfactants eventually leads to intracellular precipitates originating from precipitated surfactant or surfactant-pMDI complexes. The authors suggested that pMDI interacts directly with pulmonary surfactant lining fluids, the first line of pulmonary defense.

The ‘MDI and its oligomeric reaction products with glycols’ subgroup

In a five-day inhalation pilot study, used to select doses for a subsequent 28-day repeat-dose study, rats were exposed to 4, 10, and 20 mg 4,4’-MDI/DPG/HMWP/m3 (Ma-Hock, 2021).  Results showed effects consistent with respiratory irritation observed in previous studies with MDI substances including concentration-related changes in lavage fluid and morphological changes in lungs, as well some mild reactive changes in lung-draining lymph nodes, trachea and larynx. Moreover, significantly increased cell proliferation in large, medium and terminal bronchi in males and females and increased cell proliferation in alveoli in males (statistically significant) and females (not statistically significant) were observed. Concerning clinical pathology and other examined parameters, no systemic effect could be observed. Under the current study conditions, no observed adverse effect concentration (NOAEC) for local toxicity could not be determined with the LOAEC at 4 mg/m3. The NOAEC for systemic toxicity was 20 mg/m³.


  • Human information

As described previously, the very low vapor pressure of substances of the MDI category () prevents human exposure to vapors of MDI substances. However, certain workplace applications can result in generation of aerosols. It should be noted that the size of the aerosol droplets generated in these workplace applications differ significantly from those respirable aerosols that are generated in the laboratory studies described above. Only a very small fraction of particles generated in real-life situations are in the respirable range of humans (Ehnes et al., 2019), and thus most are filtered in the nasal passages, react with the mucus in the bronchi, and removed via mucociliary clearance. Therefore, there are few reports of the irritative effects of MDI substances in humans in the workplace.


Those reports on effects in humans that are available are very limited and considered of low relevance. In the MAK documentation for 4,4’-MDI and pMDI (Henschler, 1992), two reports from the early years of pMDI application are cited that provide limited information on pMDI-induced respiratory irritation in humans. One describes exposure of twelve workers to an aerosol mixture used to apply rigid foam insulation to the insides of railway carriages and blown by the wind to the workers standing about 20 to 40 meters away (Longley, 1964). Within a few hours, all individuals showed symptoms such as pain behind the eyes, nasal discharge, retrosternal soreness, constriction of the chest, coughing and headaches. The symptoms regressed completely within a few days. Fitzpatrick et al. (1964) reported that “vapour concentrations of 56 ppb” during spraying of a pMDI/polyol mixture did not cause irritation of workers (unspecified if respiratory or ocular) using respirators with particle filters. The MAK notes that since the saturated vapor concentration (SVC) of mMDI at 20°C is about 5 ppb, either the temperature at the sampling site was higher or the MDI was mostly present in the form of an aerosol. How much of this aerosol was retained in the mask filters is unclear.


 Conclusions – sub-lethal irritation:


Taken together, data from non-lethal acute inhalation studies indicate that a disruption of surfactant homeostasis is an early event in the pulmonary response to the deposition of MDI and occur at relatively lower MDI concentrations (NOAEC 0.7 mg/m3). Aerosols of MDI deposited in the alveolar epithelial lining fluids (surfactant) react with macromolecular nucleophiles (e.g. GSH, proteins, peptides), and when exposure concentration and/or duration is sufficient, the nucleophilic capacity of this layer becomes overwhelmed and deterioration of cell membranes and cytotoxicity occurs (e.g. total protein, LDH, apoptosis). In addition to the disruption of the surfactant homeostasis, reaction products of the MDI with alveolar macromolecules are phagocytised by activated macrophages. When activated, these macrophages release pro-inflammatory cytokines which recruit neutrophils (with a possible subsequent oxidative burst and reactive oxidative species production). The available data demonstrate early indications for toxic lung effects such as inflammation / cytotoxicity (e.g. lactate dehydrogenase (LDH), Annexin V expression) and oxidative stress (e.g. BALF glutathione (GSH) levels) at concentrations ≥8mg/m3and with consistent moderate toxicity at ≥12mg/m3. This suggests that acute exposures to MDI concentrations ≥ 12mg/m3 results in significant portal of entry cellular toxicity. 


Accordingly, the CLP classification as STOT SE 3 (H335, “May cause respiratory irritation”) for 4,4’-MDI is adopted for all substances of the MDI category.

Justification for classification or non-classification

EU classification according to CLP: H332, H335


GHS classification (GHS UN rev.2, 2007): Inhalation route (vapour): Acute Category 4.

                                                                 : Respiratory irritation STOT SE 3


Not toxic by the dermal or oral routes.