Complete Avoidance of Calcineurin Inhibitors in Renal Transplantation: A Randomized Trial Comparing Sirolimus and Tacrolimus

Authors


Abstract

Calcineurin inhibitors have decreased acute rejection and improved early renal allograft survival, but their use has been implicated in the development of chronic nephrotoxicity. We performed a prospective, randomized trial in kidney transplantation comparing sirolimus-MMF-prednisone to tacrolimus-MMF-prednisone. Eighty-one patients in the sirolimus group and 84 patients in the tacrolimus group were enrolled (mean follow-up = 33 months; range 13–47 months). At 1 year, patient survival was similar in the groups (98% with sirolimus, 96% with tacrolimus; p = 0.42) as was graft survival (94% sirolimus vs. 92% tacrolimus, p = 0.95). The incidence of clinical acute rejection was 10% in the tacrolimus group and 13% in the sirolimus group (p = 0.58). There was no difference in mean GFR measured by iothalamate clearance between the tacrolimus and sirolimus groups at 1 year (61 ± 19 mL/min vs. 63 ± 18 mL/min, p = 0.57) or 2 years (61 ± 17 mL/min vs. 61 ± 19 mL/min, p = 0.84). At 1 year, chronicity using the Banff schema showed no difference in interstitial, tubular or glomerular changes, but fewer chronic vascular changes in the sirolimus group. This study shows that a CNI-free regimen using sirolimus-MMF-prednisone produces similar acute rejection rates, graft survival and renal function 1–2 years after transplantation compared to tacrolimus-MMF-prednisone.

Introduction

Over the past 20 years, improvements in immunosuppression and patient management have reduced the rate of acute rejection and improved 1-year graft survival in kidney transplantation. However, long-term graft survival has not increased significantly (1). Several factors contribute to the persistence of late graft loss, including death with function, infections and chronic immunologic injury (2).

Treatment with the potentially nephrotoxic calcineurin inhibitors (CNIs) cyclosporine and tacrolimus may also contribute to progressive graft dysfunction, eventually leading to chronic allograft nephropathy (CAN) (3). The advent of newer immunosuppressive medications makes it possible to employ protocols that avoid CNIs and test the hypothesis that CNIs contribute to CAN. However, when employing novel protocols, it is important to remember that any increase in acute rejection rates may increase the incidence of graft failure (4). Sirolimus, an inhibitor of the mTOR signaling pathway, has been shown to decrease the incidence of acute rejection when combined with cyclosporine or tacrolimus (5–7). It has also been suggested that, by preventing fibroblast proliferation, this drug may lead to a less fibrogenic intragraft environment (8). Thus, including sirolimus in a CNI-free regimen designed to prevent CAN might be especially advantageous.

The aim of this study was to compare the complete avoidance of CNIs using a sirolimus-based immunosuppressive regimen to a tacrolimus-based regimen in kidney transplantation. We performed a prospective, open-label trial randomizing patients to receive tacrolimus, mycophenolate mofetil and prednisone or sirolimus, mycophenolate mofetil and prednisone. All patients received anti-thymocyte globulin induction. We analyzed the data for patient and graft survival, acute rejection, renal function, complications and adverse events.

Methods

This study was conducted with informed consent using a protocol approved by the Institutional Review Board of Mayo Clinic, Rochester, MN.

Patients

Inclusion criteria included living and deceased donor kidney transplant recipients at Mayo Clinic, Rochester. Patients included in this analysis were transplanted between April 2001 and January 2004. Exclusion criteria included the following: (i) recipients of multi-organ transplants; (ii) pediatric recipients; (iii) patients expected to receive a pancreas-after-kidney transplant; (iv) patients receiving an ABO-incompatible or positive crossmatch transplant; (v) pre-transplant fasting serum cholesterol level greater than 350 mg/dL or fasting serum triglyceride level greater than 500 mg/dL and (vi) pre-transplant white blood cell count of less than 3000/mm3. Approximately 12 months after enrollment began, recipients with a body mass index greater than 32 kg/m2 were excluded because of a high incidence of wound complications in obese patients using the sirolimus protocol (9). Patients were randomized to one of the two regimens immediately prior to transplantation.

Immunosuppressive regimens

All patients received Thymoglobulin® (SangStat, Fremont, CA) induction (1.5 mg/kg/day) on days 0, 1, 2, 4 and 6 and mycophenolate mofetil (CellCept®, Hoffmann-LaRoche, Inc., Nutley, NJ) at a dose of 750 mg orally twice daily. Tacrolimus (Prograf®; Fujisawa Healthcare Inc., Deerfield, IL) or sirolimus (Rapamune®; Wyeth Pharmaceuticals, Collegeville, PA) was started on post-operative day 4. Tacrolimus was initially dosed at 3 mg twice daily. Sirolimus was initially dosed at 10 mg daily for 2 d and 5 mg daily thereafter. The target 12-hour trough levels for tacrolimus was 10–12 ng/mL in the first month, 8–10 ng/mL in months 1–4 and 6–8 ng/mL thereafter. The target 24-hour trough levels for sirolimus were 15–20 ng/mL in the first 4 months and 10–15 ng/mL thereafter. Patients received 500 mg of methylprednisolone intraoperatively and tapered to a dose of 20 mg of prednisone daily at 1 month. Prednisone doses were further tapered to 5 mg daily at 3 months.

Diagnosis and treatment of rejection

All rejection episodes were proven by biopsy. We typically performed biopsies for clinical indications when the serum creatinine concentration increased by 0.2 mg/dL above baseline. Clinical rejections were generally treated with OKT3 (Orthoclone®; Ortho-Biotech Inc., Bridgewater, NJ) 5 mg/day × 7 d. In addition, we performed surveillance protocol biopsies in all recipients intraoperatively and at 4, 12 and 24 months after transplantation. Subclinical rejection was defined as histologic evidence of rejection using the Banff criteria (10) in the absence of serum creatinine alterations. Subclinical rejection episodes, including borderline changes, were treated with methylprednisolone 500 mg/day intravenously for 3 d followed by a tapered regimen. In addition to the assessment of acute rejection, biopsies were scored for chronic changes at each time point using the Banff 97 schema.

Renal function measurement

Glomerular filtration rate (GFR) was determined at 1, 12 and 24 months after transplantation using a nonisotopic iothalamate clearance test (11). Renal function also was estimated using the abbreviated Modification of Diet in Renal Disease (MDRD) equation (12).

Diagnosis and treatment of metabolic and infectious complications

Patients without diabetes mellitus prior to transplant who received pharmacologic therapy (insulin or oral agents) following transplant were diagnosed with post-transplant diabetes mellitus. Pharmacologic therapy for hyperlipidemia was started at the discretion of the attending transplant physician. Blood pressure results were obtained in the outpatient clinic using arm cuff measurements. Anti-hypertensive therapy was started at the discretion of the attending physician.

Polyoma virus infection was diagnosed by allograft biopsy. Polyoma virus nephropathy was suspected in biopsies showing focal interstitial mononuclear inflammatory cell infiltrates, dilated tubules with necrotic tubular epithelium and homogeneous intranuclear inclusion bodies. In situ hybridization for BK virus DNA was performed on paraffin-embedded sections from biopsies with suspicious light microscopy findings.

Statistical analysis

Survival and acute rejection rates were determined using Kaplan-Meier estimates. Rates between groups at specific time points were compared using Greenwood's formula and the t-test. Binomial and numerical outcomes were described using proportions and means and were compared using chi-square and t-test. A p-value of 0.05 or less was considered significant. Unless otherwise specified, numerical values are expressed as mean ± SD.

Results

Patients

Between April 2001 and January 2004, 165 patients were enrolled in this prospective study. During this period, 84 patients were randomized to the tacrolimus group and 81 to the sirolimus group. No patient was lost to follow-up and the mean follow-up period was 33 months (range 13–47 months). The demographics and selected clinical parameters of the study population are shown in Table 1. There were no significant differences between the groups regarding age, gender, frequency of retransplantation, living donor kidneys, pre-existing diabetes mellitus or sensitization. The majority (81%) of patients in this analysis were Caucasian recipients of living donor kidney transplants. Two patients in the tacrolimus group and 1 in the sirolimus group did not receive the designated study medication due to graft thrombosis within 24 hours of transplantation. These patients were not included in subsequent analyses, leaving 82 patients in the tacrolimus group and 80 patients in the sirolimus group.

Table 1. Patient demographics and selected clinical parameters
 Entire population (n= 165)Tacrolimus group (n= 84)Sirolimus group (n= 81)p-value
  1. PRA = panel reactive antibody.

Mean age (range; years)49 (19–80)48 (19–80)50 (22–73)0.4511
Gender (M:F)89:7644:4045:360.6191
Caucasian race157 (96%)79 (94%)78 (98%)0.2743
Living donor136 (83%)71 (85%)65 (81%)0.5776
Pre-emptive transplant78 (48%)45 (54%)33 (41%)0.1143
Primary transplant135 (82%)66 (79%)69 (86%)0.1143
PRA > 0%12 (7%)8 (10%)4 (5%)0.2661
Pre-existing diabetes mellitus19 (12%)8 (10%)11 (14%)0.3980

Discontinuation and other protocol alterations

Thirty patients (38%) in the sirolimus group and 13 (16%) in the tacrolimus group discontinued the assigned study medication. Discontinuations occurred in the sirolimus and tacrolimus groups at a mean of 245 d (range 23–920 d) and 342 d (range 116–783 d), respectively. Reasons for discontinuation of tacrolimus included: polyoma virus nephropathy (n = 5); allograft interstitial fibrosis (n = 3); post-transplant lymphoproliferative disorder (n = 1); neurotoxicity (n = 2); recurrent hemolytic-uremic syndrome (n = 1) and gastrointestinal intolerance (n = 1). Reasons for discontinuation of sirolimus included: wound healing complications (n = 18); pulmonary complications (n = 4); acute rejection (n = 3); recurrent focal segmental glomerulosclerosis (n = 2); polyoma virus nephropathy (n = 1); severe hypertriglyceridemia (n = 1) and thrombocytopenia (n = 1). Unless otherwise stated, those patients discontinuing the study medication were included in the intent-to-treat analysis for all outcomes.

At 1 month the mean trough level of tacrolimus was 12.7 ± 3.9 ng/dL and of sirolimus was 17.5 ± 5.2 ng/dL. At 1 year, the mean trough level of tacrolimus was 7.9 ± 1.3 ng/dL and of sirolimus was 13.9 ± 3.2 ng/dL. The mean mycophenolic acid level at 1 year was 2.4 ± 1.4 μg/mL in the tacrolimus group and 3.0 ± 2.6 μg/mL in the sirolimus group (p = 0.15).

Patient survival

One-year Kaplan-Meier patient survival rates were not significantly different between the tacrolimus group (96%) and the sirolimus group (98%) (p = 0.42; Figure 1A). Seven patients died in the tacrolimus group, 5 with a functioning allograft. The causes of death were: myocardial infarction (n = 1); thrombotic microangiopathy and mesenteric ischemia (n = 1); gastric adenocarcinoma (n = 1); endocarditis (n = 1); bowel perforation (n = 1); post-transplant lymphoproliferative disorder (n = 1) and unknown (n = 1). These deaths occurred between 7 and 20 months following transplant. In the sirolimus group, 5 patients died, all with functioning allografts. The causes of death were: pulmonary embolism (n = 1); sudden cardiac death (n = 1); sepsis (n = 1); intracranial malignancy (n = 1) and unknown (n = 1). These deaths occurred between 3 weeks and 57 months following transplant.

Figure 1.

Kaplan-Meier estimates of patient (A) and graft (B) survival for recipients in the tacrolimus and sirolimus groups. Panel A shows that the 1-year Kaplan-Meier patient survival rates were not significantly different between the tacrolimus group (96%) and the sirolimus group (98%) (p = 0.4199) using the logrank test and panel B shows that the Kaplan-Meier estimate of rate of 1-year survival with graft for patients receiving the study medication in the tacrolimus group was 92% and 94% in the sirolimus group (p = 0.9481) using the logrank test.

Graft survival

The Kaplan-Meier estimate of 1-year graft survival was 92% in the tacrolimus group and 94% in the sirolimus group (p = 0.95; Figure 1B). The death-censored 1-year graft survival rates were 96% in the tacrolimus group and 96% in the sirolimus group (p = 0.96). Causes of graft loss in the tacrolimus group included: death with function (n = 5); severe acute cellular and humoral rejection (n = 1); recurrent hemolytic-uremic syndrome (n = 1); cytomegalovirus (CMV) infection (n = 1); recurrent IgA nephropathy (n = 1); and recurrent membranoproliferative glomerulonephritis (n = 1). Causes of graft loss in the sirolimus group included: death with function (n = 5); CAN (n = 2); primary nonfunction (n = 1); severe pyelonephritis (n = 1) and acute humoral rejection (n = 1).

Acute rejection

Twelve patients (14%) in the tacrolimus group experienced a total of 14 rejection episodes—8 clinical and 6 subclinical. Fifteen patients (19%) in the sirolimus group experienced a total of 16 rejection episodes—10 clinical and 6 subclinical (Table 2). Using Kaplan-Meier analysis, there was no difference in the rejection rates during the first year between the groups (p = 0.51). Sixteen of the 30 rejection episodes occurred within 6 months of transplantation.

Table 2. Banff 97 classification of acute rejection episodes according to study group
 Tacrolimus group (N = 82)Sirolimus group (N = 80)
  1. AHR = acute humoral rejection.

 N = 8N = 10
Clinical rejectionBorderline1Borderline1
Banff IA3Banff IA1
Banff IIA2Banff IB3
AHR2Banff IIA3
Banff IIA + AHR1
AHR1
N = 6N = 6
Subclinical rejectionBorderline4Borderline2
Banff IB2Banff IA2
Banff IB1
AHR1

Renal function

At 1 month, 79 tacrolimus-group patients and 70 sirolimus-group patients underwent renal function determinations. Iothalamate clearance at 1 month was 67 ± 18 mL/min in the sirolimus group versus 58 ± 17 mL/min in the tacrolimus group (p = 0.002; Figure 2A). When corrected for body surface area, iothalamate clearance remained significantly higher in the sirolimus group [62 ± 18 mL/min/1.73 m2 vs. 56 ± 16mL/min/1.73 m2 (p = 0.03)]. The abbreviated MDRD equation estimate of creatinine clearance was 55 ± 17 mL/min in the sirolimus group versus 47 ± 14 mL/min in the tacrolimus group (p = 0.004).

Figure 2.

(A) Glomerular filtration rate measured by iothalamate clearance at 1 month, 12 months and 24 months following kidney transplantation (intent-to-treat) and the change in iothalamate clearance from 1 month to 12 months in the tacrolimus (B) and sirolimus (C) groups. Comparisons between renal function at 1 and 12 months were made using the 2-tailed paired t-test.

At 1 year, 65 tacrolimus-group patients and 64 sirolimus-group patients underwent renal function determinations (Figure 2A). At this time point, there was no difference in the uncorrected iothalamate clearance measurements between the groups—61 ± 19 mL/min in the tacrolimus group versus 63 ± 18 mL/min in the sirolimus group (p = 0.57). After correcting for body surface area, the iothalamate clearance measurements in the groups were also similar—55 ± 17 mL/min/1.73 m2 in the tacrolimus group versus 56 ± 16 mL/min/1.73 m2 in the sirolimus group (p = 0.86). The mean iothalamate clearance in patients maintained on the initial study drug for 12 months was 54 ± 17 mL/min/1.73 m2 in the tacrolimus group (n = 45) and 57 ± 17 mL/min/1.73 m2 in the sirolimus group (n = 40; p = 0.37). The abbreviated MDRD estimates of creatinine clearance were 49 ± 14 mL/min in the tacrolimus group and 52 ± 14 mL/min in the sirolimus group (p = 0.14).

At 2 years, 45 patients in each group underwent determination of renal function (Figure 2A). Using an intent-to-treat analysis at this time point, there was no difference in the mean uncorrected iothalamate clearance measurements between the groups—61 ± 17 mL/min in the tacrolimus group versus 61 ± 19 mL/min in the sirolimus group (p = 0.84; Figure 2A). After correcting for body surface area, iothalamate clearance measurements in the groups remained similar—55 ± 14 mL/min/1.73 m2 in the tacrolimus group versus 55 ± 16 mL/min/1.73 m2 in the sirolimus group (p = 0.91). When analyzing those patients actually maintained on the initial randomized study drug for 24 months, there was no difference in iothalamate clearance [56 ± 14 mL/min/1.73 m2 in the tacrolimus group (n = 41) vs. 57 ± 17 mL/min/1.73 m2 in the sirolimus group (n = 31; p = 0.70)]. However, the abbreviated MDRD estimate of creatinine clearance was higher at 2 years in the sirolimus group compared to the sirolimus group—68 ± 20 compared to 55 ± 17 mL/min (p = 0.01).

For those patients in the tacrolimus group completing iothalamate clearance measurements at both 1 and 12 months (n = 64; Figure 2B), the mean change in GFR was −0.3 mL/min/1.73 m2 (range −45 to 35 mL/min/1.73 m2). For those patients in the sirolimus group completing iothalamate clearance measurements at both 1 and 12 months (n = 60; Figure 2C), the mean change in GFR was −5.7 mL/min/1.73 m2 (range −66 to 33 mL/min/1.73 m2). The mean iothalamate clearance measurements were similar in the tacrolimus group at 1 month and 1 year for patients completing both measurements (56 ± 16 mL/min/1.73 m2 vs. 56 ± 16 mL/min/1.73 m2; p = 0.87) when compared using a 2-tailed paired t-test. However, patients completing both measurements in the sirolimus group experienced a significant decline in renal function from 1 to 12 months. In this group, the mean iothalamate clearance was 62 ± 18 mL/min/1.73 m2 at 1 month compared to 56 ± 16 mL/min/1.73 m2 at 1 year (p = 0.01) when compared using a 2-tailed paired t-test. At 1 year, the mean serum creatinine levels in the tacrolimus and sirolimus groups were 1.6 ± 0.5 mg/dL and 1.5 ± 0.5 mg/dL, respectively (p = 0.17).

Effect of immunosuppressive regimen on allograft histology

At time zero, there were no differences in the incidence of chronic histologic lesions in the donor kidneys. In total, 7% of donor kidneys showed ci1 lesions and 22% showed cv1 lesions (data not shown). At 1 year, 51 patients in the tacrolimus group and 46 patients in the sirolimus group underwent protocol surveillance allograft biopsy (Figure 3). While there were no differences in interstitial fibrosis, tubular atrophy and glomerulopathy, there was a higher incidence of chronic vascular changes in the tacrolimus group, i.e. 43% of the tacrolimus group had at least a cv1 lesion while only 26% of the sirolimus group had a cv1 or greater lesion (p = 0.03). This change was significant and the others nonsignificant when the biopsies were analyzed for patients remaining on the initial study drug (data not shown).

Figure 3.

Chronic Banff scores of protocol surveillance biopsies performed at 12 months following transplantation in the tacrolimus (n = 51) and sirolimus groups (n = 46).

Hyperlipidemia

The mean serum cholesterol level before transplant was 183 ± 44 mg/dL in the tacrolimus group compared to 179 ± 44 mg/dL in the sirolimus group (p = 0.60, Figure 4). At 12 months, the mean serum cholesterol level in the tacrolimus group was 200 ± 33 mg/dL compared to 219 ± 53 mg/dL in the sirolimus group (p = 0.02). Before transplant, the mean serum triglyceride level in the tacrolimus and sirolimus groups was 154 ± 70 mg/dL and 180 ± 76 mg/dL, respectively (p = 0.02). At 12 months, the mean serum triglyceride level in the tacrolimus group was 174 ± 102 mg/dL compared to 246 ± 131 in the sirolimus group (p = 0.001). At 12 months, 36% of the patients in the tacrolimus arm were taking lipid-lowering medications, compared to 78% of the patients in the sirolimus arm (p < 0.0001).

Figure 4.

Mean serum cholesterol (A) and triglyceride (B) levels before transplant (pre-Tx), at 4 months and at 12 months following kidney transplantation. Comparison between the two groups was made using the 2-tailed t-test.

Hypertension

At 1 month following transplantation, the mean systolic blood pressure in the tacrolimus and sirolimus groups was 130 ± 20 mmHg and 137 ± 19 mmHg, respectively (p = 0.03; Figure 5). At 12 months, the mean systolic blood pressure was 135 ± 22 mmHg in the tacrolimus group and 137 ± 15 mmHg in the sirolimus group (p = 0.56). At 1 month following transplantation, the mean diastolic blood pressure in the tacrolimus and sirolimus groups was 73 ± 11 mmHg and 74 ± 11 mmHg, respectively (p = 0.69). At 12 months, the mean diastolic blood pressure was 77 ± 14 mmHg in the tacrolimus group and 77 ± 10 mmHg in the sirolimus group (p = 0.75). In the tacrolimus group, at 1 month, 26% of the patients were taking no anti-hypertensive medications, 44% were taking one anti-hypertensive medication and 30% were taking two or more anti-hypertensive medications. In the tacrolimus group, at 12 months, 34% of the patients were taking no anti-hypertensive medications, 47% were taking one anti-hypertensive medication, and 19% were taking two or more anti-hypertensive medications. In the sirolimus group, at 12 months, 21% of the patients were taking no anti-hypertensive medications (p = 0.15), 54% were taking one anti-hypertensive medication (p = 0.48), and 25% were taking 2 or more anti-hypertensive medications (p = 0.48).

Figure 5.

Mean systolic (A) and diastolic (B) blood pressure measurements at 1 month and 12 months following kidney transplantation. Comparison between the two groups was made using the 2-tailed t-test.

Other significant events

The mean white blood cell count at 1 month in the tacrolimus and sirolimus groups was 7600 ± 3200/m3 and 4400 ± 1800/m3, respectively (p = 0.0001). At 1 year, the mean white blood cell count was 7300 ± 2900/m3 in the tacrolimus group and 6100 ± 2300/m3 in the sirolimus group (p = 0.29). At 1 month, the mean hemoglobin level in the tacrolimus and sirolimus groups was 10.6 ± 1.2 g/dL and 9.9 ± 1.2 g/dL, respectively (p = 0.0001). The mean hemoglobin level in the tacrolimus and sirolimus groups at 1 year was 12.8 ± 1.7 g/dL and 12.5 ± 1.5 g/dL, respectively (p = 0.31).

Three cases of post-transplant lymphoproliferative disease were diagnosed during the study period (two in the sirolimus group and one in the tacrolimus group). There were seven cases of polyoma virus infection in the tacrolimus group and four in the sirolimus group (p = 0.37). Ten cases of systemic CMV infection were diagnosed in the tacrolimus group (12%) compared to two cases in the sirolimus group (3%; p = 0.02). All of these infections developed in CMV seronegative recipients who received kidneys from CMV seropositive donors. There were eight cases of post-transplant diabetes mellitus in the tacrolimus group (10%) and six in the sirolimus group (7.5%) (p = 0.78). The increased incidence of wound healing complications seen in the sirolimus group in the study has been described previously (9).

Discussion

The results of this study extend previous reports demonstrating that excellent patient and allograft survival with low rates of acute rejection can be achieved without CNIs using a sirolimus-based immunosuppressive regimen (13–18). The clinical acute rejection rates (10% in the sirolimus group and 13% in the tacrolimus group) are slightly higher than, but comparable to the 6% rate described by Flechner et al. in a CNI-free protocol using sirolimus, MMF and prednisone with basiliximab induction (14). The current study also demonstrates that subclinical rejection rates using the two regimens are similar.

Many previous studies of complete calcineurin inhibitor avoidance have used cyclosporine as the comparative drug (14–18). The fact that the current study used tacrolimus may be an important reason why no difference in renal function was found at 1 and 2 years. At 2 years, the uncorrected iothalamate clearance was 61 mL/min in both groups and 55 mL/min when corrected for body surface area. These iothalamate clearances are comparable to the mean corrected iothalamate clearance of 60.6 mL/min/1.73 m2 at 2 years in sirolimus-treated patients reported by Flechner et al. (17). In both instances, actual scatterplot data suggest a significant spread of iothalamate clearances with many values below 50 in sirolimus-treated patients in both studies. In the study by Flechner et al. (17), the mean iothalamate clearance for cyclosporine-treated patients at 2 years was 49.2 mL/min/1.73 m2. Thus, one might argue that the major difference between the results of the two studies is that the tacrolimus-treated patients had better renal function at 2 years compared to cyclosporine-treated patients. Several reports have suggested that tacrolimus, at the lower target levels now commonly used clinically (mean trough level was 7.9 ± 1.3 ng/dL in this study), may have minimal nephrotoxicity and may be less nephrotoxic than cyclosporine (19–21). In addition to the use of tacrolimus, the current study differs from previous reports in that sirolimus-treated patients experienced a slight, but significant decline in renal function between 1 month and 1 year after transplant. While the cause of this decrease is unclear, it likely contributed to the lack of a significant difference in renal function between the two treatment groups.

The current study also involved primarily living donor kidneys which may be less vulnerable to the nephrotoxic effects of CNI therapy compared to deceased donor grafts. Indeed, in a European multi-center trial of cyclosporine withdrawal, the kidneys with the poorest renal function appeared to show the greatest increase in function after CNI withdrawal (22). In addition, the current study may be underpowered for the endpoint given the large ranges in renal function, the use of unpaired comparisons and the relatively short follow-up period.

Nephrotoxicity is only one of many factors that may affect renal function in the first year after transplantation and it is important to control for as many of these as possible to study nephrotoxocity longitudinally. For example, previous studies have suggested that acute rejection is one of the major factors affecting subsequent renal function (4). In the current study, the acute rejection rates were similar in both groups. In contrast, the acute clinical rejection rate in a previously-reported CNI-free protocol was 17% in the cyclosporine arm compared to 6% in the sirolimus group (14). This lower rejection rate in the sirolimus group might partially explain the improvement in renal function observed compared to the cyclosporine group. Subclinical rejection has been shown to be higher in cyclosporine-treated patients—20–30% in some series (23–25). The impact of subclinical rejection is controversial, but it has been linked to an increase in interstitial fibrosis and decreased renal function. This lack of detail regarding differences in rejection rates highlights the pitfalls of performing studies designed to assess nephrotoxocity when many factors may affect renal function.

The use of iothalamate clearance to determine renal function is also unique to this study. In general, iothalamate clearance is believed to be a more accurate test of renal function than estimated formulae. Previous studies have used calculated GFR levels based on serum creatinine levels. While discrepancies between calculated GFR and iothalamate clearances are not uncommon, in the current study another possibility exists. Muscle mass may have been affected by one agent but not the other, which in turn could influence creatinine levels, and consequently MDRD GFR estimation, independent of an effect on GFR. Indirect evidence that this is a possibility comes from studies indicating that sirolimus may decrease muscle protein synthesis (26–28).

At 1 year, there was no difference in interstitial fibrosis, glomerulopathy or tubular atrophy between the two groups. However, there was a significant increase in mild chronic Banff vascular lesions. The finding that this difference did not lead to differences in renal function agrees with our previous morphometric study demonstrating that the primary histologic changes in renal allografts that correlate with function are the degree of interstitial fibrosis and glomerulopathy (29). The long-term implications of an increase in vascular disease and the possibility of progression to a more severe lesion suggest the need for longer follow-up of these patients. Indeed, the presence of vascular lesions raises concern that a subclinical nephrotoxicity of tacrolimus which may take longer to manifest as graft dysfunction may yet emerge in these patients.

Another major issue in the development of new immunosuppressive protocols is tolerability. It appears that sirolimus-based and tacrolimus-based regimens each have unique complications that will require further refinements in patient management. Wound complications and hyperlipidemia were encountered as in previous studies and limit the attractiveness of sirolimus-based protocols (9,30–32). More patients in the sirolimus arm required pharmacologic treatment for dyslipidemia at 1 year. The incidence of wound complications with sirolimus using the current protocol was high and was the primary cause of discontinuing the drug. Protocol refinements such as delayed introduction of sirolimus or a lower target dose might allow for improved wound healing while still achieving CNI-free immunosuppression. Leukopenia requiring adjustment of MMF doses was seen in both groups. However, while the mean white blood cell count was significantly lower in the sirolimus group 1 month after transplant, there was no difference in levels at 1 year. Finally, CNIs have been associated with an increase in hypertension following renal transplantation, but the current study showed no difference in blood pressure in patients in the CNI-free group.

We conclude that a CNI-free regimen using sirolimus-MMF-prednisone leads to similar excellent patient and graft survival with low acute rejection rates in the first year after transplantation when compared to a regimen of tacrolimus-MMF-prednisone. Renal function was not significantly different in the two groups, but the CNI-free group had fewer chronic vascular changes on 1 year protocol biopsies—a finding that merits further study. The current trial also demonstrates that detailed analyses such as GFR measurements, protocol biopsies to assess subclinical rejection and surveillance for polyoma virus are important to avoid misinterpretation of the effect of CNI-sparing regimens on renal allografts. The long-term impact of total avoidance of CNIs on renal function remains unclear and longer follow-up is needed to determine if total avoidance of CNIs will lead to improvements in long-term graft survival.

Acknowledgments

The authors wish to thank Cynthia Groettum for her assistance with patient enrollment and data collection for this study. This study was supported in part by research contracts from Wyeth Research, Philadelphia, PA, Genzyme Corporation, Cambridge, MA, and Roche Laboratories Inc., Nutley, NJ.

Ancillary