This study investigated the role of genetic factors (CYP1A2) in caffeine metabolism. The CYP1A2 activity was determined in 378 Danish twins following oral intake of a single dose of 200 mg caffeine and subsequent determination of the caffeine ratio (AFMU+1MU+1MX)/17DMU in a 6-h urine sample. The mean (+/- SD) caffeine ratio was 5.9 +/- 3.4. The caffeine ratio was statistically significantly higher in men compared to women, in smoking men and women compared to non-smoking persons of the same gender and in women not taking oral contraceptives compared with women on oral contraceptives. Thus, we confirmed that CYP1A2 is more active in men than in women, that it is induced by smoking and inhibited by oral contraceptives. In the subsequent analysis of heritability, we included 49 monozygotic twin pairs and 34 same gender dizygotic twin pairs concordant for non-smoking and non-use of oral contraceptives. The intraclass correlation coefficient was 0.798 (95% confidence interval, 0.696-0.900) and 0.394 (95% confidence interval, 0.109-0.680) in the monozygotic and dizygotic twins, respectively. The correlation was statistically significantly higher (P = 0.0015) in the former compared with the latter. A biometrical model for the caffeine ratio including only additive genetic factors and unique environmental factors was the overall best fitting model. Estimates based on this model gave a heritability estimate of 0.725 (95% confidence interval 0.577-0.822). Unique environmental effects seem to account for the remainder 0.275 (95% confidence interval, 0.178-0.423). Our study shows that the CYP1A2 activity is mainly governed by genetic factors.