Lung cancer and diesel exhaust: a critical review of the occupational epidemiology literature

Abstract

The diesel exhaust (DE)-lung cancer hypothesis is evaluated. Diesel power became common after World War II, exposure was to traditional diesel exhaust (TDE) before 1988. In the next, 20 years, emissions were modified to new-technology diesel exhaust (NTDE) containing 1% of pre-1988 levels of diesel particulate matter (DPM). Nearly all pre-1990 studies were cohorts with primarily pre-diesel exposures. This review focuses on the proportion of cases with >20 years since initial DE exposure; strength of association; biological gradients; roles of chance, bias, and confounding; and consistency in 13 diesel studies. Five studies had adequate latency, six had a minority of workers with >20 years' latency, and in two studies most workers had inadequate latency. This pattern suggests too few relevant studies for evaluating the DE-lung cancer hypothesis. The 16 highest exposure categories showed 7 with probable associations (relative risk [RR] > 1.5), 7 with improbable or no associations (RRs < 1.2), and 2 with possible associations (RRs 1.2-1.5). This random pattern with many weak RRs does not support the DE-lung cancer hypothesis. Ten of 34 exposure-response (E-R) analyses showed positive trends and 24 had indeterminate or negative trends. This small number of positive biological gradients does not support causality. Weight of evidence suggests 70% of studies are indeterminate, whereas 30% are positive or negative, indicating a lack of consistency. To support a traditional diesel exhaust-lung cancer hypothesis requires more studies with longer follow-up and quantitative E-R analyses.