- Top of page
We are studying two cohorts of kidney transplant recipients, with the goal of defining specific clinicopathologic entities that cause late graft dysfunction: (1) prevalent patients with new onset late graft dysfunction (cross-sectional cohort); and (2) newly transplanted patients (prospective cohort). For the cross-sectional cohort (n = 440), mean time from transplant to biopsy was 7.5 ± 6.1 years. Local pathology diagnoses included CAN (48%), CNI toxicity (30%), and perhaps surprisingly, acute rejection (cellular- or Ab-mediated) (23%). Actuarial rate of death-censored graft loss at 1 year postbiopsy was 17.7%; at 2 years, 29.8%. There was no difference in postbiopsy graft survival for recipients with versus without CAN (p = 0.9). Prospective cohort patients (n = 2427) developing graft dysfunction >3 months posttransplant undergo ‘index’ biopsy. The rate of index biopsy was 8.8% between 3 and 12 months, and 18.2% by 2 years. Mean time from transplant to index biopsy was 1.0 ± 0.6 years. Local pathology diagnoses included CAN (27%), and acute rejection (39%). Intervention to halt late graft deterioration cannot be developed in the absence of meaningful diagnostic entities. We found CAN in late posttransplant biopsies to be of no prognostic value. The DeKAF study will provide broadly applicable diagnostic information to serve as the basis for future trials.
The two major problems in kidney transplantation today are late graft failure and the shortage of donor organs. These problems are interrelated; kidney allograft failure is currently the fourth leading cause of end stage renal disease in the United States and Canada, and recipients returning to the waiting list for deceased donor kidney transplantation contribute to the organ shortage. The use of modern immunosuppressive regimens has resulted in a significant decrease in acute rejection rates during the first year after kidney transplantation. In some studies, this has been associated with improvement in long-term outcome (1,2) at least partially due to better 1-year function (3). The stability of long-term function (the slope of the GFR) has also improved perhaps due to superior control of rejection with newer protocols (4,5). In contrast, other large series, while noting a significant decrease in early rejection rates, have shown no significant improvement in long-term outcome (6,7).
There is no doubt that late graft loss continues to be a major challenge, with almost 4300 kidney transplant recipients returning to dialysis in the United States and Canada in 2005, most of them beyond 1 year after transplantation. Furthermore, improvements in graft survival have clearly plateaued with lower acute rejection rates and better early outcomes, but late graft loss remains an ongoing problem (8).
Determining the optimal treatment of recipients with late graft dysfunction (the troubled transplant) has been limited for a number of reasons: (a) deterioration of function can begin at any time following kidney transplantation and can be insidious; (b) there are likely many causes; (c) there are only a limited number of cases in any given year at a single institution; and, perhaps most importantly; (d) biopsy is usually done late in the clinical course (if at all), after the active phase of damage has occurred and only scar tissue is observable. The majority of these cases have been labeled ‘chronic rejection,’‘chronic allograft nephropathy’ (CAN) or, more recently, ‘interstitial fibrosis and tubular atrophy, no evidence of any specific etiology’ as represented by interstitial fibrosis and tubular atrophy (IF/TA) on biopsy (9). However, these terms do not represent specific entities from an etiologic, physiologic or pathologic point of view.
To overcome these diagnostic shortcomings, the Long-Term Deterioration of Kidney Allograft Function (DeKAF) study is investigating ‘the troubled transplant’ phenotype with the aim of attributing dysfunction and/or loss to specific clinico-pathologic entities. By creating a multicenter consortium (so there are a sufficient number of patients) and doing biopsies early in the course of new-onset late graft dysfunction, we hope to overcome the above issues. Our goal is to show, with the use of early biopsy and appropriate analyses, that: (1) most chronic graft dysfunction and late graft failure can be attributed to well-understood entities that injure the nephron; (2) these disease processes or other sources of injury act at the time that deterioration is detected; (3) at least some of these entities are preventable and/or treatable and (4) the nihilism associated with the concept of chronic rejection, CAN, or IF/TA should be replaced by the assumption that therapy can prevent or slow progression in many cases. Defining specific entities will potentially allow development of intervention trials.
In this first publication, we document the local pathologists’ diagnoses in two cohorts of recipients with new-onset late graft dysfunction and show that these previously stable grafts with new-onset late graft dysfunction have a relatively high rate of subsequent graft failure. We note a high prevalence of acute rejection in long-term allografts with new dysfunction. And we show that the diagnosis of CAN in late posttransplant biopsies is of no prognostic significance.
- Top of page
DeKAF is a multicenter, observational study conducted at seven transplant centers in the United States and Canada, funded by the National Institutes of Health. Our hypotheses are that: (1) there are multiple, definable entities leading to late graft dysfunction; (2) these entities can be differentiated by means of clinical, serologic, and pathologic studies and (3) progressive late graft dysfunction is due to identifiable, currently operating injurious processes and not the consequence of a past injury. Our long-term goal is to understand and reduce long-term kidney transplant deterioration. A detailed description of the study can be found at http://www.clinicaltrials.gov (NCT00270712). Institutional Review Board approval was obtained at all participating sites. We are following two distinct populations of kidney transplant recipients: a cross-sectional cohort and a prospective cohort.
The cross-sectional cohort (launched February 1, 2006) consists of recipients transplanted prior to October 1, 2005, having a last reported serum creatinine level ≤ 2 mg/dl between January 1, 2005 and January 1, 2006, and subsequently developing graft dysfunction (defined as a ≥25% increase in creatinine level or new onset proteinuria leading to a biopsy). Inclusion and exclusion criteria are outlined in Table 1. At the time of biopsy, recipients were enrolled in the study and clinical data entered into the database. Patients in the cross-sectional cohort enroll at different times after transplant and at different times after January 2006; for example, a recipient with a baseline creatinine of 1.2 mg/dL is biopsied if the creatinine reaches 1.5 mg/dL (e.g. at year 2, 4 or at any time thereafter). This cohort provides an overview of the troubled kidney irrespective of the time from transplant. There is not a control group; however, the importance of this cohort is in defining distinct clinico-pathologic entities that may occur late posttransplant (and may be difficult to determine in an inception cohort [because of the need for long-term follow-up]).
Table 1. Inclusion criteria for the prospective and cross-sectional cohorts
|Prospective cohort inclusion criteria:|
|Kidney (or kidney-pancreas) transplant recipient, no more than 10 days posttransplant.|
|No organs other than possibly a pancreas transplanted simultaneously with the qualifying kidney transplant.|
|Cross-sectional cohort inclusion criteria:|
|Kidney (or kidney-pancreas) transplant recipient prior to October 1, 2005.|
|No organs other than the qualifying kidney or simultaneous kidney-pancreas.|
|The last available serum creatinine drawn between January 1, 2005 and January 1, 2006 ≤2 mg/dL.|
|Clinically indicated biopsy.|
Prospective cohort recipients are enrolled at the time of kidney or simultaneous kidney-pancreas transplant (provided no other organs are transplanted; Table 1). At the time of transplant (and consent), clinical information is entered into the database; baseline serum creatinine level is determined at 3 months posttransplant. Those demonstrating subsequent deterioration of graft function (defined as ≥25% increase in serum creatinine level over baseline or new onset proteinuria) undergo biopsy (hereafter referred to as the ‘index biopsy’). This cohort provides clinical information on all patients; there is a control group, without graft dysfunction, that is not biopsied. The prospective cohort was launched on October 1, 2005.
For both cohorts, medical history, recipient and donor characteristics, and operative and postoperative information are collected at the time of enrollment. For both, follow-up data are collected every 6 months; additionally, in the prospective cohort, data collection occurs at 1 and 3 months posttransplant. Event-driven data collection is triggered by any of the following: a clinically indicated biopsy; a local diagnosis of treated acute rejection (biopsy-proven or treated without obtaining a biopsy); selected viral, fungal or mycobacterial infections; and graft failure (return to dialysis, retransplantation or death).
Baseline creatinine level in the prospective cohort
To exclude the effects of permanent damage due to acute rejection, the creatinine reference point is reset for each participant 6 weeks after the initiation of therapy for acute rejection. The new value for the creatinine reference point is defined to be the average of up to three serum creatinine measurements taken at least 1 week apart and drawn between 6 and 12 weeks following initiation of treatment for acute rejection.
Allograft biopsies read by the local pathologist are used to guide clinical care according to local protocols.
Representative histologic sections (H&E, silver, PAS, trichrome stains and 11 unstained sections for additional studies) are submitted to a central laboratory for analysis. All biopsies were interpreted by the same pathologist (JG), using the Banff’97 classification (10). In addition, biopsies were scored for peritubular capillary infiltrates and for inflammation in areas of atrophy (iatr) and tubulitis in areas of atrophy (tatr). Scoring of peritubular capillary infiltrates was done as per the Banff 2005 classification (11). Iatr and tatr were scored in the same manner as i and t, except being scored in areas of fibrosis and/or atrophy: tatr 1 = 1–4 cells/atrophic tubule; tatr 2 = 5–10 cells; tatr 3 = 10+ cells; iatr 0 = inflammation in up to 10% of scarred cortex; iatr 1 = 10–24%; iatr 2 = 25–49%; and iatr 3 = 50%+ of scarred cortex.
Determination of donor specific antibodies
Serum samples (2 mL) collected at the time of each biopsy are processed, frozen at –80 degrees and sent to a central laboratory (UCLA) in batches, with information regarding the patient and donor HLA types and the patient's pretransplant sensitization status.
For both cohorts, we studied time from transplant to biopsy, renal function, local pathologists’ diagnoses, actuarial graft survival, and slope of the inverse of creatinine. For the cross-sectional cohort, we compared rate of graft loss (postbiopsy) for those with versus without a local diagnosis of CAN. Rates are given as Kaplan–Meier point estimates with confidence intervals based on the equal probability confidence bands for the entire survival curve. The log-rank test was used to compare survival curves.
Slope of the inverse of creatinine was estimated using separate longitudinal random effects models for each cohort. Slope is measured from month 3 posttransplant onwards in the prospective cohort, while in the cross-sectional cohort, a piecewise-linear model is used to estimate slope from biopsy to month 6 postbiopsy and from month 6 postbiopsy onwards. In each model, random effects are included for the intercept and each slope term (one effect for the overall slope for the prospective cohort, and two slope effects in total for the cross-sectional cohort, one for each linear segment). Because follow-up is censored for graft loss or death, estimates of slopes are potentially exposed to bias.
Confidence intervals for proportions are based on the normal approximation to the binomial distribution with continuity correction.
- Top of page
Late graft dysfunction and graft loss remain major challenges after renal transplantation. Mild elevations in serum creatinine levels have been associated with increased cardiac risk; thus late graft dysfunction may also contribute to the increased cardiovascular morbidity and mortality observed after a kidney transplant.
Some causes of late graft dysfunction are well defined (e.g. those due to recurrent disease). However, the majority of cases are attributed to the nonspecific diagnoses of CAN, chronic rejection, and IF/TA. The DeKAF study was designed to challenge the concept that the majority of cases of late graft dysfunction can be attributed to a single entity with a common (but unknown) pathogenesis and predictable rate of deterioration of renal function. In particular, we question the utility of nonspecific terms such as ‘CAN’ and hope to encourage the identification of individual specific entities to explain the deterioration and loss of these grafts.
The two patient cohorts in this study provide data regarding new onset late graft dysfunction in (a) long-term ‘stable’ recipients, and (b) new recipients. For the cross-sectional cohort, we avoided limiting the patient selection to a particular period posttransplant because progressive graft dysfunction can start at any time. As enrollment increases, we will determine the rate of new onset graft dysfunction at various posttransplant intervals. We excluded recipients with advanced graft dysfunction based on the postulate that biopsy early in the course is essential to distinguish among the entities that may cause late graft loss, and that late biopsies exhibit mainly scarring not likely to be of benefit in understanding pathogenesis. The cross-sectional cohort provides information on the troubled kidney that develops graft dysfunction at any time posttransplant.
The prospective cohort has the advantage of providing a control group, as some recipients will develop late dysfunction; while others will not. We will thus be able to identify clinical risk factors for late graft dysfunction and for progression of functional deterioration for each diagnostic entity. However, long-term follow-up will be required. To date, 2-year actuarial graft survival is 92.2%; death-censored survival, 96.7%. The prospective cohort also will allow us to confirm the characterizations of individual entities (or groups of entities) as defined in the cross-sectional cohort or to learn if recipients with earlier graft dysfunction have different entities or a different prognosis for each entity; and, will allow us to determine clinical risk factors for graft dysfunction and histopathology. For this cohort, the rate of development of late graft dysfunction (index biopsy) was 8.8% between month 3 and 1 year, 18% by year 2.
To date, we have a number of findings important to both our long-term study and to other studies of late kidney allograft functional deterioration. First, stable grafts that show new onset late graft dysfunction have a high rate of subsequent graft failure. In the cross-sectional cohort, we are clearly capturing this population of ‘troubled kidneys’: even though the mean serum creatinine level was 1.4 ± 0.3 mg/dL as of January 1, 2006, death-censored graft loss within 1 year of subsequent development of graft dysfunction was 17%; at 2 years, 29.8%. In the prospective cohort, death-censored graft survival was significantly worse for those having index biopsies after 3 months posttransplant. Graft loss was extremely rare among those not requiring biopsy, indicating graft dysfunction as the precursor of graft failure.
Second, we have described the local pathologists’ interpretation of biopsies at the time of new onset late graft dysfunction. It is of interest that biopsies from both cohorts had relatively similar local diagnoses (Table 3). As expected, ‘CAN’ was common in both cohorts: 48% in the cross-sectional cohort and 27% in the prospective cohort. The central reading of these biopsies will attempt to discriminate, within the broad nonspecific diagnosis of CAN, individual entities with different risk factors, histopathological phenotypes and ultimately prognosis.
Third, a high proportion of recipients in both cohorts were diagnosed as having acute rejection. Although this was expected for the prospective cohort (mean time from transplant to biopsy, 1.0 ± 0.6 years), it was a surprising observation in the cross-sectional cohort whose new onset of dysfunction was relatively late posttransplant (7.5 ± 6.1 years). Also of note was the observation that a similar proportion of recipients in each cohort were diagnosed with antibody-mediated rejection. These unexpected findings emphasize the need to look for a specific cause of late allograft dysfunction, as failure to recognize rejection would deprive a significant number of patients of beneficial changes in their immunosuppression regimen.
Fourth, and perhaps the most important finding of our study, to date, is the observation that, in the cross-sectional cohort, the local diagnosis of CAN was of no prognostic significance (Figure 2). Those with and without CAN had identical patterns of postbiopsy graft failure. There was also no functional difference (for those with and without CAN) in the postbiopsy slope of 1/creatinine versus time. Clearly, for recipients with new onset late graft dysfunction, the nonspecific diagnosis of CAN does not define a group of patients with a different outcome.
A fifth and potentially important observation is the high frequency of inflammation in areas of interstitial atrophy and fibrosis (iatr) (69%), and tubulitis in areas of tubular atrophy (tatr) (81%). Historically, Banff scoring has excluded both these observations. However, the Banff’97 conference encouraged additional studies to evaluate adding these to the scoring system. Our data show that in biopsies with Banff ‘i’ scores = 0, both iatr (50%) and tatr (72%) are common. Mengel et al. recently reported that scoring total inflammation (i.e. inflammation throughout the graft) was better than the using current Banff ‘i’ score (which excludes inflammation in areas of atrophy) in predicting of graft outcome (12).
Part of our study design for recipients in the prospective cohort is to reset the baseline creatinine level after treatment of an acute rejection episode, and then to do a biopsy if the creatinine rises from that new baseline. Meier–Kriesche and others have shown that recipients whose creatinine does not return to baseline after rejection treatment have decreased graft survival, and our data confirms this observation (7,13,14). However, this observation provides no understanding of the etiology of the increased risk. Others have shown no relationship between serum creatinine level and the subsequent rate of deterioration of graft function over time (4,15,16). In fact Meier–Kriesche et al. noted an increased rate of subsequent rejection episodes in those whose creatinine level did not return to baseline (7).
The purpose of the DeKAF study is the identification of individual histopathologic entities that cause late graft dysfunction, with the ultimate goal of developing intervention trials—by our consortium and others—to reduce the incidence of late graft dysfunction and loss. For the prospective cohort, we have defined the rate of new onset graft dysfunction; for both cohorts, we have defined the rate of graft loss after new dysfunction, and the local pathological biopsy interpretation. We have shown that the local diagnosis of CAN is of no prognostic significance. Our data can be used as background information for future interventional studies. Development of intervention trials to prevent or minimize graft dysfunction in kidney transplant recipients will require differentiation of clinically meaningful entities to better understand the phenotype of the ‘troubled transplant’.