Risk Assessment for Polyomavirus Nephropathy Using Urine Cytology and the Detection of Decoy Cells

Abstract

Polyomavirus nephropathy (PVN) is one cause of renal allograft loss. Disease incidence in renal allograft recipients was estimated to be 4% in 2010 with an overall graft failure rate of 34%, reaching up to 90% in advanced PVN disease subgroup C characterized by marked sclerosis (unpublished data collected by the “Banff working group” of PVN with nine participating U.S. and European centers). Because therapeutic strategies with anti-PV drug regimens are limited, the overall outcome is improved by proper risk assessment and an early diagnosis. Although all patients with disease or PVN show signs of PV activation with viruria and viremia (1–10), only a small minority of all viruric or viremic patients actually develop disease, that is biopsy-confirmed definitive PVN (3, 5, 7, 10). This phenomenon is well known from other DNA viruses that establish latency in humans after primary infection, such as Epstein-Barr virus or cytomegalovirus. Thus, how can screening for PV activation and overall risk assessment for PVN in renal allograft recipients be cost-efficient and cost-effective? Urine cytology and the search for “decoy cells” can help. We read the article by Chakera et al. (11) “Detection of polyomavirus BK reactivation after renal transplantation using an intensive decoy-cell surveillance program is cost-effective” with great interest. For more than a decade, we have successfully used urinary decoy-cell screening in the management of kidney transplant recipients (2, 4, 7, 8, 12, 13), and we would like to contribute to the discussion by sharing some of our observations. Decoy cells found in voided urine samples contain intranuclear PV (mainly BK virus) inclusions that appear as homogenous and glassy viral material on standard light microscopy. Decoy cells can easily be identified and counted by cytopathologists in standard Papanicolaou-stained urine cytology specimen using ThinPrep or cytospin preparation. The analysis can be performed in any pathology laboratory. The cutoff for clinically significant PV activation and decoy-cell shedding (“decoy positivity”) is the detection of 10 or more decoy cells per slide (1, 7) with an overall positive predictive value (PPV) for biopsy-proven definitive PVN in renal allograft recipients of approximately 25% and a negative predictive value of 100%. The PPV can be increased by incorporating additional findings into the decision-making process, such as prolonged decoy-cell shedding exceeding 4 weeks (see in the next paragraph) or concurrent allograft dysfunction. The predictive values of decoy cells are very similar to those of quantitative urine polymerase chain reaction (PCR) assays for the detection of BK-virus DNA at threshold levels of 1×107 or more BK-virus copies per milliliter of urine (7). Since 2009, we have followed 212 renal allograft recipients prospectively including decoy-cell testing. We also have detailed longitudinal information on a cohort of 32 patients with biopsy-confirmed definitive PVN (Table 1). In our population of 212 patients, 51 (24%) were decoy-cell positive. Of these 51 patients, 8 (16%) presented with multiple individual flares of decoy positivity, whereas most (43/51, 84%) only experienced a single episode of decoy positivity of various durations. Of 51 “decoy-positive” patients, 26 (51%) experienced extended, continuous decoy-cell shedding for 1 month or longer; it is especially this patient group that is at very high risk for the development of PVN (see in the next paragraph). Of 51 decoy-cell positive patients followed prospectively, 14 (27%) developed biopsy-proven definitive PVN. PVN was not diagnosed in “decoy-negative” patients.TABLE 1: Decoy-cell sheddinga in selected adult kidney transplant recipients at The University of North Carolina at Chapel HillAt The University of North Carolina at Chapel Hill (UNC), we have detailed information on a cohort of 32 patients with PVN (including 14 from the prospectively followed group of 212 patients): all of them were decoy-cell positive at the time of initial biopsy diagnosis, and 28 (88%) were with prolonged continuous decoy-cell shedding for many weeks. In patients with PVN, decoy-cell positivity preceded the histologic biopsy diagnosis on average by 4 weeks, thereby serving as an initial “alarm signal” during the prodromal disease stage. Note that, during this prodromal time window, two of our patients showed negative renal biopsy results. In these two patients, typical signs of disease were only found in subsequent repeat biopsies, only then allowing for a definitive histologic diagnosis of PVN. At the other end of the spectrum, PVN disease resolution is also reflected by decoy-cell analyses that turn from decoy positive to decoy negative on repeat testing (generally conducted every 4 weeks in patients with PVN), thereby marking viral clearance and “healing.” At UNC, all patients are tested at regular intervals for the shedding of decoy cells during the first 24 months after grafting. Positive decoy-cell readings trigger tighter surveillance at 4-week intervals, and patients are additionally tested with quantitative plasma PCR assays to detect BK virus, thereby following general recommendations for patient management with slight modifications (10). Decoy-positivity and concurrent plasma PCR test results of more than 250 BK virus copies per milliliter trigger a diagnostic kidney biopsy regardless of allograft function. Thus, decoy-cell testing serves as an initial mass-screening tool followed by a second line of targeted assays with higher predictive values for PVN (PPV for quantitative plasma PCR with a threshold of ≥1×104 BK virus copies per milliliter, approximately 75% [7]). This screening protocol resulted in the definitive diagnosis of PVN in a nonsclerotic disease stage in all of our 14 prospectively followed patients; 50% (7/14) were diagnosed in early PVN stage A, and none, in disease stage C (disease staging was according to the guidelines under discussion by the Banff PV working group). No graft was lost because of PVN, avoiding high costs associated with return to dialysis. At UNC, decoy-cell testing currently costs U.S. $286 compared with U.S. $409 for plasma or urine quantitative PCR assays for BK virus. Of all prospectively followed renal allograft recipients in our cohort, 76% (161/212) never activated PV defined by negative decoy-cell test results (Table 1); PCR assays were not ordered according to our screening guidelines, and these patients never developed PVN. Assuming that other laboratory tests including urinalysis and blood draws are routinely conducted during patient visits, our testing protocol does not incur any additional specimen collection fees. During the 24 months after transplantation, multiple routine surveillance tests monitoring for the activation of PV are performed per patient following general recommendations (10). In our prospectively followed cohort of 161 patients without PV activation, 1500 negative decoy-cell screening assays were conducted between January 2009 and February 2012 with a total cost for decoy-cell analysis of U.S. $429,400. Conversely, if screening had primarily been based on quantitative PCR assays, then the costs would have been U.S. $613,000. Using decoy-cell screening resulted in a total saving of more than U.S. $184,000 (U.S. $1140 per patient) for a period of 38 months at UNC. This conservative calculation does not consider additional savings in 37 patients in our prospectively followed cohort, which were decoy-cell positive at certain time points during follow-up (Table 1) but never developed PVN, and in whom quantitative PCR tests were limited to few selected time points. In conclusion, we share the encouraging observations made by Chakera and colleagues from Oxford University (11). Routine urine cytology including the quantitation of decoy-cell shedding is a well-established and cost-effective primary mass-screening tool to search for PV activation in renal allograft recipients and to identify patients at an increased risk for the development of PVN. Decoy positivity, particularly if detected for more than 1 month, should trigger second-line testing with additional assays, such as quantitative plasma PCR analyses to detect BK virus and renal biopsy in high-risk patients. In addition, we find decoy-cell testing intriguing because it also allows for targeted and highly predictive PV “Haufen testing” on the same urine sample (7, 14), thereby streamlining patient management and the clinical decision-making process. Volker Nickeleit 1,2 Karin True3 Randal Detwiler3 Tomasz Kozlowski4 Harsharan Singh1,2 1 Department of Pathology Division of Nephropathology The University of North Carolina at Chapel Hill Chapel Hill, NC 2 Department of Pathology and Laboratory Medicine Division of Nephropathology The University of North Carolina at Chapel Hill Chapel Hill, NC 3 Department of Internal Medicine Division of Nephrology The University of North Carolina at Chapel Hill Chapel Hill, NC 4 Department of Surgery The University of North Carolina at Chapel Hill Chapel Hill, NC