A nested lung cancer case-control study was carried out using 397 12 h
urine samples originally collected from a cohort of over 26 000 women
aged 40-64 at entry who were then followed for up to 15 years. The ur
ine samples from active smokers were first identified using a simple q
ualitative method and their total nicotine metabolites/creatinine rati
os then determined by automated colorimetric methods. The results obta
ined demonstrated the capacity of nicotine metabolite estimations in a
single 12 h sample of urine to predict the subsequent risk of lung ca
ncer. The risk of lung cancer among the biochemically proven active sm
okers during this period was 7.8 times that of the non-smokers, sugges
ting that the dose-response relationship between smoking and lung canc
er is no less steep in women than in men. The smoking-related risk of
adenocarcinoma was less than that of other lung carcinomas. It is sugg
ested that this biochemical epidemiology approach to exploring the rel
ationship between smoking and lung cancer could profitably be applied
to the study of other smoking-related diseases.