Cytochrome P450 system is involved in the biotransformation of >75% of the approved drugs, where about 20% of them are metabolized by CYP2D6 enzyme. The CYP2D6 gene is extremely variable, so far >170 variants have been detected, including single nucleotide polymorphisms (SNPs) and structural variants (SVs). Structural variants include deletions, duplications and multiplications and also hybrid genes CYP2D6::CYP2D7 and CYP2D7::CYP2D6. Different variants of the CYP2D6 gene encode CYP2D6 enzymes with different activities, so depending on the genotype, patients are classified into phenotypic groups of normal, intermediate, poor and ultrarapid metabolizers. Correctly determined genotype and phenotype are crucial for the dosing of drugs whose effects primarily depend on metabolism via the CYP2D6 enzyme, especially for narrow therapeutic index drugs. Patients with different phenotypes require different doses of such drugs in order to avoid on the one hand underdosing and absence of the desired effect, and on the other hand, overdosing which leads to side effects and toxicity. Due to the variability of CYP2D6 gene, genotyping is complex. The detection of hybrid genes is particulary challenging. Extra-long Polymerase Chain Reaction (XL-PCR) and TaqMan® real-time PCR method successfully detect SNPs, but do not have the possibility of detecting hybrid genes, instead they are falsely detected as whole genes or as deletions. CNV analysis, that detects 2 or more CYP2D6 gene fragments, revolutionize the discovery of hybrid genes. This paper describes the procedure of pharmacogenomic analysis of the CYP2D6 gene, which includes the analysis of single nucleotide polymorphisms and CNV analysis, with an emphasis on the analysis of structural variants and the difficulties in genotyping hybrid genes. An overview of the frequencies of different allels, genotypes and fenotypes in the population of patients who underwent pharmacogenomic analysis in the Division for Pharmacogenomics and Therapy Individualization of the Department for Laboratory Diagnostics University Hospital Center Zagreb is presented. Also the results of genotyping and phenotyping obtained with and without CNV analysis are compared in order to gain insight into the significance of CNV analysis. The weaknesses of CNV analysis are also described. Its complexity stems from the fact that the hybrid genes themselves come in a large number of diverse variants and are often involved in non-identical duplications. Ultimately, it can be said that CNV analysis increases the quality of pharmacogenomic analysis of the CYP2D6 gene, however the method is not perfect. Therefore, the questions of optimizing CNV analysis and the possibility of introducing other molecular methods for genotyping the CYP2D6 gene into clinical practice remain open.