In recent years, significant increases in the age of waitlisted patients and utilization of higher risk donor organs have occurred. Shifts in the use of immunosuppressive regimens with increasing use of steroid avoidance regimens have also occurred (1). Although such regimens are deemed safe in selected patients, they are also associated with an increased risk for acute rejection with uncertain impact on long-term allograft function (2–5). Calcineurin inhibitor–containing regimens have been associated with excellent graft survival rates. However, concern as to whether the histologic lesions accompanying their long-term use affect graft survival adversely persists (6). Published reports of de novo calcineurin inhibitor avoidance using sirolimus-based regimens depict varying success (7–9). Expanded criteria donor kidneys are used increasingly to offer viable transplantation opportunities to candidates with poor expected waitlist survival (10–12). Increasing use of living donor kidneys from older, more comorbid donors have been reported (13–16). Also observed in this era is the phenomenon that despite decreasing acute rejection rates, long-term graft survival has not improved significantly (17).
A paradigm for the study of posttransplant renal function incorporates analyses of effects of clinically relevant covariates on baseline estimated renal function (intercept), its change over time (slope), and the relationship of slope and intercept of estimated renal function (glomerular filtration rate [GFR] or creatinine clearance) to transplant outcome (18, 19). Using this analytic framework, single-center studies, pooled data from a consortium of single centers and a European study suggest improvement in slopes in recent eras (20–22).
Given the foregoing, we hypothesized that shifting patterns in immunosuppressive regimen use and the use of higher risk donors may have adversely affected changes in renal function after renal transplantation in the US population. The availability of serial serum creatinine measurements after transplantation in national registries affords opportunities to study the evolution of estimated GFR (eGFR) after renal transplantation across eras. It is well known that that creatinine-based GFR estimating equations have their limitations as estimates of measured GFR (23). However, the universal availability of creatinine measurements in the clinic harnesses the practical value of a measure of renal function especially when serial changes in eGFR values are the variable of interest. On the basis of these premises, by using national registry data from the United States, we evaluated changes over time in the intercept and slope of eGFR after renal transplantation in the United States.
A total of 91,241 renal transplant recipients fulfilled criteria for inclusion in the analysis. For patients with at least 6-month graft survival, 1.2% of patients had no reported creatinine, and among patients with at least 12-month graft survival, 4.8% were missing at 6- or 12-month follow-up. Sixty percent of recipients received transplants from deceased donors, 24% had a primary diagnosis of diabetes, glomerulonephritis was the primary diagnosis in 26%, average recipient age was 49 years, and the average donor age was 39 years. Hypertension was the primary diagnosis in 20% of recipients. Expanded criteria donors were used in 18% of transplants, which increased from 17.5% to 18.9% during the study period. Fifty-six percent of the population was reported as non-Hispanic white, 24% African American, 13% Hispanic, and 7% from other race or ethnic groups. Retransplanted patients comprised 12% of the population. Fifty-two percent of patients had Medicare as a primary payer, and 42% had commercial insurance. Average body mass index (BMI) of recipients was 27 kg/m2. Pretransplant dialysis time exceeded 36 months in 29% of recipients. Retransplants comprised 12% of recipients. Fifty-one percent of recipients had panel reactive antibody=0%.
The eGFR at 6 months was considered the patients' renal function intercept. Unadjusted GFR intercept at 6 months averaged 54.3 mL/min/1.73 m2 (standard deviation 18.2 mL/min/1.73 m2). Adjusted GFR intercept in living donor recipients was 55.4 mL/min/1.73 m2 and in cadaveric donor recipients was 53.6 mL/min/1.73 m2. Table 1 depicts the association between 6-month modification of diet in renal disease (MDRD) eGFR and relevant covariates in adjusted analyses, and a brief summary is provided hereunder. In adjusted analyses, the dominant clinically relevant covariate that significantly influenced GFR intercept at 6 months was donor age. Adjusted 6-month eGFR was highest with donor age less than 18 years (63.5 mL/min) and lowest when donor age exceeded 65 years (44.1 mL/min/1.73 m2; P<0.001; Table 1). Recipients ages 18 to 45 years had the highest 6-month eGFRs at 55.1 mL/min1.73 m2 and those more than 65 years averaged 54.2 mL/min/1.73 m2 (P<0.001; Table 1). Recipients with BMI less than 20 kg/m2 had an adjusted mean 6-month eGFR of 61.4 mL/min. This value declined progressively with increasing BMI and averaged 49.4 mL/min/1.73 m2 in those with BMI more than 35 kg/m2 (P<0.001). African Americans had an average 6-month eGFR of 56.1 mL/min/1.73 m2 as opposed to 51.7 mL/min/1.73 m2 in whites and 55.8 mL/min/1.73 m2 in Hispanics (P<0.001). Six-month eGFR did not differ significantly by gender. GFR intercepts increased steadily from 53.1 mL/min/1.73 m2 in 2003 to 56.5 mL/min/1.73 m2 in 2008 (P<0.001). In addition, for patients who experienced acute rejection in the first 6 months, 6-month eGFR was significantly lower (44.6 mL/min/1.73 m2 vs. 54.9 mL/min/1.73 m2, P<0.001).
Using the Chronic Kidney Disease Epidemiology Collaboration (CKD-EPI) estimating equation, baseline (6 months) eGFRs were higher versus MDRD (expectedly) and increased significantly over time (63.0 mL/min in 2003 to 66.9 mL/min in 2008, P<0.001 for effect by year).
Change in eGFR From 6 to 12 Months After Transplantation
In unadjusted analyses, from 2003 to 2008, we noted a steady decrease in the rate of decline of GFR from 6 to 12 months posttransplantation (P<0.001 for trend). At each of the time points, the rate of decline of GFR was greater for deceased donor transplants than for living donor transplants. In the period from 2007 to 2008, the 6 to 12 months posttransplant eGFR slopes were positive (Fig. 1a). These results were consistent over the period ranging from 6 to 24 months posttransplant among transplants occurring between 2003 and 2007 including a statistically significant increase over time (P<0.001 for trend; Fig. 1b). eGFR slopes also significantly increased by transplant year using CKD-EPI equation with similar estimates by transplant year as reported using the MDRD equation. Primary results without imputation remained consistent.
Acute Rejection and eGFR Slope
The overall incidence of acute rejection was low in the study period, 5.7% in the first 6 months posttransplantation and 2.5% between 6 and 12 months. However, the decline in eGFR was more pronounced in those experiencing an acute rejection episode at 6 to 12 months than in those who were rejection free (P<0.001; Fig. 1c). This result remained consistent during the study period, and among patients without a reported rejection episode, the slope in eGFR increased significantly (P<0.001). There was no significant difference (P=0.47) of subsequent slope between 6 and 12 months among patients who did and did not experience acute rejection in the first 6 months posttransplantation.
Clinical Correlates of 6 to 12 Months eGFR Slope
The results of the multivariate model analyzing major correlates of eGFR slopes in the 6- to 12-month follow-up period are portrayed in Table 2. Adjusted slopes of eGFR improved by 0.29 mL/min/year during the study period (P<0.001). The principal correlates of a negative eGFR slope were older donor age (−0.43 mL/min/10 years; P<0.001), African American race (−0.72 mL/min/year; P<0.001), diabetes as primary diagnosis (−0.64 mL/min/year; P<0.001), and female sex (−0.22 mL/min/year; P<0.001). Rates of decline of eGFR were lowest in those receiving a preemptive transplant (0.53 mL/min/year relative to patients with pretransplant dialysis time >36 months). Decline in eGFR was marginally attenuated among patients with commercial insurance versus Medicare as a primary payer (also see Supplemental Digital Content 1, http://links.lww.com/TP/A296, for slopes and intercepts across all subgroups).
Changes in eGFR Slope During the Study Period
Improvements in eGFR slope during the study period noted in unadjusted analyses remained consistent in the multivariate model adjusted for clinically relevant covariates. Slopes of eGFR improved steadily for both the 6- to 12-month and 6- to 24-month follow-up periods independent of donor, recipient, and transplant characteristics. We assessed the interaction of patient characteristics with changes over time but found no significant factors that were associated with differential increases in slope.
Results of the mixed models indicated that the contribution of variation across centers was minimal for explaining changes in renal function at different intervals. The intraclass correlation (ICC) statistics for across-center variability for changes in renal function were all less than 1%: 6 to 12 months (0.4%), 12 to 24 months (0.4%), and 6 to 24 months (0.6%). The ICC associated with transplant centers for baseline (6 months) eGFR was relatively higher (3.1%) compared with changes over time but still explained a minor proportion of variability.
Immunosuppression Regimen and eGFR
We compared calcineurin inhibitor–based regimens both with and without steroids and calcineurin inhibitor-free steroid-containing regimens with sirolimus and mycophenolate mofetil. We evaluated the independent association of immunosuppressive regimen on 6-month eGFR and 6- to 12-month eGFR slope, and results are depicted in Table 3. As indicated in the Table 4, there was no significant difference in baseline eGFR by immunosuppressive regimens. The only significant differences in GFR slope were between the tacrolimus-based regimens when compared with the combination of “other” therapies, but no significant differences were indicated between the other regimens that were classified for the study in the adjusted models.
One-Year Overall Graft and Patient Survival
Although there were relatively minor differences in the proportion of patients surviving at 1 year, both overall graft and patient survival were statistically significantly different by year with progressively higher survival in the most recent years. Overall graft survival ranged from 91.3% in 2003 to 93.7% in 2008 (P<0.001 for trend), and patient survival increased from 96.0% to 96.8% during the same period (P<0.001 for trend).
Despite the increasing use of lower quality organs and steroid avoidance regimens, GFR slopes and 1-year survival rates have improved in the United States. We do not have direct explanations for these favorable phenomena. However, we do offer some plausible explanations. These results could reflect increased skill in selection, risk stratification, and medical management of these complex patients. The period studied has seen considerable improvements in the application of sensitive techniques for tissue typing, crossmatching, and assays for donor-specific antibody and refinements in the diagnosis and treatment of acute rejection episodes (24–26). Such advances have rendered the incidence of hyperacute rejection in noncrossmatch positive transplants a rare event. During the study period and the era preceding it, the transplant community has also moved away from empiric treatment of graft dysfunction, and refinements in histopathology and standardization thereof have been adopted (27). Entities such as antibody-mediated rejection can now be diagnosed and treated early with an ever-growing therapeutic armamentarium (24, 28). During the same time frame, universal and effective prophylaxis for cytomegalovirus infections has also been used, allowing the effective and safe use of antibody induction and intensified early immunosuppression (29). Furthermore, greater refinements in the management of the emerging epidemic of polyoma virus nephropathy with effective screening and appropriate titration of the intensity of immunosuppression may have mitigated the influence of this entity in mediating attrition of graft function (30, 31). Advances in organ preservation such as pulsatile machine perfusion have also contributed to our ability to improve early function of deceased donor kidneys with improvements in outcome (32, 33).
Despite these positive and encouraging trends, our results speak to the continuing influence of acute rejection on renal function and potentially graft survival (34). As shown, eGFR slopes were inferior in those patients experiencing acute rejection in the 6- to 12-month period. These results are consistent with previous observations by other groups (20–22). This influence of acute rejection on eGFR slope occurred in the face of a low overall rate of acute rejection. This phenomenon could thus reflect the fact that under modern immunosuppression, such rejection episodes, as are clinically manifest, may display a more aggressive clinical behavior and perhaps reflect a selection of more treatment-resistant rejection (34).
The relative lack of impact of the type of immunosuppression on eGFR slopes and intercepts is a central finding in our study. This finding is likely based on the growing familiarity and skill in the transplant community in using the available immunosuppressive armamentarium at the bedside. We submit this necessarily speculative observation with a few caveats.
Creatinine-based GFR estimating equations have known biases and limitations compared with directly measured GFR (insulin clearance) in transplant recipients (23). However, these considerations are predominantly of note in the interpretation of the GFR intercept and less so in the interpretation of serial changes, the slope (23). It is unlikely that these deficiencies affected the qualitative findings of this study (35). This is largely because the analysis compared changes within patients by transplant year. As such, the biases are equally distributed within individual patients. Systematic biases may impact baseline estimations but are unlikely to explain variations by transplant year. An additional legitimate concern stems from the changes that have occurred in the utilization of assays for serum creatinine, which in turn could impact the results. Unfortunately, there is no direct way of ascertaining the exact method used to measure creatinine across centers and time periods. To the extent possible within the constraints of data available in the Scientific Registry of Transplant Recipients (SRTR), we investigated the contributions of centers to the variability of GFR estimates and slopes over time—the premise being that changes in creatinine estimation methods in centers over time would influence estimates of GFR intercept and slope. In our analyses, the contribution of centers to variability of the estimates of GFR intercept was minimal. Thus, we submit that this consistency over time and across centers likely reflects a true secular trend in eGFRs rather than a mere artifact of changing creatinine measurement techniques across centers over time.
Despite the wealth of data and the large population that this database covers, there is an acknowledged deficiency of critical patient level data and a known proportion of incomplete data pertaining to GFR data in the SRTR (36, 37). In that regard, single-center studies and data derived from consortia of single centers allow the incorporation of critical patient level data such as those pertaining to blood pressure, proteinuria, antihypertensive treatment and doses, and concentrations of immunosuppressive agents (20, 21). A limitation of analyses of immunosuppressive regimens based on registry data is lack of specific dosing and concentration information, and certainly, changes in immunosuppressive dosing or concentration targeting could explain some of our observations. Further limitations of our analyses pertaining to immunosuppression are that they are by intent-to-treat only and do not take into account the fact that regimens could have been switched over time or deliberately changed in intensity over time. However, we submit that even with all the drawbacks inherent to registry-based inquiry, our results are consistent with the previous published reports from well-conducted single-center studies and collaborative studies and providing a broad assessment of national trends (20–22, 38).
Serum creatinine and creatinine-based GFR estimating equations are universally available in the context of clinical care. As such, judicious use of the creatinine and estimates of GFR therefrom could be of use in prognosticating graft survival; acknowledged caveats notwithstanding (39, 40). Thus, our results could be of significant value in yielding expected values for the design of future interventional trials in renal transplantation. Although these results could prompt cautious optimism, it has to be kept in mind that they reflect results from retrospective analyses where cause and effect cannot be readily ascribed (36). It is also important that future analyses be directed at analyzing actual long-term outcomes based on these results as opposed to extrapolations (41, 42). Evaluating whether these results translate to improved long-term survival will be a critical next step.
MATERIALS AND METHODS
We used data from the SRTR. The SRTR data system includes data on all donor, wait-listed candidates, and transplant recipients in the United States, submitted by the members of the Organ Procurement and Transplantation Network. The Health Resources and Services Administration, US Department of Health and Human Services provides oversight to the activities of the Organ Procurement and Transplantation Network and SRTR contractors.
The study population consisted of adult (age >17 years) solitary kidney transplant recipients in the United States from 2003 to 2008. Patients with less than 6 months of graft survival or follow-up information were excluded (7%) from the study based on available forms in which the first discharge serum creatinine measure is available at 6 months. For patients with graft loss between 6 and 24 months posttransplant, we imputed an eGFR of 10 mL/min/1.73 kg/m2. We used the most recent MDRD equation to estimate GFR (eGFR) at each interval (43). We excluded the top and bottom 1% of eGFR levels as they were considered outliers for the purpose of the analyses. Acute rejection episodes were identified by indications of treatment of acute rejection in the applicable follow-up periods from the United Network for Organ Sharing follow-up forms. We did not include acute rejection as a primary explanatory variable in the multivariable models but did report findings stratified by patients who experienced acute rejection during the first 6 months and between 6 and 12 months posttransplantation.
We used multivariable general linear models to assess characteristics associated with baseline eGFR and changes in eGFR from 6 to 12 months and 6 to 24 months posttransplantation. We tested changes over time based on both adjusted and unadjusted models using year of transplantation as the primary explanatory variable and incorporated models with both transplant type and stratified by deceased and living donors. To evaluate the contribution of center heterogeneity to any differences in observed changes in renal function over time, we used hierarchical mixed models. For these models, we included patient characteristics as described for multivariable general linear models and in addition, transplant center as a random effect. We quantified the contribution of across-center variation using the ICC from the mixed model. We validated the primary findings with an additional measure of eGFR, and we repeated our analyses using the recently published CKD-EPI equation (35).
We evaluated effects by discharge immunosuppressive regimens indicated in the SRTR registry, classified by combinations that represented at least 1% of the population based on primary noninduction immunosuppression medications. Differences between regimens were evaluated adjusted for multiple comparisons (Tukey test). All analyses were conducted using SAS (version 9.2, Cary, NC).
1. Meier-Kriesche HU, Li S, Gruessner RW, et al. Immunosuppression
: Evolution in practice and trends, 1994–2004. Am J Transplant
2006; 6(5 pt 2): 1111.
2. Woodle ES, First MR, Pirsch J, et al. A prospective, randomized, double-blind, placebo-controlled multicenter trial comparing early (7 day) corticosteroid cessation versus long-term, low-dose corticosteroid therapy. Ann Surg
2008; 248: 564.
3. Luan FL, Steffick DE, Gadegbeku C, et al. Graft and patient survival in kidney transplant recipients selected for de novo steroid-free maintenance immunosuppression
. Am J Transplant
2009; 9: 160.
4. Schold JD, Santos A, Rehman S, et al. The success of continued steroid avoidance after kidney transplantation in the US. Am J Transplant
2009; 9: 2768.
5. Srinivas TR, Meier-Kriesche HU. Minimizing immunosuppression
, an alternative approach to reducing side effects: Objectives and interim result. Clin J Am Soc Nephrol
2008; 3(suppl 2): S101.
6. Nankivell BJ, Borrows RJ, Fung CL, et al. The natural history of chronic allograft nephropathy. N Engl J Med
2003; 349: 2326.
7. Ekberg H, Tedesco-Silva H, Demirbas A, et al. Reduced exposure to calcineurin inhibitors in renal transplantation. N Eng l J Med
2007; 357: 2562.
8. Flechner SM, Goldfarb D, Solez K, et al. Kidney transplantation with sirolimus and mycophenolate mofetil-based immunosuppression
: 5-year results of a randomized prospective trial compared to calcineurin inhibitor drugs. Transplantation
2007; 83: 883.
9. Flechner SM, Goldfarb D, Modlin C, et al. Kidney transplantation without calcineurin inhibitor drugs: A prospective, randomized trial of sirolimus versus cyclosporine. Transplantation
2002; 74: 1070.
10. Merion RM, Ashby VB, Wolfe RA, et al. Deceased-donor characteristics and the survival benefit of kidney transplantation. JAMA
2005; 294: 2726.
11. Metzger RA, Delmonico FL, Feng S, et al. Expanded criteria donors for kidney transplantation. Am J Transplant
2003; 3(suppl 4): 114.
12. Marks WH, Wagner D, Pearson TC, et al. Organ donation and utilization, 1995–2004: Entering the collaborative era. Am J Transplant
2006; 6(5 pt 2): 1101.
13. Reese PP, Feldman HI, McBride MA, et al. Substantial variation in the acceptance of medically complex live kidney donors across US renal transplant centers. Am J Transplant
2008; 8: 2062.
14. Textor SC, Taler SJ, Driscoll N, et al. Blood pressure and renal function
after kidney donation from hypertensive living donors. Transplantation
2004; 78: 276.
15. Textor S, Taler S. Expanding criteria for living kidney donors: What are the limits? Transplant Rev (Orlando)
2008; 22: 187.
16. Heimbach JK, Taler SJ, Prieto M, et al. Obesity in living kidney donors: Clinical characteristics and outcomes
in the era of laparoscopic donor nephrectomy. Am J Transplant
2005; 5: 1057.
17. Meier-Kriesche HU, Schold JD, Srinivas TR, et al. Lack of improvement in renal allograft survival despite a marked decrease in acute rejection rates over the most recent era. Am J Transplant
18. Hunsicker LG, Adler S, Caggiula A, et al. Predictors of the progression of renal disease in the Modification of Diet in Renal Disease Study. Kidney Int
1997; 51: 1908.
19. Hunsicker LG, Bennett LE. Acute rejection reduces creatinine clearance (Ccr) at 6 months following renal transplantation but does not affect subsequent slope of Ccr [abstract]. Transplantation
1999; 67: S83.
20. Gourishankar S, Hunsicker LG, Jhangri GS, et al. The stability of the glomerular filtration rate after renal transplantation is improving. J Am Soc Nephrol
2003; 14: 2387.
21. Kasiske BL, Gaston RS, Gourishankar S, et al. Long-term deterioration of kidney allograft function. Am J Transplant
2005; 5: 1405.
22. Marcén R, Morales JM, Fernández-Rodriguez A, et al. Long-term graft function changes in kidney transplant recipients. NDT Plus
2010; 3(suppl 2): ii2.
23. Mariat C, Alamartine E, Barthelemy JC, et al. Assessing renal graft function in clinical trials: Can tests predicting glomerular filtration rate substitute for a reference method? Kidney Int
2004; 65: 289.
24. Limaye S, O'Kelly P, Harmon G, et al. Improved graft survival in highly sensitized patients undergoing renal transplantation after the introduction of a clinically validated flow cytometry crossmatch. Transplantation
2009; 87: 1052.
25. Lindemann M, Nyadu B, Heinemann FM, et al. High negative predictive value of an amplified flow cytometry crossmatch before living donor kidney transplantation. Hum Immunol
2010; 71: 771.
26. Graff RJ, Xiao H, Schnitzler MA, et al. The role of positive flow cytometry crossmatch in late renal allograft loss. Hum Immunol
2009; 70: 502.
27. Solez K, Colvin RB, Racusen LC, et al. Banff 07 classification of renal allograft pathology: Updates and future directions. Am J Transplant
2008; 8: 753.
28. Stegall MD, Gloor JM. Deciphering antibody-mediated rejection: New insights into mechanisms and treatment. Curr Opin Organ Transplant
2010; 15: 8.
29. Humar A, Snydman D; AST Infectious Diseases Community of Practice. Cytomegalovirus in solid organ transplant recipients. Am J Transplant
2009; 9(suppl 4): S78.
30. Schold JD, Rehman S, Kayle LK, et al. Treatment for BK virus: Incidence, risk factors and outcomes
for kidney transplant recipients in the United States. Transpl Int
2009; 22: 626.
31. 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.
32. Schold JD, Kaplan B, Howard RJ, et al. Are we frozen in time? Analysis of the utilization and efficacy of pulsatile perfusion in renal transplantation. Am J Transplant
2005; 5: 1681.
33. Moers C, Smits JM, Maathuis MH, et al. Machine perfusion or cold storage in deceased-donor kidney transplantation. N Engl J Med
2009; 360: 7.
34. Meier-Kriesche HU, Ojo AO, Hanson JA, et al. Increased impact of acute rejection on chronic allograft failure in recent era. Transplantation
2000; 70: 1098.
35. Levey AS, Stevens LA, Schmid CH, et al. A new equation to estimate glomerular filtration rate. Ann Intern Med
2009; 150: 604.
36. Kaplan B, Schold J, Meier-Kriesche HU. Overview of large database analysis in renal transplantation. Am J Transplant
2003; 3: 1052.
37. Meyers CM, Kirk AD. Workshop on late renal allograft dysfunction. Am J Transplant
2005; 5: 1600.
38. Serur D, Saal S, Wang J, et al. Deceased-donor kidney transplantation: Improvement in long-term survival. Nephrol Dial Transplant
39. Hariharan S, McBride MA, Cohen EP. Evolution of endpoints for renal transplant outcome. Am J Transplant
2003; 3: 933.
40. Kaplan B, Schold J, Meier-Kriesche HU. Poor predictive value of serum creatinine for renal allograft loss. Am J Transplant
2003; 3: 1560.
41. Hariharan S, Johnson CP, Bresnahan BA, et al. Improved graft survival after renal transplantation in the United States, 1988 to 1996. N Engl J Med
2000; 342: 605.
42. Meier-Kriesche HU, Schold JD, Kaplan B. Long-term renal allograft survival: Have we made significant progress or is it time to rethink our analytic and therapeutic strategies? Am J Transplant
2004; 4: 1289.
43. Stevens LA, Coresh J, Greene T, et al. Assessing kidney function—Measured and estimated glomerular filtration rate. N Engl J Med
2006; 354: 2473.