WP 142 Validation of Automated Urine White Cell Microscopy As a Predictor of Chlamydia Trachomatis and Mycoplasma Genitalium Infection

Tuesday, June 10, 2014
International Ballroom
Marcus Pond, Mr1, Sheel Patel, Dr2, Ken Laing, Dr1, Margarita Ajayi, Ms1, Achyuta Nori, Dr1, Andrew Copas, BA, MSc, PhD3, Philip Butcher, Prof1, Phillip Hay, Dr1 and S Tariq Sadiq, Dr1, 1Institute of Infection and Immunity, St George's, University of London, London, United Kingdom, 2Department of GU Medicine, St George's Healthcare NHS Trust, London, 3Department of Infection and Population Health, Centre for Sexual Health and HIV Research, London, United Kingdom

Background: Gram stained male urethral smear (GSUS), the standard method for diagnosing non-gonococcal urethritis (NGU) is highly operator dependent and a poor predictor of infection. Rapid automated urine flow cytometry (AUFC) of first void urinary white cells (UWC) may offer point-of-care performance improvements over GSUS. The performance of AUFC for predicting Mycoplasma genitalium and Chlamydia trachomatis urethral infection was evaluated and the relationship between UWC and bacterial load assessed. 

Methods: UWC were enumerated using the bench-top Sysmex UF-100 Analyser on First void urines (FVU) from male patients. Receiver operator curve (ROC) analysis was performed on an initial “Training set” of 208 samples from symptomatic men, all undergoing GSUS, to determine optimum UWC for predicting either M. genitalium or C. trachomatis infection. This threshold was subsequently validated using a fresh “Validation set” of 229 FVUs from both symptomatic and asymptomatic patients. Pathogen load, expressed as target DNA copies/mL of urine, was determined by Droplet Digital PCR.

Results: ROC analysis gave an optimal threshold of >29 UWC/µL for either C. trachomatis or M. genitalium infection, giving sensitivities and specificities of 81.5% and 85.8% respectively compared with 86.8% and 64.7% respectively for GSUS, (≥5 polymorphonuclear leukocytes/HPF). For the validation set, the UWC threshold gave sensitivities and  specificities of 70.3% and 92% overall, 70.5% and 87% in symptomatic (n=71), and 70% and 93.9% in asymptomatic (n=158) patients respectively. A stronger correlation of UWC with urinary pathogen loads was observed for M. genitalium (τ =0.426, p=<0.001) compared to C. trachomatis (τ =0.315, p=0.016).

Conclusions: AUFC offers significantly improved specificity over microscopy in the clinic for the prediction of urethral C. trachomatis and M. genitalium infection, and could provide an alternative to microscopy for asymptomatic patients. The degree of urethral inflammation exhibits a stronger associated with pathogen load for M. genitalium compared to C. trachomatis.