Protocol, management or surveillance renal allograft biopsies have provided insight into pathogenesis and guided clinical management since the 1980s (1). Routine assessment of pathology at fixed time points irrespective of renal function, has led to major advances in our understanding of the pathophysiology of graft injury and the sequence of events leading to late allograft failure.
Early detection and treatment of subclinical rejection (SCR), before renal dysfunction, improves outcomes (2–4). Decreased prevalence of SCR with potent immunosuppression reduces this benefit in standard risk recipients, but utility is retained when transplanting high-risk patients and for monitoring steroid or calcineurin inhibitor (CNI) withdrawal (5). Biopsy is undertaken in multiple different environments ranging from dedicated outpatient facilities allowing 4-hour discharge, through to expensive multidisciplinary in-patient admissions. The justification for “protocol” biopsy has been misunderstood by some as an experimental procedure, rather than as part of a standardized program of clinical surveillance to detect subclinical pathology.
The aim of this review is to summarize current evidence on the use of surveillance histology in clinical management of renal transplant recipients.
Burdick et al. found lymphocytic infiltrates in protocol biopsies at 1 and 4 weeks after transplantation from patients with normal graft function (1). Termed SCR by Rush using the Banff schema, it occurred in 30% of well-functioning grafts by 3 months after transplantation using cyclosporine-based immunosuppression (6,7). Nankivell noted SCR in 45.7% at 3 months was associated with greater interstitial fibrosis and tubular atrophy (IF/TA) by 12 months (8). Legendre confirmed the association of clinically silent persistent inflammation with chronic tubulointerstitial damage (9) and others have reported impact on survival of fibrosis and inflammation compared with fibrosis alone (10–12).
Rush et al. randomized 72 low-risk deceased donor renal transplant recipients treated with cyclosporine, azathioprine and prednisolone, to biopsy at 1, 2, 3, 6 and 12 months with steroid treatment when SCR was detected, or to a control group biopsied at 6 and 12 months and treated only for clinical episodes of rejection. Treated SCR was associated with less clinical rejection, reduced chronic tubulointerstitial scores at 6 months and lower serum creatinine at 24 months, but treatment of SCR identified at 6 months did not prevent deterioration of renal function between 6 and 24 months (7).
In a second multicenter study, Rush et al. randomized 121 patients receiving tacrolimus, mycophenolate mofetil (MMF) and prednisolone to biopsy at 0, 1, 2, 3 and 6 months, versus 0 and 6 months with treatment of SCR in both arms. Overall prevalence of SCR had fallen to 4.6% in the biopsy arm, which was attributed to the more potent immunosuppression, thus there was no benefit to early protocol biopsies with tacrolimus and MMF therapy (13).
In the third study of SCR, Kurtkoti et al. randomized 102 recipients of living-donor transplants to protocol biopsy at 1 and 3 months or to indication biopsy alone. Immunosuppression comprised a CNI (cyclosporine or tacrolimus), antimetabolite (azathioprine or MMF) and prednisolone. Histologically proven acute rejection (AR), acute SCR and borderline changes (BL), were treated with steroids. Renal function was significantly better at 6 months and 1 year in the biopsy arm, and they concluded that protocol biopsies allowed detection and treatment of early SCR with improved short-term allograft function (2).
Observational studies have shown that persistent inflammation, even below the Banff threshold diagnosis for acute rejection, results in worse graft function at 1 and 2 years (14). Furthermore, surveillance biopsies carried out as early as 14 days after transplantation showing SCR predict a 30% decrease in graft survival rates at 10 years from transplant (3). Data from nonrandomized studies in both adults and pediatric patients treated with cyclosporine or tacrolimus, MMF and steroids suggest that treatment of SCR with steroids reduces chronic inflammation (4,15).
The appropriateness of treating BL with steroids has been questioned with recent evidence that regulatory T cells are present in renal allografts with BL, demonstrated by increased Foxp3 gene expression in peripheral blood (16). Further studies will be required to identify any predictive role for Tregs and the level of Foxp3 expression in protocol biopsies.
The relationship between different immunosuppression protocols and incidence of subclinical changes has been reported in a four-way cohort study, demonstrating reduced SCR with CNI combined with sirolimus compared to MMF (17). Two recent studies comparing cyclosporine and azathioprine with tacrolimus, MMF and IL2RA, described reduced SCR with tacrolimus, and no significant differences in graft survival, renal function and subsequent rejection if borderline changes (18) or SCR (19) were untreated. Only 2 of 18 patients were treated in the first study, follow-up was short at 1 year and Banff chronicity scores were not reported (18). The second study reported rates of IF/TA between 53% and 63% in both groups as well as an association of SCR with a CAN score ≥2 in the same 6-month biopsy (19). Of greater concern is a 27% prevalence of SCR in one and 4-month protocol biopsies found in patients undergoing rapid steroid withdrawal, comparable to rates seen in the cyclosporine/azathioprine era. SCR was followed by greater IF/TA scores at 1 year with an odds ratio of 6.62 (95% CI 2.68–16.3; Ref. 5). The prospect of uncontrolled subclinical alloimmune injury in patients undergoing increasingly popular regimes of steroid withdrawal is supported by studies of “troubled” transplants where rejection appears again as a major cause of failure. A major role for surveillance pathology may be to protect steroid-free patients from the effects of SCR.
The evidence favors treating SCR when diagnosed early after transplantation. While the incidence of SCR has fallen with current potent immunosuppression, it cannot be assumed the rates of SCR will be low with reduced levels of immunosuppression. Recent studies are hampered by small numbers and short follow-up periods, which are unable to measure the long-term impact of undetected or untreated SCR on chronic fibrosis and outcomes.
Subclinical rejection in sensitized recipients
Many transplant centers now use desensitization protocols to overcome immunological barriers such as anti-HLA donor specific antibodies (DSA) or blood group incompatibility (ABOi) that previously precluded transplantation.
Gloor et al. examined the incidence of SCR in 14 positive cross-match living donor transplants who were desensitized before transplantation. Surveillance biopsies on days 3, 7, 14, 28 and 90 revealed a mean prevalence of subclinical antibody-mediated rejection of 29%, requiring treatment with plasma exchange and corticosteroids resulting in resolution of histology on 3- and 12-month biopsies and stabilization of renal function in all patients (20).
In a retrospective case series of 116 surveillance biopsies from 50 stable positive crossmatch (+XM) desensitized patients, Kraus et al. treated subclinical cellular rejection in 40% of biopsies between 0 and 2 months, falling to 17% by 9 months. No differences were found in 1-year serum creatinine with or without SCR, postulated to be due to early intervention. Elimination of circulating DSA correlated with absence of on-going antibody mediated injury in the graft and a low risk of SCR but C4d positivity was not a helpful predictor of rejection or severity of rejection (21).
Setoguchi et al. found that the incidence of subclinical cellular rejection increased with time in ABO incompatible (ABOi) recipients. BL and SCR were treated with steroids with the majority of grafts in both groups showing normal histology or mild IF/TA and no difference in GFR or graft survival. However, interstitial inflammation with IF/TA was significantly higher in ABOi transplants at 6–12 months and was associated with a worse outcome (22).
The development of preformed (21) or de novo (23,24) anti-HLA DSA has been shown to correlate with acute humoral rejection and is a significant risk factor for transplant glomerulopathy (TG) and graft loss. Posttransplant monitoring of DSA is neither a sensitive nor specific test for predicting graft pathology and thus is not a substitute for histopathology.
Surveillance biopsy in high immunological risk patients allows early detection of acute or subclinical AMR: providing either preemptive treatment with potential better outcomes or reassurance of adequate immunosuppression with normal histology. Recent studies have revealed a very different histological and clinical course when compared with standard risk transplants and allowed correlation with persistent DSA, C4d deposition and prior humoral rejection with graft histology and function (21,22).
Implantation biopsy is standard of practice as a baseline for assessing subsequent changes. Glomerulosclerosis, vascular changes, tubular atrophy and interstitial fibrosis have all been associated independently with graft outcome. Incorporating histopathology into a clinical scoring system (25) or composite score (26) has improved predictive value of pretransplant biopsies.
Delayed graft function
Grafts with delayed graft function (DGF) are at higher risk of acute rejection, and associated with inferior graft outcome. Acute rejection cannot easily be diagnosed in the setting of DGF without surveillance biopsy and is standard of practice globally. In addition to acute tubular necrosis, subclinical AR has been reported in 18–21% and BL in 12–36% of such surveillance biopsies (27,28).
Interstitial fibrosis/tubular atrophy (IF/TA)
Surveillance biopsies have allowed greater understanding of the cumulative impact of donor disease, ischemia-reperfusion injury at the time of organ procurement, acute and SCR and CNI toxicity, in the early months after transplantation. Abnormalities are present in up to 40% of implantation biopsies, dependent on donor age and quality and on organ recovery (29). Surveillance biopsies have identified donor age, DGF, severe vascular AR and SCR as independent variables as risk factors for IF/TA (8,30). Severe IF/TA has been reported in 58% of patients by 10 years and once established, was irreversible and accompanied by deterioration in graft function (8).
Graft function is an unreliable marker of the underlying severity of graft pathology. Legendre et al. reported the prevalence of IF/TA as 25% at 3 months and 50% at 2 years in 41 patients with normal graft function (9) and Serón et al. described progression of chronic lesions (Banff ci, ct, cg and cv scores) between 4 and 14 months without changes in graft function (31). In contrast to SCR, IF/TA increases with time and when superimposed on chronic transplant arteriopathy, SCR or glomerulopathy, there is significantly worse survival, with inflammation combined with fibrosis providing the worst outcome (3,10,14).
Surveillance biopsies detect early changes while graft function is still normal, allowing early intervention designed to decelerate or abrogate development of chronic fibrosis and ultimately improve graft survival. Oberbauer et al. described less IF/TA at 3 years and improved graft survival at 4 years after withdrawal of CNI at 3 months in low immunological risk patients receiving cyclosporine A and sirolimus correlating progression of IF/TA with graft survival (32).
Surveillance biopsies have shown that CNIs reduce the incidence of early SCR and modulate IF/TA (33) but the penalty from chronic CNI-nephrotoxicity needs to be understood. Nankivell et al. described pathological changes of cyclosporine nephrotoxicity, associated with microvascular and glomerular changes. An early reversible pattern of nephrotoxicity associated with acute functional nephrotoxicity and high cyclosporine levels was followed by a later chronic, persistent pattern with a median onset at 3 years, which was usually irreversible. Both cyclosporine and tacrolimus produced similar fibrogenic effects in the kidney and a similar pattern of nephrotoxicity (34).
Flechner et al. found that sirolimus combined with MMF and steroids produced better renal function and less IF/TA compared with cyclosporine-based therapy at 2 years (35). Similar improvements in IF/TA with non-CNI based regimens have also been shown by Vincenti et al., who randomized 218 renal transplant recipients to belatacept versus cyclosporine—acute rejection rates were similar but a higher GFR and less tubulointerstitial damage occurred in the belatacept groups with maximum benefit in low risk patients receiving standard criteria kidneys compared with those receiving extended criteria donor kidneys (36,37).
Recent data question the specificity of lesions related to CNI therapy. Naesens et al. found neither tacrolimus dose nor measures of systemic exposure were associated with lesions of CNI nephrotoxicity (38). Stegall et al. have subsequently described a similar prevalence of arterial hyalinosis at 5 years in tacrolimus versus a mixed group of sirolimus-treated kidney transplant recipients, though the sirolimus group was small and preimplantation histology was not reported (39).
While CNIs maintain their overall superiority as first-line immunosuppressive agents, surveillance biopsy usefully identifies CNI nephrotoxicity.
Polyoma virus associated nephropathy (PVAN or BKVAN) has a reported incidence of 3–5%. Although prevalence is low, BKVAN results in up to 45% graft loss in affected individuals (40) and early diagnosis is necessary to resolve infection and prevent chronic damage.
The most specific surrogate markers of renal parenchymal involvement with BK are persistent viuria (≥107 copies/mL) and increasing viremia (≥104 copies/mL for >3 weeks) independent of renal function (41). Early modification of immunosuppression may favor stable graft function and viral clearance thus a surveillance biopsy can help to quantify the immune risk, with reduction in immunosuppression safe with a normal biopsy. Buehrig et al. found all patients with BKVAN diagnosed by surveillance biopsies (at 3–4 and 12 months) and managed by immunosuppression reduction had a satisfactory outcome by 6 months after diagnosis, compared with late diagnosis by indication biopsies where 70% had graft loss, functional or histological deterioration (42).
Chronic antibody mediated rejection and TG
TG is an important cause of late graft failure. Surveillance biopsies have shown that glomerulopathy can be detected as early as 1 month after transplantation (43) and severely compromise graft survival (10).
Gloor et al. reported a cumulative incidence of clinical and subclinical TG increasing from 4% at 1 year to 20% by 5 years, with prior acute rejection, hepatitis C and sensitization to anti-HLA antibodies (especially Class II) being risk factors (44). This group also reported a significantly higher incidence of TG in ABOi and +XM patients compared with conventional transplants. Humoral rejection was the greatest predictor of abnormal 1-year histology but not of graft survival (45). Longer follow-up studies are required to determine the ultimate impact of early humoral rejection. So while preliminary data confirm early detection of chronic antibody mediated rejection is feasible, the current problem is the lack of proven treatment options for established TG.
Recurrent and de novo glomerulonephritis
Clinical diagnosis prompted by hematuria and/or proteinuria leads to reported recurrence rates ranging from 10 to 20%. This is likely to be an underestimate as much disease is subclinical (46) and immunofluorescence and electron microscopy are often not routinely performed. Surveillance biopsies confirm a higher prevalence of certain GN than previously reported (47,48) with unsuspected recurrent GN in approximately 1–2% of surveillance biopsies. Recurrence rates between 42% and 55% have been reported in certain GN such as membranous nephropathy and lupus nephritis (47,48).