Inconsistency between workplace and spousal studies of environmental tobacco smoke and lung cancer.

In a risk assessment released at the end of 1992, the U.S. Environmental Protection Agency (EPA) concluded that environmental tobacco smoke (ETS) is a known human lung carcinogen. The Agency reached that conclusion primarily on the basis of epidemiologic studies of self-reported never-smoking women, in which the exposure index was marriage to a smoker. However, the use of the spousal smoking exposure surrogate introduces many potential confounding factors. Such confounding and bias due to denial of active smoking are likely explanations for weak and inconsistent reported ETS-lung cancer associations. This contention is supported by the results of 14 worldwide studies of lung cancer and ETS exposure in the workplace, which in combination indicated no risk elevation. Workplace ETS-lung cancer studies are not subject to the bias and confounding introduced by the spousal smoking exposure surrogate. The EPA ignored the workplace studies in its risk assessment and extrapolated the results of spousal smoking studies to workplace and other sources of ETS exposure. In its estimate of ETS-attributable lung cancer deaths in the United States, the EPA ascribed over 70% of the deaths to nonspousal ETS exposure, primarily workplace exposure. Considered in their entirety, the ETS-lung cancer epidemiologic data do not support a causal inference or provide a scientific basis for government regulation of smoking in the workplace.