Objectives: Precision, bias, comparability and uncertainty in CZE Paragon 2000 measurement of serum protein fractions. Results: Repeatability (expressed as CV) was: 0.6% (albumin), 2.8% (alpha-1-globulin), 1.9% (alpha-2-globulin), 3.5% (beta globulin), 1.6% (gamma globulin). Reproducibility (also expressed as CV) has been: 0.7% (albumin), 3.5% (alpha-1-globulin), 2.3% (alpha-1-globulin), 5.1% (beta-globulin), 2.7% (gamma-globulin). Bias values of albumin, alpha-2-globulin and gamma-globulin ranged in interval 2–5 %, while bias of beta-globulin reached 18 % and even 122 % of bias for alpha-1-globulin. AGE method was used for comparison. Both constant and proportional parts of systematic errors participated in measurement of albumin, constant parts dominated in alpha-1- globulin and alpha-2-globulin, while proportional systematic error was typical for beta globulin. Combined expanded uncertainty (k = 2) was as follows: albumin 1.3 %, alpha-1-globulin 7.1 %, alpha-2-globulin 4.8 %, beta-globulin 10 %, gamma-globulin 5.8 %. Conclusions: CZE Paragon 2000 is very effective, quick and namely precise method, with very well accepted level of uncertainty. However large systematic differences from usually used older electrophoresis techniques required establishment of new reference intervals.

        Key words: CZE electrophoresis, AGE electrophoresis, Paragon 2000, precision, differences, uncertainty.