The wide variations in urinary bladder and colo-rectal cancer incidence in humans have been attributed in part to metabolic factors associated with exposure to carcinogenic aromatic and heterocyclic amines. Cytochrome P-4501A2 (CYP1A2), which catalyses N-oxidation, and acetyltransferase (NAT2) which catalyses N- and O-acetylation, both appear to be polymorphically distributed in human populations; and slow and rapid NAT2 phenotypes have been implicated as risk factors for these cancers. Caffeine has also been shown to undergo 3-demethylation by CYP1A2, and it is further acetylated to 5-acetylamino-6-formylamino-3-methyluracil (AFMU) by the polymorphic NAT2. In this report, we describe a metabolic phenotyping procedure that can be used to determine concomitantly the hepatic CYP1A2 and NAT2 phenotypes. For the NAT2 phenotype, we confirm the valid use of the urinary molar ratio of AFMU/1-methylxanthine, even in alkaline urines. For the CYP1A2 phenotype, the urinary molar ratio of [1,7-dimethylxanthine + 1,7-dimethyluric acid]/caffeine, taken at 4-5 h after caffeine ingestion, was identified from pharmacokinetic analyses of 12 subjects as being better correlated (r = 0.73; p = 0.007) with the rate constant for caffeine 3-demethylation than other previously suggested ratios. This procedure was then used to determine the CYP1A2 phenotype in subjects from Arkansas (n = 101), Italy (n = 95), and China (n = 78). Statistical and probit analyses of nonsmokers indicated that the CYP1A2 activity was not normally distributed and appeared trimodal. This trimodality allowed arbitrary designation of slow, intermediate, and rapid phenotypes, which ranged from 12-13% slow, 51-67% intermediate, and 20-37% rapid, in the different populations. A reproducibility study of 13 subjects over a 5 day or 5 week period showed that, with one exception, intraindividual variability did not alter this CYP1A2 phenotypic classification. Induction of CYP1A2 by cigarette smoking was also confirmed by the increased caffeine metabolite ratios observed in the Arkansas and Italian smokers (blonde tobacco). However, Italian smokers of black tobacco and Chinese smokers did not appear to be induced. Furthermore, probit analyses of Arkansas and Italian blonde tobacco smokers could not discriminate between phenotypes, apparently as a consequence of enzyme induction.