Verification of analytical performance of DHEA assay on the Abbott Architect ci8200: Experience of the biochemistry laboratory Mohammed VI Oujda

Introduction: Dehydroepiandrosterone sulfate (DHEA-S) is the main adrenal androgen produced by the adrenal cortex and also acts as a neurosteroid. DHEA-S is an excellent indicator of adrenal androgen production. Although DHEA-S has low androgenic potency, it metabolizes into more potent male hormones such as testosterone and androstenedione. Objectives: This work describes the verification/validation procedure for the biological parameter dehydroepiandrosterone sulfate (DHEA-S). This procedure is one of the main and constant concerns of biologists in quality management, aiming to produce fair and reliable measurements. Based on the recommendations of the protocol from the technical guide for accreditation in human health, SH GTA of COFRAC according to ISO 15189, we performed the verification of a microparticle immunoassay method by chemiluminescence (CMIA) for the quantitative determination of dehydroepiandrosterone sulfate (DHEA-S) in human serum and plasma on the Architect system. Results and Discussion: The verification of criteria (repeatability, reproducibility) was conducted using routine samples from patients hospitalized at the university hospital and internal and external quality controls. The study of analytical performance showed the conformity of repeatability with the supplier’s coefficients of variation (CVs) and data from learned societies (RICOS). The results obtained for intra-laboratory reproducibility were satisfactory for level 1 with a CV of 7.85% and quite satisfactory for levels 2 and 3 with a CV of 4.63% and 4.26%, respectively. Conclusion: In conclusion, the automated system can be considered suitable for medical laboratories due to its analytical performance in the determination of common biochemical parameters. In addition to laboratory staff, equipment, and environmental factors, the accuracy and reliability of results obtained during an examination are also influenced by the methods and any subsequent validation or verification.