Environmental tobacco smoke and lung cancer: a reappraisal.

It is biologically plausible that environmental tobacco smoke (ETS) has a contributory role in the induction of lung cancer in nonsmoking individuals. However, recent findings have strengthened previous assumptions that a major part of the observed increase in lung cancer risk reported from epidemiological studies on ETS-exposed nonsmokers can be related to misclassification of smoking status and inappropriate selection of controls as well as to certain confounding factors related to life-style, and possibly also to hereditary disposition. Dose-response extrapolation-supported by a more solid database for active smokers-reflects a possible increase in lung cancer that appears to be more than one order of magnitude lower than indicated by the epidemiological studies that have been used to support regulatory action in the United States. The epidemiological studies on ETS conducted so far lack the required sensitivity to confirm increases in risk of such low magnitudes. Self-reported information on exposure to tobacco smoke has been found to be unreliable, and data from interviews with proxy respondents even more so. In addition, determination of cotinine to establish smoking status is inadequate for use in this context, i.e., to assure that misclassification in the range 5-10% does not occur; due to genetically based differences in the rate of nicotine metabolism, some active smokers will not be detected. Further, due the short half-life of cotinine in the organism, a self-reported nonsmoker may, in principle, have been a lifelong heavy smoker until just before the sampling takes place. For some of the major studies, preferential inclusion of disease-prone individuals of very low socioeconomic status among cases seems to have been present to a varying extent. Due to inclusion of this group, life-style and hereditary disposition may result in a disproportionally large impact on the recorded overall lung cancer rate. Further, a possible major confounder causing inflated risk estimates, and that was not controlled for, is apparently a high intake of saturated fat, which to a varying extent is coupled to inadequate intake of anticarcinogens present in certain foods of plant origin. The one-sided preoccupation with ETS as a causative factor of lung cancer in nonsmokers may seriously hinder the elucidation of the multifactorial etiology of these tumors.