BACKGROUND The aim of this study was to examine the sensitivity and specificity of pyuria-based diagnosis of urinary tract infection (UTI) in urine collected by transurethral catheterization, and the reliability of diagnosis of pyuria in urine collected in a perineal bag. The gold standard for UTI diagnosis is significant colony counts of a single organism in urine obtained in a sterile manner. METHODS We enrolled 301 patients who underwent medical examination at the present hospital for possible UTI between January 2005 and December 2009. We collected 438 urine samples by transurethral catheterization. We investigated the accuracy of pyuria-based diagnosis of UTI using transurethral catheterization urine specimens, and the reliability of diagnosis of pyuria using bag-collected urine specimens. RESULTS The false-negative rate of UTI diagnosis based on pyuria in transurethral catheterization urine sediments was 9.0%; there was no significant difference in the false-negative rate of UTI diagnosis between boys and girls. Approximately 28% of pyuria-positive bag-collected urine specimens were pyuria negative on transurethral catheterization; this rate was significantly higher in girls than in boys (56.7% vs. 8.9%, P < 0.0001). CONCLUSIONS The absence of pyuria in transurethral catheterization urine sediments does not rule out UTI. Pyuria in bag-collected urine specimens frequently consists of urine leukocytes from external genitalia as well as from the urinary tract.