Steroid-free Immunosuppression in Kidney Transplantation: An Editorial Review

Corticosteroids have been a mainstay of immunosuppressive drug therapy in kidney transplantation for more than 40 years. Although our understanding of the immunosuppressive mechanisms of these agents has expanded greatly during that time, dosing of corticosteroids has remained surprisingly empiric. The well-recognized adverse effects of long-term steroid therapy have led to an understandable desire to develop steroid-sparing and steroid-free immunosuppressive regimens. By increasing the risk of infection and by increasing the frequency of hypertension, hyperlipidemia and glucose intolerance, these agents undoubtedly contribute to patient mortality resulting from infectious and cardiovascular diseases. In addition, steroid-related side-effects add to the long-term cost of medical care of renal transplant recipients. One recent economic analysis considered the costs of steroid-induced hypertension, diabetes mellitus, cataracts, peripheral fractures, and aseptic necrosis of the hip in kidney transplant recipients and estimated that the cumulative 10-year cost of these side-effects was $5300 per transplanted patient (1). Unfortunately, virtually all controlled trials of steroid avoidance or steroid withdrawal reported to date indicate that elimination of steroids increases the risk of acute and/or chronic rejection in a substantial minority of patients who cannot be identified easily by clinical or immunologic parameters. Thus, the wisdom of offering steroid-free immunosuppression to kidney transplant recipients rests on the balance between the cost-saving benefits of eliminating steroids and the risk of allograft rejection. Recent reviews (2, 3) and meta-analyses (4, 5) have summarized the benefits and risks of steroid withdrawal derived from multiple controlled and uncontrolled studies performed during the past two decades. The purpose of this review is, instead, to identify key questions that have evolved from these experiences in an effort to identify issues relevant to the design of future steroid-free immunosuppressive protocols. The rationale for eliminating corticosteroids from the immunosuppression regimens administered to transplant recipients is based largely on the desire to eliminate side-effects, but also on the presumption that other immunosuppressants are potent enough to prevent allograft rejection in the absence of steroids. The mechanisms whereby corticosteroids inhibit the immune system remain incompletely understood. It has long been known that corticosteroids alter the distribution of circulating lymphocytes, causing sequestration of CD4+ cells in the reticuloendothelial system, thereby reducing egress to inflammatory sites (6). More recently, it has become clear that corticosteroids inhibit the proliferation and function of lymphocytes by blocking the expression of various lymphokines and cytokines. Owing to their hydrophobic nature, glucocorticoids easily diffuse into cells and bind to cytoplasmic receptors. The steroid-receptor complex translocates into the nucleus, where it binds to a number of DNA sites collectively referred to as glucocorticoid response elements (GRE). Binding of the steroid-receptor complex to GRE may inhibit the transcription of cytokine genes by acting on distant genes to induce expression of transcription inhibitory factors (7, 8). Recent evidence suggests that corticosteroids inhibit the action of transcription factors such as activating protein-1 (AP-1) and nuclear factor-κB (NF-κB) (9, 10). Activated glucocorticoid receptors may bind to activated NF-κB and prevent it from binding to κB sites on proinflammatory genes (11, 12). Corticosteroids also increase transcription of the gene for IκBα which binds to activated NF-κB. The IκBα protein probably causes dissociation of NFκB from target genes, resulting in its movement back to the cytoplasm (13). The major consequence of these intracellular effects of corticosteroids is an inhibition of the production of interleukin-1 (IL-1) and IL-6 by antigen-presenting cells such as macrophages, monocytes and dendritic cells. Because IL-1 is a primary costimulus for helper T cell activation, while IL-6 is a major inducer of B cell activation, corticosteroid administration has the potential to inhibit both the cellular and humoral arms of the immune response. Steroids also inhibit expression of other proinflammatory cytokines, including interferon-γ, and tumor necrosis factor. Finally, recent studies indicate that corticosteroids inhibit the expression of IL-2 by impeding the effects of DNA binding proteins that normally stimulate IL-2 gene transcription (6, 14). Thus, in contrast to most other available immunosuppressants that exert fairly specific effects on the activation or proliferation of lymphocytes, the effects of corticosteroids are multiple and influence activation, proliferation, and distribution of lymphocytes. While some of these effects are redundant with those of calcineurin inhibitors, antimetabolites and mTOR inhibitors, others (e.g. inhibition of NFκB) are unique. It is possible that these unique effects are required to prevent allograft rejection in some patients and, as a corollary, that withdrawal of steroids in such patients poses a risk of allograft rejection. In 1998, the American Society of Transplant Physicians (ASTP) and the American Society of Transplant Surgeons (ASTS) cosponsored an interdisciplinary conference to review the evidence regarding the benefits and risks of steroid-free immunosuppression available at that time. In preparation for the conference, members of the two societies were surveyed to determine existing attitudes about steroid-free immunosuppression. As shown in Figure 1, 77% of the survey respondents indicated that steroid-free immunosuppression remained a goal of future immunosuppression regimens for organ transplant recipients. When asked what incremental risk of acute rejection would be outweighed by the benefits of eliminating corticosteroids, over 70% of the respondents indicated that, compared to patients maintained on steroids, an incremental rejection risk of 10% or less would be acceptable (Figure 2). Percentage of ASTP/ASTS respondents answering the survey question: ‘Is steroid-free immunosuppression a goal in kidney transplantation?’ Percentage of ASTP/ASTS respondents answering the survey question: ‘What incremental risk of acute rejection is acceptable when comparing a steroid-free immunosuppression regimen to one in which patients are maintained on corticosteroids?’ Early experiences with steroid avoidance in patients receiving cyclosporine monotherapy indicated that approximately 50% of patients sustained episodes of acute rejection 15-17). Acute rejection associated with steroid withdrawal occurred in approximately 30% of patients subsequently maintained on azathioprine and/or cyclosporine (4, 5), with even higher rates reported from that era when corticosteroids were withdrawn early after transplantation (18). More importantly, results from several single studies 19-21) and a meta-analysis (5) of studies performed mostly in the cyclosporine-azathioprine era indicated that steroid withdrawal was associated with a significant risk of long-term graft failure that may be appreciated only after 3–5 years of follow-up, even in the absence of clinically evident episodes of acute rejection. Clearly, these rates of acute and chronic rejection were perceived to be unacceptable by most transplant physicians and surgeons, such that interest in steroid-free protocols declined greatly by the by the mid-1990s. Two large randomized trials have examined the safety of withdrawing steroids approximately 3 months following transplantation in renal allograft recipients maintained on cyclosporine and mycophenolate mofetil (22, 23). In both studies, the incidence of acute rejection was higher in the group of patients withdrawn from steroids. In the US multicenter randomized prospective trial, Kaplan-Meier estimates of the cumulative incidence of acute rejection within the first post-transplant year were 30.8% in the steroid withdrawal group and 9.8% in the steroid maintenance group (22). A key finding in the US trial was the observation that the risk of acute rejection following steroid withdrawal was 4-fold higher in African Americans than it was in Caucasian patients. In Caucasian patients, the incremental risk of acute rejection at 12 months post-transplant was less than 10% higher than the control group, but remained statistically significant. A conceptually similar, prospective randomized European multicenter study also examined the safety and benefits of steroid withdrawal 3 months post-transplant in patients receiving cyclosporine and mycophenolate mofetil (23). Patients were randomized at the time of transplantation into a control group that received standard doses of steroids or to a group that received lower doses of steroids for 3 months before they were withdrawn. At 12 months, the incidence of acute rejection in the ‘low/stop’ group was 25% vs. 15% in the control group (p = 0.008). Interestingly, there was no difference in the rate of acute rejection when the comparison was limited to patients who had received induction antibody therapy. Blood pressure and lipid levels were significantly better in the ‘low/stop’ group at last follow-up. In a recent uncontrolled, multicenter pilot study performed in Canada, complete avoidance of corticosteroids was attempted in 57 patients treated with daclizumab induction therapy, mycophenolate mofetil, and cyclosporine (24). Fourteen patients (25%) experienced acute rejection episodes, of which 13 were readily reversed with steroids. At last follow-up, serum creatinine concentration among patients experiencing acute rejection was not significantly different from patients without acute rejection. In a similar uncontrolled Danish study, 68 patients were treated with antithymocyte globulin induction, cyclosporine, and mycophenolate mofetil without the use of corticosteroids. The incidence of acute rejection was 15% after a median follow-up of 488 days (25). Preliminary results from another smaller, multicenter, randomized trial indicate that the use of induction therapy with basiliximab followed by maintenance treatment with cyclosporine and mycophenolate mofetil could facilitate early withdrawal of prednisone without a statistically significant increase in the risk of acute rejection within 6 months compared to a control group maintained on prednisone (26). Experience from uncontrolled studies suggests that steroid withdrawal or steroid avoidance in patients receiving mycophenolate mofetil and tacrolimus is associated with relatively low rates of acute rejection compared to historical controls (27). Preliminary results from two centers indicate that the use of tacrolimus and mycophenolate mofetil may facilitate early withdrawal of steroids (28, 29). In one uncontrolled study of 52 patients maintained on this combination of immunosuppressants and withdrawn from steroids 7 days post-transplant, the incidence of acute rejection within 12 months of transplantation was 25% (28). Even better results have been reported in preliminary results from a controlled trial of patients receiving an anti-interleukin-2-receptor antibody, tacrolimus, and mycophenolate mofetil (29). In 54 patients constituting the steroid withdrawal arm of the study, corticosteroids were discontinued after post-transplant day 3. The incidence of acute rejection at 6 months post-transplant was 13.2% in the prednisone-free group and 8.3% in 40 control patients maintained on prednisone. A majority of the patients enrolled in the latter study were recipients of kidneys from living donors. In a recently reported uncontrolled study, the use of daclizumab, tacrolimus and mycophenolate mofetil allowed complete avoidance of corticosteroids, with no episodes of acute rejection after 9 months of follow-up in a series of 10 pediatric transplant recipients, another group of patients generally regarded to be at high risk for rejection in steroid-free protocols (30). Experience with steroid-free immunosuppression in patients receiving sirolimus-based immunosuppression thus far has been limited to uncontrolled trials. However, experiences at two centers indicate low rates of rejection when steroids are eliminated in patients receiving sirolimus in conjunction with either cyclosporine microemulsion or tacrolimus, respectively. At one center, 156 patients treated with sirolimus and cyclosporine microemulsion have been withdrawn from steroids at varying times after transplantation ranging from 1 week to more than two years (31). To date, the overall incidence of acute rejection following steroid withdrawal has been less than 7%. Two years following attempted steroid withdrawal, approximately three-quarters of patients remain steroid-free. Our own experience has been directed towards African American patients who are treated with sirolimus, relatively low doses of tacrolimus, and prednisone for 3 months, usually without the use of induction antibody therapy. Selected patients are withdrawn from steroids between month 3 and 5. To date, 16 patients have been weaned off prednisone successfully, with only a single episode of acute rejection (unpublished data). However, follow-up after steroid withdrawal has been relatively short (range 1–16 months), and longer follow-up is needed to determine the efficacy and safety of this regimen in African Americans and other high-risk kidney transplant patients. To summarize, experience to date suggests that the risk of acute rejection associated with steroid withdrawal in patients receiving immunosuppression regimens containing mycophenolate mofetil, tacrolimus, and/or sirolimus is lower than that observed in the cyclosporine-azathioprine era. The risk of rejection may be within an acceptable range in low risk (i.e. non-African American) patients, or in those receiving antibody induction therapy or sirolimus-based immunosuppression. However, many of the outcomes cited here come from relatively small uncontrolled studies, some of which have not yet been published in peer-reviewed journals. Moreover, the duration of follow-up in each of these studies has been insufficient to determine whether steroid withdrawal in patients receiving some of the newer immunosuppressants is associated with an increased risk of chronic allograft nephropathy. Although the community of transplant physicians and surgeons may be willing to accept an incremental risk of rejection to justify the benefits of steroid withdrawal in a majority of patients, it remains problematic that most randomized trials are designed with acute rejection as the primary end-point and with equivalence as the null hypothesis. This philosophy led to the premature termination of the US trial of steroid withdrawal in patients receiving cyclosporine and mycophenolate mofetil (22), and eliminated the possibility of long-term follow-up of patients in either arm of that study. If some patients or groups of patients (e.g. African Americans, children) are more dependent on the unique effects of corticosteroids, it seems likely that the goal of achieving equivalent rates of acute rejection may doom to failure any large randomized trial of steroid-free immunosuppression. Other end-points, such as quality of life or economic cost, may be more suitable for future steroid-free protocols. Use of acute rejection as a primary end-point in any trial is based on the premise that acute rejection is a surrogate marker for graft failure. However, it is not clear whether this principle applies to acute rejection episodes that may occur when steroids are withdrawn from patients receiving some of the newer immunosuppressants. Furthermore, it is now clear that all acute rejection episodes are not alike. Acute rejection episodes that are reversed completely (based on both clinical and pathologic criteria) may not have the same negative impact on long-term graft survival as those in which patients are left with persistent renal dysfunction following treatment. In future trials of steroid-free immunosuppression, more attention should be paid to classifying acute rejection episodes based on indications of reversibility. One of the few clinical variables predictive of acute rejection in patients withdrawn from steroids in the cyclosporine-azathioprine era was the timing of withdrawal: rates of rejection were substantially higher when steroids were withdrawn earlier than later (32). Unfortunately, some of the most important long-term side-effects of corticosteroids (e.g. cataracts and osteopenia) may be initiated by pathophysiologic mechanisms that result from high doses of corticosteroids administered early after transplantation and which may not be reversed if steroids are discontinued many months later. In addition, recent observations from rodent models suggest that corticosteroids may inhibit tolerance by blocking T-cell signals that ultimately lead to apoptosis (33). Because apoptotic signals are likely to be generated early after antigen presentation to T cells, it seems that complete avoidance of steroids may be necessary in order to achieve the tolerogenic benefit of eliminating these agents. Thus, one could argue from many points of view that the optimal benefits of steroid-free immunosuppression can be achieved only by complete avoidance or very early withdrawal of corticosteroids. However, until further results are available from some of the preliminary studies delineated above, it seems prudent to stratify the next generation of steroid-free immunosuppression trials according to the risk of immunologic rejection or risk of steroid-induced side-effects(see Table 1), because early withdrawal of steroids may pose a prohibitive risk of rejection in high-risk populations. Trials of steroid avoidance or early steroid withdrawal (less than 3 months post-transplant) should be limited to patients at low risk for rejection and/or high risk for steroid side-effects. Late steroid withdrawal (more than 3 months post-transplant) should be designed for patients at high risk for rejection and/or low risk for steroid side-effects. Immunosuppression without the use of corticosteroids remains a desirable, but elusive goal in kidney transplantation. Results from recent small, and largely uncontrolled trials suggest that the risk of acute rejection associated with steroid-free immunosuppression in patients receiving newer immunosuppressants, such as mycophenolate mofetil, tacrolimus, and sirolimus, may be lower than that observed in historical controls receiving azathioprine and/or cyclosporine. However, larger studies with longer durations of follow-up are needed to determine whether residual immunosuppression with some of the newer agents can prevent the development of chronic rejection after elimination of steroids. In the mean time, it seems prudent to suggest that steroid-free immunosuppression should not be offered to patients on a routine basis in clinical practice. Enrollment of patients into randomized trials is encouraged to further delineate the benefits and risks of steroid-free immunosuppression.