Background: The urine dipstick that detects nitrite and
leukocyte esterase, and urine sediment is commonly used to diagnose or exclude urinary tract infections (UTIs) as the source of infection in febrile patients admitted to the emergency department of Dutch hospitals. However, the diagnostic accuracy of the urine dipstick and urine sediment has never been studied in this specific situation. Methods: Urinary samples of 104 febrile consecutive patients were examined. Urine culture with ≥ 105 colonies/ ml of one or two known uropathogen was used as the gold standard. The diagnostic value of the urine dipstick, urine sediment and Gram stain at various cut-off points was determined and used to develop a new diagnostic algorithm. This algorithm was validated in a new group of sepsis patients based on systemic inflammatory response syndrome (SIRS) criteria. Results: A positive nitrite on the urine dipstick (specificity 99%) rules in UTI. This is the first step of our diagnostic algorithm. The second step is to exclude UTI by absence of bacteria in the urine sediment (sensitivity 94%). The third and last step is the number of leucocytes/high-power field (hpf) in the urine sediment. Less than 10 leucocytes/hpf makes UTI unlikely whereas ≥ 10 leucocytes/hpf indicates UTI. In contrast to urine dipstick and/or urine sediment results alone, our algorithm showed both a high sensitivity
(92%) and specificity (92%) and was validated in a new
sepsis population. Conclusion: Our accurate and fast diagnostic algorithm, which combines the selective results of urine dipstick and urine sediment, can be easily used to diagnose UTI in febrile patients at the emergency department of Dutch hospitals.