Abstract Cytochrome P450 2A6 (CYP2A6) is involved in the 7-hydroxylation of coumarin, C-oxidation of nicotine, and the metabolism of tobacco specific nitrosamines. Initially in 1995 Fernandez-Salguero et al. reported a genotyping method for three alleles: CYP2A6*1 (wild-type), CYP2A6*2 (variant 1), and CYP2A6*3 (variant 2). Later studies presented in this thesis indicated that the original genotyping method produces erroneous results for the CYP2A6*3 allele due to unspecific PCR conditions and previously unknown CYP2A6*1B allele. Furthermore, the CYP2A6*2 allele genotyping caused erroneous genotypes (CYP2A6*2/*2 was misclassified as CYP2A6*1/*2). In this work, new PCR based genotyping methods were developed for CYP2A6*2 and for several new alleles (CYP2A6*1B, CYP2A6*4A/*4D and CYP2A6*5). In population-based studies, the deletion alleles (pooled as CYP2A6*4) turned out to be more prevalent among Asians (15.1%) than Caucasians (0.5%). The frequencies of the other inactive alleles varied within 0–3% in both populations. Asians totally lacked the CYP2A6*2 allele, whereas Caucasians lacked the CYP2A6*5 allele. The frequencies of two wild-type alleles, CYP2A6*1A and CYP2A6*1B alleles were 66.5% and 30.0% in Caucasians, and 43.2% and 40.6% in Asians, respectively. Correlation studies between the phenotype, as tested by the administration of coumarin, and the genotype demonstrated that individuals with the CYP2A6*2/*2 genotype were totally defective, while CYP2A6*1/*2 subjects exhibited intermediate and CYP2A6*1/*1 subjects full capablility of producing 7-hydroxycoumarin. Upon phenotyping with nicotine, individuals with the CYP2A6*1/*2 or CYP2A6*1/*4 genotype were shown to have a lower enzyme activity (one fourth of the normal activity), compared to those with the CYP2A6*1/*1 genotype. Defective CYP2A6 activity has been hypothesised to reduce the risk of environmentally (especially tobacco smoke) induced diseases either by decreasing production of genotoxic metabolites or by preventing addiction to tobacco smoking. However, in our case-control studies on Spanish patients with liver cirrhosis (n = 83) and liver cancer (n = 90) and their controls (n = 237) no significant association between the CYP2A6 genotypes and disease proneness was found. The odds ratio (OR) for developing liver cancer was was 1.4 (95% confidence interval [CI] 0.5–3.7) for genotypes containing at least one CYP2A6*2 allele. For liver cancer the respective OR was 1.3 (95% CI 0.4–4.5). Similarly, no statistically association between CYP2A6 alleles and the risk of lung cancer was observed in our Finnish study population cinsisting of 177 cases and 1089 controls; the OR for combined CYP2A6 variant allele containing genotypes (CYP2A6*1/*2 and CYP2A6*1/*4) was 1.19 (95% CI 0.56–2.45). Our studies therefore do not indicate any major modifying role for the CYP2A6 genotypes in individual susceptibility to environmentally induced diseases.