Is the accuracy of serum albumin measurements suitable for clinical application of the test?

Background. Albumin is the major protein in plasma and the determination of its concentrations is used for the prognostic assessment of several diseases. Using the information on the biological variation of the analyte, the degree of albumin measurement uncertainty for clinical laboratories should stay within ±3.90% or ±5.85% (desirable or minimum quality level, respectively). We checked the performance of the albumin immunoturbidimetric assay (Tina-quant Albumin Gen. 2, Roche Diagnostics), carried out on the Cobas c 501 platform. Methods. To estimate the uncertainty of albumin measurement, we employed the imprecision and bias data obtained by measurements of IRMM ERM-DA 470k/IFCC Human Serum Proteins reference material. The material was measured in duplicate for three consecutive days, in two identical experiments carried out 5 months apart, using the analytical system in accordance with the manufacturer’s instructions. Results. The relative standard uncertainty (uc) of imprecision was 1.88%; the bias uc, derived from the difference (-6.18%) between the obtained mean of the means (34.9 g/L) and the target value (37.2 g/L), the bias variability (0.95%) and the uc of the certified value of reference material (1.61%), was 6.42%. Thesegave a combined uc of 6.69%. Conclusions. As the requested analytical quality for albumin measurement in serum is high, the performance of field methods should be extremely good to permit their application in clinical setting. Our results seem to demonstrate that, at least for some commercial assays, problems with standardization of these measurements still persist, even if highly specific and relatively expensive immunochemical methods are used.