SummaryBackgroundNeonatal screening for congenital adrenal hyperplasia (CAH) identifies a certain proportion of newborns with transient moderate elevation of 17‐hydroxyprogesterone (17‐OHP). These children require regular follow‐up until normalization of their 17‐OHP levels. We investigated the possibility of reducing the individuals' recall rates by using genetic methods on their original neonatal dried blood spots.Patients and MethodsWe analysed neonatal dried blood spots from 753 subjects with transiently elevated levels of 17‐OHP. The CYP21A2 gene was sequenced to detect point mutations, and the presence of CYP21A2 was further confirmed by two methods utilizing the difference between CYP21A2 and its CYP21A1P pseudogene in the sequence of exon 3 (8‐bp deletion). The accuracy of the methods was verified using samples from 70 subjects with known CYP21A2 mutations and 181 healthy children.ResultAmong the 701 successfully sequenced samples from subjects with transiently elevated 17‐OHP, 670 (95%) had no point mutations or novel variants in the CYP21A2 gene. We found no individuals carrying genotypes consistent with the diagnosis of CAH (i.e. homozygotes or compound heterozygotes for point mutations, large deletions or rearrangements). However, 21 heterozygous carriers of known point mutations that cause the classic and nonclassic forms of CAH were identified. Additionally, we detected eight heterozygous and two homozygous point variants with unknown functional significance.ConclusionAlthough CAH caused by 21‐hydroxylase deficiency could be genetically excluded with a reasonable degree of confidence in 95% of the genotyped subjects that had transiently elevated 17‐OHP, the performance of the tests was suboptimal when performed using dried blood spots and time‐consuming in comparison with the current practice of repeated measurements of 17‐OHP. The introduction of this method into clinical practice seems to be impractical at this stage.