Epidemiology of fibrous glass and lung cancer

In a recent commentary, Infante et al. [1994] summarized some of the epidemiologic literature on the relationship between glass wool and lung cancer. They concluded that this material is carcinogenic for the lung in humans. A review of the commentary in comparison with the original papers covered reveals that their analysis was incomplete and characterized by selection bias. A more careful examination of the literature cited and other relevant publications shows that the data cited by Infante et al. are subject to the errors of chance and confounding and that the association when present is inconsistent, weak, and lacks a dose-response relationship. Therefore, a conclusion of causality cannot be justified.