Challenges of CYP21A2 genotyping in children with 21-hydroxylase deficiency: determination of genotype–phenotype correlation using next generation sequencing in Southeastern Anatolia

Background/purpose Although it is known that there is generally a good correlation between genotypes and phenotypes, the number of studies reporting discrepancies has recently increased, exclusively between milder genotypes and their phenotypes due to the complex nature of the CYP21A2 gene and methodological pitfalls. This study aimed to assess CYP21A2 genotyping in children with 21-hydroxylase deficiency (21-OHD) and establish their predictive genotype–phenotype correlation features using a large cohort in Southeastern Anatolia’s ethnically diverse population. Methods The patients were classified into three groups: salt-wasting (SW), simple virilizing (SV) and non-classical (NC). The genotypes were categorized into six groups due to residual enzyme activity: null–A–B–C–D–E. CYP21A2 genotyping was performed by sequence-specific primer and sequenced with next generation sequencing (NGS), and the expected phenotypes were compared to the observed phenotypes. Results A total of 118 unrelated children with 21-OHD were included in this study (61% SW, 24.5% SV and 14.5% NC). The pathogenic variants were found in 79.5% of 171 mutated alleles (60.2%, 22.2%, and 17.6% in SW, SV and NC, respectively). Patient distribution based on genotype groups was as follows: null—16.1%, A—41.4%, B—6.0%, C—14.4%, E—22%). In2G was the most common pathogenic variant (33.9% of all alleles) and the most common variant in the three phenotype groups (SW—38.8%, SV—22.2% and NC—23.3%). The total genotype–phenotype correlation was 81.5%. The correlations of the null and A groups were 100% and 76.1%, respectively, while it was lower in group B and poor in group C (71.4% and 23.5%, respectively). Conclusion This study revealed that the concordance rates of the severe genotypes with their phenotypes were good, while those of the milder genotypes were poor. The discrepancies could have resulted from the complex characteristics of 21-OHD genotyping and the limitations of using NGS alone without integrating with other comprehensive methods.