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

Abstract

Objectives A multicenter study was organized to explore sources of variation (SVs) of reference values (RVs) for 24 major immunochemistry analytes and to determine reference intervals (RIs) for the Russian population. Methods According to IFCC Committee on Reference Intervals and Decision Limits (C-RIDL) protocol, 793 healthy volunteers were recruited in St. Petersburg, Moscow, and Yekaterinburg. Serum samples were tested for five tumor markers, 19 hormones and related tests by Beckman Coulter’s UniCel DxI 800 immunochemistry analyzer. SVs were explored using multiple regression analysis and ANOVA. Standard deviation ratio (SDR) of 0.4 was used as primary guide for partitioning RIs by gender and age. Results SDR for between-city difference was <0.4 for all analytes. Secondary exclusion of individuals was done under the following conditions: for female sex-hormones, those with contraceptives (8%); for CA19-9, those supposed to have negative Lewis blood-group (10.5%); for insulin, those with BMI≥28 kg/m2 (29.9%); for the thyroid panel, those with anti-thyroid antibodies (10.3% in males; 24.5% in females). Gender-specific RIs were required for all analytes except CA19-9, CA15-3, thyroid-related tests, parathyroid hormone, and insulin. Age-specific RIs were required for α-fetoprotein and all sex-hormones except testosterone. RIs were generally derived by parametric method after Gaussian transformation using modified Box-Cox formula. Exceptions were growth hormone, estradiol, and progesterone, for which nonparametric method was required due to bimodal distribution and/or insufficient detection limit. Conclusion RIs for major hormones and tumor markers specific for the Russian population were derived based on the up-to-date internationally harmonized protocol by careful consideration of analyte-specific SVs.