The metabolic ratios (MRs) between debrisoquin and 4-hydroxydebrisoquin in urine after a single oral dose of 10 mg debrisoquin were bimodally distributed in 757 healthy, white Swedish volunteers. Forty-one subjects (5.4%) had an MR greater than 12.6 and were classified as slow debrisoquin hydroxylators. The MR was reproducible in urine stored at +8 degrees C for 1 week and at -20 degrees C over a period of 5 years. Collection intervals of 6 or 12 hours gave the same MR. Intraindividual repeatability of the debrisoquin phenotyping test was established in 37 subjects examined twice at least 2 weeks apart. The calculated frequency of the single allele that is believed to control deficient debrisoquin hydroxylation is similar among white Swedish people, as among other white groups examined so far; however, it is significantly different from the frequency in certain Oriental groups. Detailed comparisons of the prevalence of slow debrisoquin hydroxylation in different ethnic groups are not possible due to shortcomings in current epidemiologic techniques used (small materials, the location of the antimode distinguishing rapid and slow hydroxylators unknown, and family studies missing).