Chronic allograft nephropathy (CAN) in protocol biopsies is associated with graft loss while the association between subclinical rejection (SCR) and outcome has yielded contradictory results. We analyze the predictive value of SCR and/or CAN in protocol biopsies on death-censored graft survival. Since 1988, a protocol biopsy was done during the first 6 months in stable grafts with serum creatinine <300 μmol/L and proteinuria <1 g/day. Biopsies were evaluated according to Banff criteria. Borderline changes and acute rejection were grouped as SCR. CAN was defined as presence of interstitial fibrosis and tubular atrophy. Mean follow-up was 91 ± 46 months. Sufficient tissue was obtained in 435 transplants. Biopsies were classified as normal (n = 186), SCR (n = 74), CAN (n = 110) and SCR with CAN (n = 65). Presence of SCR with CAN was associated with old donors, percentage of panel reactive antibodies and presence of acute rejection before protocol biopsy. Cox regression analysis showed that SCR with CAN (relative risk [RR]: 1.86, 95% confidence interval [CI]: 1.11–3.12; p = 0.02) and hepatitis C virus (RR: 2.27, 95% CI: 1.38–3.75; p = 0.01) were independent predictors of graft survival. In protocol biopsies, the detrimental effect of interstitial fibrosis/tubular atrophy on long-term graft survival is modulated by SCR.
Different centers have conducted prospective studies of protocol biopsies in order to characterize histological lesions in renal allografts with otherwise stable renal function (1–5). There is agreement that subclinical rejection (SCR), that is, the presence of histological lesions of rejection in well functioning grafts, peaks during the initial months of post-transplantation and declines thereafter (6,7). The incidence of SCR is rather variable between centers and is influenced by the timing of protocol biopsy, the presence of an episode of acute rejection before the protocol biopsy and immunosuppressive treatment (6,8–10). The potential influence of other variables on the incidence of SCR such as donor source (deceased vs. living) or recipient type (pediatric vs adult) has not been properly evaluated (11,12).
In protocol biopsies, chronic allograft nephropathy (CAN) is a frequent finding that progresses rapidly during the first year and slower but steadily thereafter (4,13). The presence of CAN is associated with different clinical variables such as donor age, an episode of acute rejection before biopsy, the use of calcineurin inhibitors or timing of protocol biopsy (2,3). The presence of CAN in protocol biopsies has been consistently associated with a decreased long-term renal allograft survival. Furthermore, it has been shown that CAN is an independent predictor of graft failure from renal function and proteinuria, suggesting that protocol biopsies contain useful information to predict outcome that is not present in clinical data (2,3,8).
In studies of serial protocol biopsies, the presence of SCR precedes the appearance of CAN suggesting that SCR should be also associated with a poorer long-term outcome (6,14). However, studies evaluating the predictive value of SCR on long-term graft survival have yielded contradictory results (8,11,14,15). Thus, in the present study we analyzed long-term-death-censored graft survival in a large cohort of deceased donor renal transplants with a protocol biopsy taking into consideration the presence of SCR, CAN or the association of both diagnosis in the same biopsy.
Patients and Methods
Since June 1988, a prospective protocol renal allograft biopsy is performed in patients who give their informed consent and fulfill the following criteria: (a) serum creatinine <300 μmol/L; (b) proteinuria <1g/24 h and (c) stable renal function, defined as variability of serum creatinine of less than 15% during 2 weeks before and after biopsy.
The timing of protocol biopsy has been progressively delayed in order to better focus on the study of early chronic lesions. Between 1988 and 1990 the biopsy was done during the 1st month while later on the biopsy was done between the 3rd and 6th month. For the present study, we considered patients biopsied until December 2003.
Biopsies were performed under ultrasound guidance with an automated gun, processed for routine light microscopy and stained with hematoxilin-eosine, periodic acid Schiff, Masson's trichrome and silver-methenamine. Renal lesions in protocol biopsies were blindly graded and diagnosed according to the 1997 Banff criteria (16) in the absence of any clinical information. Only biopsies containing at least one arterial section and one glomerular profile were considered. SCR was defined as the presence of i-score ≥1 and t-score ≥1. The presence of CAN was defined as ci+ct scores ≥ 2.
Protocol biopsies were not considered in the clinical management of patients.
The following variables were evaluated in each patient at the time of surgery: age and gender of the donor and the recipient, number of transplant, presence of hepatitis C virus antibodies, last panel reactive antibodies, number of HLA mismatches, cold ischemia time and immunosuppressive therapy.
After surgery, the presence of delayed graft function and acute rejection were evaluated. At the time of protocol biopsies and during follow-up serum creatinine, proteinuria, total serum cholesterol, CsA or TAC dose and levels were recorded.
Patient and death-censored graft survival were evaluated on December 2004.
Results are expressed as the mean ± standard deviation. Comparison between groups was performed by means of chi-square test for categorical data, Kruskal-Wallis test for ordinal data and analysis of variance with Scheffépost hoc test for normally distributed continuous variables. Death-censored graft survival was considered the outcome variable in the present study. Graft survival was estimated by the Kaplan-Meier method and comparison between groups was done by the log-rank test. Univariate and multivariate Cox regression analysis were employed to analyzed variables associated with death-censored graft survival. All p values were 2-tailed and a p-value < 0.05 was considered significant.
During the study period, 472 protocol biopsies were done in 463 patients. In 37 cases (7.8%), an insufficient sample for histological evaluation was obtained. Accordingly, 435 renal transplants were included in the present study. Clinical characteristics of patients are summarized in Table 1. Immunosuppressive treatment received at the time of transplantation was as follows: cyclosporine (CsA) and prednisone (n = 23); CsA and prednisone associated with polyclonal or monoclonal anti-lymphocytic antibodies (n = 171); CsA, azathioprine and steroids (n = 35); CsA, mycophenolate mofetil and steroids (n = 124); tacrolimus, mycophenolate mofetil and steroids (n = 49); a calcineurin-free regimen based on anti-lymphocytic antibodies, mycophenolate mofetil and steroids (n = 23); CsA, sirolimus and steroids (n = 10).
Table 1. Demographic characteristics of patients and clinical evolution after transplantation
Mean ± SD
ESRD = end-stage renal disease; APKD = adult polycystic kidney disease; PRA = panel reactive antibodies; HCV = hepatitis C virus; HLA = human leukocyte antigen.
Donor age (years)
38 ± 17
Donor gender (male/female)
Recipient age (years)
46 ± 14
Recipìent gender (male/female)
Etiology of ESRD
8 ± 19
HCV status (negative/positive)
3.0 ± 1.1
Cold ischemia time (h)
22 ± 6
Delayed graft function (no/yes)
Acute rejection (no/yes)
At the end of follow-up 31 patients died with a functioning graft and 81 grafts were lost. Causes of patient's death were cardiovascular disease (n = 15), neoplasia (n = 11), infection (n = 3) and chronic liver failure (n = 2). Causes of graft loss were CAN (n = 63), hepatitis C virus associated glomerulonephritis (n = 7), de novo membranous glomerulonephritis (n = 4), recurrence of primary glomerulonephritis (n = 3), nontreatment compliance (n = 3) and transplantectomy due to Kaposi's sarcoma involving the graft hilium (n = 1). Mean follow-up was 91 ± 46 months.
According to Banff criteria 265 biopsies were adequate (at least 10 glomerular and 2 arterial sections), 89 were marginal (between 7 and 9 glomerular and at least 1 arterial section) and 81 were insufficient (between 1 and 6 glomerular sections and at least 1 arterial section). Marginal samples contained 8 ± 1 and insufficient samples 4 ± 1 glomerular sections, respectively. Clinical data at the time of protocol biopsy are summarized in Table 2.
Table 2. Clinical data at the time of protocol biopsy
Mean ± SD
Time of biopsy (days)
120 ± 44
Serum creatinine (μmol/L)
142 ± 48
0.31 ± 0.22
Mean arterial pressure (mmHg)
102 ± 12
Histological diagnosis according to Banff criteria were normal (n = 186, 43%), SCR (n = 74, 17%), CAN (n = 110, 25%) and CAN associated with SCR (n = 65, 15%). The incidence of CAN in patients with SCR was 47% and in patients without SCR was 37%, p = 0.05).
The severity of SCR was not different in biopsies with or without CAN (n = 49 (75.4%), n = 13 (20.0%) and n = 3 (4.6%) versus n = 61 (82.4%), n = 11 (14.8%) and n = 2 (2.7%) for borderline changes, acute rejection type I and II, respectively). The severity of CAN was also not different in biopsies displaying CAN with or without SCR (n = 52 (80.0%), n = 12 (18.5%) and n = 1 (1.5%) vs. n = 88 (80.0%), n = 21 (19.0%) and n = 1 (1.0%) for CAN type I, II and III, respectively). Similarly, the severity of cg and cv scores was not significantly different in patients displaying CAN with or without SCR (0.24 ± 0.43 vs. 0.35 ± 0.48 for cg and 0.45 ± 0.77 vs. 0.43 ± 0.77 for cv).
Clinical characteristics of patients according to histological diagnosis
Demographic characteristics and clinical evolution of patients according to histological diagnosis are summarized in Table 3. Patients with CAN received kidneys from older donors, the number of acute rejection episodes was higher in patients with SCR and the degree of sensitation was the highest in patients with CAN and SCR in the protocol biopsy. Only two patients were treated for a clinical episode of acute rejection after protocol biopsy.
Table 3. Characteristics of patients according to histological diagnosis
CAN + SCR
SCR = subclinical rejection; CAN = chronic allograft nephropathy; m = male; f = female; PRA = panel reactive antibodies; HCV = hepatitis C virus; HLA = human leukocyte antigens; CIT = cold ischemia time; DGF = delayed graft function. 1p < 0.05 versus normal group, 2p < 0.05 versus SCR group, 3p < 0.05 versus CAN group.
Donor age (years)
36 ± 16
35 ± 19
40 ± 161,2
41 ± 171,2
Donor gender (m/f)
Recipient age (years)
45 ± 13
45 ± 16
45 ± 14
46 ± 13
Recipient gender (m/f)
7 ± 20
8 ± 22
5 ± 13
14 ± 231,3
3.0 ± 1.1
2.9 ± 1.0
3.0 ± 1.1
3.1 ± 0.9
22 ± 6
22 ± 6
23 ± 6
24 ± 7
Acute rejection (%)
The incidence of SCR was also associated with immunosuppressive treatment. Patients receiving a tacrolimus-based regimen showed the lowest incidence of SCR (16%), while patients receiving a calcineurin-free regimen showed the highest incidence (56%). On the other hand, the incidence of CAN was not statistically different in patients treated with CsA, tacrolimus or a calcineurin-free regimen (40%, 43% and 30%, respectively) (Figure 1).
At the time of protocol biopsy, patients with CAN associated with SCR showed the poorest graft function and the highest degree of proteinuria (Table 4). As shown in Table 4, trough CsA or tacrolimus plasma levels at the time of biopsy were not different in the four groups of patients.
Table 4. Clinical data at the time of protocol biopsy according to histological diagnosis
SCR = subclinical rejection; CAN = chronic allograft nephropathy; SCr = serum creatinine; MAP = mean arterial pressure; CsA = cyclosporine; TAC = tacrolimus. 1p < 0.05 versus normal group, 2p < 0.05 versus SCR group, 3p < 0.05 versus CAN group. In the case of cyclosporine and tacrolimus levels the number of patients is shown between brackets.
132 ± 42
136 ± 40
150 ± 501,2
166 ± 551,2,3
0.27 ± 0.19
0.31 ± 0.21
0.33 ± 0.241
0.42 ± 0.271,2,3
99 ± 12
102 ± 12
104 ± 121
104 ± 121
192 ± 76 (155)
192 ± 70 (59)
209 ± 82 (90)
200 ± 92 (58)
9 ± 3 (24)
10 ± 4 (5)
10 ± 4 (20)
7 ± 4 (4)
Death-censored graft survival and histological diagnosis
Death-censored graft survival was lower in patients with SCR associated with CAN than in the other groups (Figure 2). Death-censored graft survival in patients with borderline acute rejection and patients with acute rejection was not different.
Univariate and multivariate Cox regression analysis was performed to analyze clinical, analytical and histological variables associated with death-censored graft survival. For this analysis, continuous variables were binarized according to the mean and all regressions were done adjusting for the timing of protocol biopsy. In Table 5, variables associated with death-censored graft survival in the univariate Cox regression analysis are summarized. Neither immunosuppressive treatment nor acute rejection episodes before the protocol biopsy were associated with graft survival. Finally, multivariate Cox regression analysis showed that only the presence of SCR associated with CAN in the protocol biopsy and hepatitis C virus were independent predictors of death-censored graft survival (Table 5).
Table 5. Variables associated with death censored graft survival by univariate and multivariate Cox regression analysis adjusting for the timing of the protocol biopsy
RR = relative risk; CI 95%= confidence interval of 95%; PRA = panel reactive antibodies; HCV = hepatitis C virus; SCr = serum creatinine at the time of protocol biopsy; CAN + SCR = chronic allograft nephropathy with subclinical rejection in the protocol biopsy.
CAN + SCR
The association between histological diagnosis and graft loss was confirmed considering only hepatitis C virus negative patients and graft loss due to CAN as the outcome variable (relative risk [RR]: 2.04, 95% confidence interval [CI]: 1.02–4.08; p = 0.04). Similarly, the association between histological diagnosis and death-censored graft survival was also confirmed when data were reanalyzed only considering the 331 patients with serum creatinine <200 μmol/L at the time of biopsy and a biopsy containing at least 7 glomeruli (RR: 1.97, 95% CI: 1.02–3.82; p = 0.04).
In the present study, we observed that death-censored long-term graft survival was significantly decreased in patients simultaneously displaying SCR and CAN in the protocol biopsy. Graft survival in patients displaying SCR or CAN was not different from patients with a normal biopsy. Furthermore, intermediate outcome variables such as serum creatinine, mean arterial pressure or proteinuria at the time of biopsy were higher in patients simultaneously displaying SCR and CAN. This result suggests that the detrimental effect of CAN on outcome is modulated by the presence of SCR.
The criteria to define SCR in different studies have not been homogeneous. Some authors have arbitrarily defined SCR as the presence of at least acute rejection grade Ia while others have also included patients with borderline changes or even nonspecific inflammatory changes (1–4,11,17). The difference between borderline changes and acute rejection depends on the number of inflammatory cells in the tubular epithelium (16), but there can be some overlap between both diagnosis, since the evaluation of tubulitis is associated with a significant inter-observer variability (18,19). In studies evaluating the predictive value of biopsies performed in clinical episodes of acute rejection, outcome in patients with borderline changes and acute rejection grade I is not different (20). Moreover, in protocol biopsy studies there is controversy whether borderline changes and acute rejection imply a different prognosis (6,8,11,15,17). In the present study, three-quarters of patients with SCR displayed borderline changes in agreement with previous studies. Furthermore, graft survival was not different in patients with borderline changes and acute rejection. Thus, for the present study, both groups were considered as SCR.
The presence of CAN in protocol biopsies has been repeatedly associated with decreased renal allograft survival (2–4). Nevertheless, studies evaluating the predictive value of SCR on renal allograft outcome are not conclusive. In a prospective randomized study performed in a single center, it has been shown that treatment of SCR evaluated by means of early protocol biopsies with steroids pulses is associated with a lower chronic damage score in 6-month protocol biopsies and with a lower 2-year serum creatinine (14). The sequential association between SCR and CAN has been confirmed in a large study of serial protocol biopsies (4,6). In the present study, we observed an association between SCR and CAN in the protocol biopsy, further suggesting a link between both entities.
In our study, patients with SCR without CAN as well as patients with CAN without SCR had an outcome that was not different from patients with a normal protocol biopsy, suggesting that neither SCR nor CAN diagnosed at an early stage are major determinants of graft survival. On the contrary, graft survival was lower in patients with SCR and CAN. Recently, Cosio et al. have observed that patients with SCR and CAN fair less well than patients only displaying CAN in a 1 year protocol biopsy study (21). However, in their study the number of patients only displaying SCR was too low for statistical analysis. Thus, our data suggest that the presence of persistent inflammation in kidneys already displaying chronic lesions is associated with a higher probability for progression to end-stage renal disease. Our data do not allow to further explore mechanisms leading to graft loss in patients with SCR and CAN, but two different mechanisms of renal damage may fit this observation. In one hand, it could be proposed that patients with SCR and CAN have a more severe inflammation than patients with SCR without CAN. However, the severity of acute lesions evaluated according to Banff criteria in both groups was not different. Unfortunately, we have not studied the phenotype of infiltrating cells or cytokine profile in order to explore qualitative differences in inflammatory cells. Alternatively, it could be argued that kidneys with CAN have a decreased capacity for tissue repair in front of different insults including persistent inflammation (22). In our study, patients with SCR and CAN showed the highest degree of sensitation and suffered from the highest incidence of acute rejection episodes before the protocol biopsy. This association favors the idea that alloimmune response maybe more intense in this group of patients. On the other hand, patients with CAN, either with or without SCR, received kidneys from older donors which favors that kidneys with CAN may be more susceptible to inflammatory mediated damage. Taken together, our results suggest that both mechanisms, enhanced alloimmune response and impaired tissue repair capacity, may contribute to decreased graft survival in patients displaying SCR with CAN. Regardless of the mechanism leading to graft loss in these patients, an epidemiological consequence of this observation is that previous studies evaluating the predictive value of CAN on outcome may have overestimated its impact while studies evaluating SCR may have underestimated its relevance.
Finally, our data raise the question whether patients displaying SCR with or without CAN could benefit from a specific treatment. Rush et al. showed that treatment of SCR with steroid pulses may slow the progression of chronic lesions (14). Moreover, tacrolimus therapy is associated with a reduced incidence of SCR, specially when it is associated with MMF (4,9,10). These data suggest that SCR can be treated with increased immunosuppression. However, in the present study we did not observe a negative impact of SCR without CAN on long-term graft survival. Accordingly, the potential benefit of increasing immunosuppression in these patients is doubtful and may be counterbalanced by adverse events associated with overimmunosuppression (23). On the contrary, our data suggest that patients displaying SCR with CAN are the most interesting group to be considered in a trial aimed to prevent progression of CAN. Recently, it has been shown that treatment with steroid pulses has no benefit in patients displaying SCR with CAN in serial protocol biopsies done from the first year (12). Thus, clinical prospective trials will be necessary to evaluate whether these patients may benefit of increasing nonnephrotoxic immunosuppression (24–26).
In summary, our data suggest that in protocol biopsies the detrimental effect of interstitial fibrosis/tubular atrophy on long-term graft survival is modulated by SCR.
This work was supported by FIS grants to F. Moreso (PI 04/0177) and D. Seron (PI 04/0086). M. Ibernon was supported by an IDIBELL grant and M Gomà by an Astellas Spain Grant.