Tobacco Documents Online

ENVIRONMENTAL TOBACCO SMOKE AND LUNG CANCER IN NONSMOKING WOMEN: A REANALYSIS James E. Enstrom, Ph.D., M.P.H. School of Public Health and Jonsson Comprehensive Cancer Center University of California Los Angeles, CA 90024-1772 April 25, 1996 ABSTRACT This reanalysis points out several serious problems with the largest case-control study of environmental tobacco smoke (ETS) and lung cancer, a supposedly well-designed population-based study. About 50% of the expected lung cancer cases among women who never smoked are missing based on reasonable assumptions, including past research by these same authors and independent case ascertainment. Cases are missing in all five geographic areas and all age groups. It appears that the deficit is due to the misclassification of never smokers as smokers because only about 5% of the otherwise eligible lung cancer cases were among never smokers. This percentage is far lower than that found in other concurrent studies and in an independent examination of medical records for lung cancer patients seen at a major hospital in the ascertainment area of the study. It is highly unlikely that their identification of cases is complete and accurate because this would imply that the lung cancer incidence rate among female never smokers had declined by 50% in the last decade, a finding that is contradicted by other evidence that this rate has remained stable. Furthermore, it appears the never smoker cases are nonrepresentative because there are too few small cell and squamous cell carcinoma cases and too many acinar cell cases when compared with expected histology distribution. Analysis of the authors' publications and grant application discloses further anomalies. The key finding of their five year study, that tobacco use by spouses is associated with a 30% excess risk of lung cancer, is actually a 90% excess risk during the first year and only a 10% excess risk during the last four years using the population controls. There is no excess risk over five years using the colon cancer controls. In addition, there is a very unusual pattern of excluded cases and controls throughout that could have a substantial effect on several of the results. All of these issues lead to conclusions about ETS and lung cancer that are different than those of the authors. They point to the need for independent confirmation, analysis, and interpretation of the original data collected for this study.