Squamous epithelial cell presence reduces accuracy of urinalysis for prediction of positive urine cultures

Abstract

Background: Diagnostic value of urinalysis specimens contaminated with squamous epithelial cells (SEC) from the genital surfaces is assumed to be limited compared to clean-catch samples. However, no studies have quantified the change in predictive value in the presence of SECs for individual urinalysis markers.

Methods: In a retrospective, single center cohort study, we analyzed all urine cultures sent from the ED over a 26-month period with corresponding urinalysis results. Cultures were classified as positive with growth of >10⁴ colony forming units of pathogenic bacteria, negative if no growth, or contaminated for all other results. UA specimens were classified as contaminated or clean based on SEC presence. Accuracy of urinalysis markers for prediction of positive cultures was calculated as an area under the curve (AUC) and was compared between contaminated and clean UA specimens.

Results: 6490 paired UA and urine cultures were analyzed, consisting of 3949 clean and 2541 contaminated samples. SEC presence was less common with male gender, older age, and smaller BMI. Urine cultures were 19.2% positive overall, and SECs were more common in contaminated cultures. AUCs for individual markers ranged from 0.557 to 0.796, with pyuria, bacteriuria, and leukocyte esterase having higher AUC in clean samples over contaminated.

Conclusion: Analysis of AUC for individual urinalysis markers showed reduced diagnostic accuracy in the presence of SECs. SEC presence also reflected much higher rates of contaminated cultures. These results support the reduced reliance on contaminated UA specimens for ruling in UTI in ED patients.