Sources of variation and establishment of Russian reference intervals for major hormones and tumor markers

Authors:

Abstract
Objectives. To study the sources of variation (SV) in reference values (RV) for 22 key immunoassays and to determine reference intervals (RIs), a multicenter study of the Russian population was conducted. Methods. In accordance with the protocol of the Committee on Reference Intervals and Limit Values (C‑RIDL) of the International Federation of Clinical Chemistry and Laboratory Medicine (IFCC), 758 healthy volunteers were included in the study, conducted in St. Petersburg, Moscow, and Yekaterinburg. Five tumor markers and 17 hormones and related analytes were analyzed in serum samples using the UniCel DxI 800 immunoassay analyzer from Beckman Coulter. SVs were analyzed using multiple regression analysis (MRA) and analysis of variance (ANOVA). The primary criterion for separating the study population by sex and age was a standard deviation ratio (SDR) of 0.4. Results. When assessing differences between groups of volunteers participating in the study in different cities, the SDR was less than 0.4. The following secondary exclusion criteria were used: for female sex hormones – contraceptive use (8%); for CA19–9—presumably Lewis negative blood type (10.5% of men and 11.3% of women); for insulin – BMI ≥28 kg/m2 (31%); for thyroid hormones – the presence of thyroid antibodies (10.3% of men and 24.5% of women); for carcinoembryonic antigen (CEA)—smoking (30% of men and 16% of women). For all analytes except CA19–9, CA15–3, thyroid function tests, parathyroid hormone, and insulin, RIs had to be separated by sex. For alpha‑fetoprotein (AFP), CEA, all female sex hormones, follicle‑stimulating hormone (FSH), and progesterone, RIs had to be separated by age for both sexes. Typically, RIs were determined parametrically after Gaussian transformation using a modified Box‑Cox formula. Exceptions were growth hormone, estradiol in postmenopausal women, and progesterone in premenopausal women, which required nonparametric calculation due to bimodal distributions and/or insufficient detection limits. Conclusions. RIs for key hormones and tumor markers specific to the Russian population were determined in accordance with a current international standardized protocol with a thorough examination of the RI for each analyte.

This publication has 32 references indexed in Scilit: