SCIENTIFIC QUESTIONS TO THE CSTEE ON PAHs
Note: Following the terms of reference, this opinion focuses on specific questions related to human health issues and does not cover ecosystem protection.
Scope of the Position Paper:
To make recommendations to the Commission on air quality standards, monitoring and assessment strategies and tools for implementation, the position paper addresses the following topics: Sources of PAH emissions, PAH concentrations in ambient air, PAH speciation, trends in emissions, ambient levels and speciation, measurement and assessment techniques.
The review is intended to describe the effects and impacts of PAH in a European context and to collate the experience of Member States in the assessment and management of risks associated with PAH, air quality standards and guidelines and means of achieving compliance.
General Comments:
The position paper is based on previous reports on the health significance of PAH exposure (see 83.: Environmental Health Criteria: report 171, 1996; report 202, 1998; WHO Regional Office for Europe, 1987; the UK EPAQS 1999, CONCAVE , IARC Vol. 34, 1984; US EPA 1994; Swedish EPA in press). It applies the traditional and generally accepted approach of these bodies to use B[a]P as a marker substance for mixtures of PAHs and extrapolation from high doses at which workers have been exposed to lower concentrations in ambient air.
The position paper carefully describes the available animal and epidemiological studies. The CSTEE agrees with the critical evaluation of these studies. It is indicated in the report that the carcinogenic effects in animals may be the result of lung overloading. These overloading effects are suggested to result in a thresholded dose-response of the carcinogenic effects. Although the CSTEE agrees with this interpretation of the carcinogenic effects observed in the rodent studies it disagrees that the carcinogenicity of PAH is thresholded. PAHs are genotoxic carcinogens and current knowledge considers such mechanisms non-thresholded. However, in ambient air PAHs usually are bound to particles. Since there is indication that bioavailability of particle bound PAH may be low, a non-linear dose response may be assumed. To avoid misinterpretation the term non-thresholded should not be used to describe such dose response.
The CSTEE also agrees that the epidemiological data including those on workers exposed to diesel exhaust only provides limited evidence of carcinogenicity to man. Due to decreasing workplace exposure it is seen unlikely that the issue will ever be resolved.
Discussing the need to set PAH limits for ambient air the CSTEE notes that exposure via ambient air is relatively small as compared to oral ingestion. Concentrations of 1 ng B[a]P/m3 ambient air will result in a daily ingestion of about 10 ng whereas ingestion via food varies between 15 and 360 ng.
Data from the Netherlands, that represent the 1976-1986 period, show that the daily intake of B[a]P via food varies from 80 to 207 ng (Vaessen et al, 1988; de Vos et al, 1990; Heisterkamp and van Veen,1997). From other countries in Europe (i.e. UK, Austria, and Italy), averaged daily B[a]P intakes of 50 to 250 ng per person are reported, ranging from 15 to 360 ng. (Dennis et al 1983; Pfannhauser 1991; Lodovici et al 1995; Turrio-Baldassarri et al 1996). Among others, US studies show that daily intake differences per individual or between individuals may span up to three orders of magnitude depending on differences in food habits and cooking methods, as well as the source of the foods (see Butler et al 1993).
Main dietary sources of PAH are cereals, oils and fats (especially vegetable margarine), milk, some leafy vegetables, e.g. kale and spinach, mussels, and prepared meat, some of which showed clear associations with conditions of local air pollution (de Vos et al 1990; Dennis et al 1991; Lodovici et al 1995). The CSTEE concludes therefore that limitation of PAH concentrations in ambient air will result in a reduction of PAH concentrations in unprocessed human food.
In rats and mice oral PAH exposure results in B[a]P- and other PAH-adducts in lung DNA, although no tumours develop at this site in animals exposed to the oral route (Culp et al 1999). The authors conclude that a high number of B[a]P adducts in lung DNA per se appears not to be sufficient to induce tumour-formation at this site. However, when mice are orally exposed to B[a]P together with other PAH in coal tar, lung tumours are induced in parallel with an about ten-fold increase in adduct number (Culp et al 1999).
Although the working group has based its conclusions on several international bodies that have used the same traditional approach, the CSTEE concludes that the position paper has not sufficiently addressed several aspects that affect accuracy of risk characterisation:
- Differentiation between free and particle bound PAH and its consequence to justify linear extrapolation from high to low doses to calculate the unit risk.
- Risk characterisation based on a non-linear model.
- Carcinogenic risks of inhalation exposure versus PAH exposure via food.
- The varying PAH concentrations in PAH mixtures with different mutagenic and carcinogenic potencies.
- Presence of other mutagenic and carcinogenic compounds in ambient air.
- Atmospheric degradation of PAH.
Considering this, proposals of air control levels for B[a]P or PAH may not be justified scientifically.
Question 1:
Does the committee agree with the view of the working group that PAH in ambient air should be considered likely carcinogens and that a limit value should be set?
The position paper is based on the previous reports indicated above. It can be concluded from these reports that PAH should be considered as likely animal carcinogens. For example, of the 33 PAHs which have been evaluated by IPCS (1998) 17 are or have been suspected of being carcinogenic in laboratory animals. More recently RIVM evaluated 17 PAH and considered that only 4 have not to be carcinogenic, the others are either accepted or suspected carcinogens. The carcinogens beyond reasonable doubt are benz[a]anthracene, benzo[b]fluoranthene, benzo[j]fluoranthene, benzo[k]fluoranthene, benzo[a]pyrene, chrysene, dibenz[a,h]anthracene, and indeno[1,2,3-cd]pyrene. Acenaphtene, acenaphtylene, fluoranthene, phenanthrene, and pyrene have been considered suspected carcinogens. Only naphthalene, anthracene, benzo[ghi]perylene, and fluorene have been considered non-carcinogenic. Recently naphthalene has been shown to be carcinogenic in rats and mice (NTP 2000). Many of the individual compounds have been classified by IARC (see Table 20 of the report).
Other PAHs like dibenzo[a,l]pyrene which may have a 100 times higher carcinogenic potency than B[a]P have not been evaluated in the position paper. Several PAH containing mixtures like coal tars and some exposure circumstances related to PAH like coal production have been investigated in adequately designed epidemiological studies and proved to be carcinogenic to humans (see Table 21).
Since ambient air contains most of the PAH compounds found in the industrial mixtures shown to cause cancer, ambient air must be considered to have a similar carcinogenic potential like industrial mixtures. However, since PAH concentrations in ambient air are much lower, a lower carcinogenic potency can be anticipated. (See response to question 2).
The CSTEE agrees with the view that as a group, PAH in ambient air should be considered likely carcinogens for man. This evaluation is based on the fact that many individual PAHs proved to be carcinogenic in a variety of experimental systems, whereas for no individual PAH the conditions have ever existed for carrying out an epidemiological study allowing for the control of any confounding effect from any other PAH. This is why no individual PAH - not even BaP - is evaluated by IARC as "sufficient evidence of carcinogenicity to man" Since the carcinogenic risk of PAH in ambient air is related to the PAH concentrations the risk can be limited by setting appropriate air limit values.
Question 2:
Does the committee agree that limit values in this case (PAH) should be based on epidemiological studies?
As pointed out in a recent review by Boffetta et al (1997) interpretation of the studies on cancer risk from PAH exposure in humans is complicated by several factors. Depending on the raw material and the combustion circumstances PAHs occur in mixtures of variable composition. Especially in occupational environments PAHs are absorbed to particles, which affect bioavailability of the individual compounds. Animal experiments indicate, however, that particles in the submicron range are potent in inducing lung cancer and that PAHs play a little role in this effect (Heinrich et al 1995, Nikula et al 1995). Epidemiological data from the exposure outside the workplace are sparse. PAHs in urban air occur in the gaseous and particle phase and originate mainly from residential heating and vehicle fuel combustion. Indoor air sources are tobacco smoke, open fires and cooking. There is strong suggestion (Dockery et al 1993, Pope et al 1995) that exposure to urban air pollution is associated with an increased risk of lung cancer in humans. However, it is not known to what extent this excess can be attributed to exposure to PAHs, related heterocyclic or other organic or inorganic compounds, to fine particles or to a combination of these.
Most available information comes from industry-based cohort and nested case-control studies, only some of which allowed for some dose-response estimates of the association between lung cancer and PAH exposure. Their usefulness in order to estimate limit values in ambient air is limited, because air composition of this workplace is hardly comparable to that in ambient air. PAH profiles and coexposure to other agents may differ significantly. Results of these studies have been analysed in Europe, US, Germany and Canada, leading to estimates of lung cancer risk/100.000 ranging between 0.1 and 9.0 (Boffetta et al 1997). The Working Group uses the highest estimate (from the coke oven study of Redmond 1976, 1983) of 9 x 10-5 at a life time exposure of 1 ng B[a]P/m3 as a starting point and recommends that the common air quality standard for B[a]P should be below this level.
The CSTEE questions the scientific justification of using the above epidemiological study for several reasons:
- Because the composition of the mixtures in ambient air and the workplace are very different both qualitatively and quantitatively the extrapolation from occupational exposure to the general population is biased to an unknown extent.
- The unit risk estimated by WHO from the coke oven worker study (and other epidemiological studies) is that of a mixture the components of which (both PAHs and other substances) contribute to the overall carcinogenic potency with interactive mechanisms which cannot be extrapolated to other mixed exposures. Other estimates give lower values of lung cancer risks (see Boffetta et al 1997)
The CSTEE also concludes that the scientific justification to use the presently available epidemiological data for deriving a common air quality standard is questionable.
Question 3:
Does the committee agree that B[a]P is an acceptable indicator for the carcinogenicity of PAH mixtures in ambient air?
In the position paper B[a]P is considered to be an appropriate marker for PAHs for most atmospheric conditions in Europe. However, there is increasing information, that PAH mixtures contain varying amounts of individual compounds with higher mutagenic or carcinogenic potencies than B[a]P (Deutsch-Wenzel et al. 1983; Devanesan et al. 1990; Cavalieri et.al. 1991).
The CSTEE also draws the attention to the presence of other potent mutagens, and possibly carcinogens, in ambient air such as nitro- and dinitro-PAH and the potent mutagen 3-nitrobenzanthrone, a Diesel-specific marker (Enya et al. 1997; Adachi et al. 1999).
Moreover, under atmospheric conditions some PAHs are rapidly converted (sometimes specifically UV-catalysed) into highly reactive derivatives such as epoxides, endoperoxides and subsequently to phenols and quinones, the toxicological relevance of which is still unknown.
The CSTEE agrees that B[a]P may be used as a semi-quantitative marker for the presence of carcinogenic PAH in ambient air or other environmental matter. However, its application for the quantitative evaluation of the carcinogenic potency of these matrices is questionable because there are also other carcinogenic PAH present in varying concentrations with differing mutagenic and/or carcinogenic potencies.
- Balancing of the carcinogenic potency of various environmental matter (vehicle exhaust, diesel exhaust extract, hard coal combustion effluents, has shown that, though most of the carcinogenic potential can be attributed to PAH, B[a]P itself contributes to a variable extent - to this effect, (Jacob 1996).
- In a number of studies relative potency factors (related to B[a]P = 1.00) for various PAH have been established showing that there are a number of PAHs which exhibit 10-20% of the activity of B[a]P and others which are even more potent. Dibenz[a,h]anthracene is so by a factor of 2 and dibenzo[a,l]pyrene is suggested to be about 100 times more active though not yet tested under comparable conditions (IPCS 1998). TEF-values are given in Table 19 of the report.
- The relative concentrations of PAH vary widely in environmental matter and also in ambient air depending on the main source of pollution (oil heating 0.5%, hard coal combustion 6.6% (IARC 1983). Data given in Table 27 of the Position Paper should be expanded by incorporation of more recently identified compounds.
- It has been shown at least for coal combustion effluents that higher boiling PAHs which mainly have not yet been identified, contribute by about 50% to the total carcinogenic potency of this matrix (Grimmer et al 1985). One of these compounds is likely to be dibenzo[a,l]pyrene.
In summary, to cover the varying PAH concentrations in PAH mixtures and the different carcinogenic potencies of the individual compounds there is increasing scientific justification to use the sum of potency factors multiplied with the actual environmental concentrations of a number of PAH as a toxicological basis for limit values. B[a]P alone should no longer be determined. The CSTEE recommends to regularly determine compounds listed in Table 27 of the report. Dibenzo[a,e]pyrene, dibenzo[a,h]pyrene and naphthalene are relevant and should also be included for determination.
Presently, there is insufficient data on the composition of PAH mixtures at the different locations whereas the potency factors become increasingly validated (Table 19 of the report). The CSTEE recommends research activities to obtain reliable information. Till then B[a]P may be used as a semi-quantitative indicator for the presence of carcinogenic PAH mixtures in ambient air.
Question 4:
Does the committee agree that these PAH compounds may act through the Ah receptor, possibly leading to up-regulation of certain enzymes that produce intermediates that are genotoxic?
It is generally accepted that one way of PAH mediated toxicity is interaction with the Ah receptor. As pointed out in the working group-report (85. - 87.) this receptor up-regulates expression of factors which control cellular growth, differentiation and metabolic PAH activation and inactivation. Since PAH metabolites interact with the DNA and are considered primary genotoxic carcinogens this will increase production of genotoxic agents and DNA adducts. However, in the Swedish EPA report on PAHs (in press) the impact of such mechanisms on the shape of the dose-response-curve has been discussed and it was concluded that the dose response of PAH carcinogenicity may be better described by an S-shaped curve.
The CSTEE agrees that part of the carcinogenic PAH mechanism is mediated through the Ah receptor that affects initiation and promotion during the carcinogenic process. The CSTEE proposes to use such observation to better describe the shape of the dose response of PAH carcinogenicity. However, it remains to be evaluated e.g. by PBPK modelling whether and to what extent such factors affect the dose response.
Question 5:
Does the committee agree that linear extrapolation is likely to overestimate the risk?
The position paper relies on the unit risk approach of WHO which is based on linear extrapolation from high to low dose resulting in life time risks of 10-4 at 1 ng B[a]P, 10-5 at 0.1 ng B[a]P and 10-6 at 0.01 ng B[a]P/ m3.
PAHs are genotoxic carcinogens and current knowledge considers such mechanisms non-thresholded and a linear extrapolation is justified. However, in ambient air PAHs are usually bound to particles and there is indication that bioavailability of particle bound PAHs is low. Moreover, PAH contaminated air also contains various highly reactive compounds that may induce chronic inflammation and contribute to carcinogenicity by formation of reactive oxygen species and cellular mediators with the consequence of increased cell proliferation and tumour promotion. Such effects are thresholded and will increase the slope of the dose response curve of PAH carcinogenicity when becoming effective. Assuming that coke oven workers are exposed to PAH and irritants a linear extrapolation from such exposure conditions to ambient air will overestimate the carcinogenic risk.
To avoid misinterpretation the term non-linear rather than threshold should be used.
The CSTEE concludes that due to
- The inconsistencies of the risk estimates provided by different epidemiological studies
- The possibly different carcinogenic mechanisms of free and particle bound PAH as well as the limited knowledge on the respective proportions of the two forms of PAH in most circumstances
- The differing compositions of PAH containing air at workplaces and in the environment and in ambient air
- The probable interaction of the carcinogenic potential of PAH with that of other compounds than PAH at the workplace and in ambient air
linear extrapolation from occupational studies to estimate risks related to ambient air most likely overestimates the carcinogenic risk.
However, studies on DNA adducts in humans suggest that at occupational exposure levels the pathways of metabolic activities of PAH to DNA damaging species become saturated (Lewtas et al 1997) suggesting that linear extrapolation from saturating exposure concentrations would underestimate tumour initiating activity of PAH at low doses. Since there is increasing evidence that DNA-adducts do not directly correlate with genotoxicity and carcinogenicity the biological relevance of such information has to be further evaluated.
Question 6:
Considering the uncertainty in the estimation of the unit risk, does the committee agree that a limit value in the range of 0.01 to 1.0 ng/m3 would provide an adequate standard of protection for public health?
The CSTEE has pointed out the large uncertainties in the estimation of the unit risk (see comments to the previous questions) so that the range of 0.01 to 1.0 ng/m3 as based on life time risks between 10-6 and 10-4 is not supported from a toxicological standpoint. However, the CSTEE recognises that any exposure to PAH is associated with a cancer risk to humans. Reduction of PAH concentrations in ambient air will reduce human exposure via inhalation. This will also reduce contamination of human unprocessed food from PAH contaminated air. The CSTEE also notes that the magnitude of food contamination is dependent on differences in food habits and cooking methods.
As shown by Table 25 of the position paper, high atmospheric concentrations of PAH e.g. - 1 ng/m3 are common in urban areas. A decrease of such levels can be achieved by applying modern technology in combustion and other technical processes. Successful application of such tools can be readily measured.
Final Conclusion
PAH mixtures in ambient air differ in concentration and in composition. B[ a] P is an important constituent of such mixture but its concentration varies. Therefore B[ a] P cannot be a sole indicator of the carcinogenic potency of such mixtures. To allow a quantitative risk assessment of PAH mixtures in ambient air a relevant number of carcinogenic PAHs must be analysed.
References
- S. Adachi, K. Kawamura, K. Takemoto, H. Suzuki and Y. Hisamatsu (1999) 7th International Inhalation Symposium: Relationship between Acute and Chronic Effects of Air Pollution, Hannover Medical School, Book of Abstracts, 67 A44.
- Boffetta P, Jourenkova N, Gustavsson P (1997) Cancer risk from occupational and environmental exposure to polycyclic aromatic hydrocarbons. Cancer Causes and Control 8, 444-472.
- Butler JP, Post GB, Lioy PJ, and Waldman JM (1993) Assessment of carcinogenic risk from personal exposure to benzo[a]pyrene in the Total Environmental Exposure Study (THEES). J.Air Waste Manage.Assoc., 43, 970-977.
- Cavalieri, E.L., Higginbotham,S. Ramakrishna,N.V.S, Devanesan,P.D., Todorvic,R., Rogan,E.G. and Salmasi,S. (1991) Comparative dose-response tumorigenicity studies of dibenzo[a,l]pyrene versus 7,12-dimethylbenz[a]anthracene, benzo[a]pyrene and two dibenzo[a,l]pyrene dihydrodiols in mouse skin. Carcinogenesis 12, 1937-1944
- Culp SJ, Gaylor DW, Sheldon WG, Goldstein LS, and Beland FA (1998) A comparison of the tumours induced by coal tar and benzo[a]pyrene in a 2-year bioassay. Carcinogenesis, 19, 117-124.
- Dennis MJ, Massey RC, McWeeney DJ, and Knowles ME (1983) Analysis of polycyclic aromatic hydrocarbons in UK total diets. Fd.Chem.Toxic., 21, 569-574.
- Dennis MJ, Massey RC, Cripps G, Venn I, Howarth N, and Lee G (1991) Factors affecting the polycyclic aromatic hydrocarbon content of cereals, fats, and other food products. Food Addit.Contam., 8, 517-530.
- de Vos RH, van Dokkum W, Schouten A, and de Jong-Berkhout P (1990) Polycyclic aromatic hydrocarbons in Dutch total diet samples (1984-1986). Fd Chem.Toxic., 28, 263-268.
- Deutsch-Wenzel R, Brune H, Grimmer G, Dettbarn G, Misfeld J (1983): Experimental studies in rat lungs on the carcinogenicity and dose-response relationship of eight frequently occurring environmental polycyclic hydrocarbons. JNCI 71, 539-544
- Devanesan PD, Cremonesi P, Nunnally JE, Rogan EG, Cavalieri EL (1990) Metabolism and mutagenicity of dibenzo[ a] pyrene. Chem Res Toxicol 3, 580-586
- Dockery DW, Pope CA, Xu X, Spengler JD, Ware JH, Fay ME, Ferris BG, Jr, Speizer FE (1993) An association between air pollution and mortality in six US cities. N Engl J Med 329, 1753-1959
- Enya T, Suzuki H, Watanabe T, Hirayama T, Hisamatsu Y (1997) 3-Nitrobenzanthrone, a powerful bacterial mutagen and suspected human carcinogen found in Diesel exhaust and airborne particulates. Environ Sci Technol 31, 2772-2776
- Grimmer G, Brune H, Deutsch-Wenzel R, Dettbarn G, Misfeld J, Abel U, Timm J: The contribution of polycyclic aromatic hydrocarbon fractions with different boiling ranges to the carcinogenic impact of emission from coal-fired residential furnaces as evaluated by topical application to the skin of mice. Cancer Lett 28, 203-211, 1985
- Heinrich U, Fahst R, Rittinghausen S et al (1995) Chronic inhalation exposure of Wistar rats and two different strains of mice to diesel engine exhaust, carbon black, and titanium dioxide. Inhal Toxicol 7, 533-556.
- Heisterkamp SH, and van Veen MP (1997) Blootstelling aan xenobiotica in voeding. Voorbeeldstoffen: Butyl benzyl phthalate (BBP), Benzo[a]pyreen en Fluorantheen. RIVM report no. 604502 002.
- IARC (1983) IARC Monographs on the Evaluation of the Carcinogenic Risk of Chemicals to Humans. Polynuclear Aromatic Compounds, Part 1, Chemical, Environmental and Experimental Data, Vol. 32, Lyon, France, p.39.
- IPCS Environmental Health Criteria 202, World Health Organisation Geneva, 1998
- Jacob J (1996): The significance of polycyclic aromatic hydrocarbons as environmental carcinogens. Pure Appl Chem 68, 301-308
- Lewtas J, Walsh D, Williams R, Dobias L: Air pollution exposure-DNA adduct dosimetry in humans and rodents: evidence for non-linearity at high doses. Mutation Res 378, 51-63, 1997
- Lodovici M, Dolara P, Casalini C, Ciapellano S, and Testolin G (1995) Polycyclic aromatic hydrocarbon contamination in the Italian diet. Fd Addit.Contam., 12, 703-713
- Nikula KJ, Snipes MB, Barr EB, Griffith WC, Henderson RF, Mauderly JL (1995) Comparative pulmonary toxicities and carcinogenicities of chronically inhaled diesel exhaust and carbon black in F344 rats. Fund Appl Toxicol 25, 80-94.
- Pfannhauser W (1991) Polyzyklische aromatische Kohlenwasserstoffe (PAKs) in der Nahrung - Ein Vergleich von Literaturdaten mit einer Östereichischen Untersuchung. Ernährung/Nutrition 15, 217-222.
- Pope CA , Thun MJ, Namboodiri MM, Dockery DW, Evans JS, Speizer FE, Heath CW Jr (1995) Particulate air pollution as a predictor of mortality in a prospective study of US adults. Am J Respirat Crit Care Med 151, 669-674
- Redmond CK (1976) Epidemiologic studies of cancer mortality in coke plant workers. In: Seventh Conference on Environmental Toxicology. Washington DC, US Environmental Protection Agency 1976, AMRL-TR-76-125, pp 93-107
- Redmond CK (1983) Cancer mortality among coke oven workers. Environ Health Perspectives 52, 67-73
- Turrio-Baldassarri L, Di Domenico A, La rocca C, Iacovella N, Rodriguez F (1996) Polycyclic aromatic hydrocarbons in Italian national and regional diets. Polycyclic Aromatic Compounds,10, 343-349.
- Vaessen HAMG, Jekel AA, and Wilbers AAMM (1988) Dietary intake of polycyclic aromatic hydrocarbons. Toxicol Environm Chem 16, 281-294.
- VCP (1990) WVC en LNV; Nederlandse Ministeries, VCP: Wat eet Nederland. Resultaten van de voedselconsumptiepeiling 1987-1988. Rijswijk.