BACKGROUND: we evaluated the performance of the flow cytometry-based UF-4000 automated urine analyzer associated with the UD-10 image analyzer (Sysmex), in comparison with optical microscopy and culture. MATERIALS AND METHODS: 2,695 consecutive urine samples of patients were included. The cell count was performed using the analyzer and the Kova cell for 316 samples, and compared according to a threshold of 10 white blood cells (WBC) and 10 red blood cells (RBC) /μL. In a second stage, the quantitative threshold of bacteria from which a bacterial alert is triggered was chosen by comparison with the culture. Finally, the reliability (versus Gram staining and culture) of the nature of Gram negative (GN?) and Gram positive (GP?) flags has been tested on 362 samples. RESULTS: the microscopy/UF4000 discrepancy rate was 8.5% for WBC, and 16% for RBC which dropped to 6.9% after switching to UD-10. The majority of these discrepancies corresponded to quantities close to the clinical threshold, mostly higher by automatic than by microscopic counts. With a chosen warning threshold of 200 germs/μL, the «GN?» and «GP?» flags resulted in 91%/86% and 79%/20% of Gram/significant cultures of GN bacilli and GP cocci, respectively. CONCLUSION: the correlation UF4000/microscopy is satisfactory for cellularities, the UD10 allowing to correct discrepancies. The «GN?» flag is reliable allowing a quick diagnostic orientation for the clinician. Finally, UF4000/UD10 has shown very good performances, notably thanks to the integration of the UD-10 image analyzer, which eliminates time consuming optical microscopy cell count.
BACKGROUND: we evaluated the performance of the flow cytometry-based UF-4000 automated urine analyzer associated with the UD-10 image analyzer (Sysmex), in comparison with optical microscopy and culture. MATERIALS AND METHODS: 2,695 consecutive urine samples of patients were included. The cell count was performed using the analyzer and the Kova cell for 316 samples, and compared according to a threshold of 10 white blood cells (WBC) and 10 red blood cells (RBC) /μL. In a second stage, the quantitative threshold of bacteria from which a bacterial alert is triggered was chosen by comparison with the culture. Finally, the reliability (versus Gram staining and culture) of the nature of Gram negative (GN?) and Gram positive (GP?) flags has been tested on 362 samples. RESULTS: the microscopy/UF4000 discrepancy rate was 8.5% for WBC, and 16% for RBC which dropped to 6.9% after switching to UD-10. The majority of these discrepancies corresponded to quantities close to the clinical threshold, mostly higher by automatic than by microscopic counts. With a chosen warning threshold of 200 germs/μL, the «GN?» and «GP?» flags resulted in 91%/86% and 79%/20% of Gram/significant cultures of GN bacilli and GP cocci, respectively. CONCLUSION: the correlation UF4000/microscopy is satisfactory for cellularities, the UD10 allowing to correct discrepancies. The «GN?» flag is reliable allowing a quick diagnostic orientation for the clinician. Finally, UF4000/UD10 has shown very good performances, notably thanks to the integration of the UD-10 image analyzer, which eliminates time consuming optical microscopy cell count.