Smoking is the principal cause of lung cancer. However, not all smokers will develop this disease. Individual susceptibility to chemically induced cancer may be explained in part by genetic differences in the activation and detoxification of procarcinogens. The activation phase of polycyclic aromatic hydrocarbon (PAH) metabolism is governed by the enzyme CYP1A1, induced by PAH when it enters the body. The extent to which PAH induces CYP1A1 activity varies greatly from one subject to another. CYP1A1 inducibility has long been associated, although inconsistently, with an increased risk of lung cancer. In 1982, Kouri corroborated Kellerman's results with a new method for measuring inducibility, but few studies have reported using this method. The glutathione S-transferases (GSTs) are involved in the detoxification phase of PAH, and the allelic deletion of GSTM1 has been also associated with an increased risk of lung cancer. We conducted a case-control study to examine the risk of lung cancer related, separately and together, to CYP1A1 inducibility, GSTM1 polymorphism and cigarette smoking in a French population. The 611 subjects were 310 incident lung cancer cases and 301 hospital control subjects. We were able to constitute a DNA bank for 552 subjects (89.5%) and gather detailed information on smoking history for all of them. Inducibility could be measured for 195 cases and 183 control subjects. Results for GSTM1 polymorphism concern 247 cases and 254 control subjects. GSTM1 polymorphism and inducibility could both be assessed for 179 cases and 166 control subjects. The odds ratio related to inducibility was 1.7 [1.0-3.0] for medium and 3.1 (1.3-7.4) for hyper inducers. The association with GSTM1 was 1.6 (1.0-2.6). With a reference category of subjects who were both low inducers and GSTM1(+), we found an odds ratio for lung cancer of 8.1 (2-31) for the subjects with both risk factors [i.e. GSTM1(-) and hyper inducers]. Our data did not reveal evidence of interaction between smoking and inducibility. On the other hand, we found an interaction of 3.6 (0.6-21) between inducibility and GSTM1.