A Randomized, Multicenter Study of Steroid Avoidance, Early Steroid Withdrawal or Standard Steroid Therapy in Kidney Transplant Recipients



This article is corrected by:

  1. Errata: Erratum Volume 8, Issue 5, 1080, Article first published online: 14 April 2008

Corresponding author: Flavio Vincenti, VincentiF@surgery.ucsf.edu


In a randomized, open-label, multicenter study, de novo renal transplant patients received no steroids (n = 112), steroids to day 7 (n = 115), or standard steroids (n = 109) with cyclosporine microemulsion (CsA-ME), enteric-coated mycophenolate sodium (EC-MPS) and basiliximab. The primary objective, to demonstrate noninferiority of 12-month GFR in the steroid-free or steroid-withdrawal groups versus standard steroids, was not met in the intent-to-treat population. However, investigational groups were not inferior to standard steroids in the observed-case analysis. Median 12-month GFR was not significantly different in the steroid-free or steroid-withdrawal groups (58.6 mL/min/1.73 m2 and 59.1 mL/min/1.73 m2) versus standard steroids (60.8 mL/min/1.73 m2). The 12-month incidence of biopsy-proven acute rejection (BPAR), graft loss or death was 36.0% in the steroid-free group (p = 0.007 vs. standard steroids), 29.6% with steroid withdrawal (N.S.) and 19.3% with standard steroids. BPAR was significantly less frequent with standard steroids than either of the other two regimens. Reduced de novo use of antidiabetic and lipid-lowering medication, triglycerides and weight gain were observed in one or both steroid-minimization group versus standard steroids. For standard-risk renal transplant patients receiving CsA-ME, EC-MPS and basiliximab, steroid withdrawal by the end of week 1 achieves similar 1-year renal function to a standard-steroids regimen, and may be more desirable than complete steroid avoidance.


The potential benefits of steroid-free maintenance following renal transplantation are well recognized (1,2). The long-term complications associated with steroid maintenance therapy include osteopenia, weight gain, hypertension, hyperglycemia and increased susceptibility to infection (3), such that aggressive steroid minimization is an attractive goal.

Modern immunosuppressive regimens have made rapid steroid discontinuation more feasible than previously reported with early immunosuppression regimens (4). In particular, the introduction of mycophenolic acid (MPA) therapy, together with the widespread adoption of antibody induction, led to a series of trials investigating different strategies to reduce steroid exposure (5–8). A multicenter, randomized trial showed that steroid withdrawal at 4 days posttransplant resulted in a similar incidence of rejection compared to standard steroids at 6 months in renal transplant patients receiving cyclosporine (CsA), mycophenolate mofetil (MMF) and basiliximab (9). In another prospective study, in which 538 renal transplant patients were given a single perioperative dose of methylprednisolone with tacrolimus, MMF and daclizumab, efficacy was similar to that seen with standard steroids: the biopsy-proven acute rejection (BPAR) rate was 16.5% in both arms (10). The safety of discontinuing steroid therapy within the first week posttransplant has been confirmed elsewhere in patients receiving a calcineurin inhibitor, MMF and induction with an IL-2 receptor antagonist or Thymoglobulin (7–12). Data concerning complete avoidance of steroids, without any perioperative i.v. methylprednisolone, is more limited but preliminary evidence from nonrandomized trials has suggested that this may be feasible (13,14).

The current study is the first randomized, controlled study to compare the safety and benefits of complete steroid avoidance versus either early steroid withdrawal or a standard-steroid regimen in a renal transplant population and the first steroid-minimization study to adopt renal function at 1 year posttransplant as the primary endpoint. The objective of this study was to assess the noninferiority of a steroid-free regimen or early steroid withdrawal versus a standard-steroids regimen in standard-risk patients receiving CsA, enteric-coated mycophenolate sodium (EC-MPS) and basiliximab. The results are potentially relevant to decision-making regarding the use and duration of steroid therapy in routine clinical practice.

Materials and Methods

Study design and conduct

In this 12-month, open-label, multicenter study de novo renal transplant recipients giving written informed consent were randomized to receive no steroids, steroids to day 7 posttransplant (steroid withdrawal), or standard steroids, all in combination with CsA microemulsion (CsA-ME, using C2 monitoring), EC-MPS and basiliximab. The study was conducted in full compliance with the amended Declaration of Helsinki following approval from the Institutional Review Committee at each center.

Patient population

Patients aged 18–75 years who received a primary kidney transplant from a non-HLA identical deceased, living-related or living-unrelated donor were eligible to enter the study. Key exclusion criteria comprised receipt of a kidney from a donor >60 years, a donor after cardiac death, previous kidney transplant and/or a transplant of any other organ, current (most recent prior to transplantation) panel reactive antibodies >20% and cold ischemia time >24 h.

Immunosuppression and concomitant therapy

Randomization into one of the three groups was undertaken in a 1:1:1 ratio using a validated system that automates the random assignment of treatment groups to randomization numbers. In the first group, patients received no i.v. or oral steroid therapy (‘steroid-free group’). In the second group, patients received steroids until day 7 (D7) posttransplant, comprising i.v. methylprednisolone (500 mg perioperatively [D1], 250 mg [D2] and 125 mg [D3]) followed by oral prednisolone at a recommended dose of 60 mg/day (D4), 40 mg/day (D5), 30 mg/day (D6) and 20 mg/day (D7). Steroids were then ceased (‘steroid-withdrawal group’). In the third group, patients received standard steroid therapy comprising i.v. methylprednisolone as in the steroid-withdrawal group, followed by oral prednisolone starting at 60 mg/day tapered to 20 mg/day by D7 and continued at 10–30 mg/day to the end of month 1, 10–20 mg/day during month 2, and 5–10 mg/day thereafter (‘standard-steroids group’). All patients received CsA-ME (Neoral®, Novartis Pharma AG, Basel, Switzerland), EC-MPS (myfortic®, Novartis Pharma AG), and basiliximab (Simulect®, Novartis Pharma AG). The initial dose of CsA-ME was 10 mg/kg/day, given within 24 h of transplantation, after which the dose was adjusted to achieve C2 target by D3 and maintain C2 level to predefined targets (ranges) as follows: 1700 ng/mL (1500–2000 ng/mL) during month 1, 1500 ng/mL (1300–1700 ng/mL) during month 2, 1300 ng/mL (1100–1500 ng/mL) during month 3, 1100 ng/mL (900–1300 ng/mL) during months 4–6 and 900 ng/mL (800–1000 ng/mL) thereafter. A single dose of EC-MPS (720–1440 mg) was given pretransplant, after which EC-MPS was administered within 24 h of transplantation at a dose of 1440 mg/day in two divided doses. Basiliximab was administered to all patients (20 mg; D1 and D4).

Acute rejection was treated with i.v. methylprednisolone according to local practice, (recommended dose; 500–1000 mg/24 h for 3 days). Treatment with antithymocyte antibodies was permitted in cases of steroid-resistant or vascular rejection. For patients with BPAR in the steroid-free and steroid-withdrawal groups continuation of oral steroid therapy throughout the remainder of the study was recommended.

Local practice determined Pneumocystis jeroveci pneumonia prophylaxis and prophylaxis for cytomegalovirus (CMV) for >3 months (except in CMV-negative recipients/negative donor), or treatment (e.g. ganciclovir) in CMV viral load-monitored (weekly) patients found to have at least one CMV-infected cell.

Study endpoints

The primary efficacy variable was glomerular filtration rate (GFR) calculated by the Nankivell formula (15) at month 12. Secondary efficacy variables included the cumulative incidence of a composite endpoint of death, graft loss or BPAR at months 3 and 12; cumulative incidence of BPAR at months 3 and 12; patient and graft survival at month 12; and the percentage of patients free from steroids at month 12. Safety variables included the incidence of adverse events and serious adverse events, blood pressure, lipid levels (total cholesterol, low-density lipoprotein [LDL]-cholesterol, high-density lipoprotein [HDL]-cholesterol and triglycerides) and blood glucose levels. The primary endpoint for measurements of bone mineral density (BMD) (measured at two different anatomical locations) was the percentage change in BMD of lumbar spine from baseline (D3–D8) to month 12; the secondary endpoint was the percentage change in BMD of hip (proximal femur) from baseline to month 12.


Study visits took place at baseline, days 1, 3, 5 and 8, week 2, and months 1, 3, 6, 9 and 12, including regular monitoring of hematology, blood chemistry, urinalysis, HbA1c, vital signs, adverse events, serious adverse events and infections. Dual x-ray absorptiometry (DXA) measurements of BMD in the lumbar spine and proximal femur were undertaken in a subgroup of patients at selected centers at baseline and at month 12. GFR calculations included patients who had died or lost their grafts, for whom GFR values were considered to be zero. In posthoc analyses, GFR values were also calculated including only those patients surviving with a functioning graft (‘observed case analysis’).

Suspected rejection episodes were to be confirmed by local biopsy according to Banff'97 criteria (16).

Statistical analysis

The intent-to-treat (ITT) population comprised all patients who were randomized, received at least one dose of EC-MPS and underwent kidney transplantation. The safety population comprised all patients who received at least one dose of EC-MPS and underwent at least one safety evaluation. An interim analysis was undertaken when all patients had completed the first 3 months of the study, including the cumulative incidence of BPAR, and the cumulative composite incidence of BPAR, patient death or graft loss, together with safety parameters (adverse events and laboratory values).

The sample size was calculated to provide 80% power to detect noninferiority of GFR (Nankivell) in either the steroid-free or steroid-withdrawal group compared to the standard-steroids group. A difference of <7 mL/min/1.73 m2 was considered to be noninferior. Based on values from previous studies (17) the standard deviation for GFR with a standard-steroids regimen was calculated to be 15 mL/min/1.73 m2. Based on these assumptions, 89 patients per treatment arm were necessary for between-group comparisons to be made at an adjusted significance level of 5%. The sample size calculation for the subpopulation of patients in whom BMD was measured (58 patients per group) was calculated to offer 80% power to detect a difference of 3% by a 2-sided t-test on the BMD of lumbar spine with a significance level of 0.05, based on a standard deviation of 5.745%.

Comparisons of GFR were performed using Wilcoxon rank-sum test. Analyses of patient and graft survival were performed using a Kaplan–Meier approach, with comparisons performed using the log-rank test. To take into account the multiplicity of testing, the analysis of BPAR rate at months 3 and 12 was conducted at a significance level of 0.0001% and 4.9999%, respectively, using the Hochberg procedure. Kaplan–Meier estimates of incidence were calculated and comparisons were performed using a z-test at the specified significance level.


Patient population

Three hundred and thirty-seven patients from 40 sites in North America, South America, Europe, Australia and Asia were randomized: steroid-free group, n = 112; steroid-withdrawal group, n = 116; standard-steroids group, n = 109. Two nontransplanted patients were excluded from the ITT population; one received EC-MPS and underwent a safety assessment so was included in the safety population.

Study populations are summarized in Table 1. Seventy-five patients experienced DGF (steroid-free [n = 26], steroid-withdrawal [n = 26] and standard-steroids [n = 23]). Thirty patients (8.9%) discontinued the study prematurely, and 82 patients (24.3%) discontinued study treatment (EC-MPS) prematurely, with adverse events the most common reason (n = 35) (Table 1). Recipient and donor characteristics were similar between groups other than a significantly higher proportion of living-related donors in the steroid-free group versus the steroid-withdrawal group (Table 2).

Table 1.  Patient disposition
 Steroid freeSteroid withdrawalStandard steroidsTotal
  1. 11 patient did not undergo transplantation or receive a safety assessment and was excluded from the safety population.

  2. 21 patient in the steroid-free group and 1 patient in the steroid-withdrawal group did not undergo transplantation and were excluded from the ITT population.

  3. 3ITT patients without delayed graft function.

  4. 4ITT patients without delayed graft function or major protocol deviation.

  5. 5Administrative problems, protocol violation, subject's condition no longer required study drug, subject withdrew consent.

Analysis populations
Safety population112 (100.0%) 1151 (99.1%)  109 (100.0%)336 (99.7%)
ITT population1112 (99.1%)  1152 (99.1%)  109 (100.0%)335 (99.4%)
ITT subpopulation385 (76.6%)89 (77.4%)86 (78.9%)260 (77.2%)
ITT patients with delayed graft function26 (23.4%)26 (22.6%)23 (21.1%) 75 (22.3%)
Per protocol population454 (48.2%)54 (46.6%)67 (61.5%)175 (51.9%)
Discontinued study prematurely
Subject withdrew consent2 (1.8%)5 (4.3%)5 (4.6%)12 (3.6%)
Death5 (4.5%)2 (1.7%)2 (1.8%) 9 (2.7%)
Lost to follow-up2 (1.8%)4 (3.4%)3 (2.8%) 9 (2.7%)
Total9 (8.0%)11 (9.5%)10 (9.2%) 30 (8.9%)
Discontinued study medication prematurely
Adverse events11 (9.8%)14 (12.1%)10 (9.2%)  35 (10.4%)
Unsatisfactory therapeutic effect5 (4.5%)9 (7.8%)2 (1.8%)16 (4.7%)
Abnormal laboratory values5 (4.5%)2 (1.7%)3 (2.8%)10 (3.0%)
Graft loss2 (1.8%)2 (1.7%)3 (2.8%) 7 (2.1%)
Death2 (1.8%)1 (0.9%)0 (0.0%) 3 (0.9%)
Lost to follow-up0 (0.0%)1 (0.9%)0 (0.0%) 1 (0.3%)
Other53 (2.7%)5 (4.3%)2 (1.8%)10 (3.0%)
Total28 (25.0%)34 (29.3%)20 (18.3%) 82 (24.3%)
Table 2.  Patient demographics and baseline characteristics. Figures in parentheses indicate percentage of patients
 Steroid freeSteroid withdrawalStandard steroidsTotal
(n = 111)(n = 115)(n = 109)(n = 335)
  1. 1p = 0.006 versus standard-steroids group.

  2. 2p = 0.014 versus steroid-withdrawal group.

Mean age ± SD (years)42.5 ± 13.34145.7 ± 12.3147.4 ± 13.1345.2 ± 13.04
Male gender72 (64.9%)84 (73.0%)70 (64.2%)226 (67.3%)
 Caucasian91 (82.0%)86 (74.8%)78 (71.6%)255 (76.1%)
 Black0 (0.0%)0 (0.0%)4 (3.7%) 4 (1.2%)
 Asian10 (9.0%) 13 (11.3%)13 (11.9%) 36 (10.7%)
 Other10 (9.0%) 16 (13.9%)14 (12.8%) 40 (11.9%)
Mean age ± SD (years)40.1 ± 14.442.1 ± 13.041.4 ± 12.141.2 ± 13.2
Age >60 years5 (4.5%)8 (6.9%)5 (4.6%)18 (5.4%)
Male gender60 (54.1%)55 (47.8%)49 (45.0%)164 (49.0%)
 Caucasian86 (77.5%)81 (70.4%)78 (71.6%)245 (73.1%)
 Black2 (1.8%)4 (3.5%)3 (2.8%) 9 (2.7%)
 Asian11 (9.8%) 10 (8.6%) 10 (9.2%) 31 (9.3%)
 Other5 (4.5%)8 (7.0%)7 (6.4%)20 (6.0%)
 Unknown7 (6.3%)12 (10.4%)11 (10.1%)30 (8.9%)
 Deceased 57 (51.4%)281 (70.4%)64 (58.7%)202 (60.3%)
 Living-related40 (36.0%)24 (20.9%)30 (27.5%) 94 (28.1%)
 Living-unrelated14 (12.6%)10 (8.7%)15 (13.8%) 39 (11.6%)
CMV R−/D+19 (17.1%)17 (14.8%)12 (11.0%) 48 (14.3%)
HLA mismatch 4–644 (39.6%)43 (37.4%)44 (40.4%)131 (39.1%)


At 12 months, 65 patients (59%) randomized to steroid-free therapy and 82 patients (71%) in the steroid-withdrawal group were free of steroids, as per protocol. In the standard-steroids group, steroids had been withdrawn in 13 patients (12%) at variance with the protocol. Among the patients who were receiving steroids at month 12 (day 241–360), the respective mean daily doses of steroids were 11.5 ± 11.3 mg, 12.6 ± 10.7 and 7.1 ± 4.2 mg.

Mean doses of CsA-ME, and mean C2 level, were similar at month 12 in all groups, as was EC-MPS dosing (Table 3). No difference was found between groups regarding the need for EC-MPS dose reductions, most of which were attributed to adverse events.

Table 3.  Nonsteroid immunosuppressive medication
 Steroid freeSteroid withdrawalStandard steroids
(n = 111)(n = 115)(n = 109)
Mean CsA-ME dose ± SD at month 12 (mg/kg/day) 3.4 ± 1.1 3.6 ± 2.1 3.2 ± 1.2
Mean CsA C2 level ± SD (ng/mL)
 D51338 ± 4901460 ± 5251426 ± 530
 Month 12 790 ± 268 805 ± 291 808 ± 300
EC-MPS dose
 Mean ± SD (mg/day), months 0–121256 ± 2641278 ± 2481340 ± 206
 Median, months 0–12143714381440
 % recommended dose88%89%93%
Basiliximab administration110 (98%)112 (97%)107 (98%)

Renal function

The primary endpoint was GFR at 12 months (Nankivell) (Table 4). For the ITT population, the 95% CI for the median difference from the standard-steroids group was −7.8 to 2.7 for the steroid-free group and −8.8 to 1.6 for the steroid-withdrawal group, i.e. noninferiority could not be demonstrated for either the steroid-withdrawal group or the steroid-free group versus standard-steroids based on the prespecified threshold of 7 mL/min/1.73 m2. In the observed case analysis, i.e. excluding patients who died or lost their graft, median GFR at 12 months was similar between groups (Table 4) with demonstrated noninferiority for both steroid-minimization arms compared to the standard-steroids group. Median GFR (Nankivell) values were stable in each group between months 6 and 12 (Figure 1). Median serum creatinine in the observed case analysis was 149 μmol/L, 149 μmol/L and 141 μmol/L, respectively, in the steroid-free, steroid-withdrawal and standard-steroids groups at month 3, and 130 μmol/L, 135 μmol/L and 141 μmol/L at month 12 (all N.S.).

Table 4.  Renal function at 12 months. GFR calculated using Nankivell formula
ITT population1Steroid freeSteroid withdrawalStandard steroids
(n = 111)(n = 115)(n = 109)
GFR (mL/min/1.73 m2)
 Mean ± SD56.0 ± 24.455.1 ± 22.358.8 ± 20.5
 Median (range)58.6 (44.3–72.7)59.1 (40.7–69.3)60.8 (49.1–70.7)
 95% CI versus standard steroids2−7.8 –2.7−8.8–1.6 
Observed case analysis3Steroid freeSteroid withdrawalStandard steroids
(n = 84)(n = 78)(n = 85)
  1. 1Including patients who died or lost their graft.

  2. 295% confidence interval (CI) for median difference to standard steroids group.

  3. 3Excluding patients who died or lost their graft.

GFR (mL/min/1.73 m2)
 Mean ± SD64.7 ± 17.261.6 ± 18.562.1 ± 17.6
 Median (range)64.7 (25.6–127.3)62.6 (2.1–112.4)63.4 (1.7–92.0)
 95% CI versus standard steroids2−3.8–6.4−6.5–4.9 
Figure 1.

Median glomerular filtration rate (GFR, Nankivell) in patients randomized to steroid-free therapy, steroid withdrawal at day 7, or standard steroids. ITT population. Vertical bars indicate 95% confidence intervals.

In a posthoc analysis, renal function was also analyzed in terms of actual steroids received, such that patients randomized to the steroid-free group but in whom steroids were administered were grouped with those randomized to steroid withdrawal who required steroids after day 7 (n = 124), and those who remained steroid-free were grouped with steroid-withdrawal patients who did not receive steroids after day 7 (n = 101). There were no significant differences in GFR at month 12 between these groups, although the size of this posthoc population (n = 225) meant that it was not powered to detect differences.


By month 12, nine patients (2.7%) had died and nine (2.7%) lost their grafts, with no significant differences between groups. At month 12, the incidence of the composite endpoint of BPAR, graft loss or death was significantly higher only in the steroid-free group versus the standard-steroids group (Table 5). The incidence of BPAR was 31.5%, 26.1% and 14.7%, respectively (p = 0.004 for the steroid-free group and p = 0.046 for the steroid-withdrawal group, both versus the standard-steroids group) (Table 5). BPAR occurred significantly earlier in the steroid-free group (p = 0.003) and the steroid-withdrawal group (p = 0.030) compared to the standard-steroids group (Kaplan–Meier estimates, log-rank test) (Figure 2). At month 3, the composite endpoint had occurred in significantly more patients in the steroid-free group (n = 31, 27.9%; p = 0.002) and the steroid-withdrawal group (n = 25, 21.7%; p = 0.032) than the standard-steroids group (n = 12, 11.0%). BPAR was also more frequent at month 3 in the steroid-free patients (n = 28, 25.2%, p < 0.001) and the steroid-withdrawal group (n = 21, 18.3%, p = 0.017) versus the standard-steroids group (n = 8, 7.3%). In subpopulation analyses, there were no significant differences in the incidence of any efficacy endpoint between the steroid-withdrawal and standard-steroids groups in patients with immediate graft function, in patients with DGF (data not shown), or in the per-protocol population, although the study was not powered to detect differences in these subgroups (Table 5).

Table 5.  Efficacy endpoints at month 12. Figures in parentheses show percentage of patients
 Steroid freeSteroid withdrawalStandard steroidsp (steroid free vs. standard steroids)p (steroid withdrawal vs. standard steroids)
  1. 1ITT patients without delayed graft function.

  2. 2ITT patients without delayed graft function or major protocol deviation.

ITT populationn = 111n = 115n = 109 
BPAR, graft loss or death40 (36.0%)34 (29.6%)21 (19.3%)0.007N.S.
BPAR35 (31.5%)30 (26.1%)16 (14.7%)0.0040.046
Graft loss4 (3.6%)2 (1.7%)3 (2.8%)N.S.N.S.
Death5 (4.5%)2 (1.7%)2 (1.8%)N.S.N.S.
ITT subpopulation1n = 85n = 89n = 86 
BPAR, graft loss or death27 (31.8%)23 (25.8%)14 (16.3%)0.020N.S.
BPAR24 (28.2%)21 (23.6%)12 (14.0%)0.025N.S.
Graft loss2 (2.4%)1 (1.1%)2 (2.3%)N.S.N.S.
Death3 (3.5%)1 (1.1%)0 (0.0%)N.S.N.S.
Per protocol population2n = 54n = 54n = 67 
BPAR, graft loss or death16 (29.6%)12 (22.2%) 8 (11.9%)0.021N.S.
BPAR15 (27.8%)10 (18.5%) 7 (10.4%)0.018N.S.
Graft loss1 (1.9%)1 (1.9%)1 (1.5%)N.S.N.S.
Death2 (3.7%)1 (1.9%)0 (0.0%)N.S.N.S.
Figure 2.

Cumulative incidence of biopsy-proven acute rejection (BPAR) in patients randomized to steroid-free therapy, steroid withdrawal at day 7, or standard steroids (Kaplan–Meier, ITT population). The value p = 0.003 for the steroid-free group and p = 0.03 for the steroid-withdrawal group, both versus the standard-steroids group (log-rank test).

Of note, significantly fewer biopsies were performed in the standard-steroids group (n = 130) than in either the steroid-free patients (n = 189, p = 0.007) or the steroid-withdrawal patients (n = 209, p < 0.001) and the mean number of biopsies per patient was also significantly higher in the steroid-minimization groups (Table 6). More patients in the steroid-free group were diagnosed with Grade IA rejection compared to the standard-steroids group (20 patients [18.0%] vs. 9 [8.3%], p = 0.032) (Table 6). When Grade IA rejection events were excluded, differences in all other grades of rejection were not significantly different between the three groups. Moderate to severe rejection episodes (Banff grade >2) were uncommon in all treatment groups. Steroid-resistant BPAR episodes were reported in four patients; steroid-free (n = 1), standard-steroids (n = 1), and steroid-withdrawal (n = 2). The median number of days to first BPAR episode among those patients who experienced BPAR was 14 (steroid-free), 53 (steroid-withdrawal) and 104 (standard-steroids), respectively (p = 0.003 for the steroid-free group and p = 0.022 for the steroid-withdrawal group, both vs. the standard-steroids group).

Table 6.  Number of biopsies and severity of rejection episodes at 12 months (ITT population).
 Steroid free (n = 111)Steroid withdrawal (n = 115)Standard steroids (n = 109)p (steroid free vs. standard steroids)p (steroid withdrawal vs. standard steroids)
Mean number of biopsies per patient1.71.81.2<0.050.001
Severity of BPAR
 Mild (Grade IA) 20 (18.0%)16 (13.9%)9 (8.3%)0.032N.S.
 Mild (Grade IB)10 (9.0%)6 (5.2%)4 (3.7%)N.S.N.S.
 Moderate (Grade IIA) 2 (1.8%)3 (2.6%)1 (0.9%)N.S.N.S.
 Moderate (Grade IIB) 1 (0.9%)2 (1.7%)1 (0.9%)N.S.N.S.
 Severe (Grade III)0 (0%)1 (0.9%)1 (0.9%)N.S.N.S.

In the steroid-free group, patients who experienced acute rejection had significantly lower CsA C2 levels than those who remained rejection-free, but this was not the case in the other two treatment groups. Among the patients randomized to steroid-free therapy, BPAR (up to D29 posttransplant) was reported in 15/60 patients (25%) with C2 <1500 ng/mL compared to 3/40 patients (7.5%) with C2 >1500 ng/mL (p = 0.033).

Adverse events

All but three patients reported at least one adverse event or infection (Table 7). There were no differences in the frequency of gastrointestinal adverse events between groups, although nausea and vomiting were reported as adverse events most frequently in the steroid-free group. Pyrexia was also observed more often in steroid-free patients. Infections occurred in 71 steroid-free (63.4%), 73 steroid-withdrawal (63.5%) and 71 standard-steroids patients (65.1%). CMV infection developed in 12, 8 and 9 patients in the three treatment arms. The incidence of serious adverse events in patients was similar in all groups (58 [51.8%] (steroid-free), 61 [53.0%] (steroid-withdrawal) and 59 patients [54.1%] (standard steroids)). Serious hematological abnormalities were observed in 17 patients; steroid-free (n = 4), steroid-withdrawal (n = 6) and standard-steroids (n = 7). Malignancies were reported in four patients (two in the steroid-free group [basal cell carcinoma, malignant neoplasm], one in the steroid-withdrawal group [basal cell carcinoma] and one in the standard-steroids group [basal cell carcinoma]). No posttransplant lymphoproliferative disease was observed.

Table 7.  Adverse events occurring in ≥10% of patients regardless of relation to study drug (safety population). Figures in parentheses indicate percentage of patients
 Steroid free (n = 112)Steroid withdrawal (n = 115)Standard steroids (n = 109)
Any adverse event111 (99.1%) 114 (99.1%) 108 (99.1%) 
Abdominal pain19 (17.0%)20 (17.4%)18 (16.5%)
Anemia47 (42.0%)39 (33.9%)26 (23.9%)
Constipation44 (39.3%)44 (38.3%)39 (35.8%)
Diarrhea25 (22.3%)34 (29.6%)28 (25.7%)
Hypercholesterolemia15 (13.4%)11 (9.6%) 13 (11.9%)
Hyperkalemia34 (30.4%)30 (26.1%)26 (23.9%)
Hypomagnesemia16 (14.3%)14 (12.2%)22 (20.2%)
Leukopenia20 (17.9%)19 (16.5%)15 (13.8%)
Nausea65 (58.0%)46 (40.0%)52 (47.7%)
Pyrexia39 (34.8%)27 (23.5%)17 (15.6%)
Tremor20 (17.9%)19 (16.5%)24 (22.0%)
Vomiting50 (44.6%)35 (30.4%)27 (24.8%)

Adverse events with a suspected relation to study drug (EC-MPS) were reported in 67 patients (59.8%), 63 patients (54.8%) and 59 patients (54.1%) in the three treatment groups. EC-MPS was discontinued due to adverse events in 11 patients (9.8%), 14 patients (12.1%) and 10 patients (9.2%) in the steroid-free, steroid-withdrawal and standard-steroids groups, respectively; one, two and one patients discontinued EC-MPS due to gastrointestinal events.

Steroid-related adverse events

Table 8 presents data at month 12 relating to parameters associated with steroid exposure. At baseline and month 12 the number of patients with diabetes mellitus (defined as fasting glucose >6.9 mmol/L) was similar between groups; however, the number of patients starting antihyperglycemic medication during the study was significantly lower in the steroid-free group than the standard-steroids arm (4.5% vs. 14.7%, p = 0.010). The median increase in BMI was lower in both steroid-sparing groups than in the standard-steroids group, reaching statistical significance in the withdrawal group (p = 0.008), and differences in weight gain were not attributable to differences in diarrhea, which occurred with similar frequency in all groups. Total cholesterol, LDL-cholesterol and HDL-cholesterol did not differ significantly between groups, but lipid-lowering medication was taken by a smaller proportion of steroid-withdrawal patients than standard-steroids patients (p = 0.020). Median triglyceride level was lower in the steroid-withdrawal group than the steroid-free group at month 6 (p = 0.025) and month 12 (p = 0.030). No differences were observed in terms of systolic or diastolic blood pressure between groups. Mean systolic blood pressure at study end was 132.8 ± 20.2 mmHg, 134.9 ± 17.8 mmHg and 132.5 ± 16.6 mmHg in the steroid free, steroid-withdrawal and standard-steroids group, respectively. Over the study systolic blood pressure decreased in all groups with lowest decrease in the withdrawal group (−5.8 mmHg change from baseline; p = 0.040 vs. standard-steroids group). Mean diastolic blood pressure at the end of study was 79.7 ± 12.2 mmHg, 80.3 ± 10.0 mmHg and 78.5 ± 10.5 mmHg in the three groups, respectively.

Table 8.  Parameters associated with steroid exposure at month 12. Continuous variables are shown as median values unless otherwise stated
 Steroid free (n = 112)Steroid withdrawal (n = 115)Standard steroids (n = 109)
  1. 1p = 0.010; 2p = 0.008; 3p = 0.030; 4p = 0.018 (all vs. standard-steroids group).

De novo use of antihyperglycemic medication5 (4.5%)114 (12.2%)16 (14.7%)
Change in body mass index (kg/m2)1.030.8821.88
Total cholesterol (mmol/L) [range]4.9 [2.5–7.5]4.9 [3.0–6.8]5.0 [3.0–9.7]
Triglyceride level (mmol/L) [range]1.6 [0.4–8.1]1.63[0.4–7.3]1.9 [0.6–11.0]
Use of lipid-lowering medication59 (53.2%)42 (36.5%)457 (52.3%)
Use of antihypertensive medication83/104 (79.8%)86/106 (81.1%)89/101 (88.1%)
Mean change in spine bone mass density (% change)−1.80 ± 6.72−1.77 ± 6.03−1.53 ± 9.93
Mean change in hip bone mass density (% change)1.21 ± 8.25 (n = 38)0.64 ± 8.36 (n = 39)−0.23 ± 8.10 (n = 36)

In a subprotocol, BMD measurements were made in 113 patients (38 steroid-free patients, 39 steroid-withdrawal patients and 36 standard-steroids patients). Spine BMD decreased in all three randomized groups by between a mean of 1.5% and 1.8% from baseline without significant differences between treatment arms. Hip BMD increased by a mean of 1.2% and 0.6% in the steroid-free and steroid-withdrawal groups, respectively, and decreased by 0.2% in the standard-steroids arm, but the differences were not significant.


These results indicate that in renal transplant patients receiving CsA-ME, EC-MPS and steroids with basiliximab induction, steroid withdrawal at day 7 posttransplant achieves comparable 1-year renal function to a standard-steroids regimen, which became statistically significant on the observed-case analysis. One-year graft and patient survival rates were similar, although the proportion of patients experiencing BPAR following early steroid withdrawal was approximately 10% higher than in the standard-steroids group, and occurred earlier. At 12 months, 65 patients (59%) randomized to steroid-free therapy and 82 patients (71%) in the steroid-withdrawal group were free of steroids, as per protocol.

In this open-label study, investigators appeared more likely to suspect rejection and perform a biopsy in patients with rising creatinine in the steroid-free or steroid-withdrawal groups, since there were significantly more biopsies performed in these groups. Only the incidence of rejections graded IA differed significantly between treatment groups. Regardless, the majority of rejection episodes in both the steroid-free and steroid-withdrawal groups were mild, and it has been shown previously that BPAR does not affect outcomes if it is not sufficiently severe to impair the recovery of renal function to baseline levels (18). Renal function at 1 year was stable in both steroid-minimization groups, and comparable to that seen in the standard-steroids group, suggesting that the increase in BPAR did not translate into inferior long-term function.

Posthoc central reading of biopsies was considered because of its ability to correct for bias in grading by local pathologists, but was not undertaken. In a dedicated analysis, the grading distribution, biopsy frequency and potential regional effect were assessed. No difference in gradings was detected between the regions, although a higher frequency of biopsies was found in the Latin America and Asia Pacific regions compared to North America and Europe. Many more biopsies were performed in the ‘experimental’ arms (i.e. steroid-free and steroid-withdrawal arms), a bias that could not be overcome by central reading.

Administration of steroids during the first week posttransplant in patients receiving a calcineurin inhibitor, MPA therapy and an IL-2 receptor antagonist has been reported to be associated with BPAR rates of 15–20% (7–12), whereas the only previous multicenter prospective study of complete steroid avoidance showed a 25% incidence of BPAR at 1 year (13). There is preliminary evidence to suggest that the anti-inflammatory action of steroids may help to limit ischemia reperfusion injury following transplantation (2,19), an effect that could potentially contribute to reduced risk of BPAR seen with short-term steroids compared to a steroid-free regimen. In a study by the Stanford group kidney transplant recipients treated without steroids showed a time lag before serum creatinine decreases posttransplant compared to those given steroid therapy (20), which would tend to support the hypothesis that a short course of perioperative steroids may provide a useful initial anti-inflammatory effect for reducing the ischemia-reperfusion injury. However, since the steroid-free patients with rejection in our study had lower CsA exposure than the other groups, the immunosuppressive effect of early steroids also appears important, at least when the immunosuppressive intensity of other agents falls below a minimum threshold.

Within the relatively short 1-year timeframe of this trial, some evidence of potential clinical benefits from steroid-minimization was seen in the steroid-free and steroid-withdrawal groups. Reduced de novo use of antidiabetic medication and lipid-lowering medication, reduced triglyceride level, less weight gain and a trend to increased hip BMD were observed in one or both of the steroid-minimization groups versus the standard-steroids group. Differences between the randomized groups are likely to have been muted by reintroduction of steroids to a sizeable proportion of patients in the steroid-free and steroid-withdrawal groups and discontinuation of steroids in over 10% of those randomized to maintenance steroids. The difference between reintroduction of steroids in the steroid-free and steroid-withdrawal groups may reflect the higher number of patients in the steroid-free group in whom steroids were initiated in response to a diagnosis of rejection. Nevertheless, the modest improvements in antidiabetic medication, weight gain, triglycerides and BMD seen at 1 year in our steroid-free and steroid-withdrawal populations are encouraging. Similar findings of an increased incidence of acute rejection with a modest improvement in metabolic profile associated with steroid withdrawal was recently reported in the follow-up of a blinded trial of rapid steroid withdrawal in patients receiving tacrolimus and MMF as maintenance therapy (22).

We are aware of the methodological limitations of the study. Despite donor age >60 years being an exclusion criteria, this applied to approximately 5% of patients although it was evenly spread between cohorts. In the standard-steroids group, 12% of patients were steroid-free against protocol. While reasons for steroid withdrawal were not recorded, this is likely to be due to intolerance in the majority of cases.

In conclusion, these results suggest that in a standard-risk renal transplant population receiving CsA-ME, EC-MPS and basiliximab, withdrawal of steroids by the end of the first week posttransplant may offer a favorable risk–benefit balance, with comparable 1-year renal function to a standard-steroids regimen. This strategy appears preferable to complete avoidance of steroid therapy, particularly since a high proportion of patients randomized to no steroids subsequently required steroid introduction. Longer follow-up is required, however, to characterize the safety and the benefit of steroid sparing beyond 1 year.


With grateful thanks to the FREEDOM Study investigators. Argentina: L. Toselli, P. Novoa. Australia: R. Walker, G. Russ, S. Campbell, M. Jose. Brazil: M. Mazalli, H. Tedesco. Canada: A. Salazar, J. Zaltman, J. Lawen, P. Campbell, A. Shoker, S. Paraskevas. Germany: I. Hauser, W. Arns. Italy: F. Schena. Malaysia: Z. Morad. New Zealand: H. Pilmore. Singapore: V. Anantharamar. Spain: A. Fructuoso, J. Grinyo, J. Baltar. Taiwan: P. Lee, S. Chu, W. Yang. United Kingdom: R. Moore, A. Ready, M. Shehata, S. Kashi. United States: J. Leone, K. Bodziak, F. Vincenti, H. Sollinger, M. Peskovitz, W. Bennett, D. Holt, F. Ferguson. The FREEDOM Study was funded by Novartis Pharma AG. We have the following conflicts of interest to declare with regards to Novartis concerning activity within the last 3 years: FV and IH have received travel grants and/or educational grants.