Risk estimates for silicosis: comparison of animal and human studies

Estimates of human risks of occupational exposures to chemicals based on the results of animal studies conventionally apply simple scaling and safety (uncertainty) factors to the no-observed-adverse-effect level (NOAEL) in animals to estimate acceptable occupational exposure limits in humans.We have conducted a risk assessment for lung fibrosis from inhaled crystalline silica that follows the traditional approach of extrapolation from animal studies, and we compare this result with observed human risks based on epidemiological studies. To our knowledge animal inhalation studies of crystalline silica, that use a sufficient range of exposures to estimate NOAEL directly, have not been performed, so we have applied bio-mathematical modelling to the available animal data to estimate the NOAEL for inflammation. The resulting estimate for the rat is 0.1mg.m-3. Conventional scaling and extrapolation methods recommended by the US EPA have then been applied to estimate a human acceptable average exposure limit. This resulted in an estimate of about 0.001mg.m-3. The risk estimates were compared with human risk estimates for fibrosis based on epidemiological data. These comprised ACGIH summary conclusions on the risk estimates provided by epidemiological studies, and the risks demonstrated by one epidemiological study with unusually detailed exposure information. The average exposure limits implied by the risk estimates from the epidemiological studies ranged from 0.01mg.m-3 to about 0.05mg.m-3, some 9 to 45 times higher than the limits derived from the animal studies. The conventional uncertainty factors applied in the animal-based risk estimates may be over-precautionary. Extension of the biomathematical model to extrapolate from animals to humans would provide a sounder basis for extrapolation than the present uncertainty factors.