Rapidly discriminating culture-negative urine specimens from patients with suspected urinary tract infections by UF-5000

Aim: A rapid and reliable method of discriminating such specimens would be very useful. Materials & methods: We analyzed 566 urine specimens from patients with suspected urinary tract infections using a fully automated urine particle analyzer (UF-5000) and evaluated its performance for culture-negative urine specimens. Results: Using the algorithm cutoff values of bacteria less than 30/μl and/or white blood cell less than 200/μl, we obtained a sensitivity of 97.8%, a specificity of 74.6%, a positive predictive value of 46.9%, a negative predictive value of 99.3%, an agreement of 78.9% with the culture method and reduced 61% unnecessary urine culture. Regarding the discrimination of bacterial Gram groups, 67.7% (63/93) of cases were correctly analyzed using the UF-5000 bacteria information, with a Cohen's kappa concordance coefficient of 0.775 (χ2 = 31.65, p < 0.001). Conclusion: The performance of UF-5000 for rapidly discriminating culture-negative specimens was quite acceptable for clinical use.