How low can you go? Analytical performance of five automated testosterone immunoassays.
Level 4 - case-series / case-control
Laboratory-based analytical method comparison study using clinical specimens without prospective clinical outcome design.
PubMed 29763574 · doi:10.1016/j.clinbiochem.2018.05.008
What was done
Researchers evaluated the analytical performance of five automated total testosterone immunoassays (Abbott ARCHITECT i2000 SR, SIEMENS ADVIA Centaur, SIEMENS IMMULITE 2000, Beckman Coulter DxI 800, and Roche MODULAR E170). They assessed functional sensitivity (FS at 20% coefficient of variation [CV]), 5-day imprecision, assay interference (hemolysis, icterus, and lipemia), limit of blank, dilution linearity, and recovery using NIST Standard Reference Material 971. Accuracy was compared against an in-house liquid chromatography-tandem mass spectrometry (LC-MS/MS) method using patient samples from men, women, boys, and girls.
What was found
FS at 20% CV was 0.14 nmol/L for ARCHITECT, 0.36 nmol/L for DxI, 0.77 nmol/L for E170, 1.23 nmol/L for Centaur, and 3.49 nmol/L for IMMULITE. Five-day total imprecision CVs were ≤ 9.0% for all methods, and all met claims for interference and limit of blank. Deviations from target recoveries in dilution linearity ranged from 3.4% (ARCHITECT) to 14.3% (DxI). NIST SRM 971 recoveries ranged from 79.2% to 149.2% (DxI, male and female, respectively). Compared to LC-MS/MS, more immunoassays under-recovered in adult samples and over-recovered in pediatric samples; slopes ranged from 0.71 (IMMULITE, women) to 1.35 (DxI, boys). Combined average percent bias was higher in boys (28.0%), girls (25.7%), and women (22.8%) than in men (11.6%).
Why it matters
While automated immunoassays perform adequately at adult male testosterone levels, significant inaccuracy and variable functional sensitivity persist at lower concentrations, indicating that LC-MS/MS remains preferable for pediatric and female samples.
Limits
The abstract does not state the total number of patient samples analyzed. Accuracy was benchmarked against a single in-house LC-MS/MS assay, and clinical diagnostic misclassification rates were not evaluated.
Cited by
- supports Standard clinical hormone assays used historically have lacked the sensitivity to accurately detect normal low circulating testosterone levels in women.