Performance of flow cytometry to screen urine for bacteria and white blood cells prior to urine culture, "Beyond the Abstract," by Callen Giessen and John C. Lieske, MD

BERKELEY, CA ( - In our laboratory, urine samples are often screened by manual Gram stain to determine if a culture should be performed. In the current study, this approach was confirmed to be quite sensitive (SE = 0.94), however it lacked specificity (SP = 0.68) as an indicator of a positive bacterial culture (> 105 cfu/mL). Therefore, roughly one-third of samples “positive” for bacteria by Gram stain had a negative bacterial culture.

The Gram-stain screening strategy has many factors that could contribute to this low specificity. Numerous pre-analytic features can result in a false-positive Gram stain result. These factors include, but are not limited to, barriers inherent in collecting a truly clean urine specimen, differences in slide plating technique, subjective interpretation of stain results, and contamination of slides and/or loops. Importantly, in our laboratory protocol, only samples with no bacteria on the Gram stain slide are considered “negative” and therefore not sent for culture. Thus it is likely some samples which display small amounts of bacteria by Gram stain, possibly due to contamination from pre-analytic variables, are unnecessarily cultured.

Recent advances in flow cytometry and staining strategies now allow quantification of bacteria present urine samples. Therefore, we evaluated the ability of one such platform, the Sysmex UF-1000i™(Sysmex America, Inc., Mundelein, IL, USA), to predict positive bacterial culture growth (> 105 cfu/mL). Results were compared head-to-head with our traditional Gram stain approach. At a cutoff of 288.9 bacteria/µL, flow cytometry had equal sensitivity (SE = 0.94) to our current Gram stain approach, but with improved specificity (SP = 0.84). Since flow cytometry can also quantitate WBCs, we assessed the relationship between the amount of pyuria and a positive culture. Results demonstrated that the WBC count was both less sensitive (SE = 0.89) and specific (SP = 0.79) than bacteria alone. By receiver operating curve analysis, the AUC for bacteria was 0.95 and AUC for WBCs was 0.90. An approach that considered both parameters and the interaction between them was not markedly better than bacteria alone (AUC for bacteria and WBCs = 0.96).

In certain patient populations, urine bacterial growth at lower thresholds (e.g., > 104 cfu/mL) can be clinically significant. Our study found that both the Gram stain and UF-1000i™ screening methods had reduced SE and SP for this cut-off. Thus, it is very important to consider what patient populations are appropriate for implementation of a flow cytometry screening approach as a triage for culture. It is also possible that a flow cytometry approach might perform better or worse in different patient populations than the one we studied, and it would be important to establish appropriate bacterial cutoffs specific to one’s own center. With these caveats in mind, the current study suggests implementation of a triage procedure which includes the bacteria count from the UF-1000i™ could reduce unnecessary cultures by as much as 55% when compared to a Gram stain screening approach. Even greater reductions could result if no screening strategy is currently in place.

Written by:
Callen Giessen,a and John C. Lieske, MDa, b as part of Beyond the Abstract on This initiative offers a method of publishing for the professional urology community. Authors are given an opportunity to expand on the circumstances, limitations etc... of their research by referencing the published abstract.

aRenal Laboratory, Department of Laboratory Medicine, Mayo Clinic, Rochester, MN
bDivision of Nephrology and Hypertension, Department of Internal Medicine, Mayo Clinic, Rochester, MN

Performance of flow cytometry to screen urine for bacteria and white blood cells prior to urine culture - Abstract

More Information about Beyond the Abstract