ABSTRACT Urine culture is the gold standard for identifying microorganisms in urine. However, the process is time-consuming and usually takes days to get the results, which impedes timely treatment. This study aimed to evaluate the performance of UF-5000 in excluding bacterial urinary tract infection (UTI) and detecting the presence of Gram-negative bacteria. A total of 1,522 urine specimens were subjected to routine urine analysis and culture. Bacteria, white blood cell counts, and UF-5000 bacteria information flags were compared with urine culture. Bacteria forward scatter (B_FSC), and bacteria fluorescent light intensity (B_FLH) parameters were assessed to determine the optimal angle for discriminating bacterial growth patterns, which was verified in the validation cohort. The optimal cut-off values were 42.2/µL and 100.2/µL to rule out UTI for bacteria at ≥10 4 CFU/mL and ≥10 5 CFU/mL, respectively. The agreement of UF-5000 bacterial classification and urine culture was fair (Kappa = 0.227). Regarding the discrimination of bacterial groups, “Gram Neg?” was associated with a specificity of 94.0% and a positive predictive value of 87.2% in the detection of Gram-negative bacteria. By estimating the B_FSC and B_FLH parameters, the best angle was 28° for distinguishing Gram-negative and Gram-positive bacteria. Among the 136 urine specimens with the low angle pattern (<28°) in the validation cohort, 127 specimens were confirmed to contain Gram-negative bacteria. The UF-5000 analyzer is a suitable and rapid tool to exclude negative urine specimens, and bacterial information flags for Gram-negative bacteria could be trusted by clinicians. IMPORTANCE The strength of our study relied on being the first study assessing the bacteria forward scatter (B_FSC) and bacteria fluorescent light intensity (B_FLH) research parameters with a UF-5000 urine analyzer and establishing the best angle for distinguishing Gram-negative and Gram-positive bacteria. When the bacterial scatter plot angle is less than 28°, the possibility of Gram-negative bacterial infection is more than 80%. Meanwhile, we find that UF-5000 bacterial information flags have a significant advantage in detecting Gram-negative bacteria with a specificity of over 90% and a positive predictive value of over 80%.
Read full abstract