Scholarly article on topic 'Risk assessment for consumer exposure to toluene diisocyanate (TDI) derived from polyurethane flexible foam'

Risk assessment for consumer exposure to toluene diisocyanate (TDI) derived from polyurethane flexible foam Academic research paper on "Environmental engineering"

CC BY-NC-ND
0
0
Share paper
OECD Field of science
Keywords
{"Risk assessment" / "Consumer exposure" / "Toluene diisocyanate" / "Polyurethane flexible foam"}

Abstract of research paper on Environmental engineering, author of scientific article — Scott M. Arnold, Michael A. Collins, Cynthia Graham, Athena T. Jolly, Ralph J. Parod, et al.

Abstract Polyurethanes (PU) are polymers made from diisocyanates and polyols for a variety of consumer products. It has been suggested that PU foam may contain trace amounts of residual toluene diisocyanate (TDI) monomers and present a health risk. To address this concern, the exposure scenario and health risks posed by sleeping on a PU foam mattress were evaluated. Toxicity benchmarks for key non-cancer endpoints (i.e., irritation, sensitization, respiratory tract effects) were determined by dividing points of departure by uncertainty factors. The cancer benchmark was derived using the USEPA Benchmark Dose Software. Results of previous migration and emission data of TDI from PU foam were combined with conservative exposure factors to calculate upper-bound dermal and inhalation exposures to TDI as well as a lifetime average daily dose to TDI from dermal exposure. For each non-cancer endpoint, the toxicity benchmark was divided by the calculated exposure to determine the margin of safety (MOS), which ranged from 200 (respiratory tract) to 3×106 (irritation). Although available data indicate TDI is not carcinogenic, a theoretical excess cancer risk (1×10−7) was calculated. We conclude from this assessment that sleeping on a PU foam mattress does not pose TDI-related health risks to consumers.

Academic research paper on topic "Risk assessment for consumer exposure to toluene diisocyanate (TDI) derived from polyurethane flexible foam"

Contents lists available at SciVerse ScienceDirect

Regulatory Toxicology and Pharmacology

journal homepage: www.elsevier.com/locate/yrtph

Regular Article

Risk assessment for consumer exposure to toluene diisocyanate (TDI) derived from polyurethane flexible foam

Scott M. Arnold a'*, Michael A. Collins b, Cynthia Grahamc, Athena T. Jolly d, Ralph J. Parode Alan Poolef, Thomas Schuppg, Ronald N. Shiotsukah, Michael R. Woolhisera

a The Dow Chemical Company, Toxicology & Environmental Research and Consulting, 1803 Building, Midland, MI 48674, United States bInternational Isocyanate Institute, Inc., Bridgewater House, Whitworth Street, Manchester Ml 6LT, UK c Huntsman, 8600 Gosling Road, The Woodlands, TX 77381, USA

d Occupational and Environmental Medicine, 850 Penns Way, West Chester, PA 19382, USA e BASF Corporation, 1609 Biddle Avenue, Wyandotte, MI 48192, United States fDow Europe GmbH, Bachtobelstrasse 3, CH-8810 Horgen, Switzerland

gBASF Polyurethanes GmbH, Product Safety, Ecology and Toxicology, Elastogranstrasse 60, 49448 Lemfoerde, Germany h Bayer MaterialScience LLC, 100 Bayer Road, Pittsburgh, PA 15205, United States

ARTICLE INFO

ABSTRACT

Article history: Received 11 May 2012 Available online 1 August 2012

Keywords: Risk assessment Consumer exposure Toluene diisocyanate Polyurethane flexible foam

Polyurethanes (PU) are polymers made from diisocyanates and polyols for a variety of consumer products. It has been suggested that PU foam may contain trace amounts of residual toluene diisocyanate (TDI) monomers and present a health risk. To address this concern, the exposure scenario and health risks posed by sleeping on a PU foam mattress were evaluated. Toxicity benchmarks for key non-cancer endpoints (i.e., irritation, sensitization, respiratory tract effects) were determined by dividing points of departure by uncertainty factors. The cancer benchmark was derived using the USEPA Benchmark Dose Software. Results of previous migration and emission data of TDI from PU foam were combined with conservative exposure factors to calculate upper-bound dermal and inhalation exposures to TDI as well as a lifetime average daily dose to TDI from dermal exposure. For each non-cancer endpoint, the toxicity benchmark was divided by the calculated exposure to determine the margin of safety (MOS), which ranged from 200 (respiratory tract) to 3 x 106 (irritation). Although available data indicate TDI is not carcinogenic, a theoretical excess cancer risk (1 x 10 7) was calculated. We conclude from this assessment that sleeping on a PU foam mattress does not pose TDI-related health risks to consumers.

© 2012 Elsevier Inc. All rights reserved.

1. Introduction

Polyurethanes (PU) are polymers made by reacting diisocya-

nates (monomers with two isocyanate (NCO) groups) with polyols

or chemically related compounds (Fig. 1). These polymers are man-

ufactured in industrial settings for subsequent use in consumer

products such as furniture, automotive interiors, bedding, carpet

underlay, insulation and coatings. It is thought that any NCO groups present in the PU foam following curing are attached to

large molecular weight heterogenous polymers that prevents their release by either evaporation or diffusion (Vangronsveld et al., 2012). The levels of attached NCO groups decline rapidly, probably by reacting with ambient atmospheric moisture (Cole et al., 1987). Fully cured PU products are considered toxicologically inert (USEP-A, 2011a) since they contain neither unreacted TDI nor biologically available NCO groups (Dieterich et al., 1993).

Despite these generally held beliefs, there are reports that unreacted TDI is present in cured PU foam. Gagné et al. (2003) and

Abbreviations: BHR, bronchial hyperreactivity; BMDS, USEPA benchmark dose software; CSF, cancer slope factor; FEV, forced expiratory volume; FEV1, forced expiratory volume in one second; FVC, forced vital capacity; GPMT, guinea pig maximization test; LADD, lifetime average daily dose; LMS, linearized multistage; LLNA, local lymph node assay; LOAEL, lowest observable adverse effect level; MOS, margin of safety; MDL, method detection limit; NTP, national toxicology program; NESIL, no expected sensitization induction level; NOAEL, no observed adverse effect level; NOEL, no observed effect level; OEL, occupational exposure limit; POD, point of departure; PU, polyurethanes; RfC, reference concentration; TDI-GSA, TDI-guinea pig serum albumin conjugates; TWA, time-weighted-average; TDA, toluene diamine; TDI, toluene diisocyanate.

* Corresponding author. Fax: +1 989 638 2425.

E-mail addresses: smarnold@dow.com (S.M. Arnold), office@globalisocyanates.co.uk (M.A Collins), cynthia_graham@huntsman.com (C. Graham), athena.jolly@gmail.com (AT. Jolly), ralph.parod@basf.com (R.J. Parod), apoole@dow.com (A. Poole), thomas.schupp@basf.com (T. Schupp), ron.shiotsuka@bayer.com (R.N. Shiotsuka), mwoolhi-ser@dow.com (M.R. Woolhiser).

0273-2300/$ - see front matter © 2012 Elsevier Inc. All rights reserved. http://dx.doi.org/10.1016Zj.yrtph.2012.07.006

Fig. 1. Generic reaction of TDI with a polyol to form a polyurethane.

Krone et al. (2003) reported the presence of low concentrations of ''free TDI'' (i.e., residual unreacted TDI monomer) in PU products, even after prolonged aging. However, the presence of free TDI is unlikely. The absence of TDI emissions [limit of detection (LOD) of 0.2-0.5 ng/g] from PU foam spiked with TDI (Hugo et al., 2000) suggests that results reported by Gagné et al. (2003) are likely due to either degradation of the foam by the extraction procedure and/or reaction of the derivatization agent with other extractable, low-molecular weight components (e.g., oligourea) in the foam. In addition, the reliability of the Krone et al. (2003) results have been questioned based on the absence of both control samples and the positive identification of isocyanates, use of dimethyl sulfoxide as solvent, and the unusual ratios of putative TDI isomers in different samples (Cleet, 2005). The TDI commonly used in foam manufacture is an isomer mixture of 2,4- and 2,6-TDI (80:20). Toxicity tests have predominantly been performed on this isomer mixture, and in this manuscript TDI is used to refer gener-ically to the isomer mixture or single isomers.

While few concerns are expressed about the safety of PU foams, this assessment quantitatively addresses the concern that unre-acted TDI monomer may escape from the polymer matrix and pose a health risk. Sleeping on a PU foam mattress was the exposure scenario selected for this assessment since it represents a situation that would accentuate the potential for consumer exposure.

2. Methodology

It is generally accepted that risk assessment can be performed in the following four steps (National Research Council, 1983):

• Hazard assessment

• Dose-response assessment

• Exposure assessment

• Risk assessment

Risk is a function of both toxicity (hazard and dose-response) and exposure. The sections below detail the risk assessment process for TDI. By combining appropriate analytical data and exposure modelling with the toxicity benchmarks provided herein, similar risk characterizations can be made for other TDI-based PU products.

2.1. Toxicity assessment

Toxicity refers to the inherent property of every chemical to cause adverse health effects at some level of exposure. It can be determined by evaluating responses in experimental animals exposed to chemicals under defined laboratory conditions (most commonly) or in humans exposed to chemicals in their environment. Traditionally, toxic effects have been broadly divided into non-cancer and cancer endpoints. In this document, only toxicity endpoints relevant to TDI (i.e., skin irritation; skin sensitization; respiratory sensitization; lung irritation and decrement not related to asthma; and carcinogenicity) are considered. Dose-response data on the potential adverse health effects from TDI in PU products are likely the same as that for the monomer itself. Comprehensive information about the toxicity of TDI monomer is available in

various reviews or agency documents (Bolognesi et al., 2001; Canadian Government, 2008; Collins, 2002; ECHA, 2011; European Commission Joint Research Centre, 2000; IARC, 1999; National Research Council, 2004; Ott, 2002; USEPA, 1995).

For non-cancer endpoints, toxicity benchmarks were identified by dividing each point of departure (POD) by a combined uncertainty factor. PODs (e.g., No Observed Adverse Effect Level, NOAEL) were selected following a review of relevant dose-response data for each endpoint. Combined uncertainty factors were calculated as the product of two or more factors that compensate for uncertainties associated with the POD (e.g., inter- and intra-species variability, extrapolation from less than lifetime-to-lifetime exposures and weakness of the toxicological database). Uncertainty factors were derived from regulatory and other expert guidance and refinements to these factors based on the toxicological database for TDI. For cancer, the cancer slope factor (CSF) was derived using the USEPA Benchmark Dose Software (BMDS) (http://www.epa.-gov/ncea/bmds/dwnldu.html) version 2.2 and the USEPA preferred linearized multistage (LMS) cancer model to describe the relationship between risk and dose below the experimentally observed range. The CSF is defined as the upper 95% confidence limit on the slope of the risk-dose relationship with units of risk per mg/ kg bw/day.

2.2. Exposure assessment

Exposure is a function of the mass of material in contact with the body as well as the exposure conditions under which contact occurs. In the case of dermal irritation and sensitization where contact alone is a key parameter, the appropriate dose-metric is mass per unit area (European Commission, 2010; Kimber et al., 2008; Loveless et al., 2010). Otherwise, exposures are commonly expressed in units of mg/kg bw/day, the amount of material contacted (mg) per mean body weight (kg) of the subject per unit of time (day). Exposures can be measured empirically or modelled based on the physical-chemical properties of the chemical, physiological characteristics of the receptor population (e.g., absorption rate, inhalation rate, body weight), and contact conditions (i.e., frequency, duration, and route) for the exposure scenario under study. Selected exposure variables typically represent a combination of average and upper-bound values designed to estimate the exposure experienced by individuals at the upper end (e.g., 90th percentile) of the exposure distribution. A common exposure metric for cancer endpoints is the Lifetime Average Daily Dose (LADD).

If there were to be human exposure to TDI monomer from PU products, in principle it would be most likely to occur via inhalation (release of TDI to air) and dermal contact (transfer of TDI to the skin). One scenario which would provide maximal opportunity for such exposure via these routes would be sleeping on a flexible PU foam mattress. A high index (119) PU foam was selected for this assessment. The index refers to the stoichiometric ratio of TDI to polyol used to manufacture the foam, with a value of 100 indicating equal amounts of TDI and polyol. The index of commercial PU foam bedding typically ranges between 105 and 115. The high index PU foam used for this assessment was conservatively selected to represent PU foam most likely to contain residual unreacted TDI. The magnitude of other potential TDI exposures (e.g., ingestion of

food contacting PU products, contact with medical devices) are not considered herein as they are expected to be relatively minor routes of exposure; in contact with traces of water, TDI would polymerize forming polyurea (Mormann et al., 2006).

2.2.1. Dermal exposure

The amount of TDI available for dermal exposure during sleep was based on results from migration studies by Vangronsveld et al. (2012). In this study, migration of TDI was determined by placing a migration collection system (i.e., filter paper disks impregnated with a TDI trapping agent) in contact with the surface of PU flexible foam, then clamping the foam/migration collection assembly together (foam compression ~25% by height) for periods of either 8 or 24 h. At the end of the test period, the cell was disassembled, and the migration collection system was analyzed for TDI.

2.2.2. Inhalation exposure

The TDI concentration in air during a sleep period was determined based on results from emission studies by Vangronsveld et al. (2012). These studies were carried out on PU flexible foam using three different emission cells - the FLEC® cell, the micro-chamber (i-CTE™) and a flow through cell. The TDI emitted to the air passing through each test cell was measured using a glass fiber filter impregnated with a TDI trapping agent. Both the FLEC® cell and i-CTE™ are used by commercial product emission testing laboratories and have been referenced by international standard methods for emission testing (i.e., prENV 13419-2, ISO/DIS 16000-10, ASTM 7143-05). The FLEC® cell was determined to be the most appropriate for the derivation of the inhalation exposure concentration since it afforded the most sensitive measurement of emitted TDI.

2.3. Risk characterization

The assessment of risk depends on the type of adverse health effect. For non-cancer effects, a general approach is to divide the toxicity benchmark by the calculated exposure value to derive a Margin of Safety (MOS). The target MOS is 1 since the associated uncertainty factors (e.g., inter- and intra-species variability) are included within the derivation of the toxicological benchmark. For clarity, this term is different than the term ''Margin of Exposure'' (MOE), which is defined in this context as the ratio of the POD (such as a NOAEL) and the estimated exposure dose or concentration. The target MOE then reflects the magnitude of the uncertainty factor(s) associated with the particular toxicological endpoint. For genotoxic carcinogens, regulatory agencies conservatively assume that cancer risks exist below the experimentally derived NOAEL (i.e., there is no threshold) and approach zero only when exposures do. To address potential cancer concerns, the LADD was multiplied by the CSF for TDI to determine an upper bound on the excess cancer risk for a lifetime of sleeping on a foam mattress and the product was compared to the acceptable excess cancer risk of 10~6 commonly targeted by regulatory agencies for the general population.

3. Results

3.1. Toxicity assessment

3.1.1. Skin irritation

Studies in rabbits have reported that TDI produces moderate dermal irritation when applied to intact or abraded skin (Duprat et al., 1976; Knapp and Baker, 1974). In a mouse Local Lymph Node Assay (LLNA) examining the dermal sensitization potential of TDI,

Woolhiser et al. (1998) reported irritation at the TDI application site. Other rabbit and mouse studies have reported variable results ranging from slight irritation (Wazeter et al., 1964) to corrosion (unpublished study by Suberg et al., 1984). While in isolation these studies are insufficient to characterize the dermal irritation potential of TDI due to the lack of documentation and use of non-standard methodology, the overall weight of evidence indicates that TDI is irritating to the skin of experimental animals.

Studies in humans (Daftarian et al., 2002; Huang et al., 1991) have reported skin effects associated with occupational exposures to TDI. While it is often unclear if the skin effects are attributable to primary irritation or sensitization, there is a suggestion that irritant dermatitis is more common than allergic contact dermatitis (Daftarian et al., 2002). In subjects with occupational skin disease, approximately 2% of the 360 patients investigated showed evidence of skin irritation when 15 mg of a 1.5% or 2% TDI solution was applied to the skin for two days (Kanerva et al., 1999). Given the test chamber area (50 mm2), a 1.5% solution corresponds to a skin exposure of 600 ig TDI/cm2. This value is considered the Lowest Observable Adverse Effect Level (LOAEL) and is divided by a combined uncertainty factor of 3 to derive a NOAEL of 200 ig TDI/cm2 for human skin irritation. The combined uncertainty factor is based on a factor of 3 for LOAEL to NOAEL conversion and a factor of 1 for intraspecies variability. The former reflects the extreme exposure conditions (i.e., two days contact under occlusive dressing), the low incidence of irritation observed by Kanerva and coworkers, as well as the lack of human skin irritation reported by others. An intraspecies factor of 1 reflects the fact that the LOAEL was derived from a relatively large group of potentially sensitive individuals attending a clinic for skin problems.

3.1.2. Skin sensitization

The complexities of skin sensitization thresholds are not well understood making it difficult to assess the relevance of the comparatively low thresholds to TDI seen in animals to the apparently high thresholds predicted by human experience. The no observed effect level (NOEL) for the induction of sensitization derived from animal models appears to be lower than that in humans. Behind this observation are extreme exposure conditions associated with animal protocols and the absence of well documented cases of skin sensitization in the workplace. This contention is consistent with the animal and human data summarized below.

In animals, TDI is a well recognized skin sensitizer (Auletta, 1984; Duprat et al., 1976; Hilton et al., 1995; Karol et al., 1981; van Och et al., 2000; Thorne et al., 1987; Woolhiser et al., 1998; Zissu et al., 1998). Two studies that followed guideline methodologies and included dose-response information were those by Aul-etta (1984) and Hilton et al. (1995). Auletta (1984) performed a dose-response sensitization study in Guinea pigs. While irritation was observed at an induction dose of 30 mM TDI (0.52% weight/ volume), sensitization reactions were not seen at any challenge dose, demonstrating a NOEL for induction. Laboratory experience suggests that the topical application of 50 il can cover up to 15 cm2 of skin. Thus, the NOEL of 30 mM in Guinea pigs corresponds to a dermal sensitization threshold of about 17,000 ng/ cm2. Hilton et al. (1995) determined an EC3 of 0.02% (w/v)1 for dermal sensitization using the mouse LLNA, which has emerged as a preferred method for sensitization testing due to the pragmatic and scientific advantages it affords (Loveless et al., 2010). Based on the LLNA protocol (i.e., 25 il applied to 1 cm2 of skin on each ear), the EC3 of 0.02% corresponds to a dermal exposure of 5000 ng TDI/

1 EC3 is defined as the concentration leading to a threefold increase of the baseline lymph node cell proliferation in the LLNA using thymidine labeling (Kimber et al., 2001).

In humans, allergic contact dermatitis (ACD) due to TDI is considered rare and has been associated with poor workplace hygiene (Huang et al., 1991). In a flexible foam plant, 26 TDI exposed workers reporting skin symptoms underwent patch testing with a 2% TDI solution, none developed a reaction to the TDI antigen (Daftar-ian et al., 2002). In addition, only 1% of dermatology patients with a suspected sensitivity to diisocyanates demonstrated an allergic skin reaction to TDI (Liippo and Lammintausta, 2008). In a series of 360 dermatology patients skin tested for ACD to plastics and glues, 1.5 or 2% TDI (i.e., 4.5 x 105ng/cm2 or 6 x 105ng/cm2) was found to elicit an allergic response in only 0.8% of cases (Kanerva et al., 1999). This study provides prevalence rates for a dermatology clinic population, where patients were tested for occupational skin disease using a modified European series that included a panel of 26 standard allergens (Bruynzeel et al., 1995) and 20-30 other compounds relevant to their work and hobbies. As the standard series itself is expected to have false positive rate of at least 5% (Nethercott, 1990), the prevalence of 0.8% for TDI does not indicate a meaningful level of TDI skin sensitization in this cohort.

A number of approaches have been advanced in recent years (e.g., ECETOC, 2008; ECHA, 2010; Felter et al., 2003; Griem et al., 2003; Kimber et al., 2001; Safford, 2008) that use relative potency data (i.e., EC3) obtained from mouse LLNA studies to derive a ''no expected sensitization induction level'' (NESIL). Different features of these approaches are described in Table 1. Aside from common uncertainty factors that account for interspecies and intraspecies, these approaches take into consideration factors that are more specific to dermal exposure and sensitization testing. Taking these approaches into consideration for the current PU foam scenario, a benchmark toxicity value (NESIL) for the general population of 333 ng/cm2 was derived by dividing the NOEL (EC3) of 5000 ng/ cm2 by a total uncertainty factor of 15 that was based on multiple considerations. The LLNA appears to be a sensitive indicator of dermal sensitization potential relative to human experience, and a factor of 3 was used to account for potential interspecies variability in the EC3 value. In addition, despite the rarity of skin sensitization in humans, a factor of 5 was selected for intraspecies variability to account for potential uncertainties associated with the human development of and predisposition to ACD. A factor of 1 was selected for matrix effects since the availability of TDI in acetone:olive oil (LLNA) is likely greater than that of TDI, which must passively migrate through multiple molecular barriers (e.g., fabric, sweat) before contacting the skin. Factors that might be used to reflect enhanced penetration, longer exposure, dermal integrity, or occlusive exposure are not applicable to the scenario of sleeping on a mattress. Uncertainty factors of 1 were assigned for considerations such as data quality, dose-response profile, and reliability (ECHA, 2010) since the LLNA data were reported in a peer-reviewed publication from a reliable laboratory and were supported by similar data from other investigators. The calculated NESIL value is judged conservative given the relatively rare induction of ACD to TDI in humans. The discrepancy between the animal data and human experience must be considered when evaluating their respective relevance for risk assessment.

3.1.3. Respiratory sensitization

The ability of TDI to induce respiratory sensitization in susceptible individuals is a known adverse health effect in humans. While various animal models (i.e., induction by inhalation or dermal exposures) have been constructed that measure respiratory responses (i.e., immediate-onset and/or delayed-onset) following an inhalation challenge with TDI (e.g., Botham et al., 1988; Karol, 1983; Pauluhn and Mohr, 1998), there is still no validated model that adequately reflects the respiratory sensitization process and response in humans. In particular, the role that dermal contact

with TDI plays in the development of respiratory sensitization remains an open question. Data on this issue have been considered (Graham et al., 2002) and it was concluded that while animal and human data suggest the immune system can be activated by topical exposures to TDI, it is unclear whether dermally-mediated activation is sufficient to initiate respiratory sensitization in humans. Given this mechanistic uncertainty, as well as the absence of clinical data demonstrating that TDI skin exposures lead to an asthmatic response, this assessment focuses on only the potential risk of respiratory sensitization induced by the inhalation of TDI. Data in animals and humans relevant to this topic are presented below. Although dose-response relationships are not clearly identified, the limited data indicate that the induction of respiratory sensitization is a threshold phenomenon, although the NOEL is dependent on the data set examined.

The induction of respiratory sensitization was evaluated by exposing Guinea pigs (head only) to TDI vapor at concentrations of 20 ppb for 70 days or 120-7600 ppb for a week followed by an inhalation challenge with TDI-guinea pig serum albumin conjugates (TDI-GSA) (Karol, 1983). Respiratory responses were evaluated by measuring increases in respiratory rate and antibody production after challenge with 1% TDI-GSA. The NOEL for induction was 120 ppb TDI based upon antibody production. Using a mouse model, Matheson et al. (2005) reported respiratory tract responses following the daily inhalation of 20 ppb TDI for 6 weeks. After a two week non-exposure period, animals were challenged via inhalation to 20 ppb TDI for 1 h. TDI-treated mice demonstrated enhanced airway inflammation and some hyperreactivity to methacholine (PENH), elevated IgE and IgG antibody levels, and increased Th1/Th2 cytokine expression in lung tissue. These responses support a LOEL for induction of 20 ppb in this mouse model.

Many clinical and epidemiologic studies have been carried out since 1950 in order to evaluate the risk of developing occupational asthma. Recent studies have shown a downtrend in the global and regional incidences of occupational asthma in general and of isocy-anate related asthma in particular since the 1990s (Bakerly et al., 2008; Buyantseva et al., 2011; Paris et al., 2012; Vandenplas et al., 2011), perhaps indicative of the increasing emphasis on workplace exposure controls. In a large Canadian study of 223 companies using diisocyanates between 1984 and 1988 (Tarlo et al., 1997), the calculated overall annual incidence of occupational asthma was higher (0.7%) for companies with at least one case of occupational asthma and time-weighted-average (TWA) concentrations >5 ppb than for all potentially exposed employees (0.2%) of the companies under study (Ott et al., 2003). Detailed reviews (Ott, 2002 and Ott et al., 2003) have also shown an association between the declining incidence of diisocyanate asthma and decreasing levels of airborne TDI in the workplace. However, even in workplaces with TWA TDI concentrations below 5 ppb, short-term concentrations exceeded 20 ppb during some work activities and occasionally exceeded 80 ppb (Ott et al., 2003). These high short-term exposures may explain the cases of diisocyanate asthma seen with TWA TDI concentrations below 5 ppb, since studies with detailed exposure data indicate that high, short-term exposures to TDI can lead to respiratory sensitization (Bugler et al., 1991; Ott et al., 2000; Weill etal., 1981). This hypothesis is consistent with the occupational experience that respiratory sensitiza-tion does not occur when concentrations of 5 ppb and below are maintained (DFG, 2003; Diller, 1998). In a review of the critical data for the TDI OEL, it was stated that ''If the exposure concentrations of TDI are kept below 10-20 ppb, generally no new cases of TDI asthma are observed'' (AGS, 2006). These considerations have led to the acceptance of 5 ppb (8-h TWA) and 20 ppb (short term limit) by many regulatory authorities for control of occupational exposures to TDI.

Table 1

Uncertainty factor scheme for the dermal sensitization endpoint.

Uncertainty factor (UF) ECHA ECETOC Safford (2008) Griem et al. Felter et al. (2003) TDI-PU flexible foam

(2010) (2008) (DST/TTC - animal use) (2003) (Ansad and Anead)

Interspecies (animal to human) 10(EC3 - Not Log10 NESIL lg/cm2 = 1.16 EC3 » NOAEL, EC3 lower bound potency 3 (EC3 » NOAEL,

LOAEL-NOAEL)a warranted Log10 EC3 lg/cm2 - 0.64 (~4-1 fold factor) Variability = 3 classification » default NOAEL Variability = 3)

Intraspecies (variability) 5-10 10 10 10 10 5 (occurrence rare, variability low)

Matrix 1-10 <1-10 1-10 - 1-10 1 (acetone: stringent conditions)

Case-by-case exposure conditions 10 1-10 1-10 10 1-10 n/a

(repeated, duration, dermal integrity,

humidity, product use)

Dose-response, reliability 1-10 - - - - 1

Data quality 1-10 - - - - 1

Typical total UF (assumes quality data) 500-1000 100 100-400 300 100 15

EC3 is defined as the concentration leading to a threefold increase of the baseline lymph node cell proliferation in the LLNA using thymidine labeling (Kimber et al., 2001). DST - dermal sensitization threshold & TTC - threshold of toxicological concern. NESIL - no expected sensitization induction level.

ANSAD - acceptable non-sensitizing area dose & ANEAD - acceptable non-eliciting area dose. NOAEL = No observable adverse effect level. LOAEL = Lowest observable adverse effect level.

a ECHA TGD Appendix R.8-10. EC3 data generally correlate well with human skin sensitisation thresholds derived from historical predictive testing; however, there are cases where this correlation is poor and the two values may differ by 10-fold or more. In view of this variation, the default AF of 10 for interspecies variation (see Section R.8.4.3) should be used, unless there is evidence (e.g., from a close analogue of the substance in question) of good correlation between the EC3 and human NOAEL/ LOAEL.

Based on the available human data, 5 ppb was selected as the NOAEL for the induction of respiratory sensitization in workers. The NOAEL was adjusted to 1.8 ppb (i.e., 5 ppb x 5/7 x 10/20) to account for the difference in exposure frequency between working (5 days/week) and sleeping (7 days/week) as well as the volume of air inhaled performing 8 h of light to moderate work (10 m3) and daily activities (20 m3) (USEPA, 2011b). The benchmark toxicity value of 0.9 ppb (6400 ng/m3) for the induction of respiratory sen-sitization in the general population was derived by dividing 1.8 ppb by a total uncertainty factor of 2 for intraspecies variability (ECHA, 2010).

3.1.4. Respiratory tract effects - irritation and lung decrement not related to asthma

While asthma is thought to be mediated by immunological mechanisms, the respiratory tract irritation initiated by TDI appears to involve activation of the transient receptor potential Ankyrin 1, an ion channel receptor located in sensory nerves lining the airways (Taylor-Clark et al., 2009) which, in response to noxious stimuli, mediates reflex respiratory responses including, cough, rhinitis, bronchoconstriction, and neurogenic inflammation (Jordt et al., 2004). These sensory mediated effects may lead to an accelerated decline in pulmonary function. As no animal studies could be found that distinguished between thresholds for asthma and those for irritation and pulmonary function, the point of departure for these endpoints relies on epidemiology studies of worker populations.

Studies by Henschler et al. (1962) in human volunteers reported that acute exposures to TDI resulted in mild respiratory tract irritation at 50 ppb TDI, but not at 20 ppb. Other investigators have reported results of inhalation challenge tests in healthy volunteers as well as persons with asthma or bronchial hyperreactivity (BHR) with no known isocyanate exposure. Different exposure protocols were used but, in general, TDI concentrations did not exceed 20 ppb. Baur (1985) reported mild irritation in healthy subjects and more severe respiratory symptoms, cough and chest tightness, in 4 of 15 asthmatics. Other studies detected no evidence of irritation, inflammatory response, or changes in forced expiratory volume (FEV) in subjects with bronchial hyperreactivity (BHR), in

non-isocyanate asthmatics, or in healthy controls after TDI challenge (Chester et al., 1979; Fabbri et al., 1987; Moller et al., 1986). In a study of 17 human volunteers without previous exposures to isocyanates, 12 subjects exhibited moderate BHR after exposure to TDI at 5 ppb for 6 h followed by a 20 min exposure at 20 ppb. While none of the subjects had significant respiratory symptoms, some marginal changes in airway caliber and epithelial permeability were found that were attributed to a possible pharmacological effect. In conclusion, TDI concentrations of 20 ppb for up to 30 min have not been found to produce notable effects in healthy individuals. Persons diagnosed with non-TDI related asthma or BHR have been found to be more sensitive to inhalation exposures to TDI, experiencing symptoms and changes in specific airway resistance at TDI concentrations of 10-20 ppb. It has been noted that persons diagnosed with TDI-induced asthma can be more sensitive, reacting to lower concentrations of TDI (O'Brien et al., 1979).

Acute bronchial irritation following exposure above the safe occupational level can lead to inflammation and airflow limitations, which are reversible. Such transient declines in lung function across a working shift were observed in the earlier TDI literature (Gandevia, 1963). It can be postulated that daily exposure sufficient to cause chronic lower airway inflammation or massive overexposure leading to airway remodeling can impair respiratory function and lead to the permanent lung function decrement also reported in the earlier literature (Adams, 1975; Holness et al., 1984; Peters and Wegman, 1975; Wegman et al., 1974).

Two key indicators of pulmonary function employed in epide-miological studies are the forced expiratory volume in one second (FEV1) and forced vital capacity (FVC). A large number of studies focusing on pulmonary function have been conducted in TDI manufacturing units as well as in foam production plants and spray painting facilities. These have included both cross-sectional and longitudinal studies. Occupational experience has demonstrated that long-term exposures above 20 ppb may result in a significant decline in FEV1 and Peak Expiratory Flow Rate over time; whereas, concentrations of 5 ppb and below seem to be safe against airway function impairment and sensitization. Reports of respiratory effects observed between 5 and 20 ppb have been negative (Musk

et al., 1985) or inconclusive (DFG, 2003; Diller, 1998). Based on pulmonary function results from larger and more recent longitudinal studies where the TWA TDI concentrations were 65 ppb, Ott and co-workers (2002, 2003) concluded that ''these studies do not provide evidence of an accelerated rate of decline in FEV1 among TDI-exposed employees''. This conclusion is supported by the results of a prospective study of 251 TDI polyurethane foam production workers in the UK followed over a period of 17 years (Clark et al., 2003). The decreases in FEV1 and FVC seen in this study were unrelated to TDI exposure level and were comparable to those measured in other populations not exposed to TDI. Consistent with these observations, the German Commission for the Investigation of Health Hazards of Chemical Compounds in the Work Area (DFG, 2003) concluded that workplace exposures to TDI should be limited to a TWA of 5 ppb (35 x 103 ng/m3) to prevent accelerated deterioration of lung function in exposed workers.

For non-occupational settings, the US Environmental Protection Agency (USEPA, 1995) calculated a ''Reference Concentration" (RfC)2 for TDI using five years of worker exposure data reported by Diem et al. (1982). In the Diem et al. (1982) study, workers were divided into two groups, those with low and high cumulative exposures to TDI. When comparing the two groups, the authors reported a significant decline in lung function among never smokers in the high exposure group, but not among previous or current smokers. Although offering no explanation for these discrepancies, USEPA (1995) judged the TWA exposure (0.9 ppb TDI) of never smokers in the low cumulative exposure group to be the study NOAEL and the TWA exposure (1.9 ppb) for never smokers in the high exposure group to be the LOAEL. EPA used the former value as the basis for its RfC calculation. However, there are concerns that 0.9 ppb may be overly conservative. For example, Diem et al. (1982) suggested that the effect observed in never smokers may be due to the peak exposures, noting that TWA concentrations in the high exposure group exceeded 5 ppb 15% of the time but only 2% of the time in the low exposure group. Indeed, when the same worker data were reviewed by Weill et al. (1981), it concluded that the time spent above 20 ppb was well correlated with the annual decline in pulmonary function. The association of pulmonary effects with high, short-term peak exposures is consistent with the conclusion by Diem et al. (1982) and by others in more recent literature reviews (Ott et al., 2000; Ott, 2002; Ott et al., 2003) that an 8-h TWA of 5 ppb adequately protects against pulmonary lung function decrements caused by TDI. Despite these concerns, the RfC of 70 ng/m3 (0.01 ppb) based on a general population NOAEL of 2000 ng/m3 (0.3 ppb) was conservatively selected as worst-case benchmark tox-icity value for pulmonary irritation and decrements in lung function in the general population.

3.1.5. Carcinogenicity

TDI is classified as a Group 2B carcinogen (possibly carcinogenic to humans) by the International Agency for Research on Cancer (IARC, 1999), as a Category 2 carcinogen (suspect human carcinogen) in the European Commission (1998), as reasonably anticipated to be a human carcinogen by the US National Toxicology Program (NTP, 2011), and as an A4 carcinogen (not classifiable as a human carcinogen) by the American Conference of Governmental Industrial Hygienists (ACGIH, 2010). These classifications are based on the increased tumor incidences observed by the NTP (1986) when TDI in corn oil was administered directly into the stomach of rodents by oral gavage. However, this study was flawed both technically (i.e., mishandling of the test material) and conceptually

2 The RfC is defined as ''an estimate (with uncertainty spanning perhaps an order of magnitude) of a daily inhalation exposure of the human population (including sensitive subgroups) that is likely to be without an appreciable risk of deleterious effects during a lifetime''.

(i.e., gavage exposures) resulting in the formation of toluene diamine (TDA), a known animal carcinogen, both prior to and after TDI administration (NTP, 1986; Appendix I; Dieter et al., 1990). Given the qualitative similarity between the carcinogenic responses seen in rodents exposed to TDI and TDA (NCI, 1979), the NTP (Dieter et al., 1990) concluded that the degradation of TDI to TDA could explain the carcinogenic effects noted with TDI. Quantitative support for this conclusion comes from two studies. In the first (Tim-chalk et al., 1994), rats were gavaged with either 60 mg/kg TDI in corn oil (same dose as used by NTP) or 3 mg/kg TDA. Urinary analyses indicated that both produced comparable metabolic profiles of free TDA, N-acetylated TDA, and TDI/TDA conjugates. This finding is consistent with about 5% of the TDI gavage dose (i.e., 3 mg/kg of the 60 mg/kg TDI dose) being converted to TDA. In the second (Sielken et al., in press), a statistical comparison of the carcinogenic responses seen with TDI (NTP, 1986) and TDA (NCI, 1979) support the conclusion that the carcinogenic responses to TDI are consistent with 5% of the gavaged TDI being transformed to TDA either before and/or after exposure. The NTP (Dieter et al., 1990) dismissed concerns over its flawed study by stating that TDA would be similarly formed if exposures occurred via inhalation. This misconception persists in the scientific community despite data to the contrary.

The reactivity of TDI and its propensity to form TDA is different in pure aqueous versus complex biological systems. Whereas the formation of ureas and polyureas is the predominant reaction pathway in water at neutral pH, conjugation with biomolecules dominates in complex biological systems (Day et al., 1997; Mor-mann et al., 2006; Seel et al., 1999). The reactions of TDI in biological systems can be influenced by the pH of the in vivo environment. The pH neutral and macromolecule-rich environments associated with physiological exposures (i.e., inhalation, dermal, buccal) to TDI favor conjugation with macromolecules with no detectable free TDA (Mormann et al., 2006; Rosenberg and Savolainen, 1985; Timchalk et al., 1994). In contrast, the introduction of TDI directly into the acidic environment of the stomach (i.e., bolus dose by gavage) favors the formation of free TDA, which can be detected systemically (Jeffcoat, 1988; Kennedy and Brown, 1998; Timchalk et al., 1994). A testament to the influence of pH on the conversion of TDI to TDA is the laboratory practice of using acid hydrolyses to convert TDI/TDA conjugates in biological fluids to free TDA (Skarping et al., 1994). The in vivo conversion of TDI to TDA and the subsequent induction of a carcinogenic response only under aphysiological (i.e., gavage) exposure conditions is consistent with the observations that (a) lifetime inhalation exposures of rodents to TDI vapor at a maximum tolerated concentration of 150 ppb did not elicit a carcinogenic response (Loser, 1983; Owen, 1984), (b) free TDA was not detected in rats following a 6-h inhalation exposure to TDI vapor at 2 ppm (Timchalk et al., 1994), a concentration 400-fold higher than the TDI TLV, (c) free TDA was not detected in the urine of TDI exposed workers before subjection to acid hydrolysis (Skarping et al., 1994), and (d) three epidemio-logical studies with updates, representing the combined long-term mortality experience of more than 17,000 PU foam production workers, failed to find an association between occupational exposure to diisocyanates and an increased risk of cancer (Hagmar et al., 1993a,b, updated by Mikoczy et al., 2004; Schnorr et al., 1996; Sorahan and Pope, 1993, updated by Sorahan and Nichols, 2002).

The data summarized above indicate that potential physiological exposures to TDI such as those reported here (i.e., inhalation and dermal contact) likely result in the formation of conjugates by the reaction of TDI with biological macromolecules (Brown and Burkert, 2002; De Marzo et al., 2000; Lange et al., 1999; Mraz et al., 1998, 2000; Ogawa et al., 2006; Valstar et al., 2004; Wisnewski et al., 2006; Ye et al., 2006), with no detectable free

TDA(Mormann et al., 2006; Rosenberg and Savolainen, 1985; Skar-ping et al., 1994; Timchalk et al., 1994). Since inhalation exposures in rodents and humans are not associated with carcinogenic effects and do not produce detectable levels of free TDA in either species, the cancer risk posed by the inhalation of TDI is considered negligible and is not further evaluated herein.

Although dermal contact with TDI is expected to result in the formation of polyureas on the skin surface and conjugates between TDI and skin macromolecules, long-term studies of dermally applied TDI have not been conducted. While acute exposure studies in rats indicate a small fraction of dermally applied radiolabelled TDI (0.9%) can reach the systemic circulation (Hoffmann et al., 2010), it is unlikely to be in the form of free TDA (Rosenberg and Savolainen, 1985). Given the classification of TDI as a potential human carcinogen by some authoritative bodies, the following calculation was done with the conservative assumption that the TDI that reaches the systemic circulation will be converted into TDA. The following calculation of the excess cancer risk posed by dermal exposures to TDI is based on this theoretical assumption and is evaluated here by extrapolation to the only available tumor data, the NTP gavage study.

Using tumor data from the TDI gavage study in rodents (Dieter et al., 1990; NTP, 1986), the carcinogenic potency of TDI was calculated using the USEPA benchmark dose methodology (http:// www.epa.gov/ncea/bmds/dwnldu.html) version 2.2 and the LMS cancer model incorporated in this software. CSF calculations were performed on the tumor rates adjusted for intercurrent mortality. While CSF calculations are frequently used for genotoxic (non-threshold) carcinogens, this approach is used here as a health protective assumption and not to imply that TDI is genotoxic after dermal administration. CSF calculations were performed for all tumors observed in rats and mice, and the CSF with the highest unit risk was conservatively selected to represent the carcinogenic potency of TDI. The representative CSF (1.7 x 10~8 perng TDI/kg bw/day) was derived from combined fibroma and fibrosarcomas in the male rat. This tumor type may be relevant for dermal exposures. The dosages were 0, 30 or 60 mg/kg bw/day administered 5 days per week. The CSF was adjusted using a rat to human allometric scaling factor of 4, and a correction factor of 7 days/5 days (continuous exposure) (ECHA, 2010) and subsequently divided by the gastrointestinal absorption efficiency of TDI (20%; Timchalk et al., 1994) to yield a final CSF of 4.8 x 10~7 per ng TDI/kg bw/day.

3.2. Exposure characterization

This assessment determines the amount of TDI monomer that might migrate or be emitted from PU foam based on the results of a novel set of studies described by Vangronsveld et al. (2012). These studies do not rely on solvent extraction of PU polymers that can result in chemical degradation of the polymer and the release of molecules bound in the polymer matrix. The TDI levels measured by Vangronsveld et al. (2012) were subsequently combined with conservative (health protective) exposure factors (e.g., inhalation rate, contact area, exposure duration and frequency) to determine upper-bound dermal and inhalation exposure values (for non-cancer endpoints) and the LADD (for the cancer endpoint) associated with sleeping on a flexible PU foam mattress.

3.2.1. Skin irritation, sensitization and LADD

TDI was not detected migrating from foam with a method detection limit (MDL) of 0.16 ng TDI (combined isomers)/cm2 of foam surface area. Based on one-half the MDL, the mass of TDI assumed contacting the skin for the assessment of irritation and sensitization was 0.08 ngTDI/cm2. Since the LADD is based on a lifetime exposure, three age groups were considered: <2 years of age, 2 to <16 years of age, and 16-70 years of age (USEPA, 2005).

The LADD (i.e., potential internal dose of TDI) was calculated by combining the external dose to the skin (0.08 ng TDI/cm2), an 8-h dermal absorption value for 2,4-TDI in rat skin of 0.9% (Hoffmann et al., 2010), and age-specific variables as outlined in Table 2. The calculated LADD likely overestimates any potential exposure since human skin is generally recognized as being 3-fold to 10-fold less permeable to chemicals than rat skin and since the PU mattress and much of the skin would typically be covered by fabric which would impede TDI migration and thereby reducing the extent of any exposure.

3.2.2. Respiratory tract sensitization and lung decrement

TDI was not detected in emissions from PU flexible foam (Vangronsveld et al., 2012). MDLs of 0.003 ng/cm2 and 0.002 ng/cm2 were reported for 2,4-TDI and 2,6-TDI, respectively. Summing these values gives an estimated minimum detectable total TDI emission rate of 0.005 ng/cm2 (50 ng TDI emitted/m2 of foam). An upper-bound air concentration (2.2 ng TDI/m3) was determined based on an emission rate of 25 ng TDI/m2 (i.e., one-half the MDL; USEPA, 1989) and the assumptions in Table 3. This concentration was reduced to an effective upper-bound concentration of 0.44 ng TDI/m3 (2.2 ng TDI/m3 x 4m3/20m3) to account for the fact that the toxicity benchmark is based on the 20 m3 of air inhaled in 24 h while only 4 m3 of air is inhaled daily during 8 h of sleep (USEPA, 2011b).

3.3. Risk characterization

3.3.1. Non-cancer endpoints

The MOSs for the non-cancer endpoints ranged from 200 to 3 million, well above the target MOS of 1 (Table 4). As indicated earlier in Section 3.1.2, estimates of dermal sensitization threshold determined from experimental animal studies and that from the workplace experience do not provide a consistent profile of potency since dermal sensitization is rare in humans. Regardless, a conservative NESIL was calculated based upon animal data. Therefore, skin irritation, skin sensitization, respiratory sensitization and lung decrement effects from potential dermal and inhalation exposures to TDI while sleeping on a mattress poses a negligible non-cancer risk.

3.3.2. Cancer endpoints

The modelled lifetime exposure to sleeping on a PU foam mattress was combined with the CSF and conservative age-dependent adjustment factors of 10 and 3 (for <2 years of age and 2-16 years of age, respectively) to yield a total excess cancer risk of 1 x 10~7, which is 10-fold lower than the lower boundary of the typical excess cancer risk range (1 x 10~6to1 x 10~4) targeted by regulatory agencies. This conservative excess cancer risk is calculated using a CSF derived from a technically and conceptually flawed TDI gavage study that resulted in the production of TDA, a known rodent carcinogen, and ignores the observation that TDI is not carcinogenic by inhalation, the typical exposure route in humans (Section 3.1.5).

4. Discussion

A risk assessment framework for the evaluation of TDI-derived PU products was developed to evaluate the potential consumer health risks associated with ''sleeping on a flexible PU foam mattress." While this scenario is thought to represent a situation accentuating potential consumer exposure to PU foam (and thus TDI if present), additional exposures can also be hypothesized. Although no evidence exists to suggest such exposure, TDI emissions from PU carpet underlayment could contribute to the inhalation exposure, although migration of TDI would not affect dermal

exposures since skin contact with the underlayment is physically precluded. Such additional sources of theoretical exposure to TDI were not included in this assessment; however, sufficient information is presented herein to adjust the estimated risks for more complex exposure scenarios. For the assessment, toxicity benchmarks for TDI relevant endpoints were combined with migration and emission data from a TDI-based PU flexible foam commonly used in bedding. The results from the emission and migration tests showed no detectable TDI. Therefore, for the purposes of deriving a measure of emission rate from this foam sample, non-detectable results were assigned a numeric value of 50% of the respective MDLs.

Human data are preferred for the identification of benchmark toxicity values. However, the occurrence of skin irritation and allergic contact dermatitis in workers has been reported only rarely. Therefore, toxicity benchmarks for these endpoints were based on data in animals. An important aspect of the animal data used for these benchmarks is that they demonstrate a dose (threshold) below which skin irritation and sensitization do not appear to occur (Selgrade et al., 2006) and in principle this conclusion is considered applicable to humans as well. However, the thresholds derived from animal experiments appear to be lower than those of humans, based on limited occurrences reported after several decades of workplace experience. This discordance is particularly evident for skin sensitization where occupational experience shows that TDI (and indeed other diisocyanates) are relatively weak inducers of allergic contact dermatitis while the mouse LLNA indicates TDI is a strong skin sensitizer (ECETOC, 2008). Thus, the effect levels for skin irritation and sensitization derived in this assessment are conservative relative to human effect levels. It is beyond the scope of this assessment to resolve these apparent discrepancies, but there appears to be a need to elucidate the difference in mechanism of dermal sensitization between the rodent sensitization assays (mouse LLNA, GPMT) and human allergic dermatitis. Until further clarity is achieved, the derived threshold for dermal sensitization should be used with caution for human risk assessment. There appears to be less of a discrepancy for dermal irritation to TDI. The toxicity benchmarks used herein should be re-assessed as new scientific information for TDI becomes available and modified as appropriate.

Respiratory sensitization and decrements in lung function are the primary adverse health effects associated with TDI. While the threshold for induction of TDI-asthma has not been empirically determined for humans (Ott, 2002), they have been characterized in animals (Karol, 1983). The downward trend in recent years in the incidence of diisocyanate occupational asthma (Vandenplas et al., 2011; Buyantseva et al., 2011; Paris et al., 2012) supports results from animal models indicating that the respiratory sensitiza-tion and lung function decrements produced by diisocyanates are threshold events and that current occupational exposure guidelines adequately protect against both. Methodologies for assessing thresholds for respiratory effects in humans continue to evolve. Advances in this field have been particularly difficult due to the complexity of sensitization and the factors that contribute to the development of allergic disease (see Table 1). While some aspects of sensitization can be quantitatively addressed (e.g., occlusion, product use), other factors like biological and genetic variation, allergic susceptibility, mechanisms of chemical allergy remain a challenge.

Studies of workers have not reported carcinogenic effects related to TDI exposure; these observations are consistent with the absence of tumors in rats and mice exposed to TDI vapor for a lifetime. For these reasons, the cancer risk posed by the inhalation of TDI was not considered in this assessment. However, because regulatory agencies continue to classify TDI as a potential human carcinogen based on a flawed study in which rodents were administered TDI by gavage, this assessment used these studies to consider the unlikely possibility that dermal exposures to TDI will cause cancer in humans. As suggested by the gavage study authors (Dieter et al., 1990) and supported by subsequent analyses (Sielken et al., in press), the tumors likely resulted from the conversion of a small amount of TDI to TDA (a known animal carcinogen) due to improper test sample handling and the aphysiological exposure conditions associated with the study protocol. TDI in contact with the skin is unlikely to be converted to TDA since the relatively pH-neutral conditions at and below the skin surface favor the formation of polyureas and conjugation of TDI with macromolecular components of the stratum corneum. Thus, the excess cancer risk calculated via this exposure route will overestimate the real risk, if any.

Table 2

Assumptions for the determination of the lifetime average daily dose (LADD) and potential cancer risk for dermal exposure of TDI derived from sleeping on a flexible foam mattress.

Parameter <2 years of age 2 to <16 years of age 16-70 years of age Unit References

TDI concentration available for dermal 0.08 0.08 0.08 ng/cm2-event Section 3.2.1

exposure per event

Body surface area in contact with foam 3050 10,300 12,800 cm2 USEPA, 2011b, one-half the 95th percentile value for children aged 1 to <2 years, 11 to <16 years, and adult males aged 40 to <50 years

Percent dermal absorption 0.9% 0.9% 0.9% percent Hoffmann et al. (2010)

Body weight 11.4 56.8 80 kg USEPA, 2011b, mean body weight for children aged 1 to <2 years, 11 to <16 years, and adults

Events per day 1 1 1 events/day The sleep period is taken as one event

Average daily dose 0.19 0.13 0.12 ng/kg/day Calculated

Days exposed 365 365 365 days/year Assumed

Years exposed 2 13 55 years Assumed

Years in a lifetime 70 70 70 years Assumed

LADDa 0.0055 0.024 0.091 ng/kg/day Calculated

Cancer slope factor (CSF) 4.8 x 10-7 4.8 x 10-7 4.8 x 10-7 (ng/kg-day)-1 Calculated, Section 3.1.5, from oral gavage data

Age dependent adjustment factor 10 3 1 USEPA, 2005

Population cancer riskb 2.6 x 10-8 3.5 x 10-8 4.3 x 10-8 Calculated

Total cancer risk 1.0 x 10-7 Calculated, the sum of all three age groups

xbody surface area x % absorptionx-

body weight

Population cancer risk = CSF x Age dependent adjustment factor x LADD.

Table 3

Assumptions for the determination of the TDI inhalation exposure concentration.

Parameter Value Unit Notes

Estimated amount of TDI emitted per area 25a ng/m2 Vangronsveld et al. 2012, emission per 8 h

Surface area of mattress 10.4b m2 calculated

Total amount of TDI released from foam mattress 260 ng 10.4 m2 x 25 ng/m2

Volume of roomc 30 m3

Air exchanges per 8 h 4 # 0.5 exchanges per hour, USEPA, 2011b

Concentration of TDI in bedroom air 2.2 ng/m3 260 ng/(30 m3 x 4)

Adjustment for breathing rate while sleeping 0.2 - 4 m3/day breathing rate during 8 h of sleep (USEPA, 2011b)

divided by standard breathing rate of 20 m3/day

Inhalation exposure concentration 0.44 ng/m3 2.2 ng/m3 x 0.2

a The results from the emission tests showed no detectable TDI. Therefore, for the purposes of deriving a measure of emission rate from this foam sample, non-detectable results were assigned a numeric value of 50% MDL for the cell type used in the method with the highest sensitivity used for the emission testing. b The surface area (m2) of all sides of a king sized mattress [2 m (length) x 2 m (width) x 0.3 m (depth)] was determined.

c The definition of a "standard" room size for a bedroom varies widely (e.g. from 16 m3 to 41 m3) in European risk assessment (European Chemicals Bureau, 2003) and Jennings et al. (1987) guidance respectively. For this assessment a room volume of 30 m3 is used, based upon the large size mattress used would likely be in a relatively larger size bedroom.

Table 4

Summary of non-cancer hazard endpoints and MOS (margin of safety) calculations for TDI.

Effect Endpoint Endpoint Uncertainty Toxicological TDI MOS

value factor Benchmark concentration

Skin irritation Human LOAEL 600,000 ng/ cm2 5000 ng/cm2 12,800 ng/m3 2000 ng/m3 3 200,000 ng/cm2 0.08 ng/cm2 3 x 106

Skin sensitization Respiratory sensitization (inhalation induction) Respiratory tract effects (lung decrement) Mouse EC3 Human occupational NOAEL Human occupational NOAEL 15 2 30 333 ng/cm2 6400 ng/m3 70 ng/m3 0.08 ng/cm2 0.44 ng/m3 0.44 ng/m3 4000 15,000 200

LOAEL = Lowest observable adverse effect level.

EC3 = the concentration leading to a threefold increase of the baseline lymph node cell proliferation in the LLNA. (Local Lymph Node Assay) using thymidine labeling (Kimber et al., 2001). NOAEL = No observable adverse effect level.

The bioavailability of TDI in cured PU foam was measured using data from emission and migration studies to estimate TDI exposures via inhalation and dermal contact, respectively. Data obtained from these studies were judged a better reflection of the physiological conditions associated with human exposures to PU foam than solvent-based extraction procedures that can catalyze the decomposition of TDI-based polymers in PU foam to produce TDI (Vangronsveld et al., 2012). During the design of the migration studies, consideration was given to using artificial sweat as a solvent to mimic the environment of the human skin surface. However, this procedure was deemed inappropriate since it might have resulted in an underestimate of migration due to the low solubility of TDI monomer in aqueous solutions and its transformation into polyureas. Instead, Vangronsveld et al. (2012) used a solid matrix coated with a diisocyanate derivatizing agent to quan-titate directly the presence of TDI migrating to the foam surface without the potential limitations of intervening media.

Several conservative (health protective) exposure assumptions were incorporated into this risk assessment. The modeled exposure scenario used a large mattress relative to the size of the room, an assumption that increases potential inhalation exposures to TDI. Furthermore, although unrealistic, it was assumed there were no intervening fabrics (e.g., mattress cover, clothing) that would normally limit dermal exposures to a reactive chemical like TDI. In addition, it was assumed that individuals would always sleep on a PU foam mattress and that the mattress was an infinite source of free TDI even though these levels, if present, would be expected to decrease over time. A potential non-conservative assumption was the 8 h sleep period, which is consistent with the time course of the emission (i.e., 8 h) and migration (i.e., 8 and 24 h) analytical

studies. Adjustment was not made for additional time one might spent on the mattress since kinetic data were not available due to the absence of detectable TDI levels in both studies. However, even a doubling of sleep time (i.e., risk) from 8 to 16 h would not affect our conclusions given the range in MOSs listed in Table 4 (200 to 3 x 106).

5. Conclusion

A quantitative risk assessment for consumer exposure to TDI while sleeping on a flexible PU foam mattress is developed that is based on a careful consideration of the toxicological properties of TDI, results from migration and emission studies from a foam typically used for flexible PU mattresses, and conservative exposure assumptions. We conclude from this assessment that sleeping on a PU foam mattress for a lifetime does not pose TDI-related health risks to consumers. Using this framework, similar risk evaluations can be developed for other PU products or applications based on the toxicity information, appropriate analytical data, and exposure modelling approaches specific to the application.

Conflict of interest statement

SMA, CG, RJP, AP, TS, RNS and MRW are employed by companies producing TDI or other diisocyanates. MAC and ATJ are consultants to the diisocyanates industry. The paper was produced during the course of employment. This work was conducted under the auspices of the International Isocyanate Institute, Inc. The opinions

and conclusions are those of the authors and not necessarily those of the Institute.

Acknowledgments

The assistance of Dr. Manfred Giersig, Bayer MaterialScience over the duration of this project is gratefully recognized.

References

ACGIH, 2010. 2010 TLVs and BEIs: based on the documentation of the threshold limit values for chemical substances and physical agents and biological exposure indices' American Conference of Governmental Industrial Hygienists, Cincinnati.

Adams, W.G.F., 1975. Long-term effects on the health of men engaged in the manufacture of tolylene di-isocyanate. Br. J. Ind. Med. 32, 72-78.

AGS, 2006. Begründung zu 4-Methyl-m-phenylendiisocyanat in TRGS 900. Ausschuss für Gefahrstoffe, Bundesanstalt für Arbeitsschutz und Arbeitsmedizin (BAuA), Dortmund, Germany. <http://www.baua.de/nn_5846/ de/Themen-von-A-Z/Gefahrstoffe/TRGS/pdf/900/900-4-methyl-m-phenylendiisocyanat.pdf> (accessed August 2011).

Auletta, C.S., 1984. A closed-patch repeated insult dermal sensitization study in guinea pigs with TDI, MDI, p-TMXDI, IPDI, m-TMXDI, HMDI and m-TMI (modified Buehler method) with cover letter 100592. EPA Doc. 86-930000053. Bio/dynamics Inc. Submission to TSCA on behalf of American Cyanamid Company.

Bakerly, N.D., Moore, V.C., Vellore, A.D., Jaakkola, M.S., Robertson, A.S., Burge, P.S., 2008. Fifteen-year trends in occupational asthma: data from the Shield surveillance scheme. Occup. Med. (Lond.) 58 (3), 169-174.

Baur, X., 1985. Isocyanate hypersensitivity. III Report No. 10349. International Isocyanate Institute, Manchester, UK.

Bolognesi, C., Baur, X., Marczynski, B., Norppa, H., Sepai, O., Sabbioni, G., 2001. Carcinogenic risk of toluene diisocyanate and 4,4'-methylenediphenyl diisocyanate: epidemiological and experimental evidence. Crit. Rev. Toxicol. 31, 737-772.

Botham, P.A., Hext, P.M., Rattray, N.J., Walsh, S.T., Woodcock, D.R., 1988. Sensitisation of guinea pigs by inhalation exposure to low molecular weight chemicals. Toxicol. Lett. 41, 159-173.

Brown, W.E., Burkert, A.L., 2002. Biomarkers oftoluene diisocyanate exposure. Appl. Occup. Environ. Hyg. 17 (12), 40-45.

Bruynzeel, D., Andersen, K., Camarasa, J., Lachapelle, J., Menne, T., White, I., 1995. The European standard series. European Environmental and Contact Dermatitis Research Group (EECDRG). Contact Derm. 33 (3), 145-148.

Bugler, J., Clark, R.L., Hill, I.D., McDermott, M., 1991. The acute and long-term respiratory effects of aromatic di-isocyanates: a five year longitudinal study of polyurethane foam workers. III Report No. 10848. International Isocyanate Institute, Manchester, UK.

Buyantseva, L.V., Liss, G.M., Ribeiro, M., Manno, M., Luce, C.E., Tarlo, S.M., 2011. Reduction in diisocyanate and non-diisocyanate sensitizer-induced occupational asthma in Ontario. J. Occup. Environ. Med. 53 (4), 420-426.

Canadian Government, 2008. Screening assessment for the challenge toluene diisocyanate. Benzene, 2,4,-diisocyanato-1-methyl-(2,4-TDI), CAS RN 584-84-9; benzene, 1,3,-diisocyanato-2-methyl-(2,6-TDI), CAS RN 91-08-7; benzene, 1,3,-diisocyanatomethyl-(TDI mixed isomers), CAS RN 26471-62-5. Environment Canada; Health Canada. <http://www.ec.gc.ca/substances/ese/eng/challenge/ batch1/batch1_TDI.cfm> (accessed August 2011).

Chester, E.H., Martinez-Catinchi, F.L., Schwartz. H.J., Horowitz, J., Fleming, G.M., Gerblich, A., McDonald, E.W., Brethauer, R., 1979. Patterns of airway reactivity to asthma produced by exposure to toluene di-isocyanate. Chest 75, 229-231.

Clark, R.L., Bugler, J., Paddle, G.M., Chamberlain, J.D., Allport, D.C., 2003. A 17-year epidemiological study on changes in lung function in toluene diisocyanate foam workers. Int. Arch. Occup. Environ. Health 76, 295-301.

Cleet, C., 2005. Isocyanates in flexible polyurethane foams. Letter to the editor. Bull. Environ. Contam. Toxicol. 74(1), 1-3 (authors' reply, 4-7).

Cole, K.C., Van Gheluwe, P., Hebrard, M.J., Leroux, J., 1987. Flexible polyurethane foam. I. FTIR analysis of residual isocyanate. J. Appl. Polym. Sci. 34, 395-407.

Collins, M.A., 2002. Toxicology of toluene diisocyanate. Appl. Occup. Environ. Hyg. 17 (12), 846-855.

Daftarian, H.S., Lushniak, B.D., Reh, C.M., Lewis, D.M., 2002. Evaluation of self-reported skin problems among workers exposed to toluene diisocyanate (TDI) at a foam manufacturing facility. J. Occup. Environ. Med. 44 (12), 1197-1202.

Day, B.W., Jin, R., Basalyga, D.M., Kramarik, J.A., Karol, M.H., 1997. Formation, solvolysis, and transcarbamoylation reactions of bis(s-glutathionyl) adducts of 2,4- and 2,6-diisocyanatotoluene. Chem. Res. Toxicol. 10 (4), 424-431.

De Marzo, N., Jovine, L., Rizzotti, P., Cassetti, P., Boschetto, P., Miotto, D., Saetta, M., Maestrelli, P., Mapp, C.E., 2000. Modification of serum proteins in guinea pigs immunized and challenged with toluene diisocyanates. Scand. J. Work Environ. Health 26 (2), 153-160.

DFG, 2003. Toluene diisocyanate. In: Greim, H. (Ed.), Occupational Toxicants: Critical Data Evaluation for MAK Values and Classification of Carcinogens, vol. 20. Wiley-VCH, Weinheim, pp. 291-338.

Diem, J., Jones, R., Hendrick, D., Glindmeyer, H., Butcher, B., Salvaggio, J., Dharmarajan, V., Weill, H., 1982. Five-year longitudinal study of workers

employed in a new TDI manufacturing plant. Am. Rev. Respir. Dis. 126, 420428.

Dieter, M.P., Boorman, G.A., Jameson, C.W., Matthews, H.B., Huff, J.E., 1990. The carcinogenic activity of commercial grade toluene diisocyanate in rats and mice in relation to the metabolism of the 2,4- and 2,6-TDI isomers. Toxicol. Ind. Health 6 (6), 599-621.

Dieterich, D., Grigat, E., Hahn, W., Hespe, H., Schmelzer, H.G., 1993. Principles of polyurethane chemistry and special applications. In: Oertel, G. (Ed.), Polyurethane handbook. Chemistry - raw materials - processing - application - properties, 2nd ed. Carl Hanser Verlag, Munich, pp. 11-53 (Chapter 2).

Diller, W.F., 1998. Critical review of the human toxicology of TDI. III Report No. 11198. International Isocyanate Institute, Manchester, UK.

Duprat, P., Gradiski, D., Marignac, B., 1976. Irritating and allergizing power of two isocyanates; toluene diisocyanate (TDI), diphenylmethane diisocyanate (MDI). Eur. J. Toxicol. 9 (1), 41-53.

European Chemicals Bureau, 2003. Technical guidance document on risk assessment in support of: Commission Directive 93/67/EEC on risk assessment for new notified substances; Commission regulation (EC) 1488/94 on risk assessment for existing substances; Directive 98/8/EC of the European Parliament and of the Council concerning the placing of biocidal products on the market, second ed. Parts I & II. European Commission, Joint Research Centre, Italy.

ECETOC, 2008. Potency values from the local lymph node assay: application to classification, labelling and risk assessment. ECETOC Document No. 46. European Centre for Ecotoxicology and Toxicology of Chemicals, Brussels, Belgium.

ECHA, 2010. Guidance on information requirements and chemical safety assessment. Chapter R.8: characterisation of dose [concentration]-response for human health, version 2, December 2010. European Chemicals Agency, Helsinki. <http://guidance.echa.europa.eu/docs/guidance_document/

information_requirements_r8_en.pdf?vers=16_12_10> (accessed August 2011).

ECHA, 2011. m-Tolylidene diisocyanate (EC No: 247-722-4, CAS No: 26471-62-5) [REACH online dossier]. European Chemicals Agency, Helsinki. <http:// apps.echa.europa.eu/registered/data/dossiers/DISS-9eaa0e2f-2626-4d70-e044-00144f67d031/DISS-9eaa0e2f-2626-4d70-e044-00144f67d031_DISS-9eaa0e2f-2626-4d70-e044-00144f67d031.html> (accessed December 2011).

European Commission, 1998. Commission Directive 98/98/EC of 15 December 1998 adapting to technical progress for the 25 time Council Directive 67/548/EEC on the approximation of laws, regulations and administrative provisions relating to the classification, packaging and labelling of dangerous substances. Off. J. Eur. Commun. 41, L355, 30 December, 386-389.

European Commission Joint Research Centre, 2000. IUCLID CD-ROM year 2000 edition: public data on high volume chemicals. Report EUR 19559 EN. European Chemical Bureau, Ispra (VA), Italy.

European Commission, 2010. Risk characterisation of local effects. European Commission, Joint Research Centre, Institute for Health and Consumer Protection. Ispra, Italy. <http://ihcp.jrc.ec.europa.eu/our_activities/health-env/ risk_assessment_of_Biocides/guidance-documents> (accessed February 2012).

Fabbri, L.M., Boschetto, P., Zocca, E., Milani, G., Pivirotto, F., Plebani, M., Burlina, A., Licata, B., Mapp, C.E., 1987. Bronchoalveolar neutrophilia during late asthmatic reactions induced by toluene diisocyanate. Am. Rev. Respir. Dis. 136, 36-42.

Felter, S.P., Ryan, C.A., Basketter, D.A., Gilmour, N.J., Gerberick, G.F., 2003. Application of the risk assessment paradigm to the induction of allergic contact dermatitis. Regul. Toxicol. Pharmacol. 37 (1), 1-10.

Gagné, S., Lesage, J., Ostiguy, C., Van Tra, H., 2003. Determination of unreacted 2,4-toluene diisocyanate (2,4TDI) and 2,6-toluene diisocyanate (2,6TDI) in foams at ultratrace level by using HPLC-CIS-MS-MS. Analyst 128 (12), 1447-1451.

Gandevia, B., 1963. Studies of ventilatory capacity and histamine response during exposure to isocyanate vapour in polyurethane foam manufacture. Br. J. Ind. Med. 20, 204-209.

Graham, C., Ott, M.G., Hoffmann, H-D., Shiotsuka, R.N., Davies, P., Landry, T.D., Collins, M.A., 2002. Respiratory sensitization: the role of dermal contact with MDI and TDI. GIL Report 2002/D. Gilbert International Ltd., Manchester, UK.

Griem, P., Goebel, C., Scheffler, H., 2003. Proposal for a risk assessment methodology for skin sensitization based on sensitization potency data. Regul. Toxicol. Pharmacol. 38 (3), 269-290.

Hagmar, L., Welinder, H., Mikoczy, Z., 1993a. Cancer incidence and mortality in the Swedish polyurethane foam manufacturing industry. Br. J. Ind. Med. 50, 537543.

Hagmar, L., Stroemberg, U., Welinder, H., Mikoczy, Z., 1993b. Incidence of cancer and exposure to toluene diisocyanate and methylene diphenyldiisocyanate: a cohort based case-referent study in the polyurethane foam manufacturing industry. Br. J. Ind. Med. 50, 1003-1007.

Henschler, D., Assmann, W., Meyer, K.O., 1962. Zur Toxikologie der Toluylendiisocyanate. Arch. Toxikol. 19, 364-387.

Hilton, J., Dearman, R.J., Basketter, D.A., Kimber, I., 1995. Identification of chemical respiratory allergens: dose-response relationships in the mouse IgE test. Toxicol. Methods 5 (1), 51-60.

Hoffmann, H., Leibold, E., Ehnes, C., Fabian, E., Landsiedel, R., Gamer, A., Poole, A., 2010. Dermal uptake and excretion of 14C-toluene diisocyante (sic) (TDI) and 14C-methylene diphenyl diisocyanate (MDI) in male rats. Clinical signs and histopathology following dermal exposure of male rats to TDI. Toxicol. Lett. 199, 364-371.

Holness, D., Broder, I., Corey, P., Booth, N., Mozzon, D., Nazar, M., Guirguis, S., 1984. Respiratory variables and exposure-effect relationships in isocyanate-exposed workers. J. Occ. Med. 26 (6), 449-455.

Huang, J., Wang, X., Chen, B., Ueda, A., Aoyama, K., Matsushita, T., 1991. Immunological effects of toluene diisocyanate exposure on painters. Arch. Environ. Contam. Toxicol. 21, 607-611.

Hugo, J.M., Spence, M.W., Lickly, T.D., 2000. The determination of the ability of polyurethane foam to release toluene diisocyanate into air. Appl. Occup. Environ. Hyg. 15 (6), 512-519.

IARC, 1999. IARC Monographs on the Evaluation of Carcinogenic Risks to Humans, vol. 71, Part 3. Re-evaluation of some organic chemicals, hydrazine and hydrogen peroxide. International Agency for Research on Cancer, Lyon.

Jeffcoat, A.R., 1988. Absorption and disposition of orally administered 2,4-toluene diisocyanate in the Fischer 344 rat. Project Report No. 2. Report RTI/3662/02P. Research Triangle Institute, NC.

Jennings, P., Hammerstrom, K., Colemen Adkins, L., Chambers, T., Dixon, D., 1987. Methods for assessing exposure to chemical substances. Methods for assessing consumer exposure to chemical substances. Report No. EPA 560/5-85-007, 7, 71. US Environmental Protection Agency, Office of Toxic Substances, Washington, DC.

Jordt, S.E., Bautista, D.M., Chuang, H.H., McKemy, D.D., Zygmunt, P.M., Hogestatt, E.D., Meng, I.D., Julius, D., 2004. Mustard oils and cannabinoids excite sensory nerve fibres through the TRP channel ANKTM1. Nature 427 (6971), 260-265.

Kanerva, L., Jolanki, R., Alanko, K., Estlander, T., 1999. Patch-test reactions to plastic and glue allergens. Acta Derm. Venereol. 79 (4), 296-300.

Karol, M.H., Hauth, B.A., Riley, E.J., Magreni, C.M., 1981. Dermal contact with toluene diisocyanate (TDI) produces respiratory tract hypersensitivity in guinea pigs. Toxicol. Appl. Pharmacol. 58, 221-230.

Karol, M.H., 1983. Concentration-dependent immunologic response to toluene diisocyanate (TDI) following inhalation exposure. Toxicol. Appl. Pharmacol. 68, 229-241.

Kennedy, A.L., Brown, W.E., 1998. Investigation of the chemistry of TDI in biological materials. III Report No. 11320. International Isocyanate Institute, Manchester, UK.

Kimber, I., Basketter, D.A., Berthold, K., Butler, M., Garrigue, J.-L., Lea, L., Newsome, C., Roggeband, R., Steiling, W., Stropp, G., Waterman, S., Wiemann, C., 2001. Skin sensitization testing in potency and risk assessment. Toxicol. Sci. 59,198208.

Kimber, I., Dearman, R.J., Basketter, D.A., Ryan, C.A., Gerberick, G.F., McNamee, P.M., Lalko, J., Api, A.M., 2008. Dose metrics in the acquisition of skin sensitization: thresholds and importance of dose per unit area. Regul. Toxicol. Pharmacol. 52 (1), 39-45.

Knapp, R., Baker, R.G., 1974. Primary skin irritation test with toluene diisocyanate in albino rabbits: report to BASF Wyandotte Corporation. IBT No. 601-05008. Industrial Bio-Test Laboratories, Inc., Northbrook, Illinois 60062, USA.

Krone, C.A., Ely, J.T.A., Klingner, T., Rando, R.J., 2003. Isocyanates in flexible polyurethane foams. Bull. Environ. Contam. Toxicol. 70, 328-335.

Lange, R.W., Day, B.W., Lemus, R., Tyurin, V.A., Gagan, V.E., Karol, M.H., 1999. Intracellular S-glutathionyl adducts in murine lung and human bronchoepithelial cells after exposure to diisocyanatotoluene. Chem. Res. Toxicol. 12, 931-936.

Liippo, J., Lammintausta, K., 2008. Contact sensitization to 4,4'-diaminodiphenylmethane and to isocyanates among general dermatology patients. Contact Dermatitis 59 (2), 109-114.

Loser, E., 1983. Long-term toxicity and carcinogenicity studies with 2,4/2,6-toluene-diisocyanate (80/20) in rats and mice. Toxicol. Lett. 15, 71-81.

Loveless, S.E., Api, A.M., Crevel, R.W., Debruyne, E., Gamer, A., Jowsey, I.R., Kern, P., Kimber, I., Lea, L., Lloyd, P., Mehmood, Z., Steiling, W., Veenstra, G., Woolhiser, M., Hennes, C., 2010. Potency values from the local lymph node assay: application to classification, labelling and risk assessment. Regul. Toxicol. Pharmacol. 56 (1), 54-66.

Matheson, J.M., Johnson, V.J., Vallyathan, V., Luster, M.I., 2005. Exposure and immunological determinants in a murine model for toluene diisocyanate (TDI) asthma. Toxicol. Sci. 84, 88-98.

Mikoczy, Z., Welinder, H., Tinnerberg, H., Hagmar, L., 2004. Cancer incidence and mortality of isocyanate exposed workers from the Swedish polyurethane foam industry: updated findings 1959-98. Occup. Environ. Med. 61 (5), 432-437.

Moller, D.R., McKay, R.T., Bernstein, I.L., Brooks, S.M., 1986. Persistent airways disease caused by toluene diisocyanate. Am. Rev. Respir. Dis. 134,175-176.

Mormann, W., Vaquero, R.L., Seel, K., 2006. Interactions of aromatic isocyanates with N-acetyl-L-cyteine(sic) under physiological conditions: formation of conjugates, ureas and amines. EXCLI J. 5, 191-208.

Mraz, J., Simek, P., Nohova, H., Bouskova, S., Galova, E., Linhart, I., Smejkal, J., 1998. Identification of 2,4-toluenediisocyanate (2,4-TDI) adducts with globin and albumin. In: Abstr. Paper, 4th Int. Symp. on Biological Monitoring in Occup. Environ. Health, Seoul, Korea, 23-25 September, 1998.

Mraz, J., Simek, P., Nohova, H., Duskova, S., Galova, E., Hornychova, M., Linhart, I., 2000. Studies on diisocyanate adducts with blood proteins as potential biomarkers of exposure. In: Paper presented at Isocyanate 2000: First International Symposium on Isocyanates in Occupational Environments, Stockholm, 19-21 June, 2000, pp. 68-70.

Musk, A.W., Peters, J.M., Bernstein, L., 1985. Absence of respiratory effects in subjects exposed to low concentrations ofTDI and MDI: a reevaluation. J. Occup. Med. 27 (12), 917-920.

National Research Council, 1983. Risk assessment in the Federal Government: managing the process. National Academies Press, Washington, DC.

National Research Council, 2004. Toluene 2,4- and 2,6-diisocyanate: acute exposure guideline levels. In: Acute Exposure Guideline Levels for Selected Airborne Chemicals, vol. 4. Appendix 4. National Academies Press, Washington, DC.

NCI, 1979. Bioassay of 2,4-diaminotoluene for possible carcinogenicity. National Cancer Institute Carcinogenesis Technical Report Series No. 162; DHEW Publication No. (NIH) 79-1718; PB 293593. US Department of Health, Education & Welfare, Public Health Service, National Institutes of Health.

Nethercott, J.R., 1990. Practical problems in the use of patch testing in the evaluation of patients with contact dermatitis. Curr. Probl. Dermatol. 2 (4), 97123.

NTP, 1986. Toxicology and carcinogenesis studies of commercial grade 2,4(80%)-and 2,6(20%)-toluene diisocyanate in F344/N rats and B6C3F1 mice (gavage studies).US Department of Health and Human Services. National Toxicology Program. NTP TR 251; NIH Publication No. 86-2507. Research Triangle Park, NC.

NTP, 2011. Report on Carcinogens, 12th ed., US Department of Health and Human Services, Public Health Service, National Toxicology Program.

O'Brien, I.M., Newman-Taylor, A.J., Burge, P.S., Harries, M.G., Fawcett, I.W., Pepys, J., 1979. TDI-induced asthma II. Inhalation challenge tests and bronchial reactivity studies. Clin. Allergy 9, 7-15.

Ogawa, M., Oyama, T., Isse, T., Yamaguchi, T., Murakami, T., Endo, Y., Kawamoto, T., 2006. Hemoglobin adducts as a marker of exposure to chemical substances, especially PRTR class I designated chemical substances. J. Occup. Health 48 (5), 314-328.

Ott, M.G., Klees, J.E., Poche, S.L., 2000. Respiratory health surveillance in a toluene di-isocyanate production unit, 1967-97: clinical observations and lung function analyses. Occup. Environ. Med. 57 (1), 43-52.

Ott, M.G., 2002. Occupational asthma, lung function decrement, and toluene diisocyanate (TDI) exposure: a critical review of exposure-response relationships. Appl. Occup. Environ. Hyg. 17 (12), 891-901.

Ott, M.G., Diller, W.F., Jolly, A.T., 2003. Respiratory effects of toluene diisocyanate in the workplace: a discussion of exposure-response relationships. Crit. Rev. Toxicol. 33 (1), 1-59.

Owen, P.E., 1984. The toxicity and carcinogenicity to rats of toluene diisocyanate vapour administered by inhalation for a period of 113 weeks. Addendum Report Volume 1. III Report No. 10233. International Isocyanate Institute, Manchester, UK.

Paris, C., Ngatchou-Wandji, J., Luc, A., et al., 2012. Work-related asthma in France: recent trends for the period 2001-2009. Occup. Environ. Med.. http:// dx.doi.org/10.1136/oemed-2011-100487.

Pauluhn, J., Mohr, U., 1998. Assessment of respiratory hypersensitivity in guinea pigs sensitized to toluene diisocyanate: a comparison of sensitization protocols. Inhal. Toxicol. 10,131-154.

Peters, J.M., Wegman, D.H., 1975. Epidemiology of TDI-induced respiratory disease. Environ. Health Perspect 11, 97-100.

Rosenberg, C., Savolainen, H., 1985. Detection of urinary amine metabolites in toluene diisocyanate exposed rats. J. Chromatogr. 323, 429-433.

Safford, R.J., 2008. The dermal sensitisation threshold - a TTC approach for allergic contact dermatitis. Regul. Toxicol. Pharmacol. 51 (2), 195-200.

Schnorr, T.M., Steenland, K., Egeland, G.M., Boeniger, M., Egilman, D., 1996. Mortality of workers exposed to toluene diisocyanate in the polyurethane foam industry. Occup. Environ. Med. 53 (10), 703-707.

Seel, K., Walber, U., Herbold, B., Kopp, R., 1999. Chemical behaviour of seven aromatic diisocyanates (toluenediisocyanates and diphenylmethanediisocyanates) under in vitro conditions in relationship to their results in the Salmonella/ microsome test. Mutat. Res. 438,109-123.

Selgrade, M., Boykin, E., Haykal-Coates, N., Woolhiser, M., Wiescinski, C., Andrews, D., Farraj, A., Doerfler, D., Gavett, S., 2006. Inconsistencies between cytokine profiles, antibody responses, and respiratory hyperresponsiveness following dermal exposure to isocyanates. Toxicol. Sci. 94,108-117.

Sielken, R.L., Bretzlaff, R.S., Valdez-Flores, C., Parod, R., in press. Statistical comparison of carcinogenic effects and dose-response relationships in rats and mice for 2,4-toluene diamine to those ascribed to toluene diisocyanate. Hum. Ecol. Risk. Assess.

Skarping, G., Dalene, M., Lind, P., 1994. Determination of toluenediamine isomers by capillary gas chromatography and chemical ionization mass spectrometry with special reference to the biological monitoring of 2,4- and 2,6-toluene diisocyanate. J. Chromatogr. A 663,199-210.

Sorahan, T., Nichols, L., 2002. Mortality and cancer morbidity of production workers in the UK flexible polyurethane foam industry: updated findings, 1958-98. Occup. Environ. Med. 59 (11), 751-758.

Sorahan, T., Pope, D., 1993. Mortality and cancer morbidity of production workers in the United Kingdom flexible polyurethane foam industry. Br. J. Ind. Med. 50, 528-536.

Suberg, H., Löser, E., Kaliner, G., 1984. Toluylendiisocyanat (TDI): Untersuchungen zur primär reizenden/ätzenden Wirkung an der Kaninchenhaut. Bericht Nr: 12533. Bayer AG, Wuppertal, Germany.

Tarlo, S.M., Liss, G.M., Dias, C., Banks, D.E., 1997. Assessment of the relationship between isocyanate exposure levels and occupational asthma. Am. J. Ind. Med. 32, 517-521.

Taylor-Clark, T.E., Kiros, F., Carr, M.J., McAlexander, M.A., 2009. Transient receptor potential ankyrin 1 mediates toluene diisocyanate-evoked respiratory irritation. Am. J. Respir. Cell Mol. Biol. 40 (6), 756-762.

Thorne, P.S., Hillebrand, J.A., Lewis, G.R., Karol, M.H., 1987. Contact sensitivity by diisocyanates: potencies and cross-reactivities. Toxicol. Appl. Pharmacol. 87, 155-165.

Timchalk, C., Smith, F.A., Bartels, M.J., 1994. Route-dependent comparative metabolism of [14C]toluene 2,4-diisocyanate and [14C]toluene 2,4-diamine in Fischer 344 rats. Toxicol. Appl. Pharmacol. 124,181-190.

USEPA, 1989. Risk assessment guidance for Superfund. Volume I: human health evaluation manual (Part A). Interim final. Report No. EPA/540/1-89/002. Section 5.3.3. US Environmental Protection Agency, Office of Emergency and Remedial Response, Washington, DC.

USEPA, 1995. 2,4-/2,6-Toluene diisocyanate mixture (TDI) (CASRN 26471-62- 5). Integrated Risk Information System (IRIS) Database. US Environmental Protection Agency. <http://www.epa.gov/iris/subst/0503.htm> (accessed August 2011).

USEPA, 2005. Supplemental guidance for assessing susceptibility from early-life exposure to carcinogens. Report No. EPA/630/R-03/003F. Risk Assessment Forum. US Environmental Protection Agency, Washington, DC.

USEPA, 2011a. Toluene diisocyanate (TDI) and related compounds: action plan [RIN 2070-ZA14]. US Environmental Protection Agency. <http://www.epa.gov/oppt/ existingchemicals/pubs/actionplans/tdi.pdf> (accessed February 2012).

USEPA, 2011b. Exposure factors handbook: 2011 edition. National Center for Environmental Assessment, Washington, DC; EPA/600/R-09/052F. Available from the National Technical Information Service, Springfield, VA, and online at <http://www.epa.gov/ncea/efh>.

Valstar, D.L., Schijf, M.A., Nijkamp, F.P., Bloksma, N., Henricks, P.A.J., 2004. Glutathione-conjugated toluene diisocyanate causes airway inflammation in sensitised mice. Arch. Toxicol. 78 (9), 533-539.

Vandenplas, O., Lantin, A.C., D'Alpaos, V., Larbanois, A., Hoet, P., Vandeweerdt, M., Thimpont, J., Speybroeck, N., 2011. Time trends in occupational asthma in Belgium. Respir. Med. 105 (9), 1364-1372.

Vangronsveld, E., Berckmans, S., Spence, M., 2012. Measurement of Toluene Diisocyanate (TDI) emission to air and migration to a surface from flexible polyurethane foam. Accepted for Publication in Annals Occup. Hygiene.

van Och, F.M.M., Slob, W., de Jong, W.H., Vandebriel, R.J., van Loveren, H., 2000. A quantitative method for assessing the sensitizing potency of low molecular weight chemicals using a local lymph node assay: employment of a regression method that includes determination of the uncertainty margins. Toxicology 146 (1), 49-59.

Wazeter, F.X., Buller, R.H., Geil, R.G., 1964. Toluene diisocyanate (TDI) and polymethylene polyphenylisocyanate (PAPI): acute dermal toxicity studies (LD50) in the albino rabbit. Unpublished report by the International Research and Development Corporation.

Wegman, D., Pagnotto, L., Fine, L., Peters, J., 1974. A dose-response relationship in TDI workers. J. Occup. Med. 16 (4), 258-260.

Weill, H., Butcher, B., Dharmarajan, V., Glindmeyer, H., Jones, R., Carr, J., O'Neil, C., Salvaggio, J., 1981. Respiratory and immunologic evaluation of isocyanate exposure in a new manufacturing plant. DHHS (NIOSH) Publication No. 81-125. US Department of Health and Human Services, Morgantown, West Virginia, USA.

Wisnewski, A.V., Redlich, C.A., Ye, Y-M., Park, H-S., 2006. Biophysical/chemical and immunological changes in human albumin exposed to TDI. Poster presented at ATS 2006: Int. Conf. American Thoracic Society, San Diego, 19-24 May, 2006.

Woolhiser, M.R., Hayes, B.B., Meade, B.J., 1998. A combined murine local lymph node and irritancy assay to predict sensitization and irritancy potential of chemicals. Toxicol. Methods 8 (4), 245-256.

Ye, Y.-M., Kim, C.-W., Kim, H.-R., Kim, H.-M., Suh, C.-H., Nahm, D.-H., Park, H.-S., Redlich, C.A., Wisnewski, A.V., 2006. Biophysical determinants of toluene diisocyanate antigenicity associated with exposure and asthma. J. Allergy Clin. Immunol. 118 (4), 885-891.

Zissu, D., Binet, S., Limasset, J.-C., 1998. Cutaneous sensitization to some polyisocyanate prepolymers in guinea pigs. Contact Derm. 39, 248-251.