CYP1A2, a member of the cytochrome P450 superfamily (CYPs), is involved in the metabolic activation of several carcinogens, among them aromatic and heterocyclic amines, nitroaromatic compounds, mycotoxins and estrogens. Several drugs are also metabolized by CYP1A2. Individual differences in CYP1A2 activity may thus influence individual susceptibility to cancer risk and the therapeutic efficacy of some drugs. In humans, CYP1A2 has been detected only in the liver, where it seems to be regulated by at least two mechanisms, one controlling constitutive levels of expression and another regulating inducibility. Wide interindividual differences in CYP1A2 activity have been described. They may be due to factors such as gender, race, genetic polymorphisms, and exposure to inducers. Higher activity has been shown in men than in women. Wide variation across racial/ethnic groups has been reported. Overall, slow and intermediate CYP1A2 metabolizers represent about 50% of Caucasians, while their frequency in Japanese subjects seems to be much lower. No nucleotide differences that could explain the phenotypic variability of the CYP1A2 gene have been found in any exons, exon-intron junctions, or 5'-flanking regions of the gene. However, two genetic variants have been identified which seem to be associated with CYP1A2 inducibility only. Induction of CYP1A2 activity has been reported as a consequence of cigarette smoking, dietary factors, several drugs, chronic hepatitis, and exposure to polybrominated biphenyls and 2,3,7,8-tetrachlorodibenzo-p-dioxin. Several epidemiological studies have been conducted into the relationship between CYP1A2 activity, alone or in combination with other CYPs, and cancer risk. In the absence of a genotypic assay, only the CYP1A2 phenotype can be assessed at present. Many compounds have been tested as in vitro probes to assess CYP1A2 activity in humans. Currently, caffeine has the best potential for use in epidemiological studies: metabolites of caffeine after coffee consumption are measured as an index of CYP1A2 activity. Variable results have been obtained with caffeine-based methods, the use of some caffeine metabolite ratios having given bimodal or trimodal distributions while others have suggested normal or unimodal distributions. Although the epidemiological studies are limited because only phenotyping data are available, there is a suggestion of increased risk of colon cancer and bladder cancer in subjects with rapid CYP1A2 activity. A higher level of 4-aminobiphenyl-haemoglobin adducts has also been found in moderate smokers with rapid CYP1A2 phenotype than in subjects with slow activity.