Polymorphism for CYP2D6 was determined genetically as part of a hospital-based case-control study. The cases were males with a histologically confirmed lung cancer diagnosis, < 75 years old, and no previous cancer diagnosis. Male controls were matched for age, hospital and residence area. This study includes 301 cases and 310 controls. A DNA bank was established for 547 patients (89.5%), and genotypes for CYP2D6 were differentiated by the Heim and Meyer method for the DNA samples of 249 cases and 271 controls. Among the cases, the frequencies of homozygous for the wild-type (EM), heterozygous (HEM) and homozygous for the mutant alleles (PM) were 62%, 32% and 7%; among the controls: 57%, 37% and 6%. Using EM as the reference, and adjusting for age, hospital and residence, we estimated the odds ratios for the HEM group and the PM group at 0.8 (95% CI [0.5-1.2]) and 1.1 (95% CI [0.5-2.4]) respectively. The PM frequency among the cases of adenocarcinoma was twice as high as among the controls (OR = 1.8, 95% CI [0.7-4.9]). This result was not observed among squamous and small cell carcinoma (OR = 0.7, 95% CI [0.3-1.8]). Twelve different case-control studies on CYP2D6 and lung cancer have so far been performed; the ORs they estimate range from 0.1 to 2.0, with a median value of approximately 0.6. This result lends some support to the hypothesis that belonging to the PM group is associated with a slight protective effect against lung cancer, but does not take into account the possibility that results may vary according to histologic type. In this context, the suggestion of a positive relationship between CYP2D6 and adenocarcinoma seems to us to merit investigation.