Background and objective: CYP2C19 is clinically important in Korea because of the relatively high incidence of poor metabolizers in the population. To fully understand the genetic mechanism of the CYP2C19 defect in poor metabolizers, all variants need to be studied simultaneously. The aim of this study was to investigate the usefulness of CYP2C19 haplotypes as a marker of CYP2C19 enzyme activity in Koreans. Methods: We analysed the single nucleotide polymorphisms and haplotypes of the CYP2C19 gene in 150 healthy Koreans and found three major (frequency > 0·1) haplotypes (H1, H2 and H3). One oral dose of 40 mg omeprazole (Losec®) was administered to 30 subjects grouped as H1/H1, H2/H2, H1/H2, H1/H3 and H2/H3. The pharmacokinetics of omeprazole and its metabolites, 5-hydroxyomeprazole and omeprazole sulphone, in those groups was analysed. Results and discussion: The area under the plasma concentration-time curve (AUC0→∞) and elimination half-life (T1/2) of omeprazole were significantly greater in the H2/H2 and H2/H3 groups than in the H1/H1 group (P < 0·05), whereas the metabolic ratios of omeprazole to 5-hydroxyomeprazole were also markedly higher. Conclusion: Although a specific SNP of CYP2C19 may be predictive of enzyme activity, haplotyping is more reliable for identifying poor metabolizers in populations with variant alleles other than CYP2C19*2 and *3 alleles.