We compared 104 samples of urine from hospital patients. We monitored reproducibility of estimations using Sysmex UF-100 for erythrocytes, leukocytes and epithelial cells within 1 month. Assess ed values ranged between 5–7% (CV). Correlation of results of UF-100 with microscopy and dipstick (Combur-10-Roche, Supertron) was 0.66–0.72 for erythrocytes, 0.61 for leukocytes and 0.997 for epithelial cells. Stability at room temperature was at least 3 hours for leukocytes, casts and epithelial cells and 2 hours for erythrocytes and bacteria. A high number of falsely positive results of casts in urinary samples without proteinuria was ascertained for Sysmex. We only found a low correlation of the number of erythrocytes and leukocytes and proteinuria in the hospitalized patients population. We estimated the necessity of microscopic reanalyses – 25% for leukocytes and 21% for erythrocytes. ROC analysis showed a cut-off 8/ml (leukocytes and erythrocytes) to have 100% negative predictive value for UF-100 if microscopy was used as reference.

        Key words: UF-100, microscopy, dipstick, urinary sediment, comparability.