Assessment of Albuminuria - comparison of immunoturbidimetry and high pressure liquid chromatography method
Abstract
Objective: The aim of our study was the implementation of HPLC method for assessing albuminuria, confirming or refusing the hypothesis about the existence of co-eluting proteins, and comparison of urine albumin concentrations assessed using HPLC and immunoturbidimetric methods in patient samples. Material and Methods: We developed the HPLC method under these chromatographic conditions: multilinear gradient of mobile phase consisting of water (solvent A), acetonitrile (solvent B), 100 mM NaH2PO4 (solvent C), 1 % trifluoroacetic acid (solvent D), flow rate 2 mL/min, temperature 22 °C, UV detection in the wavelength 280 nm, chromatographic column Zorbax 300 SB-C3, liquid chromatograph Agilent 1200 (Agilent Technologies, USA).
We analyzed two mixtures of urine. The first one was prepared from 30 patient urine samples with immunoturbidimetrically physiological albuminuria, to which we added albumin standard.
The second mixture was prepared from 30 patient urine samples with mild albuminuria. We analyzed and compared albuminuria in a group of 301 patients (161 diabetics and 140 nondiabetics, 182 males and 119 females, age 44.1 +- 25.3 years).
Results: Transferrin, α-1-acid glycoprotein, α-1-antitrypsin, α-1-antichymotrypsin and hemopexin do not interfere with albumin in the presented HPLC method, whereas the elution curve of prealbumin splits into several peaks, of which a few interfere with albumin. With regard to very low prealbumin concentrations in serum and in urine, we can assume that this interference is not of clinical importance.
In mixture 1 we did not find a significant difference between the albuminuria assessed using both methods (79.1 mg/L immunoturbidimetrically vs. 82.5 mg/L HPLC), while in mixture 2 we measured over 26 % greater albuminuria using HPLC (79.5 mg/L immunoturbidimetrically vs. 99.9 mg/L HPLC). These results suggest that immunounreactive albumin really exists.
We found a statistically significant difference between the methods in patient urine samples (entire patients' group: [median +- SEM]: 21.1 +- 38.2 mg/L immunoturbidimetrically vs. 30.4 +- 42.2 mg/L HPLC, p < 0.0001, Mann-Whitney test), which was most remarkable in the group of diabetic patients with immunoturbidimetrically physiological albuminuria. Conclusion: Our results prove that the HPLC method for albumin detection is more sensitive than immunoturbidimetry and sufficiently specific and there are no significant interferences with other proteins.