It has recently been shown that the most common mutation (named m1) in both Caucasian and Japanese poor metabolizers (PM) of S-mephenytoin is a single base pair mutation (G-->A) in exon 5 of the CYP2C19 gene. In Japanese, a second defective allele of CYP2C19 named m2 consists of a G-->A mutation in exon 4. In the present study, we have investigated the inheritance of the CYP2C19 wild type allele (wt) and the two defective alleles (m1 and m2) in families of 11 Danish PM probands. The study was carried out for two principal reasons. First, we wanted to confirm the autosomal recessive inheritance of the defective alleles, and second, we wanted to examine the specificity and sensitivity of the CYP2C19 genotyping test. Individuals were phenotyped by measuring the ratio of S/R mephenytoin excreted in the urine after administration of mephenytoin, and genotyping was carried out by a PCR-based DNA amplification procedure. The genotypes of nine of the 11 probands were consistent with their phenotypes. Eight were homozygous m1/m1, and one was heterozygous m1/m2. The genotypes of two putative PM probands (wt/m1) were not consistent with their phenotypes. On the basis of extended phenotyping (additional late urine collections (24-36 h) and acidification of urine), one of these could probably be reclassified as an extensive metabolizer (EM) while the other was considered to be a true PM. This suggests the presence of an additional unknown mutant allele in the latter. Seven of the 41 phenotyped relatives in the 11 families were phenotyped as PMs, and with the exception of the father of family 10, their genotypes (m1/m1) were consistent with their phenotypes. Extended phenotyping (acidification of urine) suggested that the father of family 10 in fact is an EM and hence that his genotype (wt/m1) is concordant with his phenotype. Thus, the specificity of genotyping tests for PM was 100%, while the sensitivity was 15/16 or 94%. Our study provides unequivocal evidence for autosomal recessive inheritance of the PM trait.