In a molecular epidemiological study of lung cancer cases (n = 81) and noncancer controls (n = 67), polycyclic aromatic hydrocarbon (PAH)-DNA adducts were evaluated in peripheral blood leukocytes from all subjects and in a smaller number of lung tissue specimens collected prior to or at surgery. Sister chromatid exchanges (SCE) in lymphocytes were also studied in a subset of cases and controls. Questionnaire, medical record, or tumor registry data provided a family history of cancer, as well as information on cigarette smoking, dietary and occupational exposure to PAHs, and other factors related to SCEs. In both cases and controls PAH-DNA adducts in leukocytes measured by an ezyme-linked immunosorbent assay were not significantly related to age, sex, ethnicity, amount of cigarette smoking, passive smoking, dietary charcoal, or caffeine consumption. Nor did family history of cancer or histological type of cancer significantly affect adduct levels. However, when subjects were stratified by smoking status (current, former, and nonsmoker), lung cancer cases who were current smokers had significantly higher levels of covalent adducts than current smoker controls. A seasonal variation was observed in PAH-DNA binding, with a peak in adduct levels during July–October. This peak corresponds to that seen in a prior study of aryl hydrocarbon hydroxylase inducibility by other investigators. The finding of significant levels of PAH-DNA adducts in former smokers and non-smokers supports an earlier observation that this marker is not smoking specific but reflects a pervasive and variable “background” exposure to PAH. These results are consistent with a genetically determined enhancement of PAH-DNA adduct formation in leukocytes of lung cancer cases which is evident in current smokers. The results in lung tissue are limited by the small number of samples. Adduct levels were not significantly increased in lung tissue of smokers compared with nonsmokers. An inverse linear correlation was seen between adduct values in lung tissue and age of the donors. SCEs were significantly related to pack years of smoking. However, there was no difference in the frequency of SCE between cases and controls; nor were SCE and DNA adducts significantly correlated in this small sample.


This work was supported by USPHS Grants 5 R01CA-35809 and 5 R01CA-39174 from the Division of Extramural Activities, National Cancer Institute; by National Cancer Institute Grant P01CA-21111; and by Grant 41387 from the American Cancer Society.

This content is only available via PDF.