In the previous issue of this journal Hellström & Vitols (2004) reviewed and compared the properties of the proton pump inhibitors. They and Schaffalitzky de Muchadell (2004) agreed that the proton pump inhibitors are comparable with respect to inhibition of gastric acid secretion. The choice of proton pump inhibitors will primarily depend on the price of the recommended dose.Two years after the first proton pump inhibitor, omeprazole was registered, we demonstrated that this drug is metabolized by the polymorphic cytochrome P450 enzyme CYP2C19 (Andersson et al. 1990). The other proton pump inhibitors have subsequently been shown also to be metabolized by CYP2C19. In a MiniReview in this journal, Klotz et al. (2004) now discuss the importance of this polymorphic enzyme for the dosing of the 5 proton pump inhibitors prescribed today. They nicely review the pharmacodynamic effects of the proton pump inhibitors in the 3 genotypes of CYP2C19. In several studies, especially from Japan (Furuta et al. 2002 and other references in the review by Klotz et al., 2004), a higher cure rate in patients with zero (poor metabolizers) or one (heterozygous extensive metabolizers) compared to extensive metabolizers with two functional CYP2C19 genes. The latter genotype group needs a higher dose of the proton pump inhibitor to compensate for their more rapid drug metabolism. I also agree with Klotz et al (2004) claiming ''initial genotyping for this enzyme and higher dosage in extensive metabolizers is likely to improve the clinical efficacy of proton pump inhibitors''.Initially the proton pump inhibitors were used for short periods in time, a few weeks, but today they are used for long-term treatment especially to treat gastro-oesophageal reflux disease. We found in 72 patients treated for more than one year with 20 mg omeprazole per day, that heterozygous extensive metabolizers had 3 times higher serum gastrin than extensive metabolizers with two functional CYP2C19 alleles (PΩ 0.0001) (Sagar et al. 2000). Also concentrations of pepsinogen I and chromogranin A were significantly different in the two genotypic groups. In another long-term omeprazole study, we demonstrated that serum vitamin B 12 levels decreased in heterozygous extensive metabolizers, but not in extensive metabolizers with two functional CYP2C19 genes (Sagar et al. 1999). These results further strengthen the utility of determination of the CYP2C19 genotype before treatment (especially long-term) with proton pump inhibitors to adjust the dose for the individual patient. This will aid to maintain a sufficient acid inhibition with a good response and to minimize the risk to develop enterochromaffin-like cell hyperplasia and atrophic gastritis.