Establishing normal values of total testosterone in adult healthy men by the use of four immunometric methods and liquid chromatography-mass spectrometry

The total testosterone (T) cutoffs clinically adopted to define late-onset hypogonadism (LOH) do not consider the differences that exist between different analytical platforms, nor do they consider the body mass index (BMI) or age of the patient. We aimed at providing method, age and BMI-specific normal values for total T in European healthy men. A total of 351 eugonadal healthy men were recruited, and total T was measured with four automated immunometric assays (IMAs): ARCHITECT i1000SR (Abbott), UniCel DxI800 (Beckman Coulter), Cobas e601 (Roche), IMMULITE 2000 (Siemens) and liquid chromatography-tandem mass spectrometry (LC-MS/MS). Reference ranges (RRs) were calculated for each method. Passing and Bablok regression analysis and Bland-Altman plot showed an acceptable agreement between Abbott and LC-MS/MS, but a poor one between LC-MS/MS and the other IMAs. Age-specific T concentrations in non-obese (BMI <29.9 kg/m2) men were greater than in all men. The total T normal range, in non-obese men aged 18-39 years, measured with LC-MS/MS was 9.038-41.310 nmol/L. RRs calculated with LC-MS/MS statistically differed from the ones calculated with all individual IMAs, except Abbott and among all IMAs. Statistically significant differences for both upper and lower reference limits between our RRs and the ones provided by the manufacturers were also noticed. We calculated normal ranges in a non-obese cohort of European men, aged 18-39 years, with four commercially available IMAs and LC-MS/MS and found statistically significant differences according to the analytical method used. Method-specific reference values can increase the accuracy of LOH diagnosis and should be standardly used.