Reassessment of the Access Testosterone chemiluminescence assay and comparison with LC‐MS method

Aim of the study To reassess the imprecision and Limit of Quantitation, to evaluate the cross‐reaction with dehydroepiandrosterone‐sulfate (DHEAS), the accuracy toward liquid chromatography‐mass spectrometry (LC‐MS) and the reference interval of the Access Testosterone method, performed by DxI immunoassay platform (Beckman Coulter). Material and Methods Imprecision was evaluated testing six pool samples assayed in 20 different run using two reagents lots. The cross‐reaction with DHEAS was studied both by a displacement curve and by spiking DHEAS standard in two serum samples with known amount of testosterone. The comparison with LC‐MS was evaluated by Passing‐Bablock analysis in 21 routine serum samples and 19 control samples from an External Quality Assurance (EQA) scheme. The reference interval was verified by an indirect estimation on 2445 male and 2838 female outpatients. Results The imprecision study showed a coefficient of variation (CV) between 2.7% and 34.7% for serum pools from 16.3 and 0.27 nmol/L. The value of Limit of Quantitation at 20% CV was 0.53 nmol/L. The DHEAS showed a cross‐reaction of 0.0074%. A comparison with LC‐MS showed a trend toward a slight underestimation of immunoassay vs LC‐MS (Passing‐Bablock equations: DxI=−0.24+0.906 LCMS in serum samples and DxI=−0.299+0.981 LCMS in EQA samples). The verification of reference interval showed a 2.5th‐97.5th percentile distribution of 6.6‐24.3 nmol/L for male over 14 years and