BK virus allograft nephropathy has emerged in the late 1990s and this decade as a major complication of kidney transplantation (1–4). This virus infection in an immunosuppressed kidney transplant recipient leads to a nephropathy that markedly reduces graft survival. The sudden appearance and rapid increase of this virus infection echoes the pattern of emergence of cytomegalovirus infection and Epstein-Barr virus infections posttransplant in prior decades (5). Most data on BKVAN to date have arisen from single- or limited-center series (6–10). Many of these studies provide conflicting information regarding risk factors for BKVAN. For example, some studies suggest that tacrolimus (TAC) or mycophenolate use increase the risk for BKVAN (11–13), whereas other studies do not (8, 14, 15).
In 2004, the Organ Procurement Transplant Network (OPTN) transplant database in the United States added BK virus infection treatment as a question on its postkidney transplant follow-up form, thus allowing for a large national sample of treatment of BK virus (TBKV) data to be collected. In the present study we sought to determine the association of TBKV with various recipient, donor and transplant characteristics already collected in the OPTN database.
The present case-cohort study included primary and solitary kidney transplant recipients between January 1, 2003 through December 31, 2006, from the OPTN database, with at least 6 months of graft survival, and who were reported on discharge maintenance regimen using cyclosporine A (CsA), TAC, azathioprine (AZA), or mycophenolate mofetil (MMF). Recipients reported to have received multiple induction agent types (e.g., basiliximab and rabbit-ATG) or who received OKT3 (only 62 patients in this study time period, with only one TBKV, within 570 days posttransplant) were excluded from the analysis.
The OPTN has been collecting information regarding TBKV treatment since June 30, 2004 on the kidney recipient follow-up forms submitted by transplant centers across the United States. Each center defined the need for TBKV at their discretion. Overall incidence of TBKV, as indicated by time to first report of TBKV, was calculated at every 6-month interval posttransplant, using Kaplan-Meier method to prevent possible underreporting (16) and stratified by study variable. We chose 2003 as the starting year of transplant to capture cases reported from a prior transplant year (run-in period) and to assess yearly trends.
We included various recipient, donor, and transplant characteristics or variables collected by the OPTN and as described in our previous studies (17–19). Recipient variables included gender, age at transplant, ethnicity, treatment of acute rejection (AR) within 6 months of transplant, delayed graft function as indicated by a need for dialysis within 1 week of transplant, antibody induction type (basilixumab, daclizumab, thymoglobulin, ATG, alemtuzumab, or no induction), discharge maintenance regimen (CNI-based: CsA, TAC, or none; anti-metabolite-based: AZA, MMF, or none), and whether mTOR-i (mammalian target of rapamycin)-inhibitor or steroids were used as part of the discharge regimen. Donor variable considered in the analysis was the type of donor (living vs. deceased). Transplant variables included in the analysis were donor to recipient cytomegalovirus serostatus, center volume as defined by the number of kidney transplants in the previous year (<100 vs. 100+) and transplant year. Cumulative incidence of TBKV within 2 years of transplant calculated at each 6-month interval was calculated using the Kaplan-Meier (K-M) method and compared among different groups of variables using the log-rank test.
Since BK virus infection can occur at any time posttransplant, risk factors associated with a report of TBKV were then determined using multivariate Cox proportional hazard regression analysis. To ensure comparable follow-up among various induction types and immunosuppressive regimens, records with follow-up beyond 24 months were censored. The time of 24 months was chosen because it represented the median follow-up time for the induction group with the shortest follow-up. Additionally, to determine if TBKV report was associated with higher risk of subsequent graft loss, we fitted a time-dependent Cox regression model incorporating TBKV and the other variables listed above.
A P value of less than 0.05 was considered statistically significant and all reported P values were two tailed. For variables with missing data, the median or most frequent category was used. Analyses were performed using SAS statistical software version 9.1 (SAS Institute, Cary, NC).
The present study included 48,292 primary and solitary kidney transplants from the OPTN database, with at least 6 months of survival, with follow-up as of July 18, 2008. Of these, 1474 transplants have reported TBKV within 24 months of transplant to the OPTN, for a K-M incidence of 3.45%. Overall cumulative rate of TBKV, calculated at every 6-month interval between transplant and 60 months posttransplant using the K-M methods, increased steadily with time, to 6.6% at 60 months (Fig. 1). The 95% CI range at each time point was fairly tight, suggesting a high degree of accuracy of the point estimate.
Table 1 summarizes the K-M cumulative incidence of TBKV at 24 months of transplant for the study variables and the log-rank test P value for comparing the incidence within 24 months among various groups of the study variables. Cumulative incidence of TBKV was significantly different among the various recipient age groups. The highest incidence of TBKV was observed in pediatric recipients (age 0–17 years), followed by the older recipients (>55 years). Cumulative incidence of TBKV was significantly higher in male or African-American, those receiving a deceased donor transplant, higher degrees of human leukocyte antigen (HLA) A, B, and DR mismatch, more recent transplant year, lower center transplant volume, certain induction agents (R-ATG and alemtuzumab), and with TAC-based or MMF-based discharge maintenance regimens. Use of m-TOR inhibitor was associated with a lower cumulative incidence.
In contrast to the cumulative incidence figures shown in Table 1, Table 2 provides the results of the multivariate Cox proportional hazards regression model to determine the adjusted risk factors for time to TBKV. The use of R-ATG, TAC-based regimens, MMF-based regimens, African-American recipient, and more recent transplant year were all associated with significantly higher risk for TBKV. Female gender, Hispanic recipient and recipient of other ethnic group, living donor, and higher center volume were associated with lower risk for TBKV. Comparisons between different groups of variables within a group, not just to baseline group, by linear testing revealed significantly higher risk for TBKV with (a) rabbit-ATG use versus IL-2R Ab (P<0.001); (b) use of CsA-based regimens versus regimens with no CNI (P=0.0213); (c) black recipients versus Hispanic (P<0.001) or other race (P=0.0095); (d) Asian recipients versus Hispanic recipients (P=0.0092). Among HLA mismatches, only A locus mismatch remained significantly associated. Alemtuzumab use did not remain as a significant risk factor in the adjusted model, suggesting a confounding effect in univariate models of simultaneous emergence of this drug and our event of interest. Surprisingly, presence of delayed graft function, a known risk factor for other infections such as septicemia (20), was associated with a lower risk for TBKV in the Cox model, even though the K-M incidence of TBKV was not significantly different. The high incidence of TBKV in the youngest and oldest recipients observed by the Kaplan-Meier method suggested that recipient age had a nonlinear relationship with TBKV, so we fitted both linear and quadratic terms of recipient age in the Cox model. Both terms were significant in the Cox model, suggesting that pediatric recipients and the oldest recipients were associated with a significantly higher risk of TBKV. AR, reported within first 6 months posttransplant, increased the risk for TBKV at any time point significantly by an adjusted hazard ratio (AHR) of 2.54 (P<0.001). Because the OPTN does not collect exact dates for infection or rejection events within a time interval, we were also interested in whether a prior AR could increase risk for subsequent TBKV. We therefore performed an additional Cox regression analysis after removing all cases with TBKV reported within first 6 months posttransplant. In this version of the analysis, using 48,051 records, AR within first 6 months posttransplant remained as a significant risk factor for subsequent TBKV (AHR 2.56, P<0.001). All other variables in this version of the Cox regression model also retained similar AHR values and significance (full data not shown). Perhaps most importantly, the adjusted hazard ratio for transplant year increased significantly at all posttransplant time points with each more recent year.
Table 3 provides the estimated rate of TBKV based on injectable induction agent used at time of transplant, calculated at every 6-month interval between transplant and 24 months posttransplant. Cumulative rate of TBKV treatment increased with time even if no induction agent was used, and with each of the agents used.
In Table 4, we have analyzed the discharge oral maintenance regimens according to the following categories, to better characterize the influence of antimetabolite drugs: CsA-based, TAC-based, no CNI use, AZA-based, MMF-based, no antimetabolites used, steroid use and mTORi use. Some overlap can exist in these categories. AZA-based regimens were rarely used in the United States in the study time period. MMF-based regimens were associated with higher cumulative rates of TBKV reported treatment than no antimetabolite-based regimens, though rates were not as high as with Tac-based regimens.
Association of cumulative TBKV rate with transplant year is analyzed in Table 5. Again, for each transplant year, the cumulative TBKV rate increased over time. However, the 6- and 12-month cumulative incidence was also higher with each more recent transplant year, suggesting again that the rise in incidence is real, not an artifact of reporting or the K-M methods or of delayed reporting of prior events in a recently introduced form.
In Table 6, we provide results of a Cox regression model to determine if TBKV was a risk factor for accelerated graft loss within 3 years posttransplant. Even after adjustment for the many variables included in the model, TBKV remained a significant factor (AHR 1.69, P<0.001).
This report provides data that include a large series of transplant recipients and TBKV cases. The advantages of an analysis of such large registry data are the multicentered nature of the data and the ability to overcome Type II statistical errors from insufficient sample size. Thus, our report is able to confirm that use of certain more potent immunosuppressive regimens, such as TAC-based versus CsA-based, are associated with higher risk for TBKV. Similarly, R-ATG induction use, generally considered to be more long lasting and immunosuppressive over more receptor pathways than monoclonal agents, was also associated with higher TBKV risk. These results are in accordance with some of the prior single-center series data (11, 12, 14). Some of the risk factors we elucidated are not modifiable per se, such as recipient age, race, or gender. However, other risk factors may be more amenable to modification, such as type of immunosuppression and total intensity of immunosuppression regimen. In general, regimen intensity, more than any one agent use, seems to be the major contributor to viral infection risk. New findings elicitable only in a large multicenter analysis such as this include the protective effect exerted by large volume centers, performing 100 or more kidney transplants per year.
One result of concern was our finding that more recent transplant year was associated with a significantly higher rate of TBKV, even within a relatively recent time period. Furthermore, the AHR was higher with each recent year, suggesting against increased reporting as the sole cause of this higher incidence. Better detection is possible but the significance of BKV as a complication of transplantation gained prominence in 1999 (21), well before the starting year of our study (2003). These results would suggest that BK virus infection and nephropathy rates are still on the rise, despite the known consequences of this disease and the association with more potent immunosuppression use.
Higher degrees of HLA mismatch has been identified as a risk factor for TBKV in both retrospective and prospective studies (6, 22). In our study, higher mismatch for each locus was associated with higher K-M incidence of TBKV, but only A locus mismatch remained significant in the adjusted model. The increased rejection risk with HLA mismatching (23), through its treatment with escalation of immunosuppression or by direct rejection-related inflammatory responses have been proposed as potential mechanisms for this association. But other studies have not identified HLA mismatching as a risk factor for TBKV (8, 13, 24). With regards to impact on subsequent graft loss, in our analysis, only A locus mismatch in conjunction with TBKV imparted greater risk.
Our study found that pediatric recipients were at higher risk for TBKV. This new finding is consistent with similar findings for other viruses such as Epstein-Barr virus infection and posttransplant lymphoproliferative disorders (19). Similar to herpesviruses, polyoma virus is commonly acquired by immunocompetent hosts through childhood and establishes a long-lived latency in the host. Children who receive a kidney transplant in early years may not have been previously exposed to virus. These children then acquire a primary, more severe infection when immunosuppressed. Several pediatric studies have documented the incidence of BK virus nephropathy in children (14, 25), but to our knowledge no study has been able to show a comparative higher risk for children versus young-to-middle age adults.
One strength of our study was the use of Kaplan-Meier method to estimate the cumulative incidence of TBKV over time, both overall and in association with specific risk factors. This method takes into account the incomplete and varying follow-up and therefore reduces the potential biases of underreporting associated with an inflated denominator for computing the rate.
Like any registry analysis, our study has several potential limitations. The definition of TBKV is per center report and may vary between centers. The data field from the OPTN follow-up form that we used is labeled as BK virus nephropathy treatment; it is possible that some cases went untreated or had no interventions performed. Reporting bias by individual centers is also possible, especially with events that have a low prevalence still such as TBKV. However, analyzing rare events through large multicenter registries helps to uncover information that is not easily gleaned from single-center series. Ureteral stent placement or BK virus sero-mismatch have also been linked recently to higher BK virus nephropathy incidence (10, 13, 26). We could not analyze for these associations as the OPTN does not collect data on stent placement or BK virus serology at time of kidney transplant. Because later follow-up information in OPTN tends to be less complete, especially for more recent years, data on subsequent graft loss after TBKV may have greater chance of error, though most studies to date report accelerated graft loss once BK virus nephropathy develops (8, 27, 28), in accordance with our results. This accelerated graft loss may be reduced with intensive monitoring (2).
In summary, this large registry retrospective cohort analysis reveals that TBKV rates continued to rise with more recent transplant year in the time period 2003 to 2006; and that certain more potent immunosuppressive regimens increase the risk for TBKV.
These data were presented as a platform presentation at the American Transplant Congress in June 2008. The analyses are based on OPTN data as of July 18, 2008 and are subject to change due to future data submission or correction by centers. Interpretations of the data represent the opinions of the authors only. The authors acknowledge Yulin Cheng from UNOS, for putting together the analysis dataset.
1. Hirsch HH, Steiger J. Polyomavirus BK. Lancet Infect Dis
2003; 3: 611.
2. Dall A, Hariharan S. BK virus nephritis after renal transplantation. Clin J Am Soc Nephrol
2008; 3(suppl 2): S68.
3. Randhawa PS, Demetris AJ. Nephropathy due to polyomavirus type BK. N Engl J Med
2000; 342: 1361.
4. Bohl DL, Brennan DC. BK virus nephropathy and kidney transplantation. Clin J Am Soc Nephrol
2007; 2(suppl 1): S36.
5. Dharnidharka VR, Harmon WE. Management of pediatric postrenal transplantation infections. Semin Nephrol
2001; 21: 521.
6. 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.
7. Nickeleit V, Klimkait T, Binet IF, et al. Testing for polyomavirus type BK DNA in plasma to identify renal-allograft recipients with viral nephropathy. N Engl J Med
2000; 342: 1309.
8. 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.
9. Drachenberg CB, Papadimitriou JC, Mann D, et al. Negative impact of human leukocyte antigen matching in the outcome of polyomavirus nephropathy. Transplantation
2005; 80: 276.
10. Yeo FE, Yuan CM, Swanson SJ, et al. The prevalence of BK polyomavirus infection in outpatient kidney transplant recipients followed in a single center. Clin Transplant
2008; 22: 522.
11. Howell DN, Smith SR, Butterly DW, et al. Diagnosis and management of BK polyomavirus interstitial nephritis in renal transplant recipients. Transplantation
1999; 68: 1279.
12. 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.
13. 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.
14. Smith JM, Dharnidharka VR, Talley L, et al. BK virus nephropathy in pediatric renal transplant recipients: An analysis of the North American Pediatric Renal Trials and Collaborative Studies (NAPRTCS) registry. Clin J Am Soc Nephrol
2007; 2: 1037.
15. Khamash HA, Wadei HM, Mahale AS, et al. Polyomavirus-associated nephropathy risk in kidney transplants: The influence of recipient age and donor gender. Kidney Int
2007; 71: 1302.
16. Kaplan EL, Meier P. Non parametric estimation from incomplete observations. J Am Stat Assoc
1958; 53: 457.
17. Cherikh WS, Kauffman HM, McBride MA, et al. Association of the type of induction immunosuppression
with posttransplant lymphoproliferative disorder, graft survival, and patient survival after primary kidney transplantation. Transplantation
2003; 76: 1289.
18. Johnson SR, Cherikh WS, Kauffman HM, et al. Retransplantation after post-transplant lymphoproliferative disorders: An OPTN/UNOS database analysis. Am J Transplant
2006; 6: 2743.
19. Dharnidharka VR, Tejani AH, Ho PL, et al. Post-transplant lymphoproliferative disorder in the United States: Young Caucasian males are at highest risk. Am J Transplant
2002; 2: 993.
20. Abbott KC, Oliver JD III, Hypolite I, et al. Hospitalizations for bacterial septicemia after renal transplantation in the United States. Am J Nephrol
2001; 21: 120.
21. Drachenberg CB, Beskow CO, Cangro CB, et al. Human polyoma virus in renal allograft biopsies: Morphological findings and correlation with urine cytology. Hum Pathol
1999; 30: 970.
22. 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.
23. Opelz G. Impact of HLA compatibility on survival of kidney transplants from unrelated live donors. Transplantation
1997; 64: 1473.
24. Bohl DL, Storch GA, Ryschkewitsch C, et al. Donor origin of BK virus in renal transplantation and role of HLA C7 in susceptibility to sustained BK viremia. Am J Transplant
2005; 5: 2213.
25. Acott PD, Hirsch HH. BK virus infection, replication, and diseases in pediatric kidney transplantation. Pediatr Nephrol
2007; 22: 1243.
26. Thomas A, Dropulic LK, Rahman MH, et al. Ureteral stents: A novel risk factor for polyomavirus nephropathy. Transplantation
2007; 84: 433.
27. Ramos E, Vincenti F, Lu WX, et al. Retransplantation in patients with graft loss caused by polyoma virus nephropathy. Transplantation
2004; 77: 131.
28. Mylonakis E, Goes N, Rubin RH, et al. BK virus in solid organ transplant recipients: An emerging syndrome. Transplantation
2001; 72: 1587.