Prospective Randomized Trial of Maintenance Immunosuppression with Rapid Discontinuation of Prednisone in Adult Kidney Transplantation

Abstract

Rapid discontinuation of prednisone (RDP) has minimized steroid-related complications following kidney transplant (KT). This trial compares long-term (10-year) outcomes with 3 different maintenance immunosuppressive protocols following RDP in adult KT. Recipients (n=440; 73% living donor) from 3/2001–4/2006 were randomized into 1 of 3 arms: cyclosporine (CSA) and mycophenolate mofetil (MMF) (CSA/MMF, n=151); high-level tacrolimus (TAC, 8–12 μg/L) and low-level sirolimus (SIR, 3–7 μg/L) (TAC^H/SIR^L, n=149); or low-level TAC (3–7 μg/L) and high-level SIR (8–12 μg/L) (TAC^L/SIR^H, n=140). Median follow-up was ~7 years. There were no differences between arms in 10-year actuarial patient (~70%), graft (~60%), death-censored graft (~80%) survival, or in allograft function. There were no differences in the 10-year actuarial rates of biopsy-proven acute rejection (30%, 26%, and 20% in CSA/MMF, TAC^H/SIR^L, and TAC^L/SIR^H) and chronic rejection (38%, 35%, and 31% in CSA/MMF, TAC^H/SIR^L, and TAC^L/SIR^H). Rates of new-onset diabetes mellitus were higher with TAC^H/SIR^L (p=0.04), and rates of anemia were higher with TAC^H/SIR^L and TAC^L/SIR^H (p=0.04). No differences were found in the overall rates of 16 other post-KT complications. These data indicate that RDP-based protocol yield acceptable 10-year outcomes, but side effects differ based on the maintenance regimen used and should be considered when optimizing immunosuppression following RDP.

Introduction

Since the earliest renal allotransplants, corticosteroids have been the mainstay of immunosuppression (1–4), but their long-term use have been associated with well-described complications, including hypertension, hyperlipidemia, glucose intolerance (and new-onset diabetes [NODM]), loss of bone mineral density (and fracture), avascular necrosis, cataracts, skin and appearance changes, and in children – growth retardation (5–10). These side effects make long-term prednisone use unpopular among kidney transplant (KT) recipients, resulting in noncompliance (11, 12), and increasing the risk of allograft rejection, dysfunction, and loss (13). Consequently, strategies to minimize or eliminate long-term steroid use have been developed over decades.

Early trials of prednisone reduction involved late steroid withdrawal (≥3 months) and were performed in KT recipients on maintenance regimens consisting of cyclosporine (CSA), prednisone and either azathioprine (14, 15) or mycophenolate mofetil (MMF) (16–19). These studies found increased rates of acute rejection (AR) and late graft loss. More recent trials have involved rapid discontinuation of prednisone (RDP) within the first week post-KT (20–24). Individual studies, meta-analyses and registry reports have indicated that RDP increases the risk of mild AR and minimizes steroid-related complications, but has no impact on patient survival (PS) or graft survival (GS) (25–34). Maintenance immunosuppression following RDP has not yet been optimized, with different centers trying combinations of induction agents (thymoglobulin, alemtuzumab, IL-2R inhibitors, none), calcineurin inhibitors (CSA, tacrolimus [TAC]), anti-metabolites (MMF, azathioprine), mammalian target of rapamycin inhibitors (sirolimus [SIR], everolimus), and even a co-stimulation blocker (belatacept) (25–36). We started using RDP in 1999 and have recently reported our 10-year actuarial outcomes (37). In 2001, we began a prospective, randomized trial to determine if 1 of 3 maintenance immunosuppressive protocols provided better outcomes. Recipients were randomized into 1 of 3 study arms: CSA and MMF (CSA/MMF), high-level TAC and low-level SIR (TAC^H/SIR^L) or low-level TAC and high-level SIR (TAC^L/SIR^H). The rationale for SIR-based maintenance immunosuppression was data suggesting it may minimize fibrosis (38, 39). Interim (2-year) results from this trial have been reported (40). Herein, we report the actuarial 10-year results.

Methods

Study Design

From 3/2001 through 4/2006, all recipients of either a first or second (including both living donor and deceased donor) KT were invited to participate in this randomized trial; the protocol was approved by the Human Subjects Committee at the University of Minnesota. Exclusion criteria for study entry were limited to the use of maintenance prednisone within 3 months before KT. Anyone that was eligible but elected not to participate in the study was treated with our standard RDP protocol with maintenance CSA/MMF. The primary endpoint in our study included: return to dialysis, death with function, retransplant, or biopsy-proven chronic allograft nephropathy (CAN).

We enrolled a total of 440 recipients, and randomized them into 1 of 3 maintenance therapy arms: CSA/MMF (n=151); TAC^H/SIR^L (n=149); and TAC^L/SIR^H (n=140). For TAC or SIR, the low blood levels corresponded to 3–7 μg/L and the high blood levels corresponded to 8–12 μg/L (Figure 1).

Figure 1.

Diagram depicting study design, including the scheme for stratification of recipients as well as provisions for immunosuppression. Abbreviations: BID, twice daily; CSA, cyclosporine; OR, operating room; POD, post-operative day; QD, once daily; SIR, sirolimus; TAC, tacrolimus.

Immunosuppression

Our protocols have been described previously in detail (40). Briefly, all recipients received Thymoglobulin (Genzyme Corp., Boston, MA) induction at 1.25–1.5 mg/kg/dose intravenously for 5 doses. For delayed graft function (DGF, defined as need for dialysis within first post-KT week), we extended the course of Thymoglobulin (to a maximum of 10 doses). All recipients received methylprednisolone (500 mg) intraoperatively, and prednisone was administered and tapered over 5 post-KT days: 1 mg/kg on post-KT day 1, 0.5 mg/kg on post-KT days 2 and 3, and 0.25 mg/kg on post-KT days 4 and 5. Prednisone was discontinued on post-KT day 6 in all recipients, except in those experiencing DGF, who were given 5 mg/day prednisone concurrently with the extended course of Thymoglobulin, and whose prednisone was discontinued at the time of Thymoglobulin discontinuation. The remaining maintenance immunosuppression varied with respect to the study arm.

CSA/MMF

Recipients of CSA/MMF were given MMF at 1 g intraoperatively. Oral MMF (either 1 g twice daily for non-African Americans or 1.5 g twice daily for African-Americans) was started on post-KT day 1, and the dosing was adjusted as needed in the presence of gastrointestinal side effects (nausea, vomiting, or diarrhea) or myelosuppression. CSA (8 mg/kg/day, given in 2 divided doses) was started post-KT, and the dosing was adjusted to achieve blood levels of 150–200 μg/L (via high performance liquid chromatography [HPLC]) for the first 3 post-KT months. In recipients with poor early graft function or clinical DGF, the CSA dosing was delayed or slowly titrated upwards.

TAC^H/SIR^L

Recipients of TAC^H/SIR^L were given SIR at 1 mg preoperatively. SIR (2 mg/day) was given postoperatively, and the dosing was adjusted to achieve blood levels of 3–7 μg/L (via HPLC). TAC (0.03 mg/kg twice daily) was started postoperatively, and the dosing was adjusted to achieve blood levels of 8–12 μg/L (via microparticle enzyme immunoassay).

TAC^L/SIR^H

Recipients of TAC^L/SIR^H were given SIR at 1 mg preoperatively. SIR (5 mg/day) was given postoperatively, and the dosing was adjusted to achieve blood levels of 8–12 μg/L (via HPLC). TAC (0.015 mg/kg twice daily) was started postoperatively, and the dosing was adjusted to achieve blood levels of 3–7 μg/L (via microparticle enzyme immunoassay).

Recipients with ≥25% increase in serum creatinine levels underwent percutaneous allograft biopsy. AR episodes were treated with steroids or with primary antibody therapy. Steroid-resistant AR was treated with lymphocyte-depleting antibody. The majority of recipients having experienced an AR episode remained on maintenance prednisone (5 mg/day) indefinitely, with a few insisting to discontinue steroid therapy again (41).

Statistical Analysis

We studied actuarial PS, GS, and death-censored GS (DCGS) rates, and rates of biopsy-proven AR and chronic rejection (CR) for each group. CR was characterized as a clinical diagnosis confirmed by presence of interstitial fibrosis and tubular atrophy (IF/TA) on histological analysis of biopsy. Graft failure was defined by retransplant, return to dialysis, or death with functioning graft.

We tracked renal function, weight, and lipid profiles, and also the rates of selected medication use (corticosteroids, lipid-lowering and anti-hypertensive agents). We also analyzed the rates of immunosuppression-related complications.

Categorical variables were analyzed via the Chi-square test. Continuous variables were analyzed via 1-way ANOVA (for data that are normally-distributed) or the Kruskal-Wallis test (for data that are not normally-distributed), and survival, rejection, and complication rates were analyzed via Kaplan-Meier methods and compared via the Log-Rank test. Adjustment for multiple comparisons was also performed via the Log-Rank test.

Results

There were no differences between study arms in follow-up (~ 7 [5.5–8.5] years, median [interquartile range, IQR]), or in recipient age (~ 49 [40–59] years, median [IQR]). Similarly, there were no differences between study arms in other donor or recipient characteristics (Table 1).

Table 1.

Recipient and donor characteristics

	Study Arm
	Arm 1 (CSA/MMF)	Arm 2 (TACH/SIRL)	Arm 3 (TACL/SIRH)
Number [n (% of Total)]	151 (34.3)	149 (33.9)	140 (31.8)
Recipient	[median (IQR) or n (% of Arm)]
Patient f/u (months)	85.1 (69.5-105.6)	84.2 (67.7-107.6)	88.4 (68.9-104.9)
Kidney f/u (months)	82.4 (58.6-102.4)	83.7 (63.1-103.7)	86.7 (64.5-103.2)
Age (years)	50.4 (39.9-58.7)	48.1 (36.7-59.4)	48.6 (41.4-58.1)
Gender [n (% of Total)]
Male [268 (60.9)]	97 (64.2)	88 (59.1)	83 (59.3)
Female [172 (39.1)]	54 (35.8)	61 (40.9)	57 (40.7)
Race [n (% of Total)]
Caucasian [400 (90.9)]	141 (93.4)	130 (87.3)	129 (92.1)
African American [18 (4.1)]	5 (3.3)	8 (5.4)	5 (3.6)
Asian American [15 (3.4)]	3 (2.0)	7 (4.7)	5 (3.6)
Native American [7 (1.6)]	2 (1.3)	4 (2.7)	1 (0.7)
Primary Cause of ESRD [n (% of Total)]
Type 1 diabetes [103 (23.4)]	32 (21.2)	36 (24.2)	35 (25.0)
Type 2 diabetes [56 (12.7)]	16 (10.6)	17 (11.4)	23 (16.4)
Polycystic disease [55 (12.5)]	20 (13.3)	20 (13.4)	15 (10.7)
Hypertension [33 (7.5)]	17 (11.3)	9 (6.0)	7 (5.0)
IgA nephropathy [29 (6.6)]	5 (3.3)	10 (6.7)	14 (10.0)
FSGS [25 (5.7)]	7 (4.6)	8 (5.4)	10 (7.1)
Unknown [21 (4.8)]	7 (4.6)	7 (4.7)	7 (5.0)
Pre-Transplant Diabetes [n (% of Total)]
Overall [175 (39.8)]	51 (33.8)	61 (40.9)	63 (45.0)
Type 1 [105 (23.9)]	32 (21.2)	37 (24.8)	36 (25.7)
Type 2 [70 (15.9)]	19 (12.6)	24 (16.1)	27 (19.3)
Transplant Type [n (% of Total)]
Kidney alone [426 (96.8)]	147 (97.4)	145 (97.3)	134 (95.7)
Simultaneous pancreas-kidney [11 (2.5%)]	4 (2.7)	3 (2.0)	4 (2.9)
Kidney after pancreas [1 (0.3)]	0 (0)	0 (0)	1 (0.7)
Kidney and other organ [2 (0.5)]	0 (0)	1 (0.7)	1 (0.7)
Transplant # [n (% of Total)]
First [409 (93.0)]	139 (34.0)	140 (34.3)	130 (31.8)
Second [31 (7.0)]	12 (38.7)	9 (29.0)	10 (32.3)
HLA Mismatch (number)	3 (2–5)	3 (2–5)	3 (2–5)
Peak PRA > 10% [88/434 (20.3)]	33/149 (22.2)	28/147 (19.0)	27/138 (19.6)
Transplant PRA > 10% [53/399 (13.3)]	23/138 (16.7)	15/138 (10.9)	15/123 (12.2)
Donor
Age (years)	39.6 (29.7-50.0)	42.3 (33.1-48.9)	43.9 (35.4-49.8)
Gender [n (% of Total)]
Male [180/425 (42.4)]	66/146 (45.2)	57/145 (39.3)	57/134 (42.5)
Female [245/425 (57.7)]	80/146 (54.8)	88/145 (60.7)	77/134 (57.5)
Race [n (% of Total)]
Caucasian [398/424 (93.9)]	140/146 (95.9)	132/144 (91.7)	126/134 (94.0)
African American [13/424 (3.1)]	4/146 (2.7)	5/144 (3.5)	4/134 (3.0)
Native American [5/424 (1.2)]	2/146 (1.4)	1/144 (0.7)	2/134 (1.5)
Asian American [6/424 (1.4)]	0 (0)	5/144 (3.5)	1/134 (0.8)
Type [n (% of Total)]
Deceased [118 (26.8)]	43 (28.5)	36 (24.2)	39 (27.9)
Living-related [194 (44.1)]	66 (43.7)	68 (45.6)	60 (42.9)
Living-unrelated [128 (29.1)]	42 (27.8)	45 (30.2)	41 (29.3)

Abbreviations: CSA, cyclosporine; ESRD, end-stage renal disease; FSGS, focal segmental glomerular sclerosis; IQR, interquartile range; PRA, panel reactive antibody; SIR, sirolimus; TAC, tacrolimus

Superscripts: H = High blood level (8-12 μg/L); L = Low blood level (3-7 μg/L)

No overall differences between arms for any of the demographic characteristics.

Of the 440 recipients, 139 (31.6%) experienced loss of their transplanted kidney and 91 (20.7%) died over the course of study follow-up. Graft loss occurred in 48 (31.8%) with CSA/MMF, 48 (32.2%) with TAC^H/SIR^L, and 43 (30.7%) with TAC^L/SIR^H. The main causes of graft loss were death with functioning graft (48.2%) and CR (25.2%). There were no differences in the cause of kidney loss between study arms (p=0.3). Loss due to chronic rejection (or IF/TA) occurred in 17 (35.4%) with CSA/MMF, in 10 (20.8%) with TAC^H/SIR^L, and in 8 (18.6%) with TAC^L/SIR^H (p=0.1). Death with a functioning graft occurred in 20 (41.7%) with CSA/MMF, 23 (47.9%) with TAC^H/SIR^L, and 24 (55.8%) with TAC^L/SIR^H (Table 2). Among recipients that died, main causes of death included sudden death (8.8%), malignant metastasis (8.8%), congestive heart failure (7.7%), skin cancer (6.6%), kidney failure (5.5%), post-transplant lymphoproliferative disorder ([PTLD] 5.5%), cerebrovascular accident (4.4%) and infection (4.4%). There were no overall differences in the cause of recipient death between study arms (p=0.5).

Table 2.

Summary of prevalent causes of kidney loss in each randomized study arm

	Study Arm
	Arm 1 (CSA/MMF)	Arm 2 (TACH/SIRL)	Arm 3 (TACL/SIRH)
	[n (% of Arm)]
Total [139 (31.6% of all kidneys)]	48 (31.8)	48 (32.2)	43 (30.7)
Cause of kidney loss [n (% of all losses)]	[n (% of losses in Arm)]
Death with functioning graft [67 (48.2)]	20 (41.7)	23 (47.9)	24 (55.8)
CR/CAN [35 (25.2)]	17 (35.4)a	10 (20.8)	8 (18.6)a
Recurrent disease [6 (4.3)]	1 (2.1)	2 (4.2)	3 (7.0)
Recipient noncompliance [5 (3.6)]	2 (4.2)	3 (6.3)	0 (0)
Calcineurin inhibitor toxicity [4 (2.9)]	0 (0)	3 (6.3)	1 (2.3)

Abbreviations: CAN, chronic allograft nephropathy; CR, chronic rejection; CSA, cyclosporine; SIR, sirolimus; TAC, tacrolimus

Superscripts: H = High blood level (8-12 μg/L); L = Low blood level (3-7 μg/L)

No overall differences between arms in cause of kidney loss (p = 0.3 via Chi-square test).

No overall differences between arms in actuarial graft loss due to CR/CAN (p = 0.1 via Log-Rank test). However, direct comparison reveals difference approaching statistical significance between Arms 1 and 3 (^ap = 0.07 via Log-Rank test.)

Chronic rejection, allograft nephropathy, and recurrent disease were confirmed through biopsy.

There were no differences between study arms in actuarial PS (p=0.7), GS (p=0.96), and DCGS (p=0.5). The 10-year PS was 73.0% with CSA/MMF, 72.9% with TAC^H/SIR^L, and 67.0% with TAC^L/SIR^H. The 10-year GS was 63.5% with CSA/MMF, 59.9% with TAC^H/SIR^L, and 59.0% with TAC^L/SIR^H. The 10-year DCGS was 77.5% with CSA/MMF, 80.4% with TAC^H/SIR^L, and 82.1% with TAC^L/SIR^H (Table 3).

Table 3.

Patient and graft survival

	Study Arm
	Arm 1 (CSA/MMF)	Arm 2 (TACH/SIRL)	Arm 3 (TACL/SIRH)
Outcome [% of Total]	[% of Arm]
Patient survival
1-year [97.3]	97.3	98.0	96.4
5-year [89.2]	91.1	90.2	86.1
10-year [71.1]	73.0	72.9	67.0
Graft survival
1-year [94.5]	94.0	96.0	93.6
5-year [78.9]	77.6	79.6	79.7
10-year [60.8]	63.5	59.9	59.0
Death-censored graft survival
1-year [97.0]	96.6	98.0	96.4
5-year [86.3]	84.2	86.5	88.5
10-year [79.9]	77.5	80.4	82.1

Abbreviations: CSA, cyclosporine; SIR, sirolimus; TAC, tacrolimus

Superscripts: H = High blood level (8-12 μg/L); L = Low blood level (3-7 μg/L)

There were no differences in between study arms in 10-year actuarial AR (p=0.2) and CR (p=0.4) rates. The 1-year overall AR rate was 18.5% with CSA/MMF, 10.1% with TAC^H/SIR^L, and 10.7% with TAC^L/SIR^H. When comparing individual study arms, there was a difference in the 1-year AR rate between CSA/MMF and TAC^H/SIR^L (p=0.03) and a near difference when comparing CSA/MMF and TAC^L/SIR^H (p=0.054). The 1-year CR rate was 11.3% with CSA/MMF, 4.0% with TAC^H/SIR^L, and 4.3% with TAC^L/SIR^H. When comparing individual study arms, there was a difference in the 1-year CR rates between CSA/MMF and TAC^H/SIR^L (p=0.01), and between CSA/MMF and TAC^L/SIR^H (p=0.02). Also, renal function was found to be stable over time and similar when comparing study arms throughout the 10-year follow-up period (Table 4).

Table 4.

Delayed graft function, rejection, and renal function

	Study Arm
	Arm 1 (CSA/MMF)	Arm 2 (TACH/SIRL)	Arm 3 (TACL/SIRH)
Outcome [n (% of Total)]	[n (% of Arm)]
Rejection
Acute rejection
1-year [58 (13.2)]	28 (18.5)ab	15 (10.1)a	15 (10.7)b
5-year [90 (20.5)]	36 (25.0)	30 (21.6)	24 (18.0)
10-year [98 (22.3)]	40 (29.6)	33 (26.2)	25 (19.6)
Chronic rejection
1-year [29 (6.6]*	17 (11.3)cd	6 (4.0)c	6 (4.3)d
5-year [99 (22.5)]	37 (26.8)	38 (28.0)	24 (19.2)
10-year [122 (27.7)]	44 (38.2)	45 (35.3)	33 (30.7)
Renal function	[mean (± SE), in units of mg/dL or mL/min]
Serum creatinine
1-year	1.6 (± 0.06)	1.5 (± 0.04)	1.5 (± 0.04)
5-year	1.8 (± 0.16)	1.7 (± 0.10)	1.5 (± 0.07)
10-year	1.5 (± 0.08)	1.5 (± 0.15)	1.7 (± 0.15)
Creatinine clearance
1-year	62.9 (± 2.7)	64.1 (± 2.3)	64.6 (± 2.4)
5-year	65.8 (± 3.2)	63.1 (± 2.8)	66.6 (± 3.0)
10-year	61.1 (± 6.0)	59.7 (± 5.2)	65.2 (± 6.3)

Abbreviations: CSA, cyclosporine; SE, standard error; SIR, sirolimus; TAC, tacrolimus

Superscripts: H = High blood level (8-12 μg/L); L = Low blood level (3-7 μg/L)

Multiple comparisons via Log-Rank:*(p < 0.05);

Direct comparisons via Log-Rank test: ^a(p = 0.03); ^b(p = 0.05); ^c(p = 0.01); ^d(p = 0.02)

For all recipients, 89.8% remained prednisone-free at 1 year, 90.4% at 5 years and 79.4% at 10 years post-KT (actuarial rates). The 5- and 10-year prednisone use did not differ significantly between study arms. Of all recipients, 53.9% underwent a drug switch from their original maintenance immunosuppressive regimen sometime during the 10-year follow-up, including 42.4% with CSA/MMF, 63.1% with TAC^H/SIR^L, and 56.4% with TAC^L/SIR^H. Most switches were due to drug side effects. For each study arm, weight and lipid profiles remained stable over time. In addition, there were no differences between arms in mean total cholesterol and triglyceride levels (data not shown). Rates of lipid-lowering medication use were different between study arms only at 6 months (p=0.04), and at 1 (p=0.001) and 3 years (p=0.007), but not different at other time points post-KT. However, rates of lipid-lowering medication use were higher in recipients taking SIR, which may indicate an increased need for lipid control in these recipients. Rates of anti-hypertensive medication use were not different between study arms at any of the post-KT time points (Table 5).

Table 5.

Summary for use of selected medications

	Study Arm
	Arm 1 (CSA/MMF)	Arm 2 (TACH/SIRL)	Arm 3 (TACL/SIRH)
Type of medication [n (% of Total)]	[n (% of Arm)]
Corticosteroids
Last kidney f/u [73/425 (17.2)]	28/146 (19.2)	21/145 (14.5)	24/134 (17.9)
1-year rates [41/401 (10.2)]**	22/137 (16.1)	5/139 (3.6)	14/125 (11.2)
5-year rates [32/332 (9.6)]	16/112 (14.3)	8/114 (7.0)	8/106 (7.6)
10-year rates [7/34 (20.6)]	3/14 (21.4)	1/8 (12.5)	3/12 (25.0)
Lipid-lowering
Pretransplant [173/287 (60.3)]	53/95 (55.8)	62/106 (58.5)	58/86 (67.4)
6-month rates [171/326 (52.5)]*	46/110 (41.8)	64/113 (56.6)	61/103 (59.2)
1-year rates [190/328 (57.9)]**	43/100 (43.0)	76/120 (63.3)	71/108 (65.7)
3-year rates [197/313 (62.9)]**	54/105 (51.4)	74/108 (68.5)	69/100 (69.0)
5-year rates [102/167 (61.1)]	31/55 (56.4)	37/56 (66.1)	34/56 (60.7)
Anti-hypertensive
Pretransplant [357/386 (92.5)]	116/129 (89.9)	122/130 (93.8)	119/127 (93.7)
6-month rates [358/418 (85.7)]	123/143 (86.0)	123/144 (85.4)	112/131 (85.5)
1-year rates [345/411 (83.9)]	122/140 (87.1)	116/141 (82.3)	107/130 (82.3)
3-year rates [296/351 (84.3)]	107/118 (90.7)	100/123 (81.3)	89/110 (80.9)
5-year rates [152/184 (82.8)]	53/60 (88.3)	54/64 (84.4)	45/60 (75.0)

Abbreviations: CSA, cyclosporine; SIR, sirolimus; TAC, tacrolimus

Superscripts: H = High blood level (8-12 μg/L); L = Low blood level (3-7 μg/L)

Multiple comparisons via Chi-Square test: *(p < 0.05); **(p < 0.01);

Immunosuppression-related complications occurred in each study arm, but complications traditionally associated with prednisone therapy (avascular necrosis, cataracts, fractures) were rare. Only 2 of our studied complications, NODM (p=0.04) and anemia (p=0.04), exhibited differences in their incidence rates between study arms. When comparing individual study arms, there were differences in the overall rates of NODM between CSA/MMF and TAC^H/SIR^L (p=0.03), and between TAC^H/SIR^L and TAC^L/SIR^H (p=0.03). The 10-year rates of NODM were 6.3% with CSA/MMF, 19.3% with TAC^H/SIR^L, and 5.5% with TAC^L/SIR^H. When comparing individual study arms, there were differences in the overall rates of anemia between CSA/MMF and TAC^L/SIR^H (p=0.02), and a near difference when comparing CSA/MMF and TAC^H/SIR^L (p=0.055). The 10-year rates of anemia were 55.2% with CSA/MMF, 70.1% with TAC^H/SIR^L, and 67.2% with TAC^L/SIR^H (Figure 2). Other notable but not significant differences between individual study arms included rates of neutropenia, PTLD and skin cancer at 10-years post-KT (Table 6).

Figure 2.

Actuarial rates of new-onset diabetes mellitus ([NODM] left) and anemia (right) as stratified for each study arm. There are differences between arms in the rates of NODM for 10 years of post-transplant follow-up (p=0.04). When comparing individual study arms, the rates of NODM were different between Arms 1 and 2 (p=0.03), and Arms 2 and 3 (p=0.03), but not different between Arms 1 and 3 (p=0.9). There are differences between arms in the rates of anemia for 10 years of post-transplant follow-up (p=0.04). When comparing individual study arms, the rates of anemia were different between Arms 1 and 2 (p=0.02), was approaching a near difference between Arms 1 and 3 (p=0.06), and was not different between Arms 2 and 3 (p=0.7).

Table 6.

Selected complication

	Study Arm
	Arm 1 (CSA/MMF)			Arm 2 (TACH/SIRL)			Arm 3 (TACL/SIRH)
Posttransplant year	[% of Arm]
	1	5	10	1	5	10	1	5	10
Complication
Anemia*	44.2	53.2	55.2ab	53.0	63.0	70.1a	59.2	63.8	67.2b
BK Polyomavirus disease	0.7	2.3	2.3	0.7	2.3	3.4	0	2.5	3.5
Calcineurin inhibitor toxicity	2.7	3.5	3.5	1.4	3.7	3.7	5.2	6.1	6.1
Cerebrovascular accident	0.7	3.2	7.6	1.4	2.9	13.2	3.7	3.7	12.9
Deep vein thrombosis	0	0	0	0	0	0	0	0	0
Cytomegalovirus disease	8.4	18.2	19.1	6.9	11.4	16.2	4.5	10.1	13.4
Fascial dehiscence	2.0	2.0	2.0	2.0	2.0	5.2	4.4	4.4	4.4
Kidney infarction/thrombosis	0.7	0.7	0.7	1.3	1.3	1.3	0.7	0.7	0.7
Lymphocele	12.1	12.1	12.1	14.3	14.3	14.3	10.4	10.4	10.4
Myocardial infarction	4.0	7.1	9.0	1.3	5.2	10.0	5.2	6.1	8.1
New onset diabetes mellitus*	1.0	3.4	6.3c	7.1	13.9	19.3cd	2.7	5.5	5.5d
Neutropenia	23.3	30.3	30.3e	16.5	20.4	21.3e	17.1	23.7	23.7
Oral ulceration	0	0	0f	1.4	1.4	3.3fg	0	0	0g
Pneumonia	5.5	20.0	34.1	10.3	21.2	31.9	8.9	23.1	41.8
Posttransplant lymphoproliferative disorder	0	0	0h	0	1.6	1.6	1.5	2.3	3.7h
Pulmonary embolus	0	0.9	0.9	0.7	0.7	2.2	1.5	1.5	1.5
Recurrent disease	1.4	3.1	5.9	0	2.5	5.0	2.2	3.8	5.7
Skin cancer	3.5	21.7	39.2i	0.7	12.8	32.0i	1.5	15.2	24.4
Thrombocytopenia	2.7	5.2	5.2	1.4	4.4	9.0	3.7	6.1	7.4

Abbreviations: CSA, cyclosporine; SIR, sirolimus; TAC, tacrolimus

Superscripts: H = High blood level (8-12 μg/L); L = Low blood level (3-7 μg/L)

Multiple comparisons via Log-Rank: *(p < 0.05);

Direct comparisons via Log-Rank test: ^a(p = 0.06); ^b(p = 0.02); ^c(p = 0.03); ^d(p = 0.03); ^e(p = 0.08); ^f(p = 0.04); ^g(p = 0.04); ^h(p = 0.04); ⁱ(p = 0.07)

Discussion

We found no long-term differences in actuarial survival outcomes (PS, GS and DCGS) when comparing CSA/MMF, TAC^H/SIR^L, and TAC^L/SIR^H maintenance protocols. We and others have shown that RDP is associated with similar PS and GS but fewer prednisone-related side effects than chronic corticosteroid maintenance therapy (42, 43). Our current analysis also shows no differences between study arms in actuarial acute or chronic rejection rates, or long-term allograft function. Approximately 80% of all recipients remained prednisone-free at 10 years post-KT. There were differences between study arms in the overall rates of only 2 (of 18) studied complications (NODM and anemia); NODM was significantly more prevalent with TAC^H/SIR^L, and anemia with TAC^H/SIR^L and TAC^L/SIR^H.

This trial has a longer median follow-up period (~7 years) than comparable trials of RDP (34, 42–46). Data with long-term follow-up is important when comparing maintenance protocols. A consideration in the design of this study was the suggestion that SIR may be a better long-term agent due to its established anti-proliferative and anti-fibrotic properties (38, 39). Our data do not show differences in survival outcomes, though there are indicators that suggest incorporation of SIR into maintenance therapy may alter the mechanism of graft failure. For example, CR/CAN accounted for 35% of graft losses with CSA/MMF versus only 19–21% of graft losses with protocols containing SIR. However, the impact of SIR on preventing IF/TA may not be as strong as initially hypothesized; 1-year CR rates were 11% with CSA/MMF versus only 4% with protocols containing SIR (p<0.05), but the 10-year CR rates were 38% with CSA/MMF versus 35% with TAC^H/SIR^L and 31% with TAC^L/SIR^H.

It may be that our sensitivity to detect the effect of SIR in minimizing intrarenal fibrosis was limited because we did not incorporate protocol biopsies. Kumar et al. recently published 5-year results from a prospective randomized trial comparing 4 different maintenance protocols following induction with basiliximab and RDP on day 2 (n=200). The study incorporated surveillance biopsies at 7 different time points from 1–60 months post-KT (34). The 4 protocols involved CSA/MMF, CSA/SIR, TAC/MMF and TAC/SIR (n=50 recipients per protocol), using doses and blood levels that were similar when compared with our study. Recipient demographics differed from our study; Kumar et al. had more African-Americans (~50% vs. ~5% in our study) and deceased donor KTs (~80% vs. ~27% in our study). The authors reported that the 5-year incidence of CR (or IF/TA) was 54% with CSA/MMF, 16% with CSA/SIR, 38% with TAC/MMF, and 14% with TAC/SIR. When comparing individual study arms, there were differences in the incidence of IF/TA between CSA/MMF and CSA/SIR (p=0.04), and between CSA/MMF and TAC/SIR (p=0.04), suggesting that the incorporation of SIR into the maintenance regimen may decrease renal allograft fibrosis. However, there were no differences between study arms in PS or GS, or in serum creatinine levels and creatinine clearances. Interestingly, they reported that the highest tolerance of original immunosuppressive therapy was with TAC/SIR (90%) and was higher than in the other study arms (82% with CSA/MMF, 84% with CSA/SIR, and 82% with TAC/MMF). Gallon et al. reported 3-year results of a prospective randomized single-center pilot study comparing TAC/MMF (n=45) and TAC/SIR (n=37) following basiliximab induction and RDP on day 2 (44); they reported GS that was significantly lower with TAC/SIR (84%) than with TAC/MMF (98%) (p=0.04).

It is unclear what accounts for the differences in outcomes between Gallon et al., Anil Kumar et al., and our study. Although demographics were different, none of the differences suggest that the study populations were distinct enough to explain the disparity in outcomes. Furthermore, their immunosuppressive protocols indicated the use of similar TAC and SIR blood levels as in our study. Of note, both studies in Gallon et al. and Kumar et al. had smaller numbers of recipients in each study arm and our follow-up has been longer.

We found that the rates of studied drug-related complications were generally low, regardless of protocol. However, there were differences between study arms. First, NODM was more prevalent in those receiving TAC^H/SIR^L. This finding can be explained in part by the known diabetogenicity of TAC, which is believed to be more pronounced than CSA (47), and which would be accentuated with higher blood levels of TAC. Second, anemia was more prevalent in those receiving SIR. This may be explained in part by a known interaction between CSA and MMF, which has been shown to reduce the therapeutic mycophenolic acid blood levels (48–50), which would in turn reduce the risk of myelosuppression. Interestingly, incidence of skin cancer was progressively lower with increasing maintenance blood levels of SIR. This finding may highlight the emerging data on the anti-neoplastic properties of SIR, particularly as it relates to non-melanoma skin cancer (51–53).

Our study had several limitations. First, as mentioned earlier, there were no protocol biopsies. Consequently, subclinical acute rejection or chronic fibrotic changes that did not manifest clinically may not have been detected. Second, our statistical analyses were based on intention-to-treat and there was a relatively high drug conversion rate in each study arm over the duration of the trial.

In summary, outcomes at 10 years post-KT remain acceptable following RDP with most recipients remaining steroid-free. Recipients in 1 of 3 different maintenance protocol arms (CSA/MMF, TAC^H/SIR^L, or TAC^L/SIR^H) had similar long-term PS and GS, rejection rates, and renal function. However, cause of graft loss and complications appeared to vary depending on the type of immunosuppression, but this needs to be studied further and will be as more recipients reach ≥10 years post-KT. In conclusion, these results help establish the long-term efficacy of several maintenance immunosuppressive protocols used following RDP with KT. Future studies are necessary to continue to monitor for differences in late graft loss and for changes in side effect profiles attributable to the immunosuppressive regimen being used.

Abbreviations

AR

acute rejection

CAN

chronic allograft nephropathy

CR

chronic rejection

CSA

cyclosporine A

CVA

cerebrovascular accident

DCGS

death-censored graft survival

DGF

delayed graft function

ESRD

end-stage renal disease

FSGS

focal segmental glomerulosclerosis

GS

graft survival

HPLC

high performance liquid chromatography

IF/TA

interstitial fibrosis and tubular atrophy

IQR

interquartile range

KT

kidney transplant

MMF

mycophenolate mofetil

NODM

new-onset diabetes mellitus

PRA

panel reactive antibody

PS

patient survival

PTLD

post-transplant lymphoproliferative disorder

RDP

rapid discontinuation of prednisone

SE

standard error

SIR

sirolimus

TAC

tacrolimus