Monitoring for polyomavirus BK and JC in urine: Comparison of quantitative polymerase chain reaction with urine cytology

Clinical monitoring for polyornavirus infection is becoming common, but the optimal detection technique remains undefined. We compared the relative efficacy of exfoliative cytology and polymerase chain reaction (PCR) for detecting viruria in 100 urine samples. Definite decoy cells were seen in 8% and probable decoy cells in 5% of specimens. PCR showed BK virus (BKV) DNA in all these and in an additional 18% of samples. Using decoy cells as a marker of polyomavirus viruria cytology has a sensitivity of 41.9% and negative predictive value of 82.8%. The specificity and positive predictive value for viruria (not viral nephropathy) are 100%. False-negative results occurred in samples with suboptimal cellularity, vaginal contamination, and a low viral load. One patient with a false-negative urine cytology developed BKV nephropathy on follow-up. Compared with PCR, urine cytology is a less sensitive technique, which requires morphologically intact cells, and cannot distinguish BKV from JC virus.