Ann. occup. Hyg., Vol. 46, No. 2, pp. 175-185, 2002
© 2002 British Occupational Hygiene Society
Published by Oxford University Press
Article |
Time Extrapolation and Interspecies Extrapolation for Locally Acting Substances in case of Limited Toxicological Data
1Research and Advisory Institute on Hazardous Substances (FoBiG), Werderring 16, D-79098 Freiburg; 2Federal Institute for Occupational Safety and Health (BAuA), Postfach 170202, D-44061 Dortmund, Germany
Received 21 January 2001; in final form 3 August 2001.
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSION AND CONCLUSIONSREFERENCES |
|---|
In the case of substances with a limited toxicological data base there is often (i) a lack of qualified human toxicological data; and (ii) a paucity of studies with adequate exposure duration. Hence, several extrapolations have to be performed to arrive at appropriate risk assessments or derive occupational exposure limits. The present paper deals with the possibilities for extrapolating the change in effect concentrations over time (time extrapolation, e.g. from subacute to chronic exposure) and for interspecies extrapolation (from animal to human) in connection with locally acting substances (respiratory toxicants). To justify the time extrapolation factors, 46 technical reports produced by the US National Toxicology Program (NTP) involving studies with subacute, subchronic and chronic exposure duration were evaluated. On the basis of geometric mean values, decreases in effect concentrations by factors of 3.2 (subacute
subchronic), 2.7 (subchronic chronic) and 6.6 (subacute chronic) were found. Differentiation according to animal species (mouse, rat), sex or substance properties did not result in any relevant changes of the mean value. NTP studies with less than lifetime exposure periods (subacute, subchronic) in many cases showed different locations of respiratory effects compared with chronic studies, and thus offered limited possibilities for qualitative prediction of long-term respiratory effects (occurrence of effects in certain regions of the respiratory tract). With regard to interspecies extrapolation, gaseous and particulate substances were evaluated separately. With some modifications (e.g. consideration of the clearance of particles of low solubility), the 1994 US Environmental Protection Agency (EPA) model for deriving reference concentrations for humans on the basis of experimental data in animals is proposed for inhalable particulate substances. In the case of gaseous substances, the assumptions of the EPA model do not seem to consider sufficiently the local inhomogeneity in substance distribution and anatomical and histological differences between the upper respiratory tracts of rodents and humans. Considerable uncertainty would attach to a default factor for interspecies extrapolation for gaseous substances.
Keywords: time extrapolation; interspecies extrapolation; risk assessment; occupational exposure limits; respiratory toxicity
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSION AND CONCLUSIONSREFERENCES |
|---|
To achieve greatest reliability, occupational exposure limits and other toxicologically based guidance values should be derived from qualified human data on substance-dependent effects. However, since the amount of information on humans is usually limited, risk assessments must rely on experimental animal data. In addition, these data may have to be adjusted if the study design is not comparable to the human exposure scenario. Interspecies extrapolations and estimates of the change in effect of concentration over time (time extrapolation) are critical steps in the risk assessment procedure. Substance-specific data should be used whenever possible. Mechanistic data, pharmacokinetic and pharmacodynamic modelling, and structure–activity considerations may aid the assessment.
Such supporting data are frequently not available. Consequently, extrapolation methods have been proposed in regulatory toxicology. These include the use of default factors, e.g. for time extrapolation, interspecies extrapolation and, possibly, additional adjustments for variability within the human population (intraspecies extrapolation) (for recent reviews see e.g. Kalberlah and Schneider, 1998; Swartout et al., 1998; Vermeire et al., 1999; WHO, 1999).
In Germany, principles for deriving ‘provisional workplace guidance values’ (Arbeitsplatzrichtwerte, ARW) for airborne substances (Table 1) have been established by the Committee on Hazardous Substances (Ausschuss für Gefahrstoffe). ARW are occupational exposure limits derived for substances with a limited database. Within this methodology, default extrapolation factors have been established to take account of the lack of data when deriving ARW for airborne substances (Table 1). These default factors are used for systemic effects of non-carcinogenic, non-mutagenic and non-reproductive toxic agents.
|
Quantification of factors and methodical principals are based on the empirical database described by Kalberlah and Schneider (1998). Time extrapolation factors in Table 1 for systemic effects are based on our own evaluations and data from the literature (see Table 2). All of the default factors mentioned in the table should be modified in the light of substance-specific knowledge.
|
The extrapolation factors were established for systemic effects. Different factors may, however, be appropriate in cases where effects are seen in certain target organs, e.g. if local effects are observed in the respiratory tracts of experimental animals. The present work focuses on time and interspecies extrapolation of local (respiratory) effects. It has not yet been generically assessed whether (i) pulmonary, bronchial or upper respiratory tract effects progress over time in a manner similar to systemic effects; and (ii) despite the specific physiology of the respiratory tract in rodents, interspecies differences for gases and particles are compatible with the general approach for systemic effects (Table 1).
Answers to these questions would help to establish a more reliable basis from which to derive biologically plausible limit values, even with substances with a poor database.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSION AND CONCLUSIONSREFERENCES |
|---|
Time extrapolation
In order to assess the aggravation of respiratory effects over time, 46 technical reports on inhalation studies produced by the US National Toxicology Program (NTP) were evaluated. The pool of NTP studies has been chosen because it consists in a sufficient large number of peer-reviewed, high quality studies with varying study duration carried out with comparable study design. The following data have been documented by the authors:
physico-chemical and fundamental toxicological data relating to these substances;
the species, strain and sex of the exposed animals;
tested concentrations and the spacing of the time periods for testing;
the no observed adverse effect level (NOAEL) and lowest observed adverse effect level (LOAEL) for respiratory effects in subacute (14 day), subchronic (90 day) and chronic (~2 yr) studies;
the type of effect in the respiratory tract [extrathoracic (ET), tracheobronchial (TB), pulmonary (PU) region];
supplementary data from other studies, if accessible.
One set of data consisted of one pair of NOAELs (or LOAELs) for two exposure durations for one species and one sex (e.g. NOAELsubacute,mouse,female NOAELsubchronic,mouse,female). LOAELs and NOAELs were not explicitly stated in the NTP reports, but were instead derived by the authors through interpretation of the data presented by the NTP. Since not all of the substances were fully tested for both species, both sexes, all exposure periods and at comparable dose levels, and since a NOAEL was not always observed in the respective studies, the number of data sets for individual evaluations varied. A plausibility check was performed to ensure that LOAELs included in the data sets matched typical LOAEL effect sizes.
On rare occasions the lowest concentration tested in the long-term study was higher than the LOAEL in the short-term study and, in addition, no signs of adaptation were reported in the NTP report. In these cases, the LOAEL from the short-term study was assumed to be valid also for the longer-term study. To examine the quantitative effect of these modifications on the ratios obtained, the evaluation was performed with and without these modifications (see the Results section).
In several 14 day studies, high mortality rates (up to 100%) were observed after exposure periods of only a few days. In these cases, the exposure was characterized as acute, and the test concentrations were excluded from further evaluation (Table 3).
|
Further evaluation was undertaken to differentiate results with respect to the physico-chemical characteristics of the substances (water solubility, reactivity) and the affected region (ET, TB, PU). These data sets were subjected to statistical analysis in order to derive changes in NOAEL or LOAEL distributions over time. Minimum and maximum values, 10th and 90th percentiles, and median, geometric and arithmetic means were documented for all of the data sets that were either combined or differentiated according to the categories mentioned above (NOAEL, LOAEL, species, sex). Finally, the data were analysed with respect to the predictivity of the affected region. Would, for example, a substance causing ET effects after subacute exposure also reveal ET effects after chronic inhalation?
A prediction was judged to be ‘correct’ if in the long-term study the same, and no additional, region of the respiratory tract was reported to be affected. If fewer affected regions were documented in the long-term study, the short-term study was still regarded as sufficiently predictive (‘correct’).
A prediction was judged to be ‘false’ if
no respiratory toxicity at all occurred in the short-term study whereas a local effect was reported in the long-term study;
there was a change in region or if additional regions were affected in the long- term study;
no respiratory toxicity at all was reported in the long-term study, whereas effects were seen after subacute exposure.
Interspecies extrapolation
Three approaches were taken to derive a default procedure for interspecies extrapolation:
1. statistical evaluation of Agency for Toxic Substances and Disease Registry (ATSDR) toxicological profile data;
2. in-depth study of some exemplary substances for which sufficient human data are available;
3. evaluation of dosimetric approaches with regard to their suitability for use in the case of substances with a limited database.
ATSDR toxicological profile reports contain quantitative analyses of human data on health effects for a considerable number of substances, provide a certain degree of quality and have been peer reviewed. They can therefore be considered as a suitable basis for a statistical evaluation. The third approach was pursued in greater depth because it was regarded as potentially offering the best scientific input for interspecies extrapolation. Approaches 1 and 2 served as validation of the dosimetric procedure.
Evaluation of ATSDR data
Thirty-three toxicological profiles of chemicals from the ATSDR were examined with regard to the reported human versus animal NOAEL or LOAEL for respiratory toxicity based on comparable time ranges of exposure. A semi-quantitative analysis was performed on the basis of the calculated ratio (f) between species:
f = NOAELAnimal/NOAELHuman
or
f = LOAELAnimal/LOAELHuman
The analysis included all substances for which sensitivity on the part of humans as compared with rodents was deduced, as follows:
significantly higher sensitivity (f > 100)
higher sensitivity (100 
f > 5)
questionably higher sensitivity (5 f > 2)
identical or similar sensitivity (2 f > 1/2)
questionably lower sensitivity (1/2 f > 1/5)
lower sensitivity (1/5 f > 1/100)
significantly lower sensitivity (f > 1/100)
Evaluations were differentiated into subgroups: (i) gases versus particles; and (ii) acute versus chronic or subchronic exposure times (gases only). A total of 55 data sets were found: 36 for gases and 19 for particles.
In-depth study of some exemplary substances
Closer investigation revealed that comparable data allowing quantitative assessment of differences between rodents and humans only existed for an extremely small number of substances. Since it was hoped that well-known respiratory toxicants with extended databases might deliver qualified interspecies factors, an evaluation was undertaken for eight such substances:
acrolein
ammonia
chlorine
epichlorohydrin
sulphur dioxide
nitrogen oxides
formaldehyde
ozone
For each of these substances, human and animal data were related to the exposure time, the affected region of the respiratory tract and the mechanism of action (if data were available), and, most importantly, to quantitative differences in LOAELs or NOAELs.
Dosimetric approaches
Dosimetric models predict the behaviour of substances in the respiratory tract by reflecting physiological and anatomical properties of the organism (e.g. organ surface area and weight, respiratory minute volume, structure, diameter and length of the upper and lower airways), mechanisms of transport and interaction (e.g. diffusion, impaction, sedimentation, interception), and physico-chemical properties of substances (e.g. water solubility; in the case of particles: particle size).
Many dosimetric models either only describe the behaviour of certain types of substances (gases or particles) in particular parts of the respiratory tract and in selected species or demand various amounts of substance-specific data. The present investigation considered only those models that fulfilled the following criteria:
the model should be practicable without detailed substance-specific knowledge;
it should provide interspecies comparison;
all compartments of the respiratory tract should be included (ET, TB, PU);
the model should use important physiological/ anatomical parameters, such as respiratory minute volume, the surface area of respiratory tract compartments or organ weights of the particular species;
the model should differentiate between gases and particles;
in the case of particles, parameters such as the size of the particles, inhalability, deposition and clearance should be included.
Before proposing a particular model as a default procedure, plausibility was ascertained by comparison with empirical data.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSION AND CONCLUSIONSREFERENCES |
|---|
Time extrapolation
Combined data (both rodent species, both sexes, LOAEL and NOAEL data) lead to the following geometric means for time extrapolation:
factor 3.3, subacute subchronic (n = 106)
factor 2.7, subchronic chronic (n = 68)
factor 7.2, subacute chronic (n = 59)
As Table 4 shows, there are no distinct differences between the geometric mean and the median, or between the geometric mean and certain subcategories (NOAEL or LOAEL, or rat or mouse only). In addition, an evaluation of possible sex differences did not reveal any relevant deviations from the combined data. Nearly all of the assessed substances mainly acted on the extrathoracic (ET) area. The number of substances acting exclusively on either the pulmonary (PU) or tracheobronchial (TB) region was insufficient for statistical purposes. It therefore did not prove possible to perform differentiated analysis of time extrapolation factors for the PU/TB region as opposed to the ET region. Similarly, an analysis of physico-chemical properties and variations in time extrapolation did not provide any meaningful results.
|
With regard to the question of adaptive processes, possible adaptations (e.g. formic acid, isoprene, nitromethane) were revealed, these being indicated by higher NOAELs in the longer-term study compared with the short-term study. More specific conclusions would require closer examination of the original data. A time extrapolation factor of 1 was observed quite often. This can be interpreted either as a plateau effect or as a result of the experimental design (dose spacing).
Modifications of raw data were carried out in cases where higher LOAEL in the longer-term studies were observed compared with the short-term study without any sign of adaptation. These modifications showed only limited influence on the overall ratios observed. Without modifications, the following ratios (geometric means) were obtained:
factor 3.2, subacute subchronic (n = 106)
factor 2.4, subchronic chronic (n = 59)
factor 6.4, subacute chronic (n = 68)
Based on the above-mentioned definition, short-term studies often do not correctly predict the quality of effects after longer exposure:
correct prediction: subacute subchronic, 70%
correct prediction: subchronic chronic, 51%
correct prediction: subacute chronic, 57%
Interspecies extrapolation
Evaluation of ATSDR data
Table 5 shows the main results of the evaluation. In 62% (gases/liquids) and 53% (particles) of the cases, humans seemed to be more sensitive than animals. Only 17% and 27%, respectively, showed a higher sensitivity on the part of animals. After differentiation for exposure duration (only shown for gases), the basic relationship was similar, with only 16% (subacute exposure) or 18% (intermediate or chronic exposure) of data sets showing a lower NOAEL or LOAEL for animals in comparison with the reported human value.
|
The label ‘significantly higher’ was attributed to a high number of particles (32%), but was attributed less frequently (6%) to gases/liquids. It should be remembered that such a factor would account not only for interspecies toxicokinetic differences but also for toxicodynamics and possibly—in part—interhuman (intraspecies) variability.
In conclusion, the results suggest a slightly higher sensitivity on the part of humans.
In-depth study of some exemplary substances
Closer examination of the original data on eight substances produced the following overall results:
Acrolein: Only data for the ET region were (with reservations) comparable and showed a higher sensitivity on the part of humans [LOAEL = 0.15 p.p.m.; irritation after 5 x 1.5 min (Weber-Tschopp et al., 1977)] compared with the rat [NOAEL = 1.4 p.p.m.; no histological findings after 6 h exposure; slight histopathological effects at 0.65 p.p.m. after 3–6 h exposure on three consecutive days (Cassee et al., 1996)];
Ammonia: Based on the following observations, humans seem to be more sensitive to ET effects: (i) rats showed irritative effects [LOAEL = 376 p.p.m.; nasal discharge after 90 days of continuous exposure (Coon et al., 1970)] and histopathological changes (LOAEL = 150 p.p.m.; 4 weeks of continuous exposure) after subchronic exposure at concentrations 150 p.p.m. (Broderson et al., 1976); and (ii) with humans, irritation was observed after short-term exposure [LOAEL = 50 p.p.m.; single exposure lasting 2 h (Verberk, 1977)] and subchronic exposure [LOAEL = 100 p.p.m. exposure for 6 weeks (6 h/day, 5 days/week) (Ferguson et al., 1977)] to lower concentrations. Due to different exposure times, the comparisons between species are very limited. In the lower respiratory tract, however, rabbits may be somewhat more sensitive than humans in regard to pulmonary function [rabbit: LOAEL = 50 p.p.m. (Mayan and Merilan, 1972); man: LOAEL = 150 p.p.m. (Cole et al., 1977)].
Chlorine: In humans, chlorine primarily acts in the lower respiratory tract, whereas in rodents the primary target organ is the ET area. Comparison is hardly possible due to different exposure times. When an increase in effect over time is considered, humans appear to be more sensitive than rats (ET area) or monkeys (PU region): nasal lesions were seen in rats after 2 yr exposure to 0.4 p.p.m. [6 h/day, 5 days/week (Wolf et al., 1995)] compared with a LOAEL for humans of 0.5 p.p.m. with irritations after a single exposure lasting 4 h (Anglen, 1981). Monkeys exhibited no effects on pulmonary parameters after exposure to 2.3 p.p.m. for 1 year [6 h/day, 5 days/week (Klonne et al., 1987)] whereas, in humans, even a single 4 h exposure to 1 p.p.m. resulted in, for example, higher airway resistance and a change in forced vital capacity (Rotman et al., 1983).
Epichlorohydrin: Only a limited amount of data is available for species comparison. Based on Gage and Wexler (Gage, 1959; Wexler, 1971), a similar sensitivity (LOAEL ~20 p.p.m. after short-term or subacute exposure, with irritation or burning of nasal mucosa) in humans and rats may be assumed.
Sulphur dioxide: A slightly higher sensitivity on the part of humans can be deduced from the existing data in the literature, with possibly more pronounced effects in the PU region than in the ET region. Reported LOAEL concentrations are 10 p.p.m. for mice [a single 24 h exposure resulting in rhinitis (Giddens and Fairchild, 1972)] or 5 p.p.m. in humans [a change in mucous flow after a single 6 h exposure (Andersen et al., 1974)]. A comparison of guinea pig data revealed a NOAEL of 5.7 p.p.m. for pulmonary effects after 1 year of exposure (Alarie et al., 1970), with humans exhibiting changed airway resistance with a threshold of ~1 p.p.m. after only 1 h exposure (Lawther et al., 1975).
Nitrogen dioxide: There appear to be no distinct quantitative differences in sensitivity between animals and humans. A quantitative comparison was not possible because of: (i) differences in affected regions; (ii) differences in sensitivity in various animal species; and (iii) the influence of a short-term ceiling exposure on the irritative response (Englert, 1992).
Formaldehyde: Species differences vary according to which local effect is examined: in the case of sensory irritation based on respiratory depression in mice at 0.53 p.p.m. after 10 min (Barrow et al., 1980) or irritation of the trigeminus nerve at 0.12–0.24 p.p.m. in humans after short-term exposure (NRC, 1980; WHO, 1989), an interspecies factor of 2 or higher seems reasonable. A similar difference exists for histological effects with a LOAEL for rats after 1 day to 6 weeks of exposure (6 h/day) to 6 p.p.m. and a NOAEL of 2 p.p.m. (Monticello et al., 1991), compared with inflammations of the human nose after 2 h of exposure to 0.41 p.p.m. (Pazdrak et al., 1993). In contrast, with regard to DNA cross-links, animals appear to be more sensitive than humans (Heck et al., 1990).
Ozone: Due to its limited water solubility, ozone progresses to the lower respiratory tract. A change in breathing frequency, airway resistance, bronchio-reagibility and inflammation was observed, the scarcity of data preventing quantitative interspecies comparison. Similar adaptation was observed in rats and humans. Most of the data support the assumption of a similar level of sensitivity in these species (EPA, 1996b).
In conclusion, even in the case of well-known respiratory toxicants, only limited data are available for the purpose of quantitative interspecies comparison. The small data set summarized above supports the assumption of a marginally higher sensitivity on the part of humans on average.
Dosimetric approaches
At the time of the authors’ evaluation of dosimetric approaches, the only model covering gases and particles and regarded as potentially suitable for interspecies extrapolations in the case of limited data was the EPA approach (EPA, 1994). The validity of this methodology, however, differed for gases and particles, and some changes are proposed in the case of particles.
Gases: The EPA (1994) approach classifies gases (or volatile liquids) according to their water solubility and reactivity. This theoretically based dosimetric model includes systemic uptake mainly for insoluble substances of low reactivity, whereas highly soluble and/or reactive substances are assumed to act locally (mainly in the extrathoracic region) and not penetrate to the blood.
Homogeneous distribution of the inhaled chemical is assumed for the individual respiratory regions. In the case of the lower regions, it is necessary to subtract the amount already deposited in the ET region. A substance-specific mass transfer coefficient, describing the transport from the airstream into the mucosa and epithelium, is included in the calculation.
If this mass transfer coefficient is not known (as is usually the case for substances with a limited data base), the formula is reduced to a dosimetric adjustment factor considering only
respiratory minute volumes (animal/human)
regional surface areas (animal/human)
This approach is employed by US EPA/IRIS (EPA, 1998) for the derivation, at present, of 20 reference concentrations for gases, all but one assumed to act solely in the ET region. With regard to the ET region, model results for combined data from rats and mice would support the thesis of a higher sensitivity on the part of humans by a (geometric mean) factor of 5.5 compared with these rodent species.
Detailed models for individual substances demonstrate, however, that the assumption of a homogeneous distribution of inhaled chemical in the respiratory tract is not valid and that failure to incorporate the mass transfer coefficient in cases of insufficient data may lead to major mistakes in the assessment. Frederick et al. (1998) showed that, in the case of acrylic acid, effects related mainly to the olfactory epithelium within the ET and that the rat is more sensitive than humans. In addition, ozone (Overton et al., 1987) or vinyl acetate (Plowchalk et al., 1997) are examples of respiratory toxicants that exhibit a relevant local inhomogeneous distribution.
Important anatomical differences between species relate not only to surface areas but also to flow characteristics and mass transfer coefficients (Frederick et al., 1994).
For lower regions of the respiratory tract (TB, PU), even fewer data are available to validate the EPA model, and the limitations of the model, already mentioned in connection with the ET tract, remain.
In conclusion, substance-specific information and biological evidence show that the basic assumption of a uniform distribution of substances within the respiratory tract regions, i.e. one that neglects aspects such as solubility/chemical reactivity, is often not valid. This assumption, however, crucially affects the default procedure in cases of limited substance-specific data. Consequently, the justification of the EPA (1994) default procedure for gases is not sufficiently supported. Together with the ATSDR evaluation and the in-depth evaluation of some substances (see above), the overall conclusion is reduced to the view that, on average, humans may be slightly more sensitive to locally acting substances than rodents, but that this cannot be attributed primarily to differences in respiratory minute volumes and surface areas.
Particles: The semi-empirical EPA (1994) approach is based on various other models of the deposition characteristics of particles. These models relate to a much better empirical database than the model for gases. The basic model includes not only surface areas and respiratory minute volumes, but also (better validated) regional deposition efficiencies for particles. In addition, it uses particle size (mass median aerodynamic diameter) and particle-size distributions (dispersity) as input data.
However, comparison with a more recent dosimetric model for particles (EPA, 1996a) reveals some obvious discrepancies:
the latter includes the dosimetric consequences of interspecies clearance differences that were not assessed by the EPA (1994) model;
it employs slightly different data for deposition rates in humans than the EPA (1994) model;
pulmonary parameters are normalized to lung weight instead of pulmonary surface area.
Using one or the other of the models leads, however, to very similar results. Weighting the plausibility of the assumptions in both models, two modifications of the EPA (1994) approach are suggested:
use of a standard surface area for the human pulmonary region of 100 m2 instead of 54 m2 (based on Mercer et al., 1994; see discussion in Kalberlah et al., 1999);
inclusion of the clearance of particles in the lower respiratory tract [based on studies by Snipes (Snipes, 1989) and implemented by EPA (EPA, 1996a)], leading to correction of the dosimetry factor by: , particles of low solubility; , particles of medium solubility; or 1, particles of good solubility.
This correction factor should only be used if the interspecies extrapolation is based on rat or mice data. It is not applicable to guinea pig, dog or monkey data.
The factor means that, due to clearance differences, humans receive up to a 4-fold dose of particulates compared with rodents (mice and rats).
|
DISCUSSION AND CONCLUSIONS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSION AND CONCLUSIONSREFERENCES |
|---|
Time extrapolation
The onset and severity of toxic effects correlates to some extent with concentration and time (Haber, 1924; EPA, 1992). However, relationships are often far from linear. Some are purely concentration-dependent, others mainly time-dependent processes for certain substances or target organs (ten Berge et al., 1986) and may be influenced, for example, by adaptation, repair or saturation processes.
In addition, it is necessary to distinguish between ‘true’ progression of effects over time and ‘observed’ progression. Currently available short-term study designs do not include the same examination depth and the same number of animals as long-term studies. Consequently, it is possible that some effects documented in the 90 day or lifetime-exposure results are either not assessed or not reported in 14 or 28 day studies. Since this bias will occur regularly unless a specific, carefully designed, in-depth, short-term study is employed, the above extrapolation factors are related to ‘observed’ rather than ‘true’ changes in effects over time. Nevertheless, in the main, the design of the NTP studies follows OECD or EU guidelines (OECD, 1992, 1998; EC, 1996) and is therefore representative of standard investigations involving subacute, subchronic and chronic exposure.
Similarly, the spacing of the doses (short-term, long-term) may have an influence on observed differences between studies with different duration. These differences may not necessarily reflect the real slope of the time–response curve.
NTP studies are normally carried out with the focus on possible carcinogenic effects in the chronic regimen. It is therefore not surprising to discover fairly low predictivity in short-term studies of range-finding character. However, in the case of many existing chemicals, only short-term studies of similar study design may be available as a basis for extrapolation to long-term effects.
However, some spot-checks based on NTP reported data confirmed that the observed increase in effect frequency and/or severity with duration could not primarily be attributed to study design: if, for example, severity of the degeneration of the olfactory epithelium was scaled from ‘minimal’ to ‘moderate’ or ‘marked’, the frequency of higher scores was clearly increased after longer exposure at equal or lower subchronic LOAELs compared with subacute LOAELs (e.g. effects on mice with nitromethane, TR 461).
In conclusion, the evaluation supports default time extrapolation factors for local (respiratory) effects similar to those proposed elsewhere for systemic effects (see Table 1; Anonymous, 1998; Kalberlah and Schneider, 1998). It must be noted, however, that default factors should only be regarded as a best estimate in the absence of better substance-specific data or in cases where the data cannot be readily assessed. The emphasis on the geometric mean values in the above presentation of results does not preclude the use of, for example, the 90th percentile, depending on the desired level of safety.
The procedure used to derive ARW fills a gap where quantitative regulatory figures are urgently needed but data are insufficient for a better-founded approach with substance-specific data. Based on the observations presented above, time extrapolation within this procedure may be extended from systemically acting substances to respiratory toxicants in a similar way.
Interspecies extrapolation
Evaluation of the ATSDR data
In the case of the evaluation of the ATSDR data, relative inaccuracy was accepted in order to obtain a larger data pool. The results were then limited to the status of providing a tendency-based conclusion only. The main inaccuracies were due to the following:
for the same exposure time, different endpoints of respiratory toxicity may have been reported for humans than for rodents;
the ATSDR classification of effect severity was adopted without being subjected to further critical analysis;
since the studies were performed by different authors, each possibly having a different focus, they may not be directly comparable;
a more thorough investigation might have shown additional studies contradicting the ‘first sight’ comparison as shown in the ATSDR summaries.
Within these important limitations, the ATSDR evaluation contradicts the hypothesis that rodents are more sensitive to irritants or other respiratory toxicants because of their differing anatomy, their breathing rate or the obligatory nasal respiration in rodents.
In-depth study of some example substances
An in-depth survey of data for some chemicals reveals the limited comparability of the quantitative data.
A major concern is the possible bias because of the different effects of the same substance seen in humans (mostly irritational or ‘subjective’ findings) and other animals (mostly histopathologically observable effects). This possible influence on results has to be confirmed. The difference in the character of the effect may lead to the underestimation or overestimation of the interspecies factor: there may well be histological alterations in humans if you perform histopathology or lavage (which is rarely done), even at lower concentrations, or there may just be reversible irritations (with histological changes only at higher concentrations). A good example is methanol, which produces no signs of subjective irritation at or above 200 p.p.m. in humans after a 4 h exposure, even though there are subclinical changes in the nasal mucosa of humans at that concentration (Muttray et al., 2000).
However, these uncertainties reflect the unsatisfactory but typical situation in risk assessment based on animal data, where effects or effect locations have to be included that may not be exactly identical between species. Despite this, the comparisons give a qualitative and quantitative indication that can be used as a starting point, if no better information exists.
If ‘respiratory depression’ is documented for rodents, this sensory irritant reaction cannot be compared directly with human respiratory effects. In addition, it is rare for the exposure durations in the animal studies in the literature to correspond to the published human scenarios. As a result, greater uncertainty attaches to any extrapolation.
Finally, it was recognized that eight substances is not a sufficient database to derive a justified default interspecies extrapolation factor.
Nevertheless, analysis of the substance data (similar to the ATSDR evaluation) did not produce any evidence of rodents being generally more sensitive to irritants than humans. Differences between rodent and human effect levels observed with some substances may be interpreted as a tendency for a somewhat greater sensitivity towards respiratory tract irritants in humans. However, the limited comparability of observations from experimental and human studies impedes the drawing of any conclusion.
Dosimetric approaches
It should be possible to use dosimetric approaches as a generic procedure to account for interspecies differences in the risk assessment of locally acting substances, even in the absence of detailed substance-specific data. Simple comparisons of surface areas and breathing rates prove insufficient for such extrapolations because of relevant differences between human and other animals. Consequently, in the case of gases, the EPA model is only of limited use in the quantitative risk assessment of locally acting substances. The three approaches discussed here currently permit only the general conclusion that humans may be somewhat more sensitive to the respiratory effects of gases. This does not preclude the possibility of animals being more sensitive to individual chemicals, something that has to be assessed on a case-by-case basis.
Particles seem to possess certain common properties that allow the reasonable employment of a dosimetric model for the prediction of deposition and clearance, even in the absence of additional substance-specific data. The (modified) EPA (1994) approach currently represents a generally applicable dosimetric model for the estimation of important interspecies differences with regard to deposition and the clearance of particles. [Since the completion of the present report, an important additional proposal has been published (RIVM, 1999). It should be reviewed for comparison.]
Acknowledgement—This article relates to an extended report (Kalberlah et al., 1999) supported by the Federal Institute for Occupational Safety and Health (F&D project F1719). The authors greatly appreciate the cooperation and the funding of this project.
|
FOOTNOTES |
|---|
* Author to whom correspondence should be addressed. e-mail: fritz.kalberlah{at}fobig.de
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSION AND CONCLUSIONSREFERENCES |
|---|
Alarie Y, Ulrich CE, Busey WM, Swann HE, MacFarland HN. (1970) Long term continuous exposure of guinea pigs to sulfur dioxide. Arch Environ Health; 21: 769–77.[Medline]
Andersen I, Lundqvist GR, Jensen PL, Procter DF. (1974) Human response to controlled levels of sulfur dioxide. Arch Environ Health; 28: 31–9.[Medline]
Anglen DM. (1981) Sensory response of human subjects to chlorine. Doctoral dissertation, University of Michigan, Ann Arbor, MI.
Anonymous. (1998) Kriterien für die Ableitung von gesundheitsbasierten Luftgrenzwerten bei limitierter Datenlage. Bundesarbeitsblatt; 10: 74–6.
Barrow CS, Steinhagen WH, Chang JCF. (1980) Formaldehyde sensory irritation. In Chemical Industry Institute of Toxicology, editor. Proceedings of the 3rd CIIT conference on toxicology, Research Triangle Park. Washington, DC: Hemisphere Publ.
Broderson JR, Lindsey JR, Crawford JE. (1976) The role of environmental ammonia in respiratory mycoplasmosis of rats. Am J Pathol; 85: 115–30.[Abstract]
Cassee FR, Groten JP, Feron VJ. (1996) Changes in the nasal epithelium of rats exposed by inhalation to mixtures of formaldehyde, acetaldehyde, and acrolein. Fundam Appl Toxicol; 29: 208–18.[Web of Science][Medline]
Cole TJ, Cotes JE, Johnson GR, Martin H, Reed JW, Saunders MJ. (1977) Ventilation, cardiac frequency and pattern of breathing during exercise in men exposed to o-chlorobenzylidene malonitrile, CS and ammonia gas in low concentrations. Q J Exp Physiol; 62: 341–51.
Coon RA, Jones RA, Jenkins LJ, Siegel J. (1970) Animal inhalation studies on ammonia, ethylene glycol, formaldehyde, dimethylamine, and ethanol. Toxicol Appl Pharmacol; 16: 646–55.[Web of Science][Medline]
EC. (1996) Commission directive 96/54/EC of 30 July 1996 adapting to technical progress for the twenty-second time Council Directive 67/548/EEC on the approximation of the laws, regulations and administrative provisions relating to the classification, packaging and labelling of dangerous substances. Part II. Test methods. Official Journal no. L248, 30. 9.1996, pp. 112–273.
Englert N. (1992) Anorganische Gase/Schwefeldioxid, Kap. VI–1. In Wichmann HE, Schlipköter H-W, Fülgraff G, editors. Handbuch der Umweltmedizin. Toxikologie, Epidemiologie, Hygiene, Belastungen, Wirkungen, Diagnostik, Prophylaxe. Landsberg: Ecomed Verlag.
EPA. (1992) Draft report: a cross-species scaling factor for carcinogen risk assessment based on equivalence of mg/kg3/4/day. Fed Register; 57: 24152–72.
EPA. (1994) Methods for derivation of inhalation reference concentrations and application of inhalation dosimetry, review draft. Washington, DC: Office of Research and Development.
EPA. (1996a) Air quality criteria for particulate matter. I–III. Research Triangle Park, NC: US Environmental Protection Agency.
EPA. (1996b) Air quality criteria for ozone and related photochemical oxidant. I–III. Research Triangle Park, NC: US Environmental Protection Agency.
EPA. (1998) IRIS, integrated risk information system. CD-ROM Datenbank, Silver Platter: Environmental Protection Agency.
Ferguson WS, Koch WC, Webster LB, Gould JR. (1977) Human physiological response and adaption to ammonia. J Occup Med; 19: 319–26.
Frederick CB, Morris JB, Kimbell JS, Morgan KT, Scherer PW. (1994) Comparison of four biologically based dosimetry models for the deposition of rapidly metabolized vapors in the rodent nasal cavity. Inhal Toxicol; 6: 135–58.
Frederick CB, Bush ML, Lomax LG, Black KA, Finch L, Kimbell JS, Morgan KT, Subramaniam RP, Morris JB, Ultman JS. (1998) Application of a hybrid computational fluid dynamics and physiologically based inhalation model for interspecies dosimetry extrapolation of acidic vapors in the upper airways. Toxicol Appl Pharmacol; 152: 211–31.[Web of Science][Medline]
Gage JC. (1959) The toxicity of epichlorhydrin vapour. Br J Ind Med; 16: 11–14.
Giddens WE, Fairchild GA. (1972) Effects of sulfur dioxide on the nasal mucosa of mice. Arch Environ Health; 25: 166–73.[Medline]
Haber F. (1924) Zur Geschichte des Gaskrieges. In Fünf Vorträge aus den Jahren 1920–1923. Berlin: Springer Verlag, pp. 76–92.
Heck Hd’A, Casanova M, Starr TB. (1990) Formaldehyde toxicity—new understanding. Crit Rev Toxicol; 20: 397–425.[Web of Science][Medline]
Kalberlah F, Schneider K. (1998) Quantification of extrapolation factors. Final report of the research project no. 1116 06 113 of the Federal Environmental Agency. Schriftenreihe der Bundesanstalt für Arbeitsschutz und Arbeitsmedizin Dortmund, Fb 797. Bremerhaven: Wirtschaftsverlag NW.
Kalberlah F, Schneider K, Schuhmacher US, Voss J-U, Ioannidis I, Oltmanns J. (1999) Zeitextrapolation und Interspeziesextrapolation bei lokal wirksamen Stoffen mit begrenzter Datenlage. Schriftenreihe der Bundesanstalt für Arbeitsschutz und Arbeitsmedizin Dortmund, Fb 862. Bremerhaven: Wirtschaftsverlag NW.
Klonne DR, Ulrich CE, Riley MG, Hamm TE Jr, Morgan KT, Barrow CS. (1987) One-year inhalation toxicity study of chlorine in rhesus monkeys (Macaca mulatta). Fundam Appl Toxicol; 9: 557–72.[Web of Science][Medline]
Kramer HJ, van den Ham WA, Slob W, Pieters MN. (1995) Derivation of conversion factors to estimate an indicative chronic NOAEL from short-term toxicity data. Bilthoven: National Institute of Public Health and the Environment.
Lawther PJ, MacFarlane AJ, Waller RE, Brooks AGF. (1975) Pulmonary function and sulphur dioxide, some preliminary findings. Environ Res; 10: 355–67.[Medline]
Mayan MH, Merilan CP. (1972) Effects of ammonia inhalation on respiration rate of rabbits. J Anim Sci; 34: 448–52.
Mercer RR, Russell ML, Roggli VL, Crapo JD. (1994) Cell number and distribution in human and rat airways. Am J Respir Cell Mol Biol; 10: 613–24.[Abstract]
Monticello TM, Miller FJ, Morgan KT. (1991) Regional increases in rat nasal epithelial cell proliferation following acute and subchronic inhalation of formaldehyde. Toxicol Appl Pharmacol; 111: 409–21.[Web of Science][Medline]
Muttray A, Klimek L, Faas M, Schäfer D, Mann W, Konietzko J. (2000) 200 p.p.m. Methanol steigern die Sekretion proinflammatorischer Interleukine in der Nasenschleimhaut des Menschen. In Schäcke G, Lüth P, editors. Hrsg. 40. Jahrestagung der Deutschen Gesellschaft für Arbeitsmedizin und Umweltmedizin, Berlin, 15–18 Mai 2000. Lübeck.
NRC. (1980) Formaldehyde: an assessment of its health effects. Report NAS/ACT/P-881A. Washington, DC: National Research Council.
OECD. (1992) OECD guidelines for the testing of chemicals, vol. 2. Paris: OECD.
OECD. (1998) Tenth addendum to the OECD guidelines for the testing of chemicals. Paris: OECD.
Overton JH, Graham RC, Miller FJ. (1987) A model of the regional uptake of gaseous pollutants in the lung. II. The sensitivity of ozone uptake in laboratory animal lungs to anatomical and ventilatory parameters. Toxicol Appl Pharmacol; 88: 418–32.[Web of Science][Medline]
Pazdrak K, Gorski P, Krakowiak A, Ruta U. (1993) Changes in nasal lavage fluid due to formaldehyde inhalation. Int Arch Occup Environ Health; 64: 515–19.[Medline]
Plowchalk DR, Andersen ME, Bogdanffy MS. (1997) Physiologically based modeling of vinyl acetate uptake, metabolism, and intracellular pH changes in the rat nasal cavity. Toxicol Appl Pharmacol; 142: 386–400.[Web of Science][Medline]
RIVM. (1999) Multiple path particle deposition model, MPPDep version 1.11. Bilthoven: Chemical Industry Institute of Toxicology, Rijksinstituut voor Volksgezonheid en Milieu.
Rotman HH, Fliegelman MJ, Moore T, Smith RG, Anglen DM, Kowalski CJ, Weg JG. (1983) Effects of low concentrations of chlorine on pulmonary function in humans. J Appl Physiol; 54: 1120–24.
Rulis AM, Hattan DG. (1985) FDA’s priority-based assessment of food additives. II. General toxicity parameters. Regul Toxicol Pharmacol; 5: 152–74.[Web of Science][Medline]
Snipes MB. (1989) Long-term retention and clearance of particles inhaled by mammalian species. Crit Rev Toxicol; 30: 175–211.
Swartout JC, Price PS, Dourson ML, Keenan RE. (1998) A probalistic framework for the reference dose. Risk Anal; 18: 271–82.[Web of Science][Medline]
ten Berge B, Zwart A, Appelman LM. (1986) Concentration-time mortality response relationship of irritant and systemically acting vapours and gases. J Hazard Mat; 13: 301–9.
Verberk MM. (1977) Effects of ammonia in volunteers. Int Arch Occup Environ Health; 39: 73–81.[Medline]
Vermeire T, Stevenson H, Peiters MN, Rennen M, Slob W, Hakkert BC. (1999) Assessment factors for human health risk assessment: a discussion paper. Crit Rev Toxicol; 29: 439–90.[Web of Science][Medline]
Weber-Tschopp A, Fischer T, Gierer R, Grandjean E. (1977) Experimentelle Reizwirkungen von Akrolein auf den Menschen. Z Arbeitswissen; 31: 166–71.
Weil CS, McCollister DD. (1963) Relationship between short- and long-term feeding studies in designing an effective toxicity test. Agric Food Chem; 11: 486.
Wexler B. (1971) Determination of epichlorohydrin contamination in an industrial facility for the manufacturing of epoxy resins. Mater Plast (Buch); 8: 322–3.
WHO. (1989) Environmental health criteria 89. Formaldehyde. International Programme on Chemical Safety. Geneva: World Health Organization.
WHO. (1999) Environmental health criteria 210. Principles for the assessment of risks to human health from exposure to chemicals. International Programme on Chemical Safety. Geneva: World Health Organization.
Wolf DC, Morgan KT, Gross EA, Barrow, C, Moss OR, James RA, Popp JA. (1995) Two-year inhalation exposure of female and male B6C3F1 mice and F344 rats to chlorine gas induces lesions confined to nose. Fundam Appl Toxicol; 24: 111–31.[Web of Science][Medline]
Woutersen RA, Til HP, Feron VJ. (1984) Sub-acute versus sub-chronic oral toxicity study in rats: comparative study of 82 compounds. J Appl Toxicol; 4: 277–80.
CiteULike 
Connotea 
Del.icio.us What's this?
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||