Polyomavirus-associated nephropathy risk in kidney transplants: the influence of recipient age and donor gender

original article

http://www.kidney-international.org & 2007 International Society of Nephrology

Polyomavirus-associated nephropathy risk in kidney transplants: the influence of recipient age and donor gender HA Khamash1, HM Wadei1, AS Mahale1, TS Larson1, MD Stegall2, FG Cosio1 and MD Griffin1 1 2

Division of Nephrology and Hypertension, Department of Medicine, Mayo Clinic College of Medicine, Rochester, Minnesota, USA and Division of Transplantation Surgery, Department of Surgery, Mayo Clinic College of Medicine, Rochester, Minnesota, USA

Polyomavirus-associated nephropathy (PVAN) is a frequent cause of kidney transplant failure. We determined the risk factors for biopsy-proven PVAN among 1027 recent kidney transplant recipients by univariate and multivariate analyses. The rate of PVAN was determined over an univariate and multivariate analysis over an average of 30 months of follow-up of patients receiving predominantly living donor grafts with antibody induction and sequential surveillance biopsies to detect subclinical graft disease. Seventy-four transplant recipients were diagnosed with PVAN with the finding made on surveillance biopsy in 40 patients. These 40 cases did not differ from the 34 non-surveillance cases with respect to baseline clinical characteristics or initial histological features. Older recipient age and female donor gender were independent risks associated with PVAN. Factors not linked to PVAN risk included the use and type of induction agent, use of tacrolimus vs sirolimus, the number of human lympocyte antigen (HLA) mismatches, or the frequency of acute rejection. We conclude that PVAN preferentially affects older age patients and allografts from female donors but is unrelated to immunological risk, choice of immunosuppression, or rejection history. Kidney International (2007) 71, 1302–1309; doi:10.1038/sj.ki.5002247; published online 4 April 2007 KEYWORDS: gender difference; immunosuppression; renal transplantation; risk factors; polyomaviruses

Correspondence: MD Griffin, Mayo Clinic, Division of Nephrology and Hypertension, 200 First St SW, Charlton 10 Transplant Center, Rochester, Minnesota 55905, USA. E-mail: [email protected] Received 20 March 2006; revised 31 January 2007; accepted 21 February 2007; published online 4 April 2007 1302

Polyomavirus-associated nephropathy (PVAN) owing to BK virus (BKV) is now a frequent cause of kidney transplant (KTx) injury and failure worldwide.1–4 Although BKV was recognized as a commonly reactivated virus within the uroepithelium of immunosuppressed individuals over 30 years ago,5,6 the emergence of PVAN as a clinicopathological entity in 2% to 10% of KTx recipients has occurred only within the past 8 years.1–4,7–12 The primary reason for this apparent shift in the pathogenicity of BKV reactivation is generally accepted to be the introduction of highly potent immunosuppressive drug combinations into clinical practice during the same time period.1–4 Despite this assumption, however, current evidence does not favor a specific role for induction therapy or for individual oral immunosuppressants in enhancing PVAN risk.1,8,10–14 Furthermore, careful analysis of KTx populations managed no differently to those from the so-called ‘cyclosporine era’ reveals BK viremia, and PVAN rates similar to those managed by more recently adopted regimens.10,14 In addition, while some centers have observed a relationship between immunosuppression boosts for acute rejection and subsequent development of PVAN,1,9 this practice has clearly become less prevalent with the decline in acute rejection rates and should not be responsible for increased incidence of PVAN. Thus, it is worth considering the possibility that additional factors may be at play in the dramatic increase in PVAN and that identification of such factors may contribute, not just to a reduction in the burden of BKV-associated disease, but also to effective tailoring of antirejection therapy to minimize the consequences of overimmunosuppression among vulnerable KTx recipients in the current era. In common with many other centers, our KTx program experienced a surge in PVAN cases between 2000 and 2004 that coincided with the use of immunosuppressive regimens designed to achieve very low acute rejection rates.15 During that same time period, our practice was also characterized by increases in overall transplant numbers, living donation rates, recipient age, and immunological complexity as well as by the adoption of a surveillance (‘protocol’) biopsy program that has facilitated diagnosis of subclinical graft disease.12,15–17 In this study, we have had the opportunity to examine the Kidney International (2007) 71, 1302–1309

original article

HA Khamash et al.: Polyomavirus nephropathy risk factors

association between biopsy-proven PVAN and an array of potential risk factors among our entire cohort of recently transplanted patients. The results of this analysis indicate that, for the type of practice we have chosen to adopt, PVAN has preferentially affected recipients of older age and also allografts from female donors. In contrast, indices reflecting choice of immunosuppressive agent, recipient immunological risk, and actual immunologic complications were not clearly associated with heightened PVAN risk. Taken in the context of existing literature, the findings highlight the fact that PVAN risk factors differ among transplant centers in a manner that appears to reflect donor/recipient characteristics as well allocation and immunosuppression practices. RESULTS Trends in biopsy-proven PVAN among total KTx recipients

The characteristics of the overall study population are displayed in Table 1. During a mean follow-up period of 30717 months, 97 KTx recipients died (9%) and an additional 103 (10%) suffered allograft loss. A total of 74 patients received a biopsy diagnosis of PVAN (7%) at a mean of 12711 months post-transplant (range 2–63). To date, the proportions of patients transplanted in 2000, 2001, 2002, 2003, and 2004 that have developed biopsy-proven PVAN are 8%, 10.5%, 7%, 4%, and 5% respectively. Fifty-four percent of the cases (40) were diagnosed by surveillance biopsy at 4 months (27%), 12 months (23%), 24 months (2.7%), or 60 months (1.3%) post-transplant with the remaining 34 cases (46%) diagnosed on biopsies performed to investigate raised serum creatinine (non-surveillance biopsy). As illustrated in Figure 1a, surveillance biopsy-diagnosed and non-surveillance biopsy-diagnosed cases of PVAN were predominantly diagnosed within the first 2 years post-KTx. The average time from transplantation to diagnosis of PVAN was lower in those diagnosed by surveillance biopsy (10710 months)

Table 1 | Baseline characteristics of patients receiving kidney transplants between January 2000 and June 2004 Number Donor source, n (%) Living related Living unrelated Deceased

524 (51) 298 (29) 205 (20)

Type of transplant, n (%) Conventional ABO incompatible Positive cross-match

900 (85) 50 (6) 77 (8)

Recipient age (years), mean7s.d. Recipient gender, % female Recipient race, % Caucasian Pretransplant diabetes mellitus, % Pretransplant dialysis, % Donor age (years), mean7s.d. Donor gender, % female Total HLA mismatches, mean7s.d. First transplant, %

50714 43 90 33 60 41713 52 2.971.7 78

HLA, human lympocyte antigen.

Kidney International (2007) 71, 1302–1309

1027

compared with those diagnosed by non-surveillance biopsy (14711 months) but not significantly so (P ¼ 0.09). The incidence of PVAN did not differ among KTx recipients managed with Thymoglobulin compared with anti-IL-2 receptor antibody induction (Figure 1b). Although there was a trend toward higher incidence among recipients managed with tacrolimus- compared with sirolimus- or cyclosporine-based immunosuppression (Figure 1c), this did not reach statistical significance. Clinical characteristics of KTx recipients with and without PVAN

The clinical characteristics of KTx recipients with and without biopsy-proven PVAN are summarized in Table 2. As can be seen, compared to patients without PVAN, those diagnosed with PVAN were significantly older at the time of transplantation and were more likely to receive a transplant from a female donor, whereas there were no significant differences between the two groups for donor source, donor age, total HLA mismatches, frequency of acute rejection before the diagnosis of PVAN. Table 2 also indicates that the groups were closely comparable with regard to the use and/or type of induction therapy as well as for the proportionate use of tacrolimus, sirolimus and cyclosporine. In addition, there were no significant differences between the two groups for frequency of preemptive transplantation or re-transplantation, for type of transplant (conventional vs ABO incompatible/positive cross-match), for race of either the donor or recipient, or for occurrence of delayed graft function (data not shown). The clinical characteristics of patients diagnosed with PVAN by surveillance biopsy were compared with those diagnosed by non-surveillance biopsy. None of the characteristics listed in Table 2 differed significantly between these two groups of patients (data not shown). The surveillance biopsy and non-surveillance biopsy-diagnosed cases were also compared for acute and chronic tubulointerstitial abnormalities of the diagnostic allograft biopsy using relevant Banff ’97 scores (i þ t for acute infiltrates and tubulitis, ci þ ct for interstitial fibrosis and tubular atrophy) and the PVAN grading system devised by Drachenberg et al.1 (A, B, and C grades). As shown in Figure 2a, the proportion of cases with no, mild, moderate, and severe acute abnormalities were not different between the two groups (P ¼ 0.3). A trend toward more frequent occurrence of moderate or severe chronic abnormalities among the non-surveillance biopsy-diagnosed cases was present but did not reach statistical significance (P ¼ 0.15). The distribution of Drachenberg grades did not differ between the two groups (P ¼ 0.7) with the large majority of each meeting criteria for grade B disease (Figure 2b). In contrast to baseline clinical and histological features, graft survival during the first 4 years of post-KTx was clearly superior for PVAN cases diagnosed by surveillance compared with non-surveillance biopsy (Figure 3). In keeping with our previous findings,15 we interpreted these analyses as indicating that PVAN diagnosed by surveillance biopsy, while associated with improved disease outcome, does not 1303

original article

HA Khamash et al.: Polyomavirus nephropathy risk factors

15%

b

All cases Surveillance biopsy cases Non-surveillance biopsy cases

10%

Percent of total transplants

Percent of total transplants

a

n = 1027

5%

15%

Anti-CD25 (n = 99)

10% Thymoglobulin (n = 834)

5%

0% 0

12 24 36 Months post-transplant

Percent of total transplants

c

48

0

12 24 36 Months post-transplant

48

15% Tacrolimus (n = 834) 10% Sirolimus (n = 93) 5% Cyclosporine (n = 23) 0

12 24 36 Months post-transplant

48

Figure 1 | The incidence of PVAN among recent KTx recipients. Kaplan–Meier curves are shown for (a). The cumulative occurrence of biopsy-proven PVAN among all 1027 patients receiving KTxs at Mayo Clinic between January 2000 and September 2004 and followed for up to 48 months post-transplantation. Cumulative occurrence of cases diagnosed by surveillance biopsy and non-surveillance biopsy is also shown. (b) KTx recipients managed with thymoglobulin compared with anti-IL2 receptor antibody (anti-CD25). (c). KTx recipients managed with tacrolimus, sirolimus, or cyclosporine as primary oral immunosuppressive agent. For (b) and (c) the number of KTx recipients in each subgroup at 0 months post-transplant is shown. There were no statistically significant differences for PVAN incidence among the subgroups analyzed.

Table 2 | Clinical characteristics of kidney transplant recipients without and with biopsy-proven PVAN during follow-up No PVAN (n=953) Donor source (%) Living related Living unrelated Deceased Recipient age (years), mean7s.d. Recipient older than 55 years, % Recipient gender, % female Diabetes mellitus (pre- or post-KTx), % Diabetes mellitus (post-KTx only)c, % Donor age (years), mean7s.d. Donor gender, % female Total HLA mismatches, mean7s.d. Acute rejection, %e Induction (none/ATG/anti-IL2R), % Oral mmunosuppression (TAC/SRL/CYA), %

PVAN (n=74)

P-value NSa

51% 29% 20%

55% 24% 21%

50715 45% 44% 67% 22% 41713 51% 2.971.7 18% 7%/83%/10% 87%, 11%, 2%

55713 62% 35% 68% 29% 44712 66% 3.071.7 15% 7%/80%/13% 92%, 7%, 1%

o0.001b 0.007a NSa NSa NSa NSb 0.01a NSd NSa NSa NSa

CYA, cyclosporine; HLA, human lympocyte antigen; SRL, sirolimus; TAC, tacrolimus. a 2 w. b Student t-test. c Defined according to American Diabetes Association diagnostic criteria. d Non-parametric t-test (Mann–Whitney). e For PVAN cases the analysis includes biopsy-proven clinical, subclinical and borderline acute rejection episodes that occurred prior to the diagnosis of PVAN and were treated with corticosteroid with or without depleting antibody.

represent a distinctly different group with regard to clinical characteristics or histological features. Hence, further univariate and multivariate risk factor analyses were performed for the entire cohort of 74 biopsy-proven cases. 1304

Recipient age and donor gender as risk factors for PVAN

The relationships between PVAN diagnosis during follow-up and recipient age and donor gender were demonstrable by univariate Cox regression analysis (Table 3). In contrast, the Kidney International (2007) 71, 1302–1309

original article

HA Khamash et al.: Polyomavirus nephropathy risk factors

a

i+t (P = 0.3)

ci+ct (P = 0.15)

100% 80%

>4 3 or 4

60%

1 or 2

40%

0 20% 0% Surv.

b

Non-surv.

Surv.

Non-surv.

Drachenberg grade (P = 0.7) 100% C B

80% 60%

A

40% 20% 0% Surv.

Non-surv

Figure 2 | Histological characteristics of surveillance and non-surveillance biopsies diagnostic of PVAN among the study cohort. (a) Results for the sum of acute tubulointerstitial abnormalities (Banff ’97 i þ t scores) and chronic tubulointerstitial abnormalities (Banff ’97 ci þ ct scores) on PVAN diagnostic biopsies are represented graphically for cases diagnosed by surveillance biopsy (surv.) or by non-surveillance biopsy (non-surv). Results are presented for i þ t and ci þ ct as the proportions of each subgroup with no abnormality (combined score 0), mild abnormality (combined score ¼ 1 or 2), moderate abnormality (combined score ¼ 3 or 4), or severe abnormality (combined score 44). (b) Results for Drachenberg grading (A, B, or C) of PVAN diagnostic biopsies are represented graphically for cases diagnosed by surveillance biopsy (surv.) or by non-surveillance biopsy (non-surv.). The P-values shown are the result of Fisher exact tests of combined Banff ’97 score and Drachenberg grade distributions between surv. vs non-surv.

Death-censored graft survival

100%

80%

60%

40%

20%

Surveillance biopsy cases Non-surveillance biopsy cases

0% 24 12 Months post-transplant

36

48

Figure 3 | Graft survival for KTx recipients diagnosed with PVAN by surveillance biopsy and non-surveillance biopsy. Kaplan–Meier curves representing the rates of death-censored renal allograft survival (defined as ‘return to dialysis or re-transplantation’) from the time of transplantation to 48 months post-KTx are shown for recipients diagnosed with PVAN by surveillance biopsy (n ¼ 40) or by non-surveillance biopsy (n ¼ 34). Po0.01 for surveillance vs non-surveillance. Kidney International (2007) 71, 1302–1309

following variables did not relate significantly with PVAN: the occurrence of acute rejection episodes before the diagnosis of PVAN; the use and type of induction agent; the type of maintenance immunosuppressive regimen (all shown on Table 3); the type of donor (living or deceased); the recipient sex or race; the recipient history of diabetes and/or dialysis pretransplant; the type of transplant (conventional vs ABO incompatible vs positive cross-match); the occurrence of delayed graft function; number of HLA mismatches; or the graft function at 3 weeks after transplantation (data not shown). In multivariable Cox regression analysis, both recipient age and donor gender retained statistically significant and independent associations with PVAN risk, while variables potentially indicative of risk for or occurrence of acute rejection before PVAN diagnosis were, again, unrelated to PVAN risk (Table 3). Figures 4 and 5 display the relationships between recipient age at transplantation and donor gender with PVAN in the form of Kaplan–Meier plots. Figure 4 demonstrates the incidence of PVAN in recipients within five different age cohorts based on age at the time of transplantation. As can be seen, increasing age, particularly greater than 55 years, was associated with increased cumulative incidence of PVAN. The relative incidence of PVAN among recipient age cohorts was calculated by Cox regression using the group of recipients between 36 and 45 (n ¼ 212) as a reference group. Compared with this reference group, the incidence of PVAN was not statistically different in recipients younger than 35 years (n ¼ 168) (hazard ratios (HR) ¼ 1.1 (0.4–3.0), P ¼ NS) and in recipients between 46 and 55 years (n ¼ 239) (HR ¼ 2.0 (0.87–4.6), P ¼ 0.10). However, recipients aged between 56 and 65 years (n ¼ 250) and older than 65 years (n ¼ 157) had a significantly higher incidence of PVAN compared with the reference cohort (HR ¼ 2.6 (1.2–5.8), P ¼ 0.01 and HR ¼ 2.7 (1.2–6.3), P ¼ 0.02, respectively). The incidence of PVAN was higher in recipients of female kidneys (Figure 5, P ¼ 0.016 log rank) and this relationship was independent of the age of the recipient (Table 3). Quantitatively, 26 of 464 (5%) recipients of male kidneys and 48 of 478 (9%) recipient of female kidneys developed PVAN (P ¼ 0.01, w2). This association was noted in living donor recipients (4.8% vs 9.3%, P ¼ 0.014) and a similar trend was observed in recipients of deceased donor kidneys (6.3% vs 8.6%), although the latter did not reach statistical significance. It should be noted that the sample size of deceased donor recipients in this study is relatively small (n ¼ 207 and 15 cases of PVAN) and that the number of female donors is significantly lower among deceased donors (45%) than among living donors (53%, P ¼ 0.04). DISCUSSION

The study reported here provides a detailed profile of factors associated with increased risk of PVAN among patients receiving KTx at our center during the past 5 years. The results reveal both similarities to and differences from existing reports.1–4,8–10,14,18–21 In support of the clinical 1305

original article

HA Khamash et al.: Polyomavirus nephropathy risk factors

Table 3 | Cox regression analysis of the association between development of PVAN during follow-up and three clinical variables in patients receiving kidney transplants between January 2000 and June 2004 Univariate analysis HR (95% CI) a

Recipient age Donor female gender Acute rejection episodeb Induction therapyc Maintenance immunosuppressiond

1.3 1.8 0.9 1.0 0.7

Multivariate analysis P-value

(1.2–1.4) (1.1–2.9) (0.4–1.6) (0.65–1.5) (0.38–1.15)

HR (95% CI)

0.004 0.01 NS NS NS

1.3 1.8 1.02 1.47 0.64

P-value

(1.2–1.4) (1.1–2.8) (0.54–1.9) (0.9–2.42) (0.37–1.1)

0.001 0.02 NS NS NS

CI, confidence intervals. a Hazard ratios (HR) were calculated for every 10 years increase in age. b Entered as number of rejection episodes that occurred during the first year and, in patients with PVAN, only those that occurred before the diagnosis of PVAN. c These analyses tested no induction vs thymoglobulin induction vs anti-CD25 induction therapies. d These analyses tested tacrolimus vs sirolimus vs cyclosporine for maintenance immunosuppression.

15%

15%

> 65 years

P = 0.016

Percent of total transplants

46 − 55 years

10%

36 − 45 years 5% < 35 years

Percent of total transplants

Female donor 56 − 65 years

10% Male donor

5%

0% 0

12

24

36

48

Months post-transplant

Figure 4 | Risk of PVAN among five recipient age cohorts. Kaplan–Meier plots of the cumulative incidence of PVAN in recipients classified according to their age at the time of transplant as indicated in the figure margin (log rank ¼ 0.03). Comparisons among groups are indicated in the results section of the manuscript.

relevance of the findings, we have had the opportunity to follow a large cohort of recent recipients in which the cumulative frequency of biopsy-proven PVAN was substantial and occurred predominantly during the first and second post-transplant years. Thus, the statistical analyses can be considered robust and the results reflective of recent practice. The broad use of surveillance biopsies among this cohort has also, we believe, assured a low likelihood of missed cases or misidentification of advanced cases as chronic allograft nephropathy.21 Importantly, clinical and histological comparison of those cases diagnosed by surveillance and nonsurveillance biopsies clarify the fact that the former do not represent a distinct group of patients or constitute incidental detection of a disease process devoid of the inflammatory and fibrotic features that characterize clinically evident PVAN. Finally, the characteristics of our program during the period of observation reflect some emerging trends in KTx clinical practice including rising living donation rates, advancing 1306

0

12

24

36

48

Months post-transplant

Figure 5 | The incidence of PVAN among recipients of female donor compared to male donor kidneys. Kaplan–Meier curves are shown of the cumulative occurrence of biopsy-proven PVAN among patients receiving KTxs from female donors compared with male donors during the first 48 months post-transplant. There was a significantly higher incidence of PVAN among recipients of female kidneys (log rank ¼ 0.016).

recipient age, development of antibody reduction/desensitization strategies, and the use of non-calcineurin inhibitorbased immunosuppression.22–25 The analysis of demographic and clinical indices from this cohort indicates, quite clearly, that two baseline factors – older recipient age and female donor gender – were associated with increased PVAN risk, whereas, within the limits of the number of patients studied, other candidate factors – most notably HLA matching, immunosuppression choice, and rejection history – were not. A variety of risk factors have been linked with PVAN diagnosis and with BKV reactivation in clinical KTx centers from North America, Europe, Australia, and Asia.1,10,20,21,26–28 In common with the majority of existing studies, we do not find compelling evidence that induction therapy or individual oral immunosuppressive medications increase the incidence of PVAN.1 It is noteworthy that we Kidney International (2007) 71, 1302–1309

HA Khamash et al.: Polyomavirus nephropathy risk factors

have had the opportunity to compare PVAN among patients managed by tacrolimus- and sirolimus-based regimens of similar immunosuppressive potency and have observed no difference between the two groups.29 One notable limitation of our study is the fact that the number of cases in our cohort managed primarily with cyclosporine was likely too small to compare confidently the PVAN incidence with that of patients managed with tacrolimus or sirolimus. In contrast to our observations here, it has been reported by others that lower level of HLA matching and prior history of acute rejection are associated with higher PVAN risk.1,9,10,20 These discrepancies, which have been noted recently in a comprehensive review of existing literature by Hirsch et al.,1 may well reflect center-specific differences in overall patient demographic and immunosuppression practice. They may also, however, derive from low numbers of index cases among the patients studied or from increased likelihood of graft biopsy in recipients with a history of rejection. By examining PVAN incidence among a large, non-selected cohort of patients undergoing graft biopsy independently of acute complications we believe that such confounding factors have been eliminated from consideration in this study. It should be noted that the influence of ureteral stent placement, which has been reported as a risk factor for BK viremia by Brennan et al.13 in a prospective clinical study, could not be evaluated here as the practice was all but universal during the study period. Similarly, the predominance of Caucasian recipients in our patient populations does not allow a rigorous examination of the influence of race on BKV-associated disease. The independent effect of recipient age on PVAN incidence among our patients is in agreement with Ramos et al.,19,30 who have also described a large experience with PVAN within a carefully characterized KTx practice. The explanation for this association cannot be directly determined here but is in accordance with age-related declines in antiviral cellular immunity31 and, more specifically, with population-based studies that indicate a waning of anti-BKV immunity with age.32 In addition, recipient age may impact the overall immunosuppressive potency of antirejection therapy – particularly of three-drug regimens.33 The significance of confirming and extending this observation is two-fold. In the first place, older recipient age, as a readily identified factor before transplantation, may be used to focus post-KTx screening protocols for BKV reactivation and to modify immunosuppression dosing and target levels. In the second place, it may explain, in part, the dramatic emergence of PVAN during the past 5 years – a period in which the number of KTx recipients aged 50 years or more has increased by over 100% in the USA, whereas the number of recipients aged less than 50 years has remained static (source: www.UNOS.org). It should also be emphasized that, based on the current study as well as existing reports, pediatric and young adult KTx recipients clearly have more than a minimal risk for developing PVAN.4,28 Furthermore, there may be specific immunologic or non-immunologic risk factors that Kidney International (2007) 71, 1302–1309

original article

are operative in younger recipients that are not evident in our current analysis. The higher incidence of PVAN among recipients of female kidneys is surprising and less readily explained on the basis of existing data. Although a type I statistical error cannot be definitively ruled out, we are confident that the association of female donor gender with increased PVAN risk has been a real clinical phenomenon in our KTx practice during the study period. Having been identified in univariate analysis as one of only two variables to be statistically and independently associated with PVAN risk, the association was confirmed in multivariate analysis involving five variables with the level of significance remaining undiminished. Furthermore, Kaplan–Meier analysis demonstrated a clear, persistent divergence in incidence of PVAN among recipients of female compared with male donor kidneys throughout the first 48 months post-transplant. To our knowledge, donor gender has not been examined previously in studies of PVAN risk factors. Although focal positivity of human native kidneys and urinary tract for BKV has been reported in autopsy specimens, the numbers analyzed have been too low to determine whether such foci are more common among females.34,35 It is possible that reactivation of BKV during pregnancy results in a greater burden of latent virus among women who subsequently serve as kidney donors.36 Smaller organ size could also constitute a risk factor for the disease itself or for more frequent diagnostic sampling of the medulla and corticomedullary junction during allograft biopsy. As with recipient age, it should be emphasized that female donor gender has become significantly more common in recent years as a result of the rise in living donor relative to deceased donor numbers.24 Indeed, the consistent ratio of males to females among deceased kidney donors in the USA (1.49:1 between 1988 and 2005) is entirely reversed for living donors (0.75:1) (Source: www.UNOS.org). An important implication of our study on the basis of these two modifiers is that renal allografts at risk for PVAN will continue to increase in number in the coming years. Although not the primary focus of this study, our analyses provide some additional insight regarding the spectrum of histological changes associated with PVAN at the time of diagnosis and its relationship to subsequent clinical course. Perhaps predictably, we demonstrate here that diagnosis by surveillance biopsy portends significantly lower risk for graft loss in the short-to-mid term. Despite this, the initial active histological features of the disease, whether evaluated by elements of the Banff ’97 scoring system or by grading according to the modified-Drachenberg system (which includes grading of viral cytopathy),1 were not strikingly different between the surveillance and non-surveillance biopsy-diagnosed groups. This observation is in keeping with the recent report of Gaber et al.11 involving a detailed histological analysis and clinical outcome in 20 PVAN cases diagnosed following decline in renal function. In contrast, the study by Gaber et al.11,12 as well as our previous work indicates that the presence of advanced fibrosis on the 1307

original article

diagnostic biopsy (Drachenberg grade C) is associated with subsequent loss of renal function regardless of viral clearance on follow-up biopsy. It should be noted that, presumably owing to predominantly early diagnosis of PVAN in this cohort, grade C cases were absent from the surveillance biopsy group and infrequent among those diagnosed by clinically indicated biopsy. Although we cannot formally rule out the possibility that surveillance biopsy-diagnosed cases represent a truly distinct clinical entity, our experience to date favors the conclusion that these cases represent the subclinical phase of a disease process that is likely to progress to irreversible graft fibrosis and functional loss if left undiagnosed. There are some specific limitations to this analysis that should be acknowledged. Although surveillance biopsy capture rates at 4, 12, and 24 months have been high and in situ hybridization for BKV is performed on all graft biopsies with cytopathy or abnormal cellular infiltrates, this should not be considered to represent the results of a screening program designed to diagnose or rule out PVAN in all cases. This being said, our practice is to maintain close laboratory follow-up indefinitely on all recipients and to carry out allograft biopsy for all cases of graft functional decline. Thus, it is unlikely that the analysis has been skewed by missed cases of PVAN. A second limitation in identifying risk factors relates to the lack of stratification for donor and recipient BKV status before transplantation. On the basis of well-documented rates of seropositivity,1,30 it could be expected that 10%–20% of transplants involve BKV-negative donor and recipient. Furthermore, recent studies indicate that donor-derived virus is responsible for the large majority of cases during the first post-transplant year – a fact that may further limit the actual number of recipients at risk.35 We do not believe, however, that the risk factors we have identified are explicable by pretransplant serologic status of the donors or recipients as anti-BKV antibodies have not been found to be more prevalent among older individuals or among females.32 Finally, the risk modifiers we observe are likely, to some degree, to be specific to the patient population and practice characteristics of our own center. It is certainly possible that the prevalence of older, Caucasian recipients of living donor kidneys among our cohort and the predominant use of depleting induction therapy and tacrolimus/mycophenolate mofetil-based immunosuppression with resulting low overall rates of acute rejection have diminished the combined impact of ‘immunological’ risk factors for PVAN in this study and brought others to the fore. Thus, KTx programs with predominantly younger recipients, with recipients of more diverse ethnic backgrounds or with alternative antirejection strategies (corticosteroid-free or rapidly weaned immunosuppression) may encounter PVAN risk factors that were not operative among our cohort.9,10,18,19,36,37 In conclusion, our observation that older KTx recipients (particularly those aged 56 years or greater at the time of transplantation) and recipients of female kidneys are at heightened risk for the potentially graft-destroying complica1308

HA Khamash et al.: Polyomavirus nephropathy risk factors

tion of PVAN provides an opportunity to target preventive measures or intensified BKV screening strategies to specific patient subgroups. With the increased utilization of BKV screening and surveillance biopsies among KTx programs worldwide, it will be possible to determine the degree to which these disease modifiers operate within different practice models. MATERIALS AND METHODS Patient population The study included an analysis of 1027 adult KTx recipients (age X18 years) transplanted at our institution between 1 January 2000 and 30 September 2004. Clinical and laboratory data were extracted from electronic databases and from patient’s medical records following IRB approval. The characteristics of this population are displayed in Table 1. These KTx recipients included 50 blood group (ABO) incompatible living donor grafts, 77 sensitized (‘positive crossmatch’) living donor recipients, and 900 conventional deceased or living donor grafts. Intravenous induction was administered in the form of antithymocyte globulin (Thymoglobulins 1.5 mg/kg – 4 to 8 doses during the first 10 days post-KTx) in 83% of cases or anti-CD25 monoclonal antibody (Simulects 20 mg on days 0 and 4 post-KTx) in 10% of cases. No induction therapy was used in 7% of cases. Initial oral immunosuppression consisted of prednisone, tacrolimus, and mycophenolate mofetil in 87% of patients. Cyclosporine and sirolimus were used as alternatives to tacrolimus in 2 and 11%, respectively. Combinations of tacrolimus or cyclosporine with sirolimus were not employed as primary immunosuppression during this time period. The dose of tacrolimus was adjusted to achieve trough levels of 10–15 ng/ml during the first month; 8–12 ng/ ml during months 2 to 4, and 6–10 ng/ml thereafter. In patients receiving sirolimus, the dose was adjusted to achieve trough levels of 15–20 ng/ml during the first 4 months post-KTx and 10–15 ng/ml thereafter. In those receiving cyclosporine, the dose was adjusted to achieve trough levels of 200–250 ng/ml during the first 4 months post-KTx and 100–200 ng/ml thereafter. During the study period, immunosuppression was not tailored for older recipients during the early post-transplant months or at other times. Recipients of ABO incompatible and positive cross-match grafts additionally received one or more antibody-reducing therapies including plasmapheresis, anti-CD20 monoclonal antibody, splenectomy, and high-dose intravenous immunoglobulin as described elsewhere.38,39 For the purposes of this study, acute rejection was defined as acute cellular rejection (including borderline and all other Banff ’97 grades) or acute humoral rejection present on a renal allograft biopsy interpreted by a consultant renal pathologist. All such episodes were treated by corticosteroid intravenous boluses and oral taper with or without depleting antibody therapy regardless of whether the diagnosis was made on clinically indicated or surveillance biopsy. Renal allografts biopsies and diagnosis of PVAN Renal allograft biopsies were performed as part of a surveillance protocol (surveillance biopsies) throughout the study period as well as for diagnostic purposes in KTx recipients with unexplained decline in graft function (non-surveillance biopsies). All recipients were asked to provide consent for surveillance allograft biopsy during transplantation surgery (‘time zero’) and at 4 and 12 months post-KTx. Final capture rates for these surveillance biopsies were 95, Kidney International (2007) 71, 1302–1309

original article

HA Khamash et al.: Polyomavirus nephropathy risk factors

75, and 65% for each of the three time points. Beginning in 2004, recipients have been additionally asked to consent to surveillance biopsy at 24 and 60 months post-KTx. All biopsies were interpreted by consultant renal pathologists and consisted of serial sections stained with hematoxylin and eosin, periodic acid-Schiff, methenamine silver, and Masson’s trichrome stains. In addition to narrative reporting, each biopsy was scored for acute and chronic abnormalities according to Banff ’97 criteria.37 In situ hybridization for BKV DNA was performed on paraffin-embedded tissue sections as previously described.12,15 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.

13.

14.

15.

16. 17.

18.

19.

Data analysis Data are expressed as means7s.d. throughout the paper. Means of normally distributed data were compared by Student t-test. Data that were not normally distributed were compared by nonparametric tests. w2 and Fisher exact tests were used to compare proportions. Assessment of variables associated with the diagnosis of PVAN was performed by Cox regression and Kaplan–Meier analyses. Comparison of death-censored graft survival for surveillance and non-surveillance biopsy-diagnosed PVAN was carried out by Kaplan–Meier analysis.

20.

21.

22.

23. 24.

ACKNOWLEDGMENTS

We acknowledge the contributions of the Mayo Clinic Renal Pathology Laboratory and of the nursing staff, coordinators, and allied health personnel of the William J von Liebig Mayo Clinic Transplant Center to the care of the patients described in this report.

25. 26.

27.

REFERENCES 1. Hirsch HH, Brennan DC, Drachenberg CB et al. Polyomavirus-associated nephropathy in renal transplantation: interdisciplinary analyses and recommendations. Transplantation 2005; 79: 1277–1286. 2. Trofe J, Gaber LW, Stratta RJ et al. Polyomavirus in kidney and kidney–pancreas transplant recipients. Transpl Infect Dis 2003; 5: 21–28. 3. Mannon RB. Polyomavirus nephropathy: what have we learned? Transplantation 2004; 77: 1313–1318. 4. de Bruyn G, Limaye AP. BK virus-associated nephropathy in kidney transplant recipients. Rev Med Virol 2004; 14: 193–205. 5. Gardner SD, Field AM, Coleman DV, Hulme B. New human papovavirus (B.K.) isolated from urine after renal transplantation. Lancet 1971; 1: 1253–1257. 6. Coleman DV, Gardner SD, Field AM. Human polyomavirus infection in renal allograft recipients. Br Med J 1973; 3: 371–375. 7. Randhawa PS, Finkelstein S, Scantlebury V et al. Human polyoma virus-associated interstitial nephritis in the allograft kidney. Transplantation 1999; 67: 103–109. 8. Binet I, Nickeleit V, Hirsch HH et al. Polyomavirus disease under new immunosuppressive drugs: a cause of renal graft dysfunction and graft loss. Transplantation 1999; 67: 918–922. 9. Nickeleit V, Hirsch HH, Binet IF et al. Polyomavirus infection of renal allograft recipients: from latent infection to manifest disease. J Am Soc Nephrol 1999; 10: 1080–1089. 10. Hirsch HH, Knowles W, Dickenmann M et al. Prospective study of polyomavirus type BK replication and nephropathy in renal-transplant recipients. N Engl J Med 2002; 347: 488–496. 11. Gaber LW, Egidi MF, Stratta RJ et al. Clinical utility of histological features of polyomavirus allograft nephropathy. Transplantation 2006; 82: 196–204. 12. Wadei HM, Rule AD, Lewin M et al. Kidney transplant function and histological clearance of virus following diagnosis of polyomavirus-associated nephropathy (PVAN). Am J Transplant 2006; 6: 1025–1032.

Kidney International (2007) 71, 1302–1309

28. 29.

30.

31.

32.

33. 34.

35.

36. 37. 38.

39.

Brennan DC, Agha I, Bohl DL et al. Incidence of BK with tacrolimus versus cyclosporine and impact of preemptive immunosuppression reduction. Am J Transplant 2005; 5: 582–594. Sachdeva MS, Nada R, Jha V et al. The high incidence of BK polyoma virus infection among renal transplant recipients in India. Transplantation 2004; 77: 429–431. Buehrig CK, Lager DJ, Stegall MD et al. Influence of surveillance renal allograft biopsy on diagnosis and prognosis of polyomavirus-associated nephropathy. Kidney Int 2003; 64: 665–673. Stegall MD, Larson TS, Prieto M et al. Living–donor kidney transplantation at Mayo Clinic – Rochester. Clinical Transplants 2002: 155–161. Cosio FG, Grande JP, Wadei H et al. Predicting subsequent decline in kidney allograft function from early surveillance biopsies. Am J Transplant 2005; 5: 2464–2472. Mengel M, Marwedel M, Radermacher J et al. Incidence of polyomavirus-nephropathy in renal allografts: influence of modern immunosuppressive drugs. Nephrol Dial Transplant 2003; 18: 1190–1196. Ramos E, Drachenberg CB, Papadimitriou JC et al. Clinical course of polyoma virus nephropathy in 67 renal transplant patients. J Am Soc Nephrol 2002; 13: 2145–2151. Awadalla Y, Randhawa P, Ruppert K et al. HLA mismatching increases the risk of BK virus nephropathy in renal transplant recipients. Am J Transplant 2004; 4: 1691–1696. Vasudev B, Hariharan S, Hussain SA et al. BK virus nephritis: risk factors, timing, and outcome in renal transplant recipients. Kid Int 2005; 68: 1834–1839. Dean PG, Gloor JM, Stegall MD. Conquering absolute contraindications to transplantation: positive-crossmatch and ABO-incompatible kidney transplantation. Surgery 2005; 137: 269–273. Halloran PF. Immunosuppressive drugs for kidney transplantation. N Engl J Med 2004; 351: 2715–2729. Magee CC, Pascual M. Update in renal transplantation. Arch Intern Med 2004; 164: 1373–1388. Randall HB, Cao S, deVera ME. Transplantation in elderly patients. Arch Surg 2003; 138: 1089–1092. Kim HC, Hwang EA, Han SY et al. Polyomavirus nephropathy after renal transplantation: a single centre experience. Nephrology 2005; 10(2): 198–203. Rocha PN, Plumb TJ, Miller SE et al. Risk factors for BK polyomavirus nephritis in renal allograft recipients. Clin Transplant 2004; 18: 456–462. Haysom L, Rosenberg AR, Kainer G et al. BK viral infection in an Australian pediatric renal transplant population. Pediatr Transplant 2004; 8: 480–484. Larson TS, Dean PG, Stegall MD et al. Complete avoidance of calcineurin inhibitors in renal transplantation: a randomized trial comparing sirolimus and tacrolimus. Am J Transplant 2006; 6: 514–522. Ramos E, Drachenberg CB, Portocarrero M et al. BK virus nephropathy diagnosis and treatment: experience at the University of Maryland Renal Transplant Program. Clinical Transplants 2002: 143–153. Effros RB, Dagarag M, Spaulding C, Man J. The role of CD8+ T-cell replicative senescence in human aging. Immunolog Rev 2005; 205: 147–157. Knowles WA, Pipkin P, Andrews N et al. Population-based study of antibody to the human polyomaviruses BKV and JCV and the simian polyomavirus SV40. J Med Virol 2003; 71(1): 115–123 , Sep.. de Fijter JW. The impact of age on rejection in kidney transplantation. Drugs & Aging 2005; 22: 433–449. Boldorini R, Veggiani C, Barco D, Monga G. Kidney and urinary tract polyomavirus infection and distribution: molecular biology investigation of 10 consecutive autopsies. Arch Pathol & Lab Med 2005; 129: 69–73. Heritage J, Chesters PM, McCance DJ. The persistence of papovavirus BK DNA sequences in normal human renal tissue. J Med Virol 1981; 8: 143–150. Coleman DV, Wolfendale MR, Daniel RA et al. A prospective study of human polyomavirus infection in pregnancy. J Infect Dis 1980; 142: 1–8. Racusen LC, Solez K, Colvin RB et al. The Banff 97 working classification of renal allograft pathology. Kidney Int 1999; 55: 713–723. Fidler ME, Gloor JM, Lager DJ et al. Histologic findings of antibody-mediated rejection in ABO blood-group-incompatible living-donor kidney transplantation. Am J Transplant 2004; 4: 101–107. Stegall MD, Gloor J, Winters JL et al. A comparison of plasmapheresis versus high-dose IVIG desensitization in renal allograft recipients with high levels of donor specific alloantibody. Am J Transplant 2006; 6: 346–351.

1309