To choose a sensitive protocol to discriminate populations exposed and not exposed to inducers, five urinary metabolite ratios (MRs) [MR1 (17X + 17U)/137X, MR2 (5-acetylamino-6-formylamino-3-methyluracil [AFMU] + 1X + 1U)/17U, MR3 (17X/137X), MR4 (AFMU + 1X + 1U + 17X + 17U)/137X, and MR5 (AFMU + 1X + 1U)/17X] were calculated in 4-5 h and 0-24 h urine samples after caffeine intake. One hundred twenty-five healthy volunteers (59 nonsmokers and 66 smokers) were included in the study. All ratios showed a log-normal distribution. MR2 in the two time intervals was the only ratio nondependent on the urine flow. Differences between nonsmokers and smokers could be detected with all ratios at 4-5 h. However, only MR2 and, to a lesser extent, MR5 allowed the discrimination of higher cytochrome P450 1A2 (CYP1A2) activity in smokers in the 0-24 h sample. Although smokers had increased urinary mutagenicity in relation to nonsmokers, a significant association between MRs and urine mutagenicity was observed only with MR2 in the 4-5 h interval; this ratio/time schedule being that of higher association with tobacco consumption. The most flow-dependent ratios, MR1, MR3, and MR4, were closely correlated with each other at the two intervals. The flow dependency profile of each ratio may explain their different power to indicate both tobacco exposure and tobacco-derived mutagenicity. In conclusion, MR2 in the period of 4-5 h after caffeine intake seems preferable, especially at high urine flow rates.