How to Improve the Positive Predictive Value of Urinary Decoy Cell Surveillance for Polyomavirus BK-Associated Nephropathy in Kidney Transplant Patients

Abstract Background Polyomavirus BK-associated nephropathy (BKVN) has been a serious problem after kidney transplantation. Detection of urinary decoy cells (UDCs) and assessment of polyomavirus BK nucleic acids by polymerase chain reactions (PCRs) are currently used, noninvasive tests. PCRs have better positive predictive value (PPV) but higher cost and lower accessibility. This study investigated ways to improve the PPV of UDCs for BKVN prediction. Methods From 2000 to 2013, kidney transplant recipients with sustained UDCs for more than half a month and who had received allograft biopsies were enrolled. We analyzed the PPV of UDCs for BKVN with 2 variables: (i) the percentage changes in serum creatinine (SCr) levels and (ii) the duration of sustained UDCs by receiver operating characteristic (ROC) curve analysis; we predicted the percentage changes in SCr levels with the corresponding PPV using a linear regression model. Results BKVN was diagnosed in 26 of 68 enrolled patients. The percentage changes in SCr levels significantly deteriorated in the BKVN group during 1–2 months of UDC positivity. According to ROC curve analysis, percentage changes in SCr levels had a significant discriminating power for BKVN during 1–1.5 month, and if the percentage changes in SCr levels were >19%, the PPV of UDCs for BKVN was 50%. Conclusions An UDC surveillance program is a judicious strategy to predict BKVN in kidney transplant patients, particularly when graft renal function shows deterioration after 1 month of UDC positivity.