Suitability of the Limit Dose in Evaluating Reproductive Toxicity of Substances and Preparations

Abstract

An oral dose of 1000 mg/kg body weight/day is mentioned in Organisation for Economic Cooperation and Development (OECD) and European Union (EU) guidelines as a default maximum dose in limit tests for studies on reproductive toxicity. This paper investigated whether upper range human exposure data from the workplace are supportive of this limit dose as an upper limit of possible human exposure. To this end, published exposure data as well as data from the database MEGA of the German “Berufsgenossenschaften” were evaluated. These data indicate that exposure concentrations in the range of 500 to 2000 mg/m³ (time-weighted averages) can be considered high human exposures to volatile compounds. Inhalation exposure to aerosols and dermal exposure result in lower dose levels. By applying suitable extrapolation factors, it was concluded that occupational exposures up to 325 mg/m³ can reliably be assessed with limit tests using a dose level of 1000 mg/kg/day. The limit dose has been proposed for use in the EU as a starting point to derive specific concentration limits for hazard classification of preparations containing reproductive toxicants, with the objective to consider the potency of the substances. This analysis shows that for some groups of chemicals, instead of the limit dose, the putative maximum levels of human exposure should be taken into account when deriving concentration limits for the classification of preparations. Furthermore, possible deviations from a linear correlation between concentration in the preparation and exposure should be considered.

In the European Union (EU), the objective of classification according to Directive 67/548/EEC is to identify all the physicochemical, toxicological, and ecotoxicological properties of chemical substances and preparations. The Directive aims at the protection of consumers, workers, and the environment. Classification is hazard based and considers intrinsic properties of the substances in a qualitative way. Among other hazards, chemical substances and preparations are classified when toxic to reproduction. Directive 1999/45/EEC lays down general concentration limits for the classification of preparations containing hazardous substances. With respect to classification as toxic to reproduction, preparations with substances assigned R-phrase 60 (substances impairing fertility, categories 1 and 2) or R61 (substances causing developmental toxicity, categories 1 and 2) have to be classified themselves if the concentration of the reproductive toxicant exceeds 0.5%. The general concentration limit for preparations containing possible reproductive toxicants of category 3 (R62 or R63) is 5%.

In contrast to mutagenic or carcinogenic substances, dose thresholds are assumed to exist for reproductive toxicants below which adverse effects are not expected to occur. Triggered by a concern about overregulating weak reproductive toxicants, i.e., substances with a high threshold for reproductive effects, a discussion was initiated about deriving substance-specific concentration limits, taking into account the potency of the respective compounds.

The German delegation to the EU Technical Committee on Classification and Labelling of Dangerous Substances proposed to use the respective substance-specific NOAEL (no observed adverse effect level) for reproductive effects and to compare this NOAEL with the general limit dose of 1000 mg/kg/day for studies on reproductive toxicity (ECB 1998). Such a limit dose is mentioned in various test method descriptions in Annex V of Directive 67/548/EEC and also in the respective Organisation for Economic Cooperation and Development (OECD) guidelines. Annex VI of Directive 67/548/EEC states with respect to the classification of substances toxic to reproduction that “if a dose level of at least 1000 mg/kg orally produces no evidence of effects toxic to reproduction, studies at other dose levels may not be considered necessary. …Under normal circumstances it is considered that effects seen only at doses in excess of the limit dose would not necessarily lead to classification as ‘toxic to reproduction.’”

This has been interpreted by the authors of the German method, i.e., the German Toxicology Advisory Group of the Committee on Hazardous Substances (Ausschuss für Gefahrstoffe, AGS), as 1000 mg/kg/day being the highest exposure level of concern. If a pure substance with a NOAEL for reproductive effects of 1000 mg/kg/day or higher should not be classified, then, the German proposal concludes, a preparation containing a substance with, e.g., a NOAEL of 100 mg/kg/day at a concentration of less than or equal to 10% should also not be subject to classification.

The general rule for setting specific concentration limits for reproductive toxicants in preparations therefore is formulated:

$$\text{Concentration\hspace{0.17em}limit\hspace{0.17em}substance\hspace{0.17em}A}=\frac{{\text{NOAEL}}_{\text{A}}·100\%}{1000\frac{\text{mg}}{\text{kg}·\text{day}}}$$

This method has been used recently for deriving concentration limits for several substances in the EU (Table 1), but it is not yet accepted as a general rule.

Obviously, the limit dose of 1000 mg/kg/day has a pivotal role in this approach. Up to now there is no documented scientific reasoning for this choice of the limit dose. There have been attempts to relate limit doses to be used in animal experiments to possible human exposure levels (Doe et al. 2006), but this has been successfully accomplished only in specific circumstances. For example, Solecki et al. (2005) estimated the maximum acute human exposure to pesticides from food consumption to be ≤1 mg/kg/day. From this they concluded that effects seen in experimental animals at dose levels in excess of 500 mg/kg/day are not relevant for deriving acute reference doses for general toxic effects of pesticides. A limit dose of 1500 mg/kg/day has been proposed for carcinogenicity testing of pharmaceuticals, provided that the recommended therapeutic dose will not exceed 500 mg/day (approximately 7.7 mg/kg/day) (ICH 1997).

The oral limit dose is not only relevant for substances for which oral exposure is the main exposure pathway, but also for volatile liquids. For regulatory purposes reproductive toxicity tests with oral application are often submitted also for liquid substances with high volatility, for which exposure under work-place conditions will mostly be via inhalation.

Recognizing the enormous importance of the limit dose not only for chemical safety testing but also for regulatory purposes when setting concentration limits for reproductive toxicants in preparations, this paper elucidates whether the limit dose of 1000 mg/kg/day reflects upper human exposure levels, particularly at the workplace, where the highest human exposure is assumed to occur. Data on high workplace exposures are compiled and analyzed and experimental doses equivalent to these human exposure levels are estimated. Finally, the concepts for taking into account potency considerations when deriving substance-specific concentration limits are discussed in the light of these estimated experimental doses and by discussing exposure arising from the use of preparations.

METHODS

Literature Searches in Bibliographic Databases

Human data on high exposure levels at the workplace have been searched in bibliographic databases (PUBMED: http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?db=PubMed; TOXLINE: http://toxnet.nlm.nih.gov/; CISDOC: http://www.ilo.org/dyn/cisdoc/index html) in several ways.

1. A large number of abstracts were retrieved and scanned for relevant information on occupational exposure. Specific search terms were used, as far as available (MESH term “Occupational Exposure”; CISDOC thesaurus term “determination in air”) and combined with mg* or ppm to search for abstracts containing numerical exposure data. Relevant publications were evaluated in original.

2. In PUBMED (using MESH term “Occupational Exposure”) occupational exposure data for all substances with German occupational exposure levels (OELs) (as laid down in TRGS 900; AGS 2006a) >1000 mg/m³ were retrieved. Relevant publications were evaluated in original.

3. In PUBMED (using MESH term “Occupational Exposure”) occupational exposure data for aerosols were searched by applying the search terms aerosol*, dust*, particle* and searching for abstracts with numerical concentration data. In addition, the background documentation for the German OEL for dusts (Anonymous 2001) was evaluated for exposure data on aerosols and for specific occupational exposures: lubricating oils, release agents (oils) in the construction industry, organic dust exposure in agriculture and the food industry.

MEGA Exposure Database

In the frame of their tasks the German employers’ liability insurance associations (Berufsgenossenschaften) measure the concentration of hazardous substances, e.g., to identify and assess hazards or to propose measures to be applied. For this reason, a measurement system of hazardous substances was developed in the early 1970s by the BG Institute of Occupational Safety and Health (BGIA) in conjunction with the “Berufsgenossenschaften.” This system monitors workplaces according to national and European legislation. Data are stored in the central MEGA exposure database (“Messdaten zur Exposition gegenüber Gefahrstoffen am Arbeitsplatz”; measurement data relating to workplace exposure to hazardous substances), which is administered and developed further by the BGIA. These data are primarily collected for prevention and not for compliance purposes and therefore represent mostly normal working conditions (Gabriel 2006). At the end of 2005, the database contained data records with 1.63 million measurement values since 1972 on 760 hazardous chemicals and 330 biological working agents (Van Gelder 2005).

Hazardous substances, for which air exposures at the work-place above 600 mg/m³ are documented, were selected from the exposure database MEGA. Only values determined with standardized sampling systems and analytical methods were selected for evaluation. Exposure peaks or worst-case measurements were not included. Measurements with a sampling time of at least 1 h and exposure duration of at least 6 h were considered. The measurements are representative for the workplace and the respective exposure duration. Two data periods were selected for the statistical evaluation (1986–1995 and 1996–2005) and evaluated separately.

Estimation of Dermal Exposure

Occupational dermal exposure is highly variable and difficult to measure. As a scenario with presumably high dermal exposure, application of preparations as aerosols by spraying has been identified. Data retrieved focus on this scenario. Various ways have been used to derive rough estimates of high dermal exposures:

• use of empirical exposure parameters (surface loading, exposed area) to get a theoretical exposure estimate;

• use of measured exposure data combined with assumptions on exposure duration;

• modeling of dermal exposure during spraying by the “dermal exposure calculator” as developed within the EU research project “Risk of Derm” (van Hemmen 2004).

The “Risk of Derm” dermal exposure calculator models are spreadsheet calculation algorithms in Microsoft Excel. Several models describing typical workplace scenarios are included in the calculator (so-called “dermal exposure operations,” DEO), which are based on empirical data (van Hemmen 2004). For the calculations conducted here, the scenario “indoor spray application” (Risk of Derm DEO unit 4) was used. Several assumptions/preconditions were necessary as inputs to the model. The following assumptions were used as representing a reasonable worst-case of possible exposure situations: overhead spraying of a volatile liquid; application rate 0.1 l/min; application duration 120 min/day.

Derivation of Equivalent Human Exposure Concentrations

When comparing oral dose levels in animal studies for reproductive effects with human inhalation exposures, differences in body size and inhalation rates have to be considered. In Germany, a methodological concept has been established to derive occupational exposure levels (OELs) called “Arbeitsplatzrichtwerte” (ARW) for substances with a limited database (TRGS 901; AGS 2006b). The concept applies extrapolation factors to account for variability and uncertainty in inter- and intraspecies extrapolation and to consider less than lifetime exposure. Similar to this concept, the recently adopted “Risk Characterisation” part of the Technical Guidance Document on Risk Assessment of the European Commission (EC 2005) envisages the application of extrapolation factors in the context of assessing human health risks from existing or new notified substances or from substances contained in biocidal products in the European Union. Both methods use allometric scaling factors to consider differences in body size, when comparing systemic doses between experimental animal species and humans. As the standard reproductive toxicity study is mostly carried out with rats, a scaling factor of 4, which applies to rats, was used for calculations. In the case of studies with mice and rabbits, which are recommended for use as alternatives to rats in OECD guidelines 415 and 414, respectively, the allometric scaling factor would be 7 (mice) or 2.4 (rabbits).

Interindividual variability is generally considered larger in humans than in in-bred experimental animal species. Both concepts mentioned above apply a factor of 5 to account for variation in susceptibility in a given group of exposed workers.

According to EC (2003) a body weight of 65 kg is assumed for adult workers (average for males and females). An inhalation rate of 10 m³/day is used for a worker’s shift to take into account the higher activity during work.

Calculation of Activity Coefficients for Real Mixtures

To estimate the partial vapour pressure of a substance in a liquid preparation, activity coefficients were calculated as described in the European Commission’s Technical Guidance Document on Risk Assessment (EC 2003, TGD Part I, Appendix IG). Activity coefficients (q _i ) are correction factors that describe the deviation of partial vapour pressure from Raoult’s law according to following equation:

$$P_i=q_i{x}_i{P}_i^s$$

where P _i = partial vapour pressure of substance i; q _i = activity coefficient of substance i; x _i = molar fraction of substance i in the preparation; and P _i ^S = vapour pressure of the pure substance i.

Activity coefficients were calculated using the software tool provided at http://www.hsrc.org/hsrc/html/ssw/ssw-downloads.html, which is based on the UNIFAC method.

RESULTS

Inhalation Exposure

Occupational exposure limits are higher for gases/vapours compared to aerosols in the higher range of values. Figure 1 depicts distributions of the current legally binding OELs in Germany (TRGS 900) (AGS 2006a). With one exception (polyethylene glycols, at 1000 mg/m³), all OELs for aerosols are <20 mg/m³. In contrast, OELs for gases/vapors are as high as 9000 mg/m³, with the highest values for carbon dioxide and 1,1,1,2-tetrachloro-2,2-difluoroethane (R112a). Therefore, different maximum exposure levels can be expected for volatile compounds and aerosols. Exposure data are evaluated separately for these two types of exposures.

Exposure to Volatile Substances—Data from Literature

Table 2 summarizes the data found in the published literature on high occupational exposure levels. These data show that concentrations in the range of 500 to 1000 mg/m³ are not uncommon and that time-weighted average (TWA) concentrations as high as 3000 mg/m³ have been reported for several substances. Even higher levels may occur for chlorofluorohydrocarbons like 1,1-dichloro-2,2,2-trifluoroethane and 1,1,2-trichloro-1,2,2-trifluoroethane during degreasing and similar activities.

Exposure to Volatile Substances—Data from MEGA

All substances in MEGA with at least one measurement above 600 mg/m³ were selected for further evaluation, with 22 substances or substance groups fulfilling this criterion. Data for these substances were retrieved under the conditions set for the search. All data reported are shift-related measurements (exposure time ≥6 h). Table 3 gives the details for the distributions of data.

For most substances, the percentage of measured concentrations ≥600 mg/m³ among all available values was below 5%. Exceptions were methyl methacrylate, methylene chloride, 1,1,1-trichloroethane, trichlorofluoromethane (R11), and trichlorotrifluoroethane (R113). The 95th percentiles were <550 mg/m³ for most substances, but were as high as 1000 mg/m³ for the substances mentioned above. The highest 95th percentile value was observed for 1,1,2-trichloro-1,2,2-trifluoroethane (5703 mg/m³). Most 99th percentiles did not exceed 1500 mg/m³, with the exception of 1,1,1-trichloroethane (2800 mg/m³ for time period 1986–1995), trichlorofluoromethane (4950 mg/m³ for time period 1986–1995), and 1,1,2-trichloro-1,2,2-trifluoroethane (9640 mg/m³ for time period 1986–1995). Fifty percent of the evaluated substances showed maximum values above 5000 mg/m³ in the time period 1986–1995, whereas this concentration was exceeded by 32% of substances in the time period 1996–2005. Because these maximum values represent less than 1% of all measurements, a generally high exposure to these substances cannot be assumed.

Exposure to Aerosols

Searches for published data on high aerosol exposures at the workplace produced no convincing data for exposures regularly exceeding 20 mg/m³ (TWA) (for exceptions, see below), which is twice the general OEL for dusts (inhalable fraction) in Germany (AGS 2006a).

In a survey of occupational dust exposure, data (respirable fraction) are compiled from German workplaces for the time period 1996 to 1998 (Anonymous 2001). Industrial branches assessed included construction industry, mining, chemical industry, foundries, wood industry, ceramics and glass industries, power plants, leather, food, and paper industries, friction pad, rock and mineral product branches, as well as textile industries. Medians in these industries were well below 10 mg/m³ for various workplaces, and 90th percentiles did not exceed 20 mg/m³. An exception is the construction industry where partially higher exposure levels (median 12.2 mg/m³, 90th percentile 125 mg/m³) exist. However, large temporal variations in exposure levels have to be taken into account at construction sites. High aerosol exposures well above 10 mg/m³ may be encountered at special workplaces in the construction industry where releasing agents (oils) are applied by spraying, which is generally performed with respiratory protection, and in the metal industry with exposure to aerosols of metal-working fluids (Anonymous 2001).

Organic dusts are present at high concentrations in some agricultural workplaces (flax scutching, median concentration 93.6 mg/m³ for five farms [Krysinska-Traczyk et al. 2004]; grain handling–silo cleaning, median concentration 63.4 mg/m³ [Simpson et al. 1999]); and in the food processing industry (flour dust exposure in mills [Smith, Parker, and Hussain 2000]) (median and maximum of 8-h TWAs for cleaning personal, 18.7 and 217 mg/m³, respectively). High concentrations of dust (fine soil particles) were also measured in agriculture, especially for activities related to ground-preparation operations (geometric mean concentrations 57.3 to 98.6 mg/m³ (Nieuwenhuijsen, Kruize, and Schenker 1998).

Taken together, these data show that average exposures to aerosols are generally below 20 mg/m³. Some occupational settings have higher exposure levels, with average concentrations up to 100 mg/m³.

Dermal Exposure

Occupational dermal exposure is highly variable from work-place to workplace and difficult to measure. Conceptually, the following terms are used (Phillips and Garrod 2001; Oppl et al. 2003): potential dermal exposure means the amount of substance reaching the outer surface of the body (i.e., clothing or skin, where unprotected), actual dermal exposure is the amount of substance on skin, internal exposure means the percutaneously absorbed dose. Potential and actual dermal exposures are most often given as amount of substance per exposed area (mg/cm²).

Table 4 summarizes the results obtained from exposure estimates using the parameters surface loading and exposed body surface, from exposure measurements, and from a dermal exposure model developed within the EU research project “Risk of Derm” (van Hemmen 2004).

As a rough estimate from these data, potential dermal exposure can be estimated to be in the range of 4000 to 20000 mg/person/day for the typical case and 20000 to 40000 mg/person/day for the reasonable worst case.

Following recommendations given by Oppl et al. (2003) and EC (2003) for taking into account the protective effect of clothing (effective personal protection is assumed for the exposure intensive spraying activities considered), a 90% reduction of exposure is assumed for calculating actual dermal exposure.

Actual dermal exposure for the reasonable worst case is therefore calculated to be 2000 to 4000 mg/person/day, or, assuming a body weight of 65 kg, approximately 30 to 60 mg/kg/day.

Comparison between the Limit Dose and High Human Exposure Levels

In the previous sections, data for high inhalation and dermal exposure at the workplace were compiled. Inhalation exposure to highly volatile substances leads to the highest exposures, whereas inhalative aerosol exposure and dermal contact result in substantially lower intakes. The data from the MEGA database indicate that 1500 mg/m³ may be considered an upper exposure level, the 99th percentiles for most of the substances being below this level. Taking into account the somewhat wider ranges of exposures reported in the literature for specific workplaces, a range of 500 to 2000 mg/m³ is considered to represent high inhalation exposures at most workplaces. For very volatile substances with high inhalation exposure, percutaneous absorption is assumed to be low, so that combined exposure via both pathways is not considered further.

For comparing the limit dose with these human exposure levels, an inhalation exposure concentration is calculated, which, under workplace conditions, would lead to a human systemic dose equivalent to the oral limit dose. To this end, a scaling factor of 4 to consider differences in body size is applied. With a respiratory volume per shift (8 h) of 10 m³ and a worker’s body weight (male/female average) of 65 kg (EC 2003), the limit dose corresponds to an occupational inhalation exposure concentration of 1625 mg/m³. For studies with mice (scaling factor 7) and rabbits (scaling factor 2.4) concentrations of 930 and 2700 mg/m³, respectively, would result.

$$\text{Inhalation\hspace{0.17em}Concentration\hspace{0.17em}}=\frac{1000\frac{\text{mg}}{\text{kg}·\text{day}}·65\hspace{0.17em}\text{kg}}{4·10\frac{{\text{m}}^{\text{3}}}{\text{day}}}=1625\frac{\text{mg}}{{\text{m}}^3}$$

When a factor of 5 is applied to cover a higher human interindividual variability, in comparison to genetically more homogenous experimental animals tested in limited numbers, a concentration of 325 mg/m³ is obtained. This can be interpreted to the effect that observations made at the limit dose of 1000 mg/kg/day in the rat study allow assessing possible effects at the workplace after exposure of up to 325 mg/m³. This calculation, in the absence of substance-specific data, assumes a similar absorption rate after oral and inhalation exposure.

Exposure to several highly volatile compounds may exceed this concentration, as was shown in Tables 2 and 3. In contrast, aerosol concentrations observed at workplaces are well below this concentration, reaching 100 mg/m³ only rarely.

For comparing human dermal exposures with the limit dose in a rat study, a scaling factor (4) and a factor for consideration of interindividual variability (5) also have to be applied to the limit dose, leading to a dose of 50 mg/kg/day. This is within the range of doses estimated for human dermal exposures under reasonable worst-case conditions (30 to 60 mg/kg/day), without considering that percutaneous absorption rates may be lower than oral ones.

Exposure Considerations for the Derivation Of Potency-Derived Concentration Limits

The proposed method to derive substance-specific concentration limits introduces potency considerations by comparing the respective NOAEL for reproductive toxic effects with the limit dose. Potency is expected to correlate approximately linearly with the concentration in the preparation. To judge whether exposure intensity is also linearly correlated with the concentration, some principal considerations with respect to dermal and inhalation exposure to substances from preparations were employed.

Dermal Exposure to Preparations

A linear relationship can be assumed between dermal exposure to a substance in a preparation and the concentration of the substance in the preparation, if interactions between constituents of the preparation can be excluded. Whereas this may hold true for many preparations, there are cases where the percutaneous absorption of one substance may be increased by a second one. Chemicals like dimethyl sulphoxide (DMSO), N,N-dimethyl formamide, or N,N-dimethyl acetamide are well known skin penetration enhancers (Tsuruta 1996). Penetration enhancement by DMSO and other substances is exploited therapeutically to improve transdermal drug delivery (Williams and Barry 2004).

Inhalation Exposure to a Preparation

A linear relationship between inhalation exposure and concentration of a substance in a (liquid) preparation can be assumed if Raoult’s law is valid, i.e., the constituents behave “ideally” and do not interact in the liquid phase. This is often not the case with real preparations, where the vapour pressure of a compound is increased or lowered by interaction with other substances in the preparation (EC 2003; Krafczyk, Gmehling, and Fischer 2000).

For the example of a mixture of 1-bromopropane and methanol, Figure 2 shows that behavior can deviate substantially from the ideal case. The vapor pressure of 1-bromopropane in methanol is up to 9 times higher than expected according to Raoult’s law. Even higher factors are obtained for binary systems with limited miscibility (EC 2003).

Complete independence of human exposure from the concentration of a substance in the preparation exists in situations where volatile substances are completely released from a preparation, e.g., in the case of adhesives or paints that dry up following application. For a volatile compound, the exposure concentration will then depend solely on the amount of substance applied per time unit.

DISCUSSION

Based on the data presented in this paper, inhalation exposure to volatile substances may reach concentrations of 500 to 2000 mg/m³ (time-weighted average) at some workplaces. No definite distributions for European workplaces can be provided, but data compiled in the German database MEGA form a sound basis to evaluate the upper exposure ranges. The picture emerging from the MEGA evaluation is supported by published data, which show a somewhat broader range of exposure concentrations in specific cases.

Dermal exposure estimates are highly uncertain. Estimated exposure levels under reasonable worst-case conditions are 30 to 60 mg/kg/day. Combined uptake from both dermal and inhalation exposure is not considered relevant for the high exposure ranges, taking into account the physicochemical properties of the substances. In addition to dermal exposure from direct contact of the skin with contaminated surfaces and by immersion into liquids, exposure can also take place via the gas phase. This pertains to a situation, where a volatile compound is present in air and is readily taken up through the skin. Kezic et al. (2000), as well as Brooke et al. (1998), investigated the percutaneous uptake of solvents from the gas phase and concluded that in most cases (for halogenated aliphatic compounds such as tetrachloroethylene and aromatic solvents such as toluene and hexane), the contribution of percutaneous uptake is about 1% or less as compared to inhalation. Only for glycol ethers (e.g., 2-methoxyethanol and 2-ethoxyethanol) can a significant contribution to the overall burden be expected (Shih et al. 2000; Kezic et al. 1997; Corley et al. 1997). However, this is unlikely to exceed intake by inhalation even under unfavorable circumstances. In the context considered here, the contribution of percutaneous absorption from the gas phase is expected to be of low importance in most cases. Calculations for liquid and solid aerosols indicate exposure levels substantially below maximum inhalation exposure for volatile substances. As the focus is on reproductive toxic effects, no attention was given to the significance of such concentrations for local respiratory effects.

Very high peak concentrations were reported in some occupational settings (see Table 2). In addition, case reports indicate that accidental oral intoxications may lead to high single exposures in the gram-per-kilogram range. Un- or mislabeled bottles have been repeatedly reported as the cause for severe intoxications of children and adults, both in the home environment and at the workplace (Winek, Wahba, and Edelstein 1990; Dean and Krenzelok 1992). Confusion of chemicals with beverages led to uptakes of 75 g tetrachloroethylene (Choi et al. 2003) and 102 to 136 g ethylene glycol by adults (Theurl, battista, and Fritzer 1989). Gavage application of substances, which is often used in studies for reproductive toxicity (e.g., according to OECD guidelines 414, 415, 421, and 422), may be seen as a simulation of peak exposure situations. But exposure levels as experienced in some poisoning events are not covered by the limit dose as applied in reproductive toxicity studies.

There are several parameters that generally have to be considered when inhalation exposures are assessed. Among these are exposure concentration, exposure duration, respiratory rate, and aerosol characterization (particle size distribution). When route-to-route extrapolations were performed in the present study, the exposures were compared on a daily basis. Differences in activity between workers and (resting) laboratory animals were taken into account by using a respiratory rate of 10 m³/day for workers. Absorption has not been specifically taken into account. This equals to the assumption that, when comparing inhalation at the workplace to the experimental setting using oral application, absorption rates are the same for oral and inhalation exposures. For the comparison of dermal exposure with the experimental oral exposure, a substantial conservativism might result at least for some substances, for which dermal absorption is considerably lower than oral absorption.

Extrapolation factors were applied above to account for differences between species due to differences in body size by allometric scaling (Schneider et al. 2004) and for larger differences in susceptibility between human individuals in comparison to those observed in standard animal tests (Clewell et al. 2002). Both factors are intended to adjust for expected differences between laboratory animals and humans (for the “average substance”) (Vermeire et al., 1999). This approach using extrapolation factors should not be confused with applying uncertainty factors, which intend to compensate for lack of knowledge on toxic properties of chemicals.

According to these considerations, occupational exposure levels up to 325 mg/m³ can be assessed with sufficient reliability by using results of a reproductive toxicity study with an upper oral dose level of 1000 mg/kg/day. Higher time-weighted average exposure concentrations can be observed for highly volatile compounds (Tables 2 and 3). As oral testing will take place only with solids and liquids this would be especially relevant for liquid substances of high volatility.

Estimates for high dermal exposures result in a dose range of 30 to 60 mg/kg/day, which is comparable to the limit dose after application of extrapolation factors for allometric scaling and for interindividual differences in susceptibility (50 mg/kg/day). It must be kept in mind that dermal exposure varies considerably depending on the workplace situation. Therefore, the assessment is associated with substantial uncertainty. As varying percutaneous absorption rates were not considered, the systemically available dose should be lower than the (extrapolated) limit dose for most substances. A case-by-case evaluation should investigate whether very high exposures are anticipated for specific substances.

With respect to aerosols, occupational exposure concentrations are usually well below 325 mg/m³ and are still lower even in cases of very high exposure levels (up to 100 mg/m³). This holds true even without considering deposition and retention patterns of the aerosols in the respiratory tract, which are determined to a large extent by particle size. Detailed consideration of aerosol inhalation toxicity conditions was therefore not deemed necessary for the purpose of this study.

It should be noticed that the strategy of data collection was focused on high exposure data. Thus, the data collected are not representative of the situation at the workplace. This strategy was chosen to compare the limit dose of 1000 mg/kg/day used in reproductive toxicity studies with high occupational exposures. The question was whether the limit dose was likely to cover all exposure situations. The results indicate that it can be considered adequate in the case of dermal exposures and more than adequate in the case of exposure to aerosols. For very high exposures, which may be observed for volatile substances, the limit dose does not seem to cover all exposure situations. This finding suggests that higher experimental dose levels for substances, for which occupational exposure levels can be higher than 325 mg/m³, could be useful. This question was not addressed in the present study because other factors such as feasibility of higher experimental doses (possible metabolic saturation, acute toxicity, etc.) have to be considered. With respect to the German method for deriving specific concentration limits for classification of preparations, these considerations lead to the conclusion that the limit dose of 1000 mg/kg/day should not be used as a default starting point under all circumstances. For highly volatile liquid substances and for compounds with anticipated high percutaneous absorption, the putative maximum levels of human exposure should be checked. This may lead to higher dose levels being used as starting points when deriving substance-specific concentration limits.

Furthermore, exposure to a substance in a preparation cannot generally be anticipated to be linearly correlated to the concentration in the preparation. Both the type of use and mutual interactions of components in the preparation may lead to deviations from linear behavior. Therefore, for an adequate approach to setting specific concentration limits, not only should potency be considered, but how and whether deviations from linear correlation between concentration in the preparation and exposure should be taken into account should also be investigated.