Kidney Transplant Function and Histological Clearance of Virus Following Diagnosis of Polyomavirus-Associated Nephropathy (PVAN)

Authors


Abstract

Polyomavirus-associated nephropathy (PVAN) is managed by reduced immunosuppression with or without antiviral therapy. Data from 55 patients with biopsy-proven PVAN were analyzed for adverse outcomes and influence of baseline variables and interventions. During 20 ± 11 months follow-up, the frequencies of graft loss, major and any functional decline were 15%, 24% and 38%, respectively. Repeat biopsies were performed in 45 patients with persistent PVAN in 47%. Low-dose cidofovir, IVIG and cyclosporine conversion were used in 55%, 20% and 55% of patients. No single intervention was associated with improved outcome. Of the variables examined, only degree of interstitial fibrosis at diagnosis was associated with kidney function decline. In contrast, donor source, interstitial fibrosis, proportion of BKV positive tubules and plasma viral load at diagnosis were all associated with failure of histological viral clearance. This retrospective, nonrandomized analysis suggests that: (i) Graft loss within 2 years of PVAN diagnosis is now uncommon, but ongoing functional decline and persistent infection occur frequently. (ii) Low-dose cidofovir, IVIG and conversion to cyclosporine do not abrogate adverse outcomes following diagnosis. (iii) Fibrosis at the time of diagnosis predicts subsequent functional decline. Further elucidation of the natural history of PVAN and its response to individual interventions will require prospective clinical trials.

Introduction

Polyomavirus-associated interstitial nephropathy (PVAN) due to BK virus (BKV) reactivation has emerged as an important cause for kidney transplant (KTx) failure (1–8). Viral reactivation in the urinary tract and in the transplanted kidney is a direct result of the immunosuppressed state and may involve donor- and/or recipient-derived virus (5–7,9,10). The surge in frequency of PVAN among KTx recipients worldwide appears to have resulted from broad adaptation of highly effective combinations of immunosuppressants (1, 2, 5–7, 9,11,12). Histologic features of PVAN include mononuclear cell infiltrates indistinguishable from those of cellular rejection, tubular epithelial cell cytopathy with nuclear inclusions and nuclear staining by polyomavirus-specific immunohistochemistry or in situ hybridization (3,7,8). If undiagnosed, PVAN may progress to advanced fibrotic injury with little or no viral staining rendering the histological distinction from chronic allograft nephropathy difficult (7,13,14). Although the prevalence of PVAN varies among transplant centers, rates of 5% to 10% have been widely reported and prospective studies examining viral reactivation suggest that up to 30% of KTx recipients are at risk during the first post-transplant year (6,7,9,15,16). Importantly, it now appears that the earliest phase of PVAN (i.e. conversion from urinary viral shedding to sub-clinical invasion of the graft) can be detected by screening of blood for BKV or by surveillance (‘protocol’) biopsy (15–20) and that PVAN may be partly preventable by early adjustment of immunosuppression (15). Nonetheless, many established cases now exist and it is likely that PVAN will continue to be diagnosed in a proportion of recipients.

Reduced immunosuppressive therapy is widely recognized as the cornerstone of treatment for PVAN despite the accompanying risk for acute rejection. Antiviral strategies have also been attempted based on in vitro suppression of viral replication (21,22). Most notable are reported use of low-dose intravenous cidofovir (23–27), oral leflunomide (28) and human immunoglobulin (IVIG) (29). Evidence that such interventions accelerate clearance of BKV from the graft or protect against further loss of graft function compared to immunosuppression reduction alone is tenuous, being based on uncontrolled case series. In general, few published studies have quantified the impact of biopsy-proven PVAN on graft function from the time of diagnosis under one or more therapeutic strategies (2,4–7). We present here an outcome analysis involving a substantial cohort of PVAN cases which clarifies the impact of baseline characteristics as well as of specific treatment strategies on functional trends during the first several years following diagnosis. In addition, we examine the factors associated with early histological clearance of virus from the graft. The results provide an important benchmark for current outcomes of PVAN managed by immunosuppression reduction alone and call into question the efficacy of other available interventions as a means to further stabilize allograft function.

Methods

Study population

Following Institutional Review Board approval of the study, all biopsy-proven cases of PVAN diagnosed at Mayo Clinic between January 2002 and December 2004 were identified. Fifty-five cases occurred in transplants carried out between February 1998 and July 2004. During this period a total of 1268 KTx were carried out at the institution (giving a frequency of 4.3%). Data regarding donor characteristics, pre- and post-transplant clinical parameters, therapeutic intervention, serial laboratory values and pathology reports on the biopsies were recorded.

Overall transplant population and baseline immunosuppression regimens

During the time period studied, all KTx recipients received intravenous methylprednisolone for 4 days. The majority received induction therapy with rabbit anti-thymocyte globulin 1.5 mg/kg/day for a total of 4 to 6 doses or basiliximab 40 mg on days 0 and 4. Oral immunosuppression consisted of prednisone tapered to 5–10 mg/day by 3 months; mycophenolate mofetil 1g or 750 mg twice daily and one of the following: tacrolimus (target trough level 10–15 ng/mL for 3 months and 6–10 ng/mL thereafter), sirolimus (target trough level 15–20 mg/mL for 3 months and 8–15 ng/mL thereafter), microemulsion cyclosporine (target trough level 200–250 ng/mL for 3 months and 150–200 ng/mL thereafter).

Case definition

A diagnosis of PVAN was confirmed for any KTx recipient having a renal allograft biopsy report of characteristic histological abnormalities (tubulointerstitial infiltrates and/or epithelial cell cytopathic changes) with nuclear positivity of multiple epithelial cells by BKV-specific in situ hybridization (ISH). Biopsies were carried out for diagnostic purposes (unexplained rise in serum creatinine—‘non-surveillance biopsy’) as well as for histological surveillance (0, 4, 12 and 24 months post-transplant—‘surveillance biopsy’).

Treatment of PVAN—immunosuppression reduction

Immediate stepwise reduction in oral immunosuppression was carried out in all patients with biopsy-proven PVAN using one of two protocols: (i) Cyclosporine conversion: Mycophenolate mofetil therapy was reduced in stepwise fashion by decrements of 250 mg twice daily every 2 weeks to a final dose of 250 mg twice daily. The majority of cases remained on 250 mg twice daily indefinitely but some were reduced further to 0 mg due to evidence of persistent PVAN or were increased back to 500 mg twice daily following subsequent diagnosis of acute rejection. Tacrolimus or sirolimus were converted to microemulsion cyclosporine with target trough level of 125–175 ng/mL. 2. Reduced baseline immunosuppression: Mycophenolate mofetil dosage was reduced as described above. Tacrolimus and sirolimus target trough levels were reduced to 4–6 ng/mL and 6–8 ng/mL, respectively.

In no case was corticosteroid bolus therapy administered or prednisone dose modified. Additional reductions in target trough levels of tacrolimus, sirolimus or cyclosporine were carried out at later time points in patients with persistent PVAN on repeat biopsy.

Treatment of PVAN—ancillary therapies

  • 1Cidofovir therapy was administered at 0.25 mg/kg in 100 mL of normal saline over 1 h every 2 weeks for a total of four doses following which allograft biopsy was repeated and patients with persistent PVAN received additional doses of 0.5 mg/kg every 2 weeks for four to five doses. In four recipients, cidofovir therapy was initiated at a later time point (66, 98, 140 and 207 days post-diagnosis) and these cases were excluded from Kaplan-Meier analysis of the effect of cidofovir on functional outcome. In the rest of the cohort, the median duration from diagnosis to cidofovir initiation was 4 days (range 1 to 35 days). Cidofovir was first administered at our center to six kidney transplant recipients with PVAN during the second half of 2002. Having observed no adverse clinical effects in these cases, we elected, in January 2003, to include low-dose cidofovir in the management protocol for all newly diagnosed PVAN cases. Thus the majority of non-cidofovir-treated cases were diagnosed in 2002 and the majority of cidofovir-treated cases were diagnosed in 2003 and 2004.
  • 2Intravenous immunoglobulin (a nonsucrose-based formulation) was administered as two doses of 1.25 g/kg with 48 h between doses. In all but 2 of a total of 12 cases, IVIG was administered to patients judged to have combined acute cellular rejection and PVAN on the diagnostic biopsy. The distribution of individual therapeutic interventions within the study cohort is summarized in Table 1.
Table 1. Summary of therapeutic interventions among 55 KTx recipients with biopsy-proven PVAN
 No additional therapyCidofovir aloneIVIG aloneCidofovir + IVIGTotal
Cyclosporine conversion1880430
Reduced baseline immunosuppression5122625

Monitoring and follow-up of PVAN

Following initial diagnosis and management, graft function was followed by serial serum creatinine concentrations (SCr; weekly for 3 months, every 2 weeks for an additional 3 months, monthly thereafter). A second allograft biopsy was obtained at 3–6 months post-diagnosis. When methodology for quantitation of plasma BKV by real-time PCR (qPCR) became available, this assay was carried out at the time of diagnosis and every 2–4 weeks afterward. Serial qPCR measurements were not available on a sufficient number of patients in this cohort to allow a statistical analysis of the influence of individual therapeutic interventions on circulating viral load.

Histological procedures and analysis

All allograft biopsies were interpreted by renal pathologists. In addition to narrative reporting, each biopsy was scored for acute and chronic abnormalities using Banff ‘97 criteria (30). In situ hybridization (ISH) for BKV DNA was performed on paraffin-embedded sections as described (17). For the current study, all ISH slides were reinterpreted by one pathologist (M. L.) unaware of the clinical history. Viral involvement was expressed in two forms: (a) Total number of tubular profiles with one or more cells showing nuclear positivity for BKV. (b) Proportion of tubular profiles with one or more cells showing nuclear positivity for BKV. For the latter, viral involvement was graded as: <5%; 5–10%; 11–20% and >20%. Diagnostic biopsies were also assigned to one of the three disease stages (A, B, C) suggested by Drachenberg and colleagues on the basis of Banff '97 acute and chronic scores and extent of viral cytopathic changes (6,13,14). In the absence of a truly discriminatory test, patients were judged to have combined acute cellular rejection and PVAN on the initial diagnostic or follow-up biopsy if there was high-grade interstitial infiltration and tubulitis (Banff IB) or any degree of intimal arteritis with low-level BKV positivity by in situ hybridization (< 5% of tubular cross-sections). Follow-up biopsies were interpreted as having acute cellular rejection if diagnostic criteria for Banff IA or greater were met and in situ hybridization for BKV was negative.

Data analysis

The study endpoint was date of return to dialysis, or most recent clinical record. Date of diagnosis of PVAN was defined as the date of the first biopsy with ISH positive for BKV. Based on the recorded data, the following four adverse outcomes were assessed for each patient: (a) Graft loss—patient death or return to dialysis. (b) Major functional decline: Graft loss or sustained (persistent for at least 8 weeks) increase in SCr by ≥50% of the value at PVAN diagnosis. (c) Any functional decline: Graft loss, sustained increase in SCr≥20% or sustained increase in SCr≥0.5 mg/dL compared to the value at PVAN diagnosis. The latter criterion was included in order to account for cases in which a sustained rise of 0.5 mg/dL was <20% of baseline (e.g. a rise from 3.0 mg/dL to 3.5 mg/dL). This criterion resulted in inclusion of only one additional case. (d) Failure of early histological clearance: Persistent positive ISH on follow-up biopsy carried out 3–6 months after diagnosis. Kaplan-Meier plots were generated for the cohort with respect to adverse outcomes and were compared by log-rank test. Nadir SCr and SCr at the time of PVAN diagnosis were compared by paired Student's t-test. Cox proportional hazard models were used to determine risk of any functional decline. With 21 events, this study had 80% power to detect hazard ratios of 3.5 or higher (two-sided, alpha = 0.05). Logistic regression models were used to assess risk for failure of early histological clearance.

Results

Clinical characteristics and functional outcomes of overall group

Demographic information and baseline clinical data for the 55 KTx recipients diagnosed with PVAN are summarized in Table 2. Of note, the majority (75%) received living donor transplants and all received induction with anti-thymocyte globulin (85%) or basiliximab (15%). The predominant oral immunosuppressive regimen was prednisone, mycophenolate mofetil and tacrolimus (91%). Table 3 summarizes the clinical and histological presentation of PVAN in these patients. The median interval between KTx and diagnosis of PVAN was 11 months (range 3–65 months). There was an overall increase in SCr associated with diagnosis of PVAN (post-KTx nadir 1.5 ± 0.3 mg/dL compared to 2.1 ± 0.7 mg/dL at diagnosis, p < 0.0001). The diagnosis was made on a clinically indicated (‘non-surveillance’) biopsy in 31 patients (56%) and on surveillance biopsy at 4, 12 or 24 months post-KTx in 24 (44%). Combined initial Banff scores for acute (i + t) and chronic (ci + ct) tubulointerstitial abnormalities were 3.0 ± 1.6 and 2.2 ± 1.5, respectively. The degree of BKV staining could be accurately assessed in 48 of the diagnostic biopsies. The median number of positive tubular profiles was 14.5 (range 1–100). Almost half (48%) could be considered as having low-level viral involvement (<5%). The large majority (85%) of the biopsies fell into the Stage B category of the staging system recently endorsed by a multidisciplinary consensus group (6). At the time of PVAN diagnosis, plasma viral load was measured in 31 patients (56%) with a median of 166 000 copies/mL (range 470–1 624 500). Acute cellular rejection developed in 9 recipients (16%) between 3 and 6 months after immunosuppression reduction. Post-diagnosis follow-up for the entire cohort was 19.7 ± 11 months. During this time, 8 recipients (15%) suffered graft loss, 13 (24%) had a major decline in graft function and 21 (38%) had any graft functional decline. Rates of graft loss, major functional loss and any functional loss following PVAN diagnosis are represented as Kaplan-Meier plots in Figure 1.

Table 2. Characteristics of 55 KTx recipients subsequently diagnosed with PVAN
Age (years) mean ± SD52.9 ± 12.3
Male gender number (%)36 (65)
Cause of ESRD number (%)
 Diabetic nephropathy15 (27)
 Hypertensive nephrosclerosis7 (13)
 Glomerulonephritis15 (27)
 Polycystic kidney disease4 (7)
 Other14 (26)
Living donor transplant number (%)41 (75)
Previous kidney or other solid organ transplant number (%)19 (35)
Delayed graft function, number (%)15 (27)
Induction immunosuppression, number (%)
 Anti-thymocyte globulin47 (85)
 Basiliximab8 (15)
Oral immunosuppression prior to PVAN diagnosis, number (%)
 Prednisone55 (100)
 Mycophenolate mofetil53 (96)
 Tacrolimus50 (91)
 Cyclosporine1 (2)
 Sirolimus6 (11)
Nadir serum creatinine (mg/dL), mean ± SD1.5 ± 0.3
Table 3. Clinical and histological characteristics of PVAN in 55 KTx recipients
Months to PVAN diagnosis, median (range)11 (2.9–65.2)
Serum creatinine (mg/dL) at diagnosis, mean ± SD2.1 ± 0.7
Diagnosis made on surveillance biopsy, number (%)24 (44)
Histologic findings at diagnosis – Banff '97 scores
 i + t, mean ± SD  3 ± 1.6
 ci + ct, mean ± SD2.2 ± 1.5
Drachenberg staging of PVAN (n = 53), number (%)
 Stage A5 (9)
 Stage B45 (85)
 Stage C3 (6)
Histologic severity of BKV infection by ISH (n = 48)
 Total ISH+ve tubular cross sections, median (range)14.5 (1–100)
 <5% Tubular positivity23 (48)
 6–10% Tubular positivity10 (20)
 11–20% Tubular positivity11 (23)
 >20% Tubular positivity4 (8)
Plasma BKV (copies/mL) at diagnosis (n = 31) median (range)166 000 (470—1 624500)
Months of follow-up mean ± SD19.7 ± 11
Acute cellular rejection after PVAN diagnosis, number (%)9 (16)
Figure 1.

Adverse functional outcomes of kidney transplants from the time of PVAN biopsy diagnosis: Kaplan-Meier plots of three levels of graft functional decline are shown from the time of PVAN biopsy diagnosis for a cohort of 55 kidney transplant recipients. The three plots represent rates of occurrence of graft loss (return to dialysis; solid line), major functional decline (graft loss or ≥50% sustained rise in serum creatinine concentration (SCr); dotted line) and any functional decline (graft loss, ≥20% sustained rise in SCr or ≥0.5 mg/dL sustained rise in SCr; dashed line)

Determinants of graft functional outcome following PVAN diagnosis

Kaplan-Meier plots of the occurrence of ‘any functional decline’ following PVAN diagnosis were constructed for patients managed with and without cidofovir, cyclosporine conversion and IVIG (Figure 2). As shown, none of these individual therapeutic interventions was associated with improved graft functional outcome. Since all patients were managed by immunosuppression reduction with or without cidofovir, these two groups were compared (Table 4). The two groups had closely comparable baseline clinical characteristics, histologic findings and viral load at time of diagnosis. Mean follow-up was greater than 1 year in both groups and was longer in the immunosuppression reduction alone group. Cidofovir-treated patients were more likely to receive concomitant IVIG and less likely to be managed by conversion to cyclosporine.

Figure 2.

Rates of occurrence of any functional decline with and without specific interventions following PVAN diagnosis: Kaplan-Meier plots are shown for rate of occurrence of any functional decline (loss, ≥20% sustained rise in SCr or ≥0.5 mg/dL sustained rise in SCr) from the time of PVAN diagnosis. Individual plots compare event-free survival for cases managed with and without: A. Low-dose cidofovir infusions (excluding four patients who received cidofovir 2 or more months after PVAN diagnoses). B. Conversion to cyclosporine therapy. C. Treatment with IVIG. In each case, p value for statistical difference between the two plots by log-rank test is shown.

Table 4. Comparison of clinical and histological variables and viral load between cidofovir-treated and noncidofovir-treated kidney transplant recipients with PVAN
 Cidofovir-treated (n = 30)Noncidofovir-treated (n = 25)p
  1. * Data were available on all 30 cases in the cidofovir-treated group and on 23 of the noncidofovir-treated controls.

Age at PVAN diagnosis (years), mean ± SD54 ± 1252 ± 130.5
Male gender, number (%)22 (73)14 (56)0.18
Deceased donor transplant, number (%)10 (33)4 (16)0.14
ATG Induction, number (%)27 (90)20 (80)0.3
Time to PVAN diagnosis (months), mean ± SD16 ± 1512 ± 110.2
Surveillance biopsy diagnosis, number (%)11 (37)13 (52)0.25
Nadir serum creatinine (mg/dL), mean ± SD1.5 ± 0.31.6 ± 0.40.25
Serum creatinine at diagnosis (mg/dL), mean ± SD2.1 ± 0.52.2 ± 0.90.8
Histologic findings at diagnosis - Banff '97 scores
 i + t, mean ± SD2.8 ± 1.73.2 ± 1.40.4
 ci + ct, mean ± SD2.1 ± 1.62.3 ± 1.30.6
Drachenberg staging of PVAN, number (%)* 0.9
 Stage A4 (13)1 (4) 
 Stage B23 (77)22 (96) 
 Stage C3 (10)0 (0) 
Total ISH+ve tubular cross-sections, median (range)16 (3–50)11 (2–100)0.13
Median BK viral load at diagnosis (copies/mL)170,000117,2500.3
Cellular rejection after IS reduction, number (%)6 (20)3 (12)0.41
Conversion to cyclosporine, number (%)12 (40)18 (72)0.03
IVIG use, number (%)10 (33)2 (8)0.02
Follow-up months, mean ± SD14 ± 926 ± 100.001

In order to more clearly identify clinical predictors of declining graft function following diagnosis of PVAN, a Cox proportional hazard model was employed using ‘any functional decline’ as the primary outcome. This analysis is summarized in Table 5 in the form of hazard ratios (HR) for an array of variables including pre-transplant characteristics, histological indices, initial plasma viral load and therapeutic interventions. The HR reached statistical significance for only one of these variables—combined interstitial fibrosis and tubular atrophy scores (ci + ct) >2 on the diagnostic biopsy (HR 2.54, p = 0.05). There was no observed association between degree of BK tubular involvement, initial plasma viral load or specific therapeutic intervention (low-dose cidofovir, IVIG and cyclosporine conversion) and subsequent functional decline.

Table 5. Relationship between a panel of potential prognostic factors and risk for any functional decline and failure of viral clearance on follow-up biopsy 3 to 6 months following diagnosis of PVAN
FactorHazard ratio for any functional decline (n = 55)pOdds ratio of failure of viral clearance (n = 45)p
Age >550.69 (0.28–1.63)0.401.07 (0.33–3.33)0.90
Male gender0.77 (0.33–1.90)0.560.24 (0.06–0.85)0.03
Deceased donor source1.45 (0.52–3.60)0.468.25 (1.28–60.2)0.01
Delayed graft function1.14 (0.41–2.81)0.793.75 (0.99–16.43)0.06
Cellular rejection prior to BK0.69 (0.11–2.38)0.600.74 (0.09–4.90)0.75
Diagnosis by non-surveillance biopsy2.30 (0.93–6.45)0.074.48 (1.29–17.6)0.02
Histological indices (diagnostic biopsy)
 i + t >20.84 (0.34–2.17)0.700.86 (0.24–3.01)0.81
 ci + ct >22.54 (0.98–6.30)0.058.2 (1.74–60.3)0.02
 >5% of tubules BKV+ve by ISH1.89 (0.76–5.09)0.175.4 (1.47–22.7)0.01
Initial plasma BKV >1.5 × 105 copies/mL1.33 (0.40–6.00)0.6618 (2.4–384)0.01
Treatment strategy
 Low-dose cidofovir1.24 (0.52–3.00)0.621.50 (0.44–5.23)0.52
 IVIG0.84 (0.24–2.26)0.740.40 (0.08–1.72)0.24
 Conversion to CyA1.03 (0.42–2.65)0.951.99 (0.57–7.38)0.29

Determinants of histological outcome following PVAN diagnosis

Of the 45 follow-up biopsies performed, 21 (47%) were positive for BKV by ISH. By logistic regression analysis, a number of variables were associated with significantly increased risk of persistent BKV positivity (Table 5). These included deceased donor source (odds ratio (OR) 8.25, p = 0.01), diagnosis by non-surveillance biopsy (OR 4.48, p = 0.02), initial blood viral load >150,000 copies/mL (OR 18, p = 0.01), ci + ct >2 on the diagnostic biopsy (OR 8.2, p = 0.02) and >5% of tubules positive by ISH on the diagnostic biopsy (OR 5.4, p = 0.01). Male gender was associated with decreased risk of viral persistence (OR 0.24, p = 0.03). There was no relationship observed between specific therapeutic interventions and risk for persistent BKV positivity on follow-up biopsy.

Discussion

This study reveals factors associated with graft functional stability and early elimination of BKV from the graft in KTx recipients with PVAN. The results suggest that a sustained rise in SCr can be avoided in the majority of affected individuals up to 3 years following diagnosis. While a variety of clinical and histological factors were found to influence viral elimination from the graft, the degree of fibrosis at diagnosis was most predictive of subsequent functional decline. We found no evidence in this cohort for a beneficial effect of ancillary interventions compared to reduced immunosuppression alone.

The potential for PVAN to result in very high rates of graft loss is well documented (1–3,11). More recently, Ramos et al. reported 16.4% graft loss during an average of 12.6 months follow-up among 67 patients diagnosed between 1997 and 2001 (4). Kaplan-Meier analysis, however, indicated ongoing attrition, predicting approximately 40% loss at 3 years. Other recent reports have indicated more favorable graft survival albeit with smaller case numbers (31–34). Among the published series, the occurrence of lesser degrees of functional loss has not been examined. In the current study, we depict trends in renal function over time following biopsy diagnosis of PVAN. The results indicate that graft loss by 3 years may be avoided in up to 80% of cases and that additional increase in SCr may be prevented in 60%. Perhaps in contrast to PVAN diagnosed in the 1990s, accurate diagnosis and prompt immunosuppression reduction was invariable among the patients reported here and a notable proportion (44%) was diagnosed on surveillance biopsies. The improved functional outcome for PVAN may, therefore, be attributed to increased awareness of the risk for BKV reactivation, development of protocols for early diagnosis of BKV reactivation and immediate stepwise reduction of immunosuppression with further modifications guided by early re-biopsy. It should be emphasized, however, that longer follow-up will determine whether grafts with less severe early functional loss will eventually succumb to late acute rejection, progression of interstitial fibrosis or recurrent viral nephropathy.

Reduced immunosuppression is widely accepted as the first line of management for PVAN. Although subsequent acute rejection may occur in 10% to 30% of cases (2,3), reports of patients in whom immunosuppression was unchanged or increased indicate very poor outcomes (3,35). Individual protocols for reducing immunosuppression have not been directly compared. Conversion from tacrolimus to cyclosporine was carried out in a number of our patients as well as in other reported series with the rationale that this agent may be less likely to promote BKV reactivation (36–38). It has since become clear, however, that BK viremia and PVAN may occur at similar rates among cyclosporine and tacrolimus-treated patients (15,39). In keeping with this, we have not observed improved functional outcomes associated with cyclosporine conversion. Although IVIG may contain high levels of anti-BKV antibodies (29), our analysis of recipients receiving IVIG following diagnosis of PVAN also does not indicate a robust effect on functional stability or on viral elimination from the graft. Whether conversion from calcineurin inhibitor to mTOR inhibitor-based immunosuppression has specific functional benefit in PVAN (40) has not been tested here but may merit consideration for cases with progressive decline in GFR.

A number of published reports describe cases of PVAN treated intermittently with cidofovir at 0.25–3.0 mg/kg (2,23,24,26,27,41). Resolution of viremia and reduction of urinary viral load have been documented—typically occurring gradually over several months. In some reports, stabilization of SCr was observed in patients for whom prior immunosuppression reduction was associated with ongoing functional decline (24,26,27). In others, however, reduced immunosuppression and cidofovir were initiated simultaneously or functional decline occurred despite cidofovir therapy (2,41). Recently, Kuypers et al. compared outcomes for 8 PVAN cases managed by immunosuppression reduction and cidofovir with those of a group managed by immunosuppression reduction alone (25). Over 4–40 months follow-up, no allograft loss occurred in the cidofovir-treated recipients compared to 9 of 13 managed without cidofovir. Although the results are compatible with a true benefit for cidofovir, it should be noted that renal functional loss was already greater at the time of diagnosis in the noncidofovir-treated patients. In addition, acute rejection prior to diagnosis and complete cessation of MMF occurred more frequently in the cidofovir-treated group. Our results, in larger groups of patients, do not suggest that low-dose cidofovir provides additional stabilization of graft function or early histological clearance of virus and emphasize the need for randomized trials of cidofovir and other potential antiviral agents in newly diagnosed PVAN or BK viremia.

An array of factors was examined for influence on graft functional stability and on histological persistence of BKV. In the case of viral persistence on follow-up biopsy, there were several indices with predictive value. These included measures of viral activity (initial results of ISH and quantitative PCR of blood) and chronic graft injury (ci + ct >2) as well as baseline transplant characteristics (donor source and recipient gender). Although not the focus of the current study, further analysis of the evolution of histological changes on serial biopsies from this group of patients may shed additional light on the role of subclinical PVAN in the pathogenesis of chronic allograft nephropathy (13). Analysis of post-diagnosis decline in graft function demonstrated that only the degree of interstitial fibrosis and tubular atrophy on diagnostic biopsy was clearly predictive. This observation is in agreement with Drachenberg et al., who demonstrated that advanced interstitial fibrosis and tubular atrophy at diagnosis correlated with poorer graft outcome (13). The strong influence of fibrotic injury on functional prognosis highlights a growing recognition that screening for BKV reactivation and development of protocols for early intervention represents the best strategy for reducing the impact of PVAN on graft loss (6,15,17,19,42,43), as well as the most important context for testing the efficacy of antiviral agents. We can infer from our findings as well as from those of others, that early diagnosis, rather than the therapeutic modality used, is a more important determinant of functional outcome. Another important implication of our results is that trends in viral activity—whether measured in urine, blood or graft parenchyma (18)—may not correlate well with functional prognosis for biopsy-proven PVAN when KTx recipients with varying degrees of chronic graft injury at the time of diagnosis are studied.

Although we report one of the largest comprehensively followed series of PVAN cases to date, the study should be recognized as being limited by its retrospective nature. Patients were assigned to one or more treatment interventions in nonrandomized fashion using inclusion criteria that evolved along with our clinical experience and an emerging literature. Event numbers for functional and histological outcomes were such that important modifiers of these outcomes may not have been detected (type II statistical error). Due to the limited number of cases that were managed without cyclosporine conversion, IVIG or low-dose cidofovir, we also cannot fully exclude the possibility that all treatment strategies affected the functional outcome equally making any beneficial effects difficult to detect. In the case of low-dose cidofovir therapy, un-treated and treated patients were predominantly diagnosed before and after December 2002, respectively. While we acknowledge that this temporal pattern has the potential to introduce bias, we would emphasize that all the reported cases were diagnosed during a period of time in which clinical and histological features of PVAN were well recognized in our practice and surveillance biopsies were being routinely carried out. In addition, comparison of the two groups demonstrated close similarity for clinical and histological measures of disease severity at diagnosis. Nonetheless, while our data suggest that there is no robust effect of low-dose cidofovir to retard the progression of graft functional decline, it should not be concluded that this therapy is ineffective in the management of BKV-associated disease in the absence of well-designed prospective clinical trials.

Our findings enhance the current state of knowledge regarding PVAN by: (a) Demonstrating that there has been a reduction in overall rate of graft loss within 3 years of diagnosis. (b) Documenting the current frequency of post-diagnosis functional deterioration. (c) Confirming the central role of graft fibrosis in predicting functional decline. (d) Identifying factors associated with intra-graft persistence of BKV following initial interventions. (e) Providing a time-to-event analysis of the impact of specific interventions on functional stability that calls into question their therapeutic efficacy in unselected cases.

Ancillary