No abstract

We concede that it was unclear that the urines screened were from predominantly asymptomatic patients. The prevalence of 5% reported by Leigh and Williams referred to the presence of significant bacteriuria in pregnant women; the 12% prevalence in our paediatric population compared with that found by Cannon et al5 in a similar population.Of the urines tested, only six fell within the "doubtful significance" range of 104-105 organisms/ml, and 50% of these were detected by the Ames reagent strips. In our laboratory we normally request repeat samples in these cases. For the purpose of our evaluation, however, we strictly adhered to Kass's criteria,6 as existing data from a number of workers7'0 still support the criterion of > I05 organisms/ml for the diagnosis of significant bacteriuria in asymptomatic patients.With the Ames dipsticks, the method of reading them influenced the number of false negative results obtained. Visual reading of the strips resulted in 19% false negative results; photometrically read strips yielded 7%. This may be regarded as an unacceptably high false negative rate, but no urine screening method is 100% accurate. All our urines giving a positive (true and false) strip result are subjected to culture and microscopy before a report is issued to the clinician. We have also encouraged our clinicians to specify those urines from symptomatic patients which require culture and microscopy without screening.With urine specimens comprising up to 40% of the workload of clinical microbiology laboratories, most are forced to use some form of screening procedure. We consider the dipstick analysis to be as accurate a screening method as those currently in use, and its ability to screen out clinically insignificant urine samples rapidly reduces the number of specimens to be cultured. This rapid screening method also enables same day reporting of negative urine specimens to be made.