Justification of an Air Quality Standard for Polycyclic Aromatic Hydrocarbons

35. The Panel accept that several PAH compounds found in ambient air are genotoxic carcinogens. It is widely held that such compounds cannot be characterised by a threshold, and that therefore no absolutely safe exposure level can be defined. Epidemiological studies have shown that long-term exposure to mixtures of PAH compounds is associated with an increased risk of lung cancer. Inhalation of PAHs has also been linked with other types of cancer, but evidence on this is less consistent. Thus, we base our recommendation for a standard on the need to protect people from developing lung cancer.

36. Nobody is exposed occupationally or environmentally to single PAH compounds; exposure is always to a mixture. This has been recognised by research workers, who often have used benzo[a]pyrene concentrations as a marker of total exposure and have reported risks from exposure to mixtures in terms of the concentration of B[a]P. Current evidence suggests that for the mixtures of PAHs which occur in ambient air, the relative contribution of B[a]P to total carcinogenic potency is similar to that found in the occupational environments which have been studied epidemiologically. We therefore have concluded that a standard should be defined in terms of B[a]P concentrations.

37. The epidemiological studies that tell us most about the risks of lung cancer from PAHs have focused on aluminium smelter and coke oven workers. Risk estimates derived from these investigations are broadly compatible, but we believe the Canadian study of aluminium smelters provides the best starting point for the derivation of a standard (Armstrong et al., 1994). We note that cumulative exposure to 10-99 µg/m³.year of a mixture of PAH compounds represented by B[a]P was associated with an approximately 50% increase in the risk of lung cancer. We regard this as the lowest level at which effects have been observed, equivalent to exposure to 0.25-2.5 µg/m³ of B[a]P (acting as a marker) for 40 years (a working life time).

38. Recognising that this level is represented by a range of concentrations we have adopted a cautious approach and chosen the lower end of the range as our starting point. We have then used a modification of the method previously employed for recommending standards for other genotoxic carcinogens, benzene and 1,3-butadiene, applying 3 safety factors:-

a. For both benzene and 1,3-butadiene we were able to identify, from epidemiological studies in workers, a level of exposure at which no effects had been found whilst accepting that larger studies might have revealed an effect. This is not the case with PAH compounds and so we have applied a safety factor of 10 to the lower end of the range of the concentrations at which effects were observed. In applying this safety factor of 10 for moving from a lowest observed adverse effect level (LOAEL) to a no observed adverse effect level (NOAEL), the Panel has paid attention to guidance from the International Programme on Chemical Safety on deriving health-based exposure limits (IPCS, 1994). The guidance reported a range of factors (3, 5 or 10) that had been used in extrapolating from a LOAEL to a NOAEL and noted that the choice of factor depended on the nature of the effects and the dose-response relationship. PAH compounds are genotoxic carcinogens (see paragraph 15) for which no completely safe level of exposure can be identified and thus we feel justified in using the highest safety factor of 10.

b. In extrapolating from a working life (40 years, 5 days per week, 8 hours per day) to an entire life a factor of 10 was applied.

c. Similarly a factor of 10 was applied to take account of the range of sensitivity to carcinogens likely to exist in the general population, for example, due to young age, genetic differences, etc..

39. In summary, a composite factor of 1000 was applied and the starting figure of 0.25µg/m³ (250 ng/m³) was divided by this to derive a standard of 0.25 ng/m³ for B[a]P. As in the case of other genotoxic carcinogens (see paragraph 15) we believe that long-term exposure is more important than short duration exposure and thus recommend that the averaging time for the standard should be 1 year.

40. As mentioned in paragraphs 6 and 10, B[a]P in ambient air occurs in both particulate and gas phases, and these may be measured separately. Though the study of aluminium smelters from which the standard has been derived found that all the PAH compounds were bound to particles, we feel it would be unwise to ignore the potential contribution of the gas phase and thus recommend that the standard should apply to the total PAH concentration.

41. In coming to this recommendation for a standard, we recognise that there are more uncertainties in the underlying science than for most of the pollutants we have considered in earlier reports. In particular, the starting point for our calculation - i.e. the lower end of the lowest range of exposures that was associated with excess lung cancer in the Canadian study of aluminium smelters - might have been different had the scientists who carried out that investigation chosen different exposure cut-points for their analysis. This arbitrary aspect of the calculation is taken into account in the application of a 10-fold safety factor, which in our judgement should ensure that any risks from exposures below the recommended standard are minimal. However, it does not necessarily follow that all exposures above the standard carry a significant risk.