Crystalline silica and lung cancer: a critical review of the occupational epidemiology literature of exposure-response studies testing this hypothesis.

IARC (2009; Metals, Particles and Fibres. IARC Monographs on the Evaluaton of Carcinogenic Risks to Humans. Volume 100C. Lyon, France: IARC) concluded that crystalline silica in occupational settings is a lung carcinogen. This conclusion is based primarily on studies with exposure-response (E-R) analyses and a pooled analysis of 10 major studies with about 1000 lung cancer cases. The purpose of this review is to critically assess this cancer classification based on E-R analyses in 18 studies from eight countries with about 2000 lung cancer cases and the same database used by IARC (2009) . The most appropriate exposure-response analysis is selected from latest study with least effect from bias, confounding, and presented graphically to assist individual assessment of the weight of evidence. Strength of association is consistently weak in the majority of studies. At the highest exposure level the mean relative risk (RR) is 1.5; four studies have strong associations (RRs > 2), three have moderate strong associations (RRs 1.5-2.0), six have weak-negligible associations (RRs 1-1.5), and five have no associations (RRs ≤1.0). Biological gradients were an inconsistent finding. Three studies had clear positive E-R trends; 3 had suggestive trends; and 12 had no E-R trends, 9 of which were flat or negative. There was a negative ER slope using RRs at the highest exposure of each study. Consistent findings of weak associations and lack of E-R trends does not support a causal association. Weight of evidence from occupational epidemiology does not support a causal association of lung cancer and silica exposure, which is contrary to the IARC conclusion using essentially the same data.