Background Hormone measurements using automated immunoassays (IAs) are known to be affected by factors present in serum. Differences in concentration of these factors or composition of the serum, as can occur in some disease states, are also described as alterations in the serum matrix. This phenomenon can affect clinical decision making. In turn, liquid chromatography tandem-mass spectrometry (LC-MS/MS) is in principle not influenced by these matrix effects. In clinical laboratories, free thyroxine (fT4), cortisol and testosterone are often measured using IAs. Therefore, the goal of this study is to unravel accuracy of fT4, cortisol and testosterone measurements in samples of hemodialysis patients, known to have an unusual matrix.

Methods 30 serum samples from both hemodialysis (HD) patients and healthy controls (HC) were collected to measure fT4, cortisol and testosterone using a well standardized  isotope dilution (ID)-LC-MS/MS method and using 5 commercially available automated IAs (Alinity (Abbott), Atellica (Siemens), Unicel DXI (Beckmann Coulter), Cobas (Roche) and Lumipulse (Fujirebio)). Method comparisons (Bland Altman plots, Passing and Bablok analyses and correlation coefficients) between LC-MS/MS and IAs were performed for both the HD and HC samples.

Results The fT4, cortisol and testosterone immunoassays deviated respectively 16-31%, 13-41% and up to 35% more from the LC-MS/MS in HD samples than in HC samples. Remarkably, the fT4 results were all falsely decreased in HD samples, whereas the cortisol and testosterone results were mainly falsely increased. Correlation coefficients calculated based on the method comparison were lower in the samples from HD patients than from HC.

Discussion and conclusion Several automated IAs are less reliable in measuring fT4, cortisol and testosterone in the altered serum matrix of hemodialysis patients. Physicians and laboratory specialists should be aware of these pitfalls in this specific patient group and this phenomenon should also be studied in other patient groups characterized by an altered serum matrix.