Quantitative evaluation of urinalysis test strips

Abstract

Background: Urine test strip results are generally reported in categories (i.e., ordinal scaled), but automated strip readers are now available that can report quantitative data. We investigated the possible use of these meters to complement flow cytometry of urine and compared reflectance readings with quantitative determinations of urinary glucose and microalbumin.

Methods: We compared URISYS 2400 (Roche) quantitative reflectance data with data from the UF-100 (Sysmex) and biochemical data for 436 nonpathologic and pathologic urine samples.

Results: Reproducibility of the reflectance signal was good for high- and low-concentration urine pools for protein (0.8% and 0.9% and 1.5% and 2.2% within and between runs, respectively), leukocyte esterase (1.1% and 1.0%; 5.1% and 1.2%), hemoglobin (1.7% and 1.1%; 8.9% and 1.1%) and glucose (2.1% and 0.5%; 6.5% and 2.3%). Fair agreement was obtained between UF-100 and test strip reflectance data for erythrocytes and hemoglobin (r = -0.680) and leukocytes and leukocyte esterase (r = -0.688). Higher correlations were observed for biochemical and test strip data comparing protein and albumin (r = -0.825) and glucose data (r = -0.851). The lower limits of detection for erythrocytes and leukocytes were 8 x 10(6)/L and 19 x 10(6)/L, respectively. The protein test (n = 220) detected 86% (95% confidence interval, 78-92%) of samples with <30 mg/L albumin with a specificity of 84% (95% confidence interval, 76-91%).

Conclusions: In urine test strip analysis, quantitative hemoglobin and leukocyte esterase reflectance data are complementary with flow cytometric results and glucose and albumin results.