Outcome at 3 Years with a Prednisone-Free Maintenance Regimen: A Single-Center Experience with 349 Kidney Transplant Recipients

Authors


*Corresponding author: Arthur J. Matas, matas001@tc.umn.edu

Abstract

Historically, late steroid withdrawal after kidney transplants has been associated with an increased rejection rate. Recently, low rejection rates have been reported for recipients treated with complete avoidance or rapid elimination of steroids. However, follow-up has been short. We herein report on 3-year outcome in recipients whose prednisone was rapidly eliminated and who were maintained on a steroid-free regimen. From 10/1/1999 through 5/1/2003, 349 recipients (254 LD, 95 CAD; 319 in first 30 s) were immunosuppressed with polyclonal antibody (Thymoglobulin), a calcineurin inhibitor, either mycophenolate mofetil or sirolimus, and rapid discontinuation of prednisone.

Actuarial 3-year patient survival was 95%; graft survival, 93%. Acute rejection-free graft survival at 1 year was 94%; at 3 years, 92%. There was no difference between LD and CAD. At 2 years, the mean (± SE) serum creatinine level for LDs was 1.6 ± 0.5 mg/dL; for CAD, 1.6 ± 0.4 mg/dL. We have no new cases of PTLD or avascular necrosis; 22 recipients (6%) developed CMV. Currently, 84% of recipients remain prednisone-free. We conclude that excellent 3-year patient and graft survival can be achieved without maintenance prednisone. With such a protocol, steroid-related side-effects are minimal.

Introduction

Corticosteroids have long been a mainstay of immunosuppression after kidney transplants. Although inexpensive, steroids are associated with debilitating side effects, including hypertension, hyperlipidemia, cataracts, osteoporosis, mood and appearance changes, and, in children, growth retardation (1). Treatment of those steroid-related side effects adds to the cost of transplants (2). In addition, such side effects have been related to increased post-transplant noncompliance (3), and noncompliance has been associated with an increased incidence of acute rejection (AR), chronic rejection, and graft loss (4). Thus, a hidden cost of steroid-related side effects may be increased graft loss.

Historically, numerous attempts have been made either to avoid steroids or, in selected recipients, to gradually withdraw steroids late posttransplant. The aim of such attempts was to minimize steroid-related complications without increasing the incidence of AR episodes. In the past, cyclosporine monotherapy was associated with a high incidence of AR (5). Similarly, late steroid withdrawal in recipients on cyclosporine and prednisone [with or without azathioprine (AZA)] was associated with an increased incidence of AR and graft loss (6–8). More recently, late steroid withdrawal in recipients on cyclosporine, mycophenolate mofetil (MMF), and prednisone has also been associated with an increased incidence of AR (9,10). And, AR episodes often lead to long-term graft dysfunction and graft loss (7,8,11).

Rather than steroid avoidance or withdrawal, some centers have considered the use of low-dose steroids. Yet, in the nontransplant literature, low-dose steroids (less than 10 mg per day of prednisone) or alternate-day steroids have been associated with significant steroid-related side effects (12–16).

The recent development of new potent immunosuppressive agents has reawakened interest in trials of steroid-free immunosuppression. Several investigators have reported excellent short-term results either with complete steroid avoidance or with rapid steroid elimination (17–26). However, only one study has reported long-term outcome (17). We previously described our pilot study of rapid steroid minimization after first living donor (LD) kidney transplants (18). Herein, we report our 3-year follow-up results after first and second LD and cadaver (CAD) transplants.

Patients and Methods

From October 1, 1999, through May 1, 2003, at our institution, 349 kidney transplant recipients were immunosuppressed with a protocol incorporating rapid discontinuation of prednisone. Through September 30, 2000, only first LD transplant recipients were enrolled in the protocol (18). In October 2000, after demonstration of efficacy in our pilot trial (18), rapid discontinuation of prednisone became our protocol for all first and second LD and CAD transplant recipients. Our philosophical approach was to determine whether rapid discontinuation of steroids would be applicable to all potential recipients (rather than limited to a low-risk group). Initially, we excluded African Americans because of their reported increased incidence of AR rates in steroid withdrawal protocols (9) and because of the small numbers of African Americans who undergo transplants at our institution; we also excluded recipients of any race who had been on steroids immediately pretransplant (because, for them, our protocol would represent steroid withdrawal). Later, in light of the experience of other transplant centers, we no longer excluded African Americans.

Immunosuppression

The immunosuppressive protocol is outlined in Table 1. All recipients received Thymoglobulin (SangStat Corporation, Fremont, CA) at a dose of 1.25–1.5 mg/kg IV for five doses, with the first dose given intraoperatively. If delayed graft function occurred, additional doses (up to a maximum of 10 doses) were given. Methylprednisolone (500 mg) was given intraoperatively. Prednisone was given at 1 mg/kg on posttransplant day 1; at 0.5 mg/kg, days 2 and 3; and at 0.25 mg/kg, days 4 and 5. After day 5, prednisone was discontinued, except in recipients with delayed graft function, who were given 5 mg/day of prednisone until Thymoglobulin treatment was discontinued; at that point, prednisone was also stopped.

Table 1.  Rapid discontinuation of prednisone protocol
  1. *Additional doses for recipients with delayed graft function.

  2. **Additional doses (5 mg/day) if additional Thymoglobulin administered.

Thymoglobulin1.25 to 1.5 mg/kg, for five doses*; first dose, intraoperatively
Steroidsmethylprednisolone 500 mg intraoperatively
 prednisone 1 mg/kg on day 1 posttransplant;
  0.5 mg/kg on days 2 and 3 posttransplant; and
  0.25 mg/kg on days 4 and 5 posttransplant**
Cyclosporine or tacrolimus
Mycophenolate mofetil or sirolimus

Initially, maintenance therapy for all recipients consisted of cyclosporine (4 mg/kg bid adjusted to achieve levels of 150–200 ng/mL, by HPLC, for the first 3 months; levels of 125–150 for the second 3 months; and levels >100 thereafter) and MMF (1 g bid orally, with the first dose given IV in the operating room). In March 2001, we began a study of cyclosporine-MMF vs. tacrolimus-sirolimus in our steroid minimization protocol. We divided our recipients into three subgroups: those on cyclosporine-MMF (n = 85) vs. high-level tacrolimus (8–12 ng/mL), and low-level sirolimus (3–7 ng/dL) (n = 82) vs. low-level tacrolimus (3–7 ng/mL) and high-level sirolimus (8–12 ng/dL) (n = 72). The only exclusion criterion for enrollment into the study was current (i.e. immediately pretransplant) steroid use. An initial review of our results showed no difference between the three groups in the incidence of AR or in graft survival (19), so we have analyzed all of our rapid discontinuation data in aggregate.

Thus, in this report, we present our total experience with rapid discontinuation of steroids: our initial pilot study in first LD kidney recipients, the extension to all first and second LD and CAD kidney recipients (including those opting not to be randomized into our three-subgroup study), and the recipients in randomized study. All recipients followed the same antibody and steroid minimization protocol.

All recipients received prophylactic ganciclovir for 3 months posttransplant. Pneumocystis prophylaxis was with Bactrim; in patients with sulfa allergies, dapsone or aerosolized pentamidine was used. Fungal prophylaxis was with oral clotrimazole or nystatin for 3 months posttransplant.

Recipients with a ≥25% increase in their serum creatinine level underwent percutaneous allograft biopsy (histology and C4d study). All AR episodes were treated with steroids or with primary antibody therapy. Steroid-resistant rejection episodes were treated with antibody. Recipients with AR were maintained on prednisone (5 mg/day) long-term (although a few insisted on coming off prednisone again).

Leukopenia, common in patients not taking maintenance prednisone, was managed by a combination of tactics. Perhaps most important was resetting our definition of an acceptable white blood cell (WBC) count. Currently, if the recipient is well and the WBC count is stable and >2.0/mm3, we do not adjust the protocol. In addition, our interventions differ depending on the interval from the transplant and the rapidity of the fall in the WBC. If the WBC count falls rapidly during the patient's initial hospitalization (and while polyclonal antibody is being given), we temporarily stop the MMF or sirolimus, Bactrim, and ganciclovir, and we lower or hold the daily dose of Thymoglobulin. As the WBC rises, we add these drugs back into the protocol. Similarly, if the WBC falls rapidly after the patient's discharge from the hospital, we stop the Bactrim and ganciclovir and decrease the dose of MMF or sirolimus. If the WBC continues to fall or if it is less than 2.0/mm3 and not rising, we consider giving granulocyte-colony stimulating factor (G-CSF) (Neupogen) (300 µg subcutaneously daily for 3 days). If the WBC count falls later posttransplant, we change the Bactrim dose to ×3/week, hold the ganciclovir, and lower the MMF (to 500 mg bid) or sirolimus (to 1 mg daily). If there is no response, we consider giving G-CSF. In one of our recipients we completely discontinued the MMF because of a persistently low WBC count; this recipient subsequently experienced an AR episode (while on cyclosporine monotherapy). So now, we give G-CSF if the WBC count remains low after the MMF dose is reduced to 500 mg bid. We have also had a few recipients (n = 6) who were unable to maintain a WBC count ≥2.0/mm3 while taking low doses of MMF (500 mg bid) or sirolimus (1 mg daily), and who had only temporary responses to G-CSF; because we were uncomfortable lowering the MMF or sirolimus dose further, these six recipients started on prednisone at 5 mg daily.

Historical control group

We compared the outcome in our kidney recipients treated with our steroid minimization protocol with the outcome in a historical cohort (January 1, 1996 – December 31, 2000) of first and second LD (n = 180) and CAD (n = 310) kidney recipients treated with polyclonal antibody, a calcineurin inhibitor, either MMF or AZA, and a prednisone taper (1 mg/kg/day tapered to 0.4 mg/kg/day by 1 month, and to 0.15 mg/kg/day by 1 year).

From 1984 through the mid-1990s, our immunosuppressive protocol for CAD recipients consisted of sequential therapy (polyclonal antibody, AZA, and prednisone at the time of the transplant, with delayed introduction of cyclosporine); for LD recipients, triple therapy (cyclosporine, AZA, and prednisone at the time of the transplant). But, in the mid-1990s, we made a number of changes to our protocol. First, when MMF was approved by the Food and Drug Administration (FDA), we replaced AZA with MMF in our CAD protocol but not our LD protocol. Subsequently, after an internal review of our results showed a higher rejection rate in (non-HLA-identical) LD recipients (on triple therapy) than in CAD recipients (on sequential therapy with MMF rather than AZA), we began to use polyclonal antibody and MMF in LD recipients as well. For the purposes of our current report, these subgroups (CAD and LD treated with polyclonal antibody, cyclosporine, MMF, and prednisone) are our historical control group. Before 1999, the polyclonal antibody used was Atgam. In February 1999, Thymoglobulin was approved by the FDA; subsequently, all recipients (including all in the steroid discontinuation group) were treated with Thymoglobulin.

Data analysis

We studied patient and graft survival rates and the incidence and timing of AR episodes. Graft failure was defined by a retransplant, by return to dialysis, or by death with a functioning graft. A total of 47 recipients underwent a pancreas transplant after their kidney (PAK) transplant and, as part of their pancreas transplant, had another course of induction therapy and recycling of a steroid taper. Therefore, these 47 PAK recipients were dropped from our data analysis (censored) at the time of their pancreas transplant.

We compared patient and graft survival rates for LD and CAD recipients on the steroid discontinuation protocol with the rates for our historical control group (on prednisone maintenance).

For the steroid discontinuation group, we also determined the incidence of a number of steroid- and immunosuppression-related side effects, including cataracts, fractures, avascular necrosis, skin cancer, posttransplant diabetes mellitus (PTDM), posttransplant lymphoproliferative disorder (PTLD), and cytomegalovirus (CMV) infection. In addition, we obtained pre- and posttransplant weight, serum cholesterol, and triglyceride levels. We estimated actuarial patient, graft, and rejection-free survival rates by using Kaplan-Meier life table analyses. To compare differences between groups, we used log-rank and Wilcoxon tests.

Results

Of the 349 kidney transplants, 254 were LD and 95 were CAD transplants; 319 (91%) were first and 30 (9%) were second transplants (Table 2). Mean recipient age (± SE) was 47 ± 13 years (46 ± 12 years, LD recipients; 48 ± 13 years, CAD recipients). Most (87%) of the recipients were white, and 63% were male. The most common primary renal disease was diabetes (Table 2). All recipients were followed for a minimum of 3 months (average follow-up, 17 months; median, 16 months).

Table 2.  Recipient characteristics: rapid discontinuation of prednisone group vs. historical control group
 Rapid
discontinuation
Historical
control
  1. CAD = cadaver donor, FSGS = focal segmental glomerulosclerosis, IgA = immunoglobulin A, LD = Living donor.

Age (years)48 ± 13 (CAD);49 ± 13 (CAD);
 46 ± 12 (LD)46 ± 12 (LD)
GenderMale: 218Male 299
 (63%),(61%);
 Female: 131Female 191
 (37%)(39%)
Retransplant30 (8.6%)102 (21%)
Race
 White87%82%
 Hispanic3%3%
 African American3%7%
 Native American3%3%
 Asian3%0.2%
 Other3%4%
Primary kidney disease1% 
 Type 1 diabetes34%24%
 Polycystic kidney disease11%11%
 Type 2 diabetes10%11%
 Hypertension7%7%
 IgA nephropathy5%6%
 FSGS5%4%
 Other30%37%

For the entire group of 349 recipients, the actuarial patient survival rate at 1 year was 97%; at 3 years, 95%. We found no difference in patient survival rates between LD and CAD recipients (Table 3). A total of 12 recipients died (10 LD, two CAD): three from sepsis, two from cardiac causes, two from pulmonary complications, one from a CVA, one from graft-vs.-host disease, one from an electrolyte imbalance, and two from other causes. Of these 12 deaths, two occurred after graft failure.

Table 3.  Posttransplant outcome: rapid discontinuation of prednisone group
 12 months36 months
  1. CAD = cadaver donor, LD = living donor, NS = not significant.

  2. p = NS for all comparisons.

Actuarial patient survival
 All97%95%
 LD96%95%
 CAD98%98%
Actuarial graft survival
 All95%93%
 LD95%91%
 CAD98%95%
Actuarial death-censored graft survival
 All98%97%
 LD98%96%
 CAD99%99%
Actuarial acute rejection-free graft survival
 All94%92%
 LD94%92%
 CAD94%91%

The actuarial graft survival rate at 1 year was 95%; at 3 years, 93% (Table 3). We found no significant difference in the actuarial graft survival rate between LD and CAD recipients. A total of 19 grafts (5.9%) failed (17 LD, two CAD): 10 from death with function, three from technical complications, two from chronic rejection, one from calcineurin toxicity, one from primary nonfunction, and two from other causes. The death-censored graft survival rate at 1 year was 98%; at 3 years, 97%. Again, we found no difference in the death-censored graft survival rate between LD and CAD recipients.

The AR-free graft survival rate at 1 year was 94%; at 3 years, 92%. We found no difference in the AR-free rate between LD and CAD recipients (Table 3). A total of 24 recipients (7%) have had a biopsy-proven AR episode: 18 mild cellular, five moderate or severe cellular and/or vascular, and one C4d-positive (but without tubulitis). Median time to AR was 29 days (range, 13–544 days). Primary treatment consisted of steroids alone in 19 recipients, steroids plus antibody in four, and plasmapheresis and IVIg in one. Of these 24 recipients with AR, 18 continued on prednisone (5 mg/day) after treatment of the AR episode. Of these 18, two have had a second AR episode. Of the six recipients who discontinued prednisone after their initial rejection treatment, none have had a second AR episode.

Historical control group

Again, our historical control group (January 1, 1996–December 31, 2000) consisted of 180 first and second LD recipients and 310 first and second CAD recipients who were on polyclonal antibody, a calcineurin inhibitor, either MMF or AZA, and prednisone. Recipient characteristics are summarized in Table 2.

The outcome in our LD recipients on the rapid discontinuation of prednisone protocol is compared with the outcome in our historical control group in Table 4. We found no difference in patient survival, graft survival, or death-censored graft survival rates. However, LD recipients on the discontinuation of prednisone protocol had significantly better AR-free graft survival rates (p = 0.01) (Table 4).

Table 4.  Posttransplant outcome: living donor recipients on rapid discontinuation of prednisone vs. historical control group
 Time posttransplant
6 months12 months24 months36 months
  1. *p = NS, **p = <0.001. NS = not significant.

Patient survival*
 Rapid discontinuation (n = 254)98%96%96%95%
 Historical control (n = 180)98%97%97%95%
Graft survival*
 Rapid discontinuation96%95%94%91%
 Historical control94%92%92%88%
Death-censored graft survival*
 Rapid discontinuation98%98%97%96%
 Historical control96%94%94%93%
Acute rejection-free graft survival**
 Rapid discontinuation94%94%93%92%
 Historical control86%78%78%77%
Chronic rejection-free graft survival*
 Rapid discontinuation99%98%98%94%
 Historical control90%96%96%90%

The outcome in our CAD recipients on the rapid discontinuation of prednisone protocol is compared with the outcome in our historical control group in Table 5. CAD recipients on the rapid discontinuation of prednisone protocol had significantly better patient survival (p = 0.03), graft survival (p = 0.002), death-censored graft survival (p = 0.02), and AR-free graft survival rates (p = 0.02) (Table 5).

Table 5.  Posttransplant outcome: CAD recipients on rapid discontinuation of prednisone group vs. historical control group
 Time posttransplant
6 months12 months24 months36 months
  1. *p = 0.03, **p = 0.002, ***p = 0.02.

Patient survival*
 Rapid discontinuation (n = 95)98%98%98%98%
 Historical control (n = 300)94%91%91%90%
Graft survival**
 Rapid discontinuation98%98%98%98%
 Historical control89%86%86%83%
Death-censored graft survival***
 Rapid discontinuation97%99%99%99%
 Historical control94%92%92%90%
Acute rejection-free graft survival**
 Rapid discontinuation96%94%94%91%
 Historical control87%84%84%80%
Chronic rejection-free graft survival
 Rapid discontinuation99%99%97%97%
 Historical control98%97%96%93%

Kidney function

For LD recipients on the rapid discontinuation of prednisone protocol, the mean (±SE) serum creatinine level at 1 year was 1.6 ± 0.5 mg/dL; at 2 years, 1.6 ± 0.5; and at 3 years, 1.8 ± 0.8. For CAD recipients on that protocol, the mean (±SE) serum creatinine level at 1 year was 1.4 ± 0.5; at 2 years, 1.6 ± 0.4.

Adverse events

We noted no cases of PTLD or avascular necrosis. One recipient (whose primary disease was diabetes) developed a cataract; three developed nonmelanotic skin cancer; and 12 had fractures. A total of 22 recipients (6%) developed CMV infection. Of these 22 recipients, 10 were on prednisone at the time they developed the infection. A further four recipients developed herpetic lesions.

Four recipients (1.8%) developed new-onset diabetes (PTDM): all within the first 6 months posttransplant. No recipients maintained on cyclosporine-MMF developed PTDM. All four of the cases were in recipients maintained on tacrolimus-sirolimus (2.6%).

Weight, serum cholesterol, and triglyceride levels did not differ significantly from pretransplant levels at 12 and 24 months posttransplant (Table 6). However, of the first 300 recipients on the rapid discontinuation of prednisone protocol, 39 were started on lipid-lowering agents.

Table 6.  Weight and lipid levels: rapid discontinuation of prednisone group
 PretransplantPosttransplant
12 months24 months
Weight (kg) 78 ± 19 78 ± 17 82 ± 20
Cholesterol (mg/dL)185 ± 45191 ± 45197 ± 57
Triglycerides (mg/dL)202 ± 33201 ± 122207 ± 122

Long-term prednisone

Of the 349 recipients on the protocol, 47 underwent a PAK transplant, so were dropped from our data analysis at the time of the pancreas transplant. Of the remaining 289 recipients with functioning grafts (seven recipients died and six grafts were lost to other causes), 243 (84%) remain prednisone-free. The most common reason for starting prednisone was AR (n = 18). Other common reasons included persistent leukopenia while on a daily dose of 500 mg bid of MMF or 1 mg daily of sirolimus (n = 6), use of low-dose prednisone because of previous long-term prednisone use and current nonspecific symptoms (n = 5), long-lasting DGF (early in our series) (n = 5), recurrent disease (n = 2), persistent bone or joint pain (n = 2), and posttransplant thrombotic microangiopathy (HUS) (n = 2).

Discussion

We and others previously reported excellent patient and graft survival with no steroids or with rapid discontinuation of steroids (17–26). However, other than the study by Birkeland, no long-term follow-up has been reported until now. Our current patient and graft survival rates compare favorably with those reported with conventional immunosuppressive regimens (27), and with our historical control group treated with a prednisone taper (Tables 4 and 5). Our observed rate of AR is actually lower than that noted in many recent studies that used prednisone as part of their immunosuppressive protocol (Table 7).

Table 7.  Posttransplant outcome: recent studies using prednisone maintenance
Recipients enrolledMaintenanceOutcomeReference
  1. LD = living donor, CAD = cadaver donor, Tmg = Thymoglobulin, CSA = cyclosporine, AZA = azathioprine, P = prednisone, IL-2R = Interleukin 2-receptor monoclonal antibody, TAC = tacrolimus, SRL = sirolimus, B-PAR = biopsy-proven acute rejection, PS = patient survival, GS = graft survival.

1st LD and CADTmg, CSA, AZA, P6 months: B-PAR = 4% 
 Atgam, CSA, AZA, P6 months: B-PAR = 17%37
1st and 2nd LD and CADIL-2R antibody, CSA  
 MMF or AZA, P6 months: B-PAR = 25.7%38
1st and 2nd CADTAC, AZA, P6 months: B-PAR = 20%39
1st LD and CADCSA, SRL (5 mg), P6 months: B-PAR = 19%40
1st LD and CADCSA, SRL (5 mg), P1 year: B-PAR = 15%41
1st and 2nd CADCSA, MMF, P1 year: PS = 96% 
 (with or without antibody)1 year: GS = 90% 
 1 year: B-PAR = 20%43
1st CADTAC, MMF, P2 year: PS = 94% 
 2 year: GS = 90% 
 2 year: B-PAR = 20%43
1st and 2nd LD and CADCurrent study1 year: PS = 97% 
 1 year: GS = 95% 
 1 year: B-PAR = 6% 

Our rapid discontinuation of steroids protocol evolved from our pilot study of first LD kidney recipients to include all our first and second LD and CAD kidney recipients. Currently, our only exclusion criterion is prednisone use immediately pretransplant (within the previous 3 months), because for such recipients this would represent prednisone withdrawal — which was previously shown to be associated with an increased rate of acute rejection (9,10). Early in our experience, a few patients (e.g. those with long-lasting DGF) continued on low-dose prednisone (5 mg/day) or restarted low-dose prednisone; these two options are no longer used, now that we are more comfortable with the protocol. Of note, we no longer use a long-term prednisone taper for any of our kidney recipients (even if they are on steroids at the time of their transplant or even if it is their 3rd or 4th transplant); if prednisone is continued beyond the 5th postoperative day, the dose is 5 mg/day.

An interesting question is why steroid avoidance (or rapid discontinuation) was not associated, in our study, with an increased AR rate. After all, two large multicenter studies reported an increased AR rate after steroid withdrawal in recipients on prednisone, cyclosporine, and MMF (9,10). Part of the answer might be that cytokine receptor expression is increased by glucocorticoid-pretreated T cells (28); such an increase, in turn, enhances IL-2-stimulated T-cell proliferation. Clinically, proliferation may translate into enhanced T-cell activity when steroids are slowly withdrawn, as opposed to rapidly eliminated, and the result is a higher rate of AR. Of interest, steroids decrease the bioavailability of MMF by increasing hepatic UDP-glucoronyl transferase activity. One study showed that, when steroids were tapered or withdrawn, the MMF AUC increased (29); that is, when steroids were eliminated, there was more MMF exposure, possibly resulting in less AR. Another study showed that tacrolimus exposure also increased after steroid withdrawal (30).

We believe that some form of anti-T-cell induction therapy is necessary, when steroids are avoided or eliminated, in order to maintain acceptable AR rates. In other studies, various induction agents have been used with success, including Campath-IH (22,26), daclizumab (23,24), basiliximab (20,21), and Thymoglobulin (17,18). One study reported steroid elimination without antibody induction (25), but the incidence of AR was 29%.

In our study, the incidence of steroid-related adverse events was low. Other investigators have reported that steroid toxicity is related to both the average daily dose and the cumulative dose over time (12). The physiologic steroid dose appears to be equivalent to approximately 7 mg/day of prednisone. Low-dose regimens typically use less than 10 mg/day of prednisone; yet even with such regimens, lumbar bone mineral density is lost (13) and the risk of vertebral fractures increases (14). Furthermore, alternate-day steroid regimens do not appear to protect against bone disease (15,16).

Of note, we saw only four cases of PTDM in our study (1.8%). However, all four cases occurred in recipients maintained on tacrolimus-sirolimus (2.6%); none occurred in recipients maintained on cyclosporine-MMF. PTDM has previously been associated with both steroids (31) and tacrolimus (32). In kidney recipients on tacrolimus and steroids, rates of PTDM as high as 20% have been reported (33,34). Our study suggests that, in the absence of steroids, the risk of PTDM is low in calcineurin inhibitor-treated recipients.

Although our results are encouraging, a number of issues are unresolved. First, our longest follow-up to date is 46 months; with additional follow-up, we may begin to see some detriments to long-term steroid-free immunosuppression. Some recipient subgroups (e.g. those with glomerulonephritis) may do better with long-term low-dose steroids.

Second, whether or not steroid minimization protocols can be used in high-risk groups is unknown. Our study suggests that such protocols can be used for CAD recipients and for second transplant recipients. But 87% of our recipients were white. Other researchers have shown that steroid minimization protocols can be successful in blacks (Kaufman, personal communication).

Third, whether or not steroids need to be continued after treatment of an AR episode is unclear. We planned a prospective randomized trial to answer this question, but given the low AR rate, our single center will not have sufficient numbers.

Fourth, whether or not steroids are necessary when kidney recipients are readmitted for a pancreas transplant is unknown. Cantarovich et al. and Kaufman et al. showed that steroid minimization protocols can be successful after simultaneous kidney-pancreas transplants (35,36). In our study, we initially maintained PAK recipients on long-term steroids. Currently, kidney recipients at our center who are steroid-free at the time of their pancreas transplant receive steroids only during their induction therapy; they are not maintained on long-term oral steroids.

Finally, how should steroid-free and calcineurin inhibitor-free approaches be balanced? Steroid-free immunosuppression has obvious advantages: steroid-related side effects are eliminated. But numerous studies have now demonstrated better kidney function when calcineurin inhibitor use is either minimized or eliminated. Long-term studies are required to answer this question. One potential solution is to develop protocols that are both steroid-free and calcineurin inhibitor-free; to date, however, such protocols have been associated with significant side effects.

In conclusion, our study showed that, with currently available immunosuppressive agents, excellent 3-year patient and graft survival rates can be achieved without long-term steroid use. With our rapid discontinuation of prednisone protocol, steroid-related side effects were minimized, and the AR incidence was low. Thus, we believe that prednisone can be rapidly eliminated from routine posttransplant immunosuppression.

Acknowledgments

We thank Mary Knatterud for editorial assistance and Stephanie Daily for preparation of the manuscript. Supported by NIH Grant DK 13083 and grants from SangStat and Fujisawa.

Ancillary