Aim: This study was performed in a sample of the Dutch population to estimate the prevalence of noncoding mutations of CYP2D6 and CYP2C19 as obtained by genotyping. In addition, the predictability of the genotyping strategy was assessed. Methods: The CYP2D6 and CYP2C19 status of 765 unrelated healthy volunteers was evaluated. Dextromethorphan and mephenytoin were used for determining the phenotypes. Genotyping was performed by PCR on the most common null alleles for CYP2D6 (except for CYP2D6*5) and CYP2C19. Results: For CYP2D6, the most frequently observed null allele was CYP2D6*4, which accounted for 89% of all null alleles. The prevalence of poor metabolizers (PMs) in healthy volunteers was 5.5%, which was lower than that found previously by phenotyping (8.0%; χ2 test P=0.009). For CYP2C19*2 and CYP2C19*3, the frequencies were 13.3% and 0.2%, respectively. The S:R ratio was significantly higher in heterozygous subjects (S:R ratio 0.22) than in homozygous wild type subjects (S:R ratio 0.11). Comparison of all subjects below 45 years showed a significantly higher S:R ratio in the female ones compared to the male ones, especially in heterozygous subjects (S:R ratio 0.39 vs. 0.19; P<0.001). Conclusions: The frequencies of CYP2D6 and CYP2C19 allelic variants were in accordance with other European populations. Assessment of *3, *4, *6, *7, and *8 alleles for CYP2D6, and *2 and *3 for CYP2C19, predicted the phenotype with an accuracy of over 98.6%. A gene–dose effect was found for CYP2C19. CYP2C19 heterozygous female subjects had a decreased CYP2C19 activity that may be at least partly due to the use of oral contraceptives.