Liquid chromatography-tandem mass spectrometry-based recalibration reduces inter-platform variability in aldosterone detection across chemiluminescence immunoassay platforms.

Abstract

Background: Accurate aldosterone (ALD) measurement is vital in managing diseases such as primary aldosteronism (PA). However, inter-platform inconsistency across chemiluminescence immunoassay (CLIA) platforms complicates clinical decisions. This study conducted the largest comparison of CLIA platforms (six CLIA platforms) and one liquid chromatography-tandem mass spectrometry (LC-MS/MS) assay for ALD detection to date, systematically evaluated inter-platform consistency and explored the role of recalibration strategy based on LC-MS/MS in improving consistency among CLIA platforms.

Methods: Fifty pooled clinical plasma samples were used to evaluate ALD levels across six different CLIA platforms and one LC-MS/MS assay. Friedman's test, Spearman correlation, Passing-Bablok regression, and Bland-Altman analysis were used to evaluate the consistency among assays. In addition, CLIA platforms were recalibrated against LC-MS/MS using regression equations with five pooled clinical plasma samples as calibration materials, and consistency was evaluated before and after recalibration.

Results: LC-MS/MS yielded significantly lower ALD levels than six CLIA platforms (median: 120.75 vs. 129.29 to 216.25 pg/mL, P < 0.05). All assays correlated strongly (R ≥ 0.955), yet regression parameters revealed most slopes deviated from 1 (0.909 to 1.444) and intercepts ranged from -31.424 to 52.272 pg/mL. Bland-Altman plots demonstrated large relative mean differences (-2.396 % to 55.876 %) between assays. The recalibration process significantly reduced relative mean differences, whereas it showed limited improvement in addressing both proportional and systematic biases.

Conclusions: These findings demonstrate that the consistency among CLIA platforms and LC-MS/MS assays is suboptimal. Actionable design strategies for developing recalibration coefficients, which reduce relative mean differences, are provided for CLIA platforms, emphasizing the necessity for standardized calibration to reduce inter-platform variability in clinical ALD measurement. The persistent proportional and systematic biases underscore the urgency for optimization in calibration strategy and detection methodology.