Volume 11, Issue 4 pp. 693-697
Personal Viewpoints
Free Access

Chronic Calcineurin Inhibitor Nephrotoxicity—Lest We Forget

J. R. Chapman

Corresponding Author

J. R. Chapman

Centre for Transplant and Renal Research, University of Sydney, Westmead Hospital, WESTMEAD, NSW, Australia

Corresponding author: Jeremy R. Chapman, [email protected]Search for more papers by this author
First published: 29 March 2011
Citations: 128

Abstract

Calcineurin inhibitor (CNI) nephrotoxicity was recognized in Cambridge in the late 1970s. The vasoconstrictor impact of cyclosporine (CsA) and to a lesser extent tacrolimus, in both acute and chronic settings, results from a decrease in vasodilators and increase in vasconconstrictors while direct tubular toxicity results from blockade of mitochondrial permeability transition pores and inhibition of prolyl isomerase. A biopsy of native kidneys of recipients of CNIs reveals nephrotoxicity as the most common pathological diagnosis with chronic CNI toxicity and hypertension the primary problems. A long-term study of randomized clinical trials with up to 20 years of follow-up shows inferiority of both renal function and graft survival for continuous CsA compared to either CsA withdrawal or continuous azathioprine and prednisolone. Pathological hallmarks of chronic CNI nephrotoxicity include stripped interstitial fibrosis, arteriolar hyalinosis and glomerular sclerosis, but with the exception of nodular arteriolar hyalinosis the findings are non specific. The model for chronic renal allograft loss must be multifactorial with both immune and nonimmune factors operating dependent upon an individual's risk factors for cell and/or antibody-mediated rejection, CNI nephrotoxicity and recurrent disease. Better outcomes will require early diagnosis and individualization of therapy dependent upon the dominant mechanisms impacting each patient. The revisionist view put forward by some senior, experienced and thoughtful individuals, challenges the concept of chronic CNI nephrotoxicity as an important clinical entity. By implication, the view that appears to be promoted is as follows: we need not fear-prolonged exposure to CNIs, and in seeking better long-term solutions for transplant recipients, we have forgotten alloimmunity. It is thus apparent that we must revisit the data and again question the basis for chronic CNI nephrotoxicity in current clinical practice. This contribution to the debate will focus on the evidence that CNIs are nephrotoxic and that their impact needs to be limited if we are to improve long-term outcomes after transplantation, leaving others to promote the contrary perspective and perhaps also to reflect on the largely unproven impact of the steroid avoidance and other minimization strategies so prevalent today.

Abbreviations:

  • CKD
  • chronic kidney disease
  • CNI
  • calcineurin inhibitor
  • CsA
  • cyclosporine
  • Tac
  • tacrolimus
  • GFR
  • glomerular filtration rate
  • IF/TA
  • interstitial fibrosis/tubular atrophy
  • mTORi
  • mammalian target of rapamycin inhibitor
  • Acute CNI Nephroxicity

    The first recognition that cyclosporine (CsA) might be nephrotoxic came as clinical trials commenced in Cambridge in the late 1970s. At heavy doses of 25 mg/kg/day, selected after the data from large animal models became available, Calne and his colleagues noted nephrotoxicity and reduced the dose to 10 mg/kg/day in subsequent patients (1). A meeting in Cambridge in September 1981 revealed some of the work that this clinical observation triggered. Nephrotoxicity was observed experimentally in the OFS rat at 45 mg/kg/day and at 30 mg/kg/day in the Wistar Furth rat, especially when salt depleted, with a 3–14 fold proximal tubular concentration of CsA and histological damage but no glomerular changes (2).

    CsA alone in clinical studies at a dose of 10 mg/kg/day provided insufficient immunosuppression; thus a mid-point dose of 17.5 mg/kg/day was used in subsequent trials, revealing both immunosuppressive efficacy and acute nephrotoxicity when CsA was withdrawn abruptly at 3 months and replaced by the then conventional therapy of azathioprine and steroids (3). An improvement in renal function was seen in this study, with the serum creatinine dropping by around 50–60 umol/L within a week of conversion in most patients. The same phenomenon was seen after CsA dose reduction in individual patients with acutely elevated creatinine levels and renal biopsies that did not show rejection. The physiological explanation for acute nephrotoxicity was illustrated during the 1980s and thought to result from dose-dependent (4) CsA-induced vasoconstriction (5,6). In a seminal study, high oral doses of 50 mg/kg were administered to rats, demonstrating progressive decline in GFR associated with similarly progressive reduction in afferent arteriolar luminal diameter. The mechanism for the observed acute nephrotoxicity was thus proposed to result from arteriolar vasoconstriction leading to glomerular ischemia (7). The electron photomicrograph of the ischemic glomerulus became a useful emblem for illustration of CsA nephrotoxicity. That this was not the whole answer using therapeutic blood levels was illustrated by the clinical recognition of the tubular damage seen on histology (8) and the elevation of uric acid blood levels (3,9). The question raised in the 1980s was whether or not the reversibility of renal impairment seen at 3 months would become irreversible with time.

    The mechanisms underlying CNI nephrotoxicity have been subject to considerable study in the intervening years. Elegantly reviewed recently, the vasoconstrictor role of CsA and tacrolimus (Tac) predominates in both the acute and chronic settings leading to many of the histological changes associated with CNI therapy (10). By the late 1980s it was clear that there was a dose-dependent reduction in both GFR and effective renal plasma flow as well as reduced microsomal protein synthesis in the rat (11). In the clinic, similar studies revealed dose-dependent decreases in GFR from 63% at 5 mg/kg/day, to an 18% reduction at 1.5 mg/kg/day. This study also revealed the cyclic nature of changes in the kidney, being worst soon after administration and least prior to a dose following 2–4 h after peak blood concentrations (12).

    The mechanisms of CNI nephrotoxicity as currently understood (10) involve both a decrease in vasodilators such as prostaglandin E2 and nitric oxide as well as an increase in the vasconconstrictors thromboxane, endothelin and the renin–angiotensin system. Direct toxicity to the tubular epithelium has been demonstrated both clinically and experimentally with isometric vacuolization resulting from the presence of giant mitochondria, probably as a result of CsA blockade of mitochondrial permeability transition pores. Enlargement of the endoplasmic reticulum also occurs and is thought to occur through inhibition of prolyl isomerase leading to accumulation of unfolded protein and impairment of protein synthesis in the endoplasmic reticulum.

    Chronic CNI Nephrotoxicity

    Chronic nephrotoxicity was first identified in cardiac transplant recipients, but the permanent histological hallmarks of striped interstitial fibrosis, tubular atrophy, medial arteriolar hyalinosis and tubular micro-calcification were also seen in renal transplants and in patients treated for autoimmune disease (6,8,13,14). Analysis of transplant recipients of organs other than the kidney reported a 16.5% risk of chronic kidney disease (CKD), with 28.9% of those requiring dialysis or renal transplantation (15). CsA increased the risk of CKD more than Tac, as did hypertension, hepatitis C and diabetes, reinforcing the multifactorial pathogenesis of renal damage in these patients. The most recent analysis of nonrenal transplant patients receiving CNI and having a native kidney biopsy for acute (in 9% of biopsies) or chronic renal impairment, or proteinuria, revealed CNI nephrotoxicity as the most common pathological diagnosis, based on striped interstitial fibrosis and tubular atrophy with arteriolar hyalinosis (16). Thrombotic microangiopathy and glomerulonephritis contributed 10% and 17% of the samples respectively, arguing the need for a biopsy rather than assumption of CNI nephrotoxicity in CNI-treated patients with impaired native renal function. These data support the case for CNI toxicity as the major avoidable cause of renal impairment after nonrenal transplantation. In the patients given the highest CNI doses-–heart and lung recipients—the percentage with arteriolar hyalinosis was 64% and 70% respectively. To quote from the conclusions of this analysis ‘chronic CNI toxicity and hypertension remain the primary problems’ (16).

    One must conclude that the native kidney is frequently impaired by CNI therapy and in a proportion of patients, chronic exposure leads to CKD with requirement for dialysis or renal transplantation.

    Chronic CNI nephrotoxicity was also evident in the renal transplant. Data from the studies in the 1990s progressively concluded that the histology, with the exception of nodular hyaline insudation of the small arteriolar wall resulting from myocyte necrosis, was not specific to CNIs. Despite this lack of a sensitive and specific markers of CNI nephrotoxicity, the analysis of histology from this period demonstrated the linkage between CNI associated arteriolar lesions, glomerular sclerosis and graft loss (17,18).

    That the pathological hallmarks of CNI nephrotoxicity also occur in renal transplants today and not just from the 1980s is beyond question. The DeKAF study of deteriorating kidneys biopsied for cause, has most recently reconfirmed the role of inflammation in biopsies from patients with chronic kidney dysfunction, no matter where it is found in the kidney, as a major predictor of graft loss in the first 3 years after transplantation (19). However a different publication from this study has identified CNI nephrotoxicity as the local pathological diagnosis in 30% of failing grafts and substantial arteriolar hyalinosis as a feature of two of the major groupings (20). The Mayo clinic series of 1- and 5-year protocol biopsies has also recently questioned the frequency, progressive nature and specificity of chronic CNI toxicity (21). They demonstrate that with the use of Tac, the incidence of moderate and severe fibrosis progressed between 1 and 5 years in 23% of allografts, or about half of what was seen in our comparable series (22). Arteriolar hyalinosis was severe by year 5 in 19% of Tac-treated patients and 1 of 20 treated exclusively with sirolimus. We have also noted, in a small cohort of patients, that the predictable postdose reduction in renal blood flow measured by duplex ultrasound that occurs with CsA did not occur with Tac (23). The evidence is compelling that Tac, in modern doses and in the protocols used in this study, is associated with a reduced but not absent chronic nephrotoxic impact at 5 years.

    The studies from our group have however shown the progressive nature of pathological damage over a 10-year time frame. The 10-year graft survival of these kidneys was around 80% and the mean measured GFR around 50 mL/min/1.73 m2. Despite these good outcomes the histology was alarming, with progressive fibrosis and tubular atrophy, arteriolar hyalinosis and glomerulosclerosis (22,24). It is the timelines of the development of pathology that must be carefully put into context when considering other shorter term studies. IF/TA develops early and is associated with inflammatory pathology in the first few months, such that two-thirds of IF/TA is present by the 2-year protocol biopsy and the continuing presence of untreated inflammation leads to graft losses in the first 3 years as in the DeKAF study (20). Development of nodular arteriolar hyaline change was associated with a threshold CsA dose of 5 mg/kg/day in our series (25). There was no correlation with impaired glucose tolerance in simultaneously performed oral glucose tolerance tests, nor with hypertension, both of which could have explained development of arteriolar hyalinosis-–though perhaps not of the nodular form identified as specific to CNI use (18). The glomerulosclerosis associated with graft loss is however a late phenomenon with little evident at 2 years and really only developing substantially by 5 years. It is evident that graft failure is highly associated with the development of transplant glomerulosclerosis (25). It is thus important to look for the long-term relationships between CNI nephrotoxicity and glomerulosclerosis and the subsequent impact first on graft function and then on graft survival.

    In our series the development of arteriolar hyalinosis and the degree of IF/TA were the best correlates with the subsequent development of glomerulosclerosis (24) It is of course also true that transplant glomerulopathy, correlated with the presence of donor-specific antibodies and chronic antibody-mediated rejection, also leads to glomerular disease and thus also graft failure (26), just as recurrent glomerular disease does (27). CNI nephrotoxicity is not an exclusive cause for graft loss in the long term; in fact it is the multifactorial nature of graft damage that probably leads to most actual chronic graft failures. The recent analysis of factors associated with graft loss in patients with transplant glomerulopathy demonstrates that two features of CNI nephrotoxicity are crucial in predicting graft failure-arteriolar hyalinosis and interstitial fibrosis, reinforcing the concept of multifactorial causes of graft loss (28).

    The long-term protocol biopsy and cohort studies that we and others (29) have conducted have thus lead to an understanding of the evolution of the pathological damage that afflicts transplanted kidneys, but randomized controlled trials yield the outcome data that seal the case against CNIs with respect to chronic nephrotoxicity.

    Clinical Trial Data

    In the early 1980s a number of trials demonstrated superiority of CsA compared to azathioprine and prednisolone. Concerns with a high incidence of lymphoma in the first 33 patients treated in Cambridge, acute nephrotoxicity and the unknown effects of long-term use of the new agent led to a protocol of 3 months of CsA followed by return to the then standard therapy of azathioprine and prednisolone. These studies have been subject to metaanalysis on a number of occasions, with interest revolving around the rate of acute rejection when CsA was switched for azathioprine, or prednisolone was discontinued and whether or not the protocol for conversion and its timing post-transplantation influenced the rate (30). The interest in these studies in the context of chronic nephrotoxicity is of course in the long-term outcomes of patients randomized to long-term CsA compared to short-term CsA. The metaanalysis performed on publications up to 1996 with average of 45 months of follow-up showed more acute rejection but nonsignificant impact on long-term outcomes from CsA withdrawal. It is tempting to hypothesize that the impact of increased rejection was balanced by the impact of increased nephrotoxicity over the 4 years of follow-up.

    There has now been a number of longer term follow-up analyses of the CsA conversion studies with the longest report being the Australian trial using the ANZDATA registry to accumulate the data (31). With a median of 20 years of follow-up and only 10 of 489 patients lost to follow-up, both the intention-to-treat and on-therapy analyses showed the inferiority in the long term of continuous CsA. The mean graft survival was 14.8 years for the short-term CsA group versus 12.4 years and 12.5 years for continuous azathioprine and prednisolone and continuous CsA respectively (p < 0.01). Renal function was also substantially worse (15.3 mL/min worse at 15 years and 14.2 mL/min worse at 20 years) in the surviving continuous CsA group compared to the short-term CsA group. It was better to use the superior immunosuppressive potency of the CNIs for the first 3 months combined with the reduced toxicity of azathioprine for the long term.

    There has been one small randomized controlled trial comparing 55 CsA and prednisolone with 53 azathioprine- and prednisolone-treated patients reported with 10-year follow-up (32). That CsA was superior is no surprise, but it is interesting to note that they switched 25% of the CsA-treated patients to azathioprine for declining renal function, successfully in seven but with graft failure or death in the other seven. Renal function was always better in the azathioprine-treated patients, and though they had more rejection, the rate of decline was the same in both groups. There is no histology available to identify if this study exchanged fibrosis resulting from rejection for fibrosis resulting from CNI nephrotoxicity, but that is a tempting speculation.

    The data from mycophenolate mofetil as a long-term replacement for CNIs have been variable and disappointing, so it was with some enthusiasm that the transplant field considered the potential for inhibitors of the target of Rapamycin (mTORi)-–sirolimus and everolimus-–to act as replacements for CNI therapy. With substantial clinical trial data available it is clear that the combination therapy involving either the mTORi agent or CNI agent is more nephrotoxic than the CNI alone, but that a sequential approach to utilize the CNI initially and the mTORi subsequently may yet produce an acceptable approach (33–34).

    An Inclusive Model for Graft Loss

    The data that our group and others have accumulated from protocol biopsies have been useful to understand the timeframe of evolution of changes and the longitudinal correlations between events and histological changes. These studies are not controlled and do not compare different immunosuppressant protocols, but they do provide clear evidence of the long timelines for graft damage. Damage to the kidney from whatever cause leads to histological change which precedes measurable functional changes partly because glomerular hypertrophy compensates for lost nephrons. Thus the glomerular filtration rate (GFR) declines relatively late in relation to histological change. The GFR must itself fall substantially before the serum creatinine rises to a point where the clinician is concerned and the statistician satisfied. It is thus no surprise that ‘for cause’ renal transplant biopsies often show late and mainly nonspecific changes. Our model for chronic graft loss is inclusive and multifactorial (35). It would be simplistic to believe otherwise. The factors which may damage a kidney in the long term are immune and nonimmune. Depending upon the individual patient's risk factors, cell-mediated rejection, antibody-mediated rejection, CNI and other drug nephrotoxicity, recurrent and de novo disease such as diabetes, mix with donor age, hypertension and renovascular changes and the long-term effects of ischemia-reperfusion injury, to cause long-term deterioration. In an individual patient it may be possible to dissect one or more dominant factors, such as recurrent disease or chronic subclinical transplant rejection, leading to graft loss, but for the majority it is probable that more than one factor is relevant. The difficulty that most studies and the majority of clinical practice struggles with, is that diagnosis of the cause of graft failure and institution of therapy is usually impossible once the end-stage pathology has been reached.

    It is not surprising that clinical acute cellular rejection, which has been the target of the last 30 years of research effort, the frequency of which is down to single figure percentages, has lost its impact as a cause of graft loss. Today we face more subtle chronic causes of graft damage such as subclinical rejection and CNI nephrotoxicity which compete with chronic antibody-mediated rejection and recurrent disease to destroy the structure and function of the transplanted kidney. None of these etiologies for graft damage can be ignored.

    Conclusions

    Does continuous CNI lead to pathological changes in the kidney transplant? Yes.

    Is continuous CNI a contributory factor to graft damage and loss in many patients? Certainly in grafts lost beyond 5 years.

    Are there patients treated with 25 years of CsA without impact? Yes, but there are not many.

    Do we have an answer to the CNI conundrum? Not yet. We cannot live without CNIs because they remain the most effective class of immunosuppressant available today, but we cannot live with them because their continuous use damages kidneys and leads to their failure.

    Does dose reduction of CsA or use of Tac reduce the impact of chronic nephrotoxicity? Yes.

      The full text of this article hosted at iucr.org is unavailable due to technical difficulties.