Abstract

Objectives Automated microscopy is increasingly used to screen urine samples for suspected urinary tract infection (UTI). A 98.8% negative predictive value has been reported in adult studies. The aim of our study was to validate this method in a paediatric population.

Methods Urine samples were collected from children with known or suspected nephrourological disease attending nephrology and urology clinics over a 6-week period. Samples were tested with dipstick, the UF-100 flow cytometer (automated microscopy) and culture. A gold standard of a positive culture of morethan 10⁵ colony forming units per ml (cfu/ml) with a pathogenic organism was used and the sensitivity, specificity and likelihood ratios were calculated.

Results 280 urine samples were collected from 263 patients (143 male, median age 10.2 years, range 0.1–19.75 years). 221 (79%) were midstream or clean-catch samples. Automated microscopy identified 42 of 186 samples as requiring culture and 17 of 19 samples which had a pure growth of more than 10⁵ cfu/ml. Two patients were not identified by automated microscopy: one was treated for vulvovaginitis, and one commenced prophylactic antibiotics prior to the culture result being obtained. The sensitivity, specificity, positive and negative likelihood ratios were 0.89, 0.85, 5.98 and 0.17, respectively. This compared to 0.95, 0.72, 3.34 and 0.29, respectively, with urine dipstick.

Conclusion Automated microscopy performed comparably to urine dipstick in the diagnosis of UTI with improved specificity and likelihood ratios with slightly reduced sensitivity. The data support the use of automated microscopy for screening urine samples for culture in children, but different automated microscopy methods and algorithms require local evaluation.