On the Performance and Reliability of Mechanized Urine Teststrip Measurement in Comparison with Visual Reading

Summary: Prescreening of urine specimens by teststrips is a valuable procedure for reducing the work load of the urine analysis iaboratory: positive results for leukocytes, erythrocytes (haemoglobin), protein, and/or nitrite are widely used to select pathological specimens for subsequent microscopic examination. By stand-ardization of the measurement conditions, mechanized teststrip reading is claimed to give more reproducible results than conventional techniques. To assess their ability to improve urine prescreening, especially with regard to the comparability of the results, the practical and analytical performance of three commercially available analysers (Rapimat ® II/T from Behringwerke AG, Urotron ® RL9 from Boehringer Mannheim GmbH, and Clinitek ® 200 from Ames/Bayer Diagnostic) was compared with visual reading. Analytical criteria were assessed using routine urine samples, while reproducibility was tested by repeated analysis of three different commercial control urines (Kova®Trol from Madaus). A mean imprecision between 3 % and 11.9% was found for the mechanized dipstick reading which was comparable to that found with visual examination (4.5 % with Combur9-teststrips, Boehringer Mannheim GmbH). Due to the crude classification of the results, the different analysers äs well äs the visual technique gave quite different distributions for each of the semiquantitative parameters in the same urine samples. Even if statistical analysis was restricted to the frequency of positive results oniy, significant differences (%2-test, p < 0.001—0.05) between methods wer