Conversion From Cyclosporine to Tacrolimus in Patients at Risk for Chronic Renal Allograft Failure: 60-Month Results of the CRAF Study

Abstract

Background. 

This study compared the long-term effects of switching from cyclosporine to tacrolimus on the incidence, progression, and severity of chronic renal allograft failure in patients with elevated serum creatinine levels.

Methods. 

Patients were assigned randomly (2:1) to switch to tacrolimus or remain on cyclosporine. Tacrolimus was initiated at 1/50th of the cyclosporine dose or 0.15 mg/kg/day, whichever dose was lower, to maintain trough concentrations between 5 and 15 ng/mL. Cyclosporine doses were adjusted to achieve trough concentrations between 100 and 300 ng/mL.

Results. 

At 60 months, the median change from baseline in serum creatinine was −0.2 mg/dL in the tacrolimus group and 0.3 mg/dL in the cyclosporine group (P=0.003). Median change in estimated creatinine clearance was 1.2 mL/min in the tacrolimus group and −4.1 mL/min in the cyclosporine group (P=0.019). The incidence of new-onset diabetes, hyperglycemia, hypertension, lymphoma, and malignancies was generally low and comparable between groups. Fewer patients in the tacrolimus group than in the cyclosporine group developed new cardiac conditions (11% vs. 28%, P=0.004), had low-density lipoprotein (LDL) cholesterol values more than 130 mg/dL (29% vs. 57%, P=0.002), or developed hyperlipidemia (24% vs. 67%, P=0.046) during the 60-month follow-up period. Despite these changes, patient and graft survival were similar for both groups.

Conclusion. 

Switching from cyclosporine to tacrolimus resulted in improved renal function and a reduction in the occurrence of new-onset cardiac conditions and hyperlipidemia, with no increase in the incidence of new-onset diabetes or new-onset hyperglycemia. However, after 5 years there was no impact on patient or graft survival.

Over the past two decades, both short-term allograft survival and acute rejection rates have markedly improved, as new immunosuppression regimens have been introduced (^{1, 2}). However, similar improvements in long-term outcomes have not been observed. Apart from death with a functioning graft, the leading cause of graft loss after the first year of transplantation is a clinicopathologic entity commonly referred to as chronic allograft nephropathy (CAN) characterized by progressive renal dysfunction accompanied by variable degrees of interstitial fibrosis, tubular atrophy, and nonspecific vascular and glomerular sclerosis (³).

Although commonly thought of as a disease entity, CAN is actually nonspecific parenchymal scarring of the renal allograft that ultimately leads to chronic renal allograft failure (CRAF). For this reason, the recent Banff 2005 schema has attempted to reverse the misconception by advocating the more descriptive terminology of “interstitial fibrosis/tubular atrophy” (⁴). Registry data have been frequently used to identify some associated risk factors, including immunologic factors such as poor human leukocyte antigen matching, and nonimmunologic factors such as donor age, tissue quality, gender, organ preservation, and recipient race (^5–7). Additionally, episodes of acute rejection, particularly late rejection episodes occurring more than 3 months after transplantation (^{8, 9}), as well as coexisting conditions such as hypertension (¹⁰), infection (¹¹), and hyperlipidemia (^{8, 12}) all increase the risk of CRAF.

Calcineurin inhibitors, including tacrolimus and cyclosporine, are effective in improving short-term outcomes (i.e., acute rejection) after transplantation; however, these drugs can be nephrotoxic when used long-term (¹³). In a well-controlled trial of cyclosporine versus tacrolimus, 72.3% and 62.0%, respectively, of biopsies exhibited chronic injury changes at 2 years (¹⁴). Moreover, features of chronic cyclosporine nephrotoxicity tend to become progressively more dominant in protocol biopsies, and by 10 years, striped interstitial fibrosis and arteriolar hyalinosis with or without tubular calcification are present almost universally (¹⁵). Nonetheless, studies attempting to reduce or eliminate calcineurin inhibitors from immunosuppressant regimens have not consistently demonstrated improvements in renal function and have frequently been associated with higher acute rejection rates (^16–18).

Compared with cyclosporine, immunosuppression with tacrolimus has been shown to result in significantly fewer biopsy-confirmed episodes of acute rejection (¹⁹). Additionally, long-term use of tacrolimus is associated with a lower incidence of treatment failure (graft loss or drug discontinuation), better renal function, less dyslipidemia, and less hypertension compared with long-term use of cyclosporine (²⁰). Notably, posttransplant renal function has been correlated with long-term graft survival. Analyses from a large database of renal transplant recipients have clearly shown that serum creatinine more than 1.5 mg/dL at 6 months and 1 year after transplant correlates with lower long-term graft survival (²¹). In addition, higher serum creatinine levels at 1 year have been associated with an increased risk of cardiovascular mortality (²²). In this study, we enrolled patients who were at high risk for the development of CRAF based on elevated serum creatinine levels at study entry, and we hypothesized that switching from cyclosporine to tacrolimus might reduce the incidence of CRAF without increasing graft failure because of other risk factors.

METHODS

This open-label, randomized, parallel-group, multicenter, secondary-intervention trial was designed to compare the long-term effects of switching to tacrolimus versus cyclosporine continuation on the incidence, progression, and severity of CRAF in patients who were at risk because of declining renal function. The study was conducted at 28 investigative sites in the United States over a 5-year period. Details of the study design and results of analyses conducted at 2 years have been reported previously (²³). Methods relevant to the 5-year follow-up period are briefly described herein. The study protocol was approved by an Institutional Review Board at each study center. Study procedures were performed in accordance with the Helsinki Declaration. The first patient was enrolled on December 5, 1998, and enrollment ended on September 30, 2000. The cut-off date for inclusion in this report was July 15, 2005.

Selection and Description of Participants

Eligible patients were men and women, deceased or living donor renal transplant recipients more than 12 years of age, at least 3 months after transplant, who had received a cyclosporine-based immunosuppression regimen since transplantation. Patients were required to have had a renal biopsy within the previous 6 months (>3 months after transplant) to exclude those with moderate or severe acute rejection (grade ≥IIB Banff 93–95 (²⁴) or grade ≥IIA Banff 97 [²⁵]). Patients were determined to be at risk for CRAF based on a serum creatinine level of ≥2.0 mg/dL 3 months after transplant or later in male patients, ≥1.7 mg/dL 3 months after transplant or later in female patients, or an increase in serum creatinine of more than 30% over postdischarge nadir. Patients who were excluded from participation in the study included recipients with solid organ transplants other than the kidney; those who were dialysis dependent; those who had recurrent primary renal disease or de novo renal disease; those who had a Cockroft-Gault estimated creatinine clearance less than 25 mL/min; those who had changed their maintenance immunosuppressant therapy within 3 months of randomization; those who required antilymphocyte therapy to treat rejection found on baseline biopsy or had been treated with antilymphocyte therapy since the baseline biopsy; those who received an investigational immunosuppressant within the previous 3 months; those who were pregnant or lactating; those who were known carriers of the human immunodeficiency virus; or those who were hypersensitive to tacrolimus or any of its excipients.

Treatment Plan

As reported previously (²³), all patients were receiving cyclosporine-based immunosuppression at the time of randomization. Patients were assigned randomly (2:1) to be switched to tacrolimus or to remain on cyclosporine. Randomization took the following risk factors into account: cadaveric versus living donor, age of donor ≥55 years, treatment for acute rejection episode since transplant, current diabetes mellitus requiring treatment, and acute or chronic rejection on protocol biopsy. Dose changes to a patient’s adjunctive maintenance immunosuppressant regimen, which could have included azathioprine, mycophenolate mofetil (MMF), and/or corticosteroids, were not permitted during the study unless there was a medically justified reason for such a change. Tacrolimus was initiated at 1/50th of the cyclosporine dose or 0.15 mg/kg/day, whichever was less, and was later adjusted to maintain trough whole blood concentrations between 5 and 15 ng/mL. Cyclosporine doses were adjusted, if necessary, to achieve trough whole blood concentrations between 100 and 300 ng/mL. Acute rejection episodes were treated according to the standard medical practices at the individual study center. Cyclosporine and tacrolimus doses were adjusted as clinically warranted.

Clinical evaluations were conducted at 6 months, 12 months, and annually thereafter up to 60 months. Renal function was monitored throughout the study by evaluating serum creatinine and creatinine clearance. All episodes of renal dysfunction were evaluated with renal biopsies to assess the possible occurrence of acute rejection. Renal dysfunction was defined as an increase in serum creatinine of ≥0.5 mg/dL or a doubling of serum creatinine compared with baseline or postrandomization nadir, whichever was less. Biopsies were performed either before or within 24 hours of initiating treatment for rejection, and were analyzed by pathologists who were blinded to treatment allocation using the Banff working classification of renal allograft pathology (²⁵). Safety parameters were monitored throughout the study and included new-onset hypertension (systolic blood pressure ≥140 mm Hg or diastolic blood pressure ≥90 mm Hg), hyperlipidemia (serum cholesterol ≥200 mg/dL or LDL cholesterol ≥130 mg/dL), hyperglycemia (fasting glucose >140 mg/dL on two occasions but not treated with insulin for >30 days), diabetes mellitus (fasting glucose >140 mg/dL requiring therapy with insulin for more than 30 consecutive days), and infections. Other significant adverse events, including malignancies, lymphoma, and lymphoproliferative diseases were also monitored throughout the study.

Study Endpoints

Efficacy endpoints included graft survival, patient survival, renal function, episodes of acute rejection, graft loss, CAN, and death with a functioning graft. Safety endpoints included incidence of new-onset hyperlipidemia, hypertension requiring therapy, cardiac conditions (defined by the investigator; dysrhythmias, coronary syndromes, cardiac failure, left ventricular hypertrophy), infections, insulin-dependent diabetes mellitus, and discontinuation from the randomized immunosuppressant regimen.

Statistical Analysis

Efficacy and safety evaluations were performed for all randomized patients who received at least one dose of the assigned study drug regimen (intent-to-treat population) with available follow-up information. Patients were followed up for 60 months or until death. Once graft loss occurred, only patient survival data were collected. Mean values were compared using the t test. Median values were compared using the Kruskal-Wallis test. Comparisons of binary outcomes were made using Pearson’s chi-square test or Fisher’s exact test, where appropriate. Patient and graft survival were compared between groups using Kaplan-Meier product-limit estimates and the log-rank test. Cumulative incidence refers to the Kaplan-Meier probability of the event at the end of 60 months, whereas the actual rate refers to the total number of events through 60 months divided by the total number of patients.

RESULTS

Patient Disposition

A total of 197 patients were assigned randomly to treatment, 131 were assigned to be switched to tacrolimus, and 66 were assigned to remain on cyclosporine (Fig. 1). A total of 186 patients were followed throughout the 60-month study, 126 patients in the tacrolimus group and 60 patients in the cyclosporine group. A total of 66 patients, 45 in the tacrolimus group and 21 in the cyclosporine group completed the study.

FIGURE 1.: Patient accounting. *Includes patients who withdrew, were lost to follow-up, or for whom no additional follow-up information was available because of administrative reasons.

Patient Demographic and Baseline Characteristics

Demographic and baseline characteristics were reported previously (²³), and were similar between the two treatment groups (Table 1). Briefly, the majority of patients were white males with an average age of 45±13 vs. 49±14 years (P=0.09) in the tacrolimus and cyclosporine groups, respectively. Although the prevalence of coexisting conditions was similar for both groups, fewer patients in the tacrolimus group had hyperglycemia at baseline compared with patients in the cyclosporine group (10% vs. 22%, respectively; P=0.02). No patient had received tacrolimus before the study.

TABLE 1: Patient demographic and baseline characteristics

Histocompatibility, transplant characteristics, donor age, and proportion of patients receiving each concomitant immunosuppression regimen were similar for both treatment groups at baseline (Table 1). Moreover, the median urine protein-to-urine creatinine ratio was similar between both treatment groups. The majority of patients had received transplants from deceased donors (82 [65%] tacrolimus, 41 [68%] cyclosporine). At the baseline biopsy, CAN was reported in 114 (90%) patients in the tacrolimus group and 54 (90%) patients in the cyclosporine group; a similar severity between treatment groups. Of those patients with CAN at baseline treated with tacrolimus, 60 (47.6%) were classified as mild, 46 (36.5%) as moderate, and 8 (6.3%) as severe. A comparable number of cyclosporine-treated patients had experienced chronic rejection of similar intensity; 27 patients (45.0%) were classified mild, 23 (38.3%) as moderate, and 4 as (6.7%) severe. Moreover, a history of biopsy-confirmed acute rejection (Banff 97 grade IA or IB) was reported in 21 (17%) patients converting to tacrolimus and 4 (7%) patients remaining on cyclosporine (P=0.06).

Study Drug Exposure

Tacrolimus and cyclosporine doses were stable throughout the study. At 60 months, the median daily tacrolimus dose was 0.05 mg/kg and the median daily cyclosporine dose was 2.61 mg/kg. Median trough levels were within the target range for tacrolimus (5–15 ng/mL) and cyclosporine (100–300 ng/mL) throughout the follow-up period (Table 2). However, at 60 months, 31% of patients in the tacrolimus group and 30% of patients in the cyclosporine group had trough blood levels below the target range, and 2% of patients in the tacrolimus group and 10% of patients in the cyclosporine group had trough blood levels above the target range. The mean and median MMF concentrations decreased from baseline through 60 months, with no statistically significant differences between treatment groups at any time point. A total of 45% of patients receiving tacrolimus were on MMF at baseline compared with 48% of cyclosporine-treated patients. At 60 months, 51% of tacrolimus-treated patients were on MMF versus 42% of cyclosporine-treated patients.

TABLE 2: Tacrolimus and cyclosporine median trough levels (ng/mL) through 60 months

Efficacy Outcomes

Renal Function

At 60 months, the mean and median change from baseline in all measures of renal function was statistically significantly more favorable in the tacrolimus group compared with the cyclosporine group for patients with a functioning graft (Table 3). Data were not included in the analysis if they were obtained outside of a 70-day period before or after 60 months. The median serum creatinine level was lower in the tacrolimus group than in the cyclosporine group throughout the 60-month follow-up period (Fig. 2A). At 60 months, the median serum creatinine was 2.0 mg/dL in the tacrolimus group and 2.7 mg/dL in the cyclosporine group (P=0.007); the median change from baseline in serum creatinine was also statistically significant (P=0.003). The mean serum creatinine level was also statistically significantly lower in the tacrolimus group at month 12 (P=0.38), month 48 (P=0.084), and month 60 (P=0.027). Significantly fewer patients in the tacrolimus group (44%) than in the cyclosporine group (64%, P=0.02) had serum creatinine values above 3.0 mg/dL at any time during follow-up.

TABLE 3: Renal function test results at month 60

FIGURE 2.: Median serum creatinine levels (A) and median estimated creatinine clearance (B) over time for patients with graft function in the tacrolimus and CsA treatment arms. (A) P values based on Kruskal-Wallis test. Number of patients in the tacrolimus group at baseline (n=119), month 6 (n=111), month 12 (n=100), month 24 (n=74), month 36 (n=56), month 48 (n=48), and month 60 (n=43). Number of patients in the cyclosporine treatment group at baseline (n=55), month 6 (n=49), month 12 (n=48), month 24 (n=32), month 36 (n=27), month 48 (n=25), and month 60 (n=19). (B). P values based on Kruskal-Wallis test. Number of patients in the tacrolimus group at baseline (n=114), month 6 (n=106), month 12 (n=96), month 24 (n=72), month 36 (n=56), month 48 (n=47), and month 60 (n=42). Number of patients in the cyclosporine treatment group at baseline (n=53), month 6 (n=48), month 12 (n=46), month 24 (n=31), month 36 (n=27), month 48 (n=24), and month 60 (n=19).

To incorporate the effects of patients who discontinued treatment, two additional analyses were performed; a modified last observation carried-forward analysis and a comparison of the slope of the inverse serum creatinine between treatment groups using a t test. For the modified last observation carried-forward analysis, the last available serum creatinine value was used for patients without a value at 60 months. For patients with a graft loss, a value of 8.0 mg/dL was assigned, and serum creatinine values for all patients were capped at 8.0 mg/dL. Using the modified last observation carried-forward analysis, the median serum creatinine levels through 60 months (tacrolimus=3.3 mg/dL; cyclosporine=3.5 mg/dL; P=0.42) were lower with tacrolimus treatment and not statistically significant. A comparison of the slope of inverse serum creatinine revealed that fewer patients in the tacrolimus group (19%) than in the cyclosporine group (32%, P=0.63 for slope at 60 months) experienced a decrease from baseline in inverse serum creatinine of more than 40% at any time over follow-up.

The median estimated creatinine clearance was higher in the tacrolimus group than in the cyclosporine group throughout the 60-month follow-up period (Fig. 2B). At 60 months, the median estimated creatinine clearance was 37.15 mg/mL in the tacrolimus group and 32.10 mg/mL in the cyclosporine group (P=0.048). The change from baseline in the median estimated creatinine clearance was 1.2 in the tacrolimus group and −4.1 in the cyclosporine group (P=0.019). Significantly fewer patients in the tacrolimus group (73%) than in the cyclosporine group (89%, P=0.026) had an estimated creatinine clearance level lower than 35 mL/min at any time over follow-up. However, no similar statistical difference was observed between groups in the proportion of patients with estimated creatinine clearance values lower than 45 mL/min at any time over the 60 months (92% vs. 93%, P=1.0).

Rejection

No significant differences were observed between the two treatment groups in the incidence of acute or chronic rejection through 60 months. At least one acute rejection episode was confirmed by biopsy in 11 (9%) patients in the tacrolimus group and 5 (8%) patients in the cyclosporine group (P=0.93). Thirty-three (26%) patients in the tacrolimus group and 15 (25%) patients in the cyclosporine group had chronic rejection confirmed by biopsy (P=0.86). Among patients who had a biopsy at 60 months, the severity of biopsy-confirmed chronic rejection had increased compared with baseline (including patients with no chronic rejection at baseline) in 9 of 25 (36%) patients in the tacrolimus group and 7 of 11 (64%) patients in the cyclosporine group (P=0.16).

Patient and Graft Survival

At month 60, kidney graft loss (including patients who died with a functioning graft) was similar between treatment groups and had occurred in 69 (55%) patients in the tacrolimus group and 34 (57%) patients in the cyclosporine group (P=0.897, log-rank test; Fig. 3). The most common reason for kidney graft loss was rejection. In the tacrolimus group, there was no significant difference in graft survival at 60 months between patients with mild chronic rejection at baseline and those with moderate/severe chronic rejection at baseline (42% vs. 27%; P=0.362), whereas, among patients in the cyclosporine group, the probability of graft survival was significantly higher for patients with mild chronic rejection than for those with moderate/severe chronic rejection at baseline (58% vs. 15%; P=0.001). Twenty-three (18%) patients in the tacrolimus group and 13 (22%) patients in the cyclosporine group died during the study (P=0.60), of whom 11 and 5 patients, respectively, had experienced graft loss before death. Death with a functioning graft was considered a graft loss.

FIGURE 3.: Kaplan-Meier probability of graft survival vs. time. P values based on log-rank test (P=0.897). Death with a functioning graft was considered to be a kidney graft loss. Cumulative number of deaths or graft losses for the tacrolimus group at month 6 (n=7), month 12 (n=19), month 24 (n=37), month 36 (n=54), month 48 (n=61), and month 60 (n=69). Cumulative number of deaths or graft losses for the CsA treatment group at month 6 (n=5), month 12 (n=9), month 24 (n=19), month 36 (n=22), month 48 (n=28), and month 60 (n=34).

Safety Outcomes

Through 60 months of follow-up, no significant differences were observed between patients in the tacrolimus and cyclosporine groups in the proportion of patients with glucose more than 126 mg/dL (46.2% for tacrolimus and 44.6% for cyclosporine treatment), hemoglobin A1C more than 6.0%, total cholesterol more than 200 mg/dL, high-density lipoprotein less than 40 mg/dL, or triglycerides more than 200 mg/dL (Table 4). However, significantly fewer patients in the tacrolimus group than in the cyclosporine group had values of blood urea nitrogen more than 50 mg/dL (47% vs. 64%, P=0.036) or LDL cholesterol more than 130 mg/dL (29% vs. 57%, P=0.002) at any time over the 60-month follow-up period. The incidence of diabetes, hyperglycemia, hypertension, lymphoma, or malignancies among patients who did not have these conditions at baseline was generally low and comparable between groups. Among patients who had follow-up data available pertaining to cardiac conditions, significantly fewer patients in the tacrolimus group (13 of 123, 10.6%) developed new cardiac conditions than in the cyclosporine group (16 of 58, 27.6%; P=0.004). Conditions from most common to least common included coronary artery disease, myocardial infarction, congestive heart failure, angina, atrial fibrillation, tachycardia, angioplasty, atrial flutter, endocarditis, pericarditis, left ventricular hypertrophy, valvular disease, murmur, pacemaker inserted, and cerebrovascular accident. Additionally, a lower proportion of patients in the tacrolimus group (4 of 17, 24%) than in the cyclosporine group (six of nine, 67%; P=0.046) developed hyperlipidemia during the follow-up period. New opportunistic infections (lack of cytomegalovirus infection at baseline; 110 tacrolimus- and 50 cyclosporine-treated patients) were reported during the 60 month follow-up period by 24 (22%) patients in the tacrolimus group and 13 (26%) patients in the cyclosporine group (P=0.56).

TABLE 4: Relevant safety outcomes through 60 months

Twenty-six of 126 (21%) patients in the tacrolimus group and 14 of 60 (23%) patients in the cyclosporine group discontinued their randomized treatment for reasons other than graft loss or death. In each group, adverse events (not including refractory rejection) led to treatment discontinuation in approximately 8% of these patients.

The incidence of serious adverse events was similar for both treatment groups. Thirty-four (27%) patients in the tacrolimus group and 19 (32%) patients in the cyclosporine group experienced serious adverse events, of whom 12 (10%) and 8 (13%) patients, respectively, experienced serious adverse events that were considered possibly or probably related to use of the study drug.

DISCUSSION

Patients with functional grafts after 2 and 5 years of follow-up experienced significant improvements in renal function with tacrolimus treatment compared with patients who remained on cyclosporine. Among this cohort, the median serum creatinine level was lower, and estimated creatinine clearance was higher in the tacrolimus group than in the cyclosporine group throughout the 60-month follow-up period. Notably, compared with results of the 2-year analysis (²³), further reductions in median serum creatinine were observed with conversion to tacrolimus at 5 years (2.3 vs. 2.0 mg/dL, respectively). In contrast, the median serum creatinine level had increased at 5 years with ongoing cyclosporine use compared with results at 2 years (2.7 mg/dL vs. 2.6 mg/dL, respectively).

Our findings substantiate results of other studies comparing the long-term effects of tacrolimus and cyclosporine on renal function. In a 2-year study of 223 allograft recipients, renal function estimated by serum creatinine levels was significantly better (P=0.04) in patients treated with tacrolimus and MMF or tacrolimus and azathioprine than in patients treated with cyclosporine and MMF at 24 months (²⁶). Similarly, results of a long-term study of 412 renal transplant recipients showed a significantly lower median serum creatinine level at 5 years among patients in the tacrolimus group (1.4 mg/dL) compared with those in the cyclosporine group (1.7 mg/dL, P=0.0014) (²⁰). A European trial of 451 renal transplant patients reported lower serum creatinine levels at 5 years in patients treated with tacrolimus (1.70 mg/dL) compared with patients treated with cyclosporine (1.74 mg/dL); a significantly lower incidence of chronic rejection was observed in these tacrolimus-treated patients (²⁷). Thus, compared with cyclosporine-based immunosuppression therapy, tacrolimus-based regimens are consistently associated with improved renal function.

It has been established that episodes of acute rejection, particularly late rejection episodes, increase the risk of CRAF (^{8, 9}). In this study, switching from a cyclosporine-based to a tacrolimus-based immunosuppression regimen was not associated with an increase in the frequency of episodes of late-onset acute or chronic rejection. Overall, patients in both groups experienced similar rates of acute rejection (approximately 8% in each group, P=0.93) and chronic rejection (approximately 25% in each group, P=0.86).

Conservation of creatinine levels and favorable changes occurring in these levels within the first year after transplant correlate best with long-term graft survival (²¹). Despite improvements in renal function with tacrolimus treatment after 5 years, no differences in patient and graft survival were observed between treatment groups. However, at the time of enrollment, an average time period of 5.3 years for cyclosporine-treated patients and 4.7 years for tacrolimus-treated patients had elapsed from transplant to enrollment. Mean serum creatinine was 2.5 mg/dL for both treatment groups at baseline and approximately 90% of patients had already experienced CAN. Consequently, kidneys from both patient cohorts were compromised, and most patients were not optimal candidates for continued long-term graft survival. Improved renal function occurred probably after irreversible damage to the kidneys had been sustained; too late for any measurable change in patient or graft survival.

There were limitations to this study. Patients were recruited into this study strictly on the basis of serum creatinine concentration. Moreover, even though the protocol stated that a biopsy would be performed after 60 months, the procedure was not completed for many patients. Of the patients who completed the study with a functioning graft, 10 of 21 (47.6%) patients treated with cyclosporine and 21 of 45 (46.7%) patients treated with tacrolimus did not have biopsies. Furthermore, the study was not powered to detect subpopulations of tacrolimus-treated patients that may have shown better graft stability.

Our findings also demonstrated that switching from cyclosporine to tacrolimus was not associated with an increase in the incidence of coexisting conditions or other risk factors for CRAF. The proportion of patients who developed diabetes, hyperglycemia, hypertension, lymphoma, and malignancies was similar for both groups. Results of a previous comparative study of tacrolimus and cyclosporine demonstrated an increased risk of posttransplant diabetes of approximately 15% with use of tacrolimus compared with use of cyclosporine (²⁸). More recent studies have disputed these findings, and have demonstrated that use of tacrolimus is not associated with a significant increased risk of new-onset diabetes or hyperglycemia (^{26, 29}), possibly because of better patient management and new adjunctive therapies. Our findings concur. Additionally, no significant difference in the occurrence of infections was observed between groups.

As the study progressed, differences in new cardiac conditions and in metabolic profile favoring tacrolimus treatment became apparent. Significantly fewer patients in the tacrolimus group than in the cyclosporine group developed new cardiac conditions (11% vs. 28%; P=0.004) or had LDL cholesterol values more than 130 mg/dL over follow-up. Additionally, fewer patients in the tacrolimus group than in the cyclosporine group developed hyperlipidemia (24% vs. 67%, P=0.046) over follow-up. There was a small number of patients without hyperlipidemia at baseline n=17 in the tacrolimus group and n=9 in the cyclosporine group. Although current definitions for diabetes mellitus, hyperglycemia, hypertension, and hyperlipidemia have become more stringent since this study was initiated, these collected data are still relevant and are indicative of a worsening health status. These observations are also important to patient health because cardiovascular disease (hyperlipidemia is a risk factor) is a significant cause of death in renal transplant patients (^{22, 30}).

Both the tacrolimus and cyclosporine treatments were well tolerated. The proportion of patients who discontinued their randomized treatment because of adverse events and who experienced serious adverse events considered possibly or probably related to use of the study drug was similar for both treatment groups.

This study demonstrates the efficacy and safety of switching from cyclosporine-based to tacrolimus-based immunosuppression in patients at risk for CRAF, and corroborates findings from the 2-year report (²³). Two and 5 years after switching treatment regimens, tacrolimus resulted in improved renal function in patients with functioning grafts and a reduction in the occurrence of new-onset cardiac conditions and hyperlipidemia, without a subsequent increase in the incidence of new-onset diabetes, new-onset hyperglycemia, infections, or other significant conditions. Acute and chronic rejection rates were comparable between treatment groups. Despite improvements in renal function and overall health status observed with tacrolimus, treatment did not correlate with improved patient and graft survivals.

ACKNOWLEDGMENTS

The manuscript was drafted by the medical writing department of Astellas Pharma US, Inc. (Yolanda Cartwright, PhD, and Sharon Rogers, PhD), which also provided editorial assistance to the authors. The authors made a substantial contribution to the study’s conception and design, acquisition of data, and analysis and interpretation of data; participated in the drafting and review of the manuscript; and approved the final version.

APPENDIX

Members of the CRAF Study Group

Members of the CRAF Study Group include: Thomas Waid, MD (University of Texas Medical Center, Lexington, KY); Edward Alfrey, MD (MS Hershey Medical Center, Hershey, PA); Laura C. Mulloy, DO (Medical College of Georgia, Augusta, GA); Fuad S. Shihab, MD (University of Utah, Salt Lake City, UT); David Conti, MD (Albany Medical College, Albany, NY); Richard Freeman, MD (New England Medical Center, Boston, MA); Angelo M. deMattos, MD (Oregon Health and Science University, Portland, OR); Stephen C. Jensik, MD, PhD (Rush Presbyterian-St. Luke’s Medical Center, Chicago, IL); Stanley Jordan, MD (Cedars Sinai Medical Center, Los Angeles, CA); George C. Francos, MD (Thomas Jefferson University Hospital, Philadelphia, PA); David Van Buren, MD (Vanderbilt University, Nashville, TN); Larry Chan, MD (University of Colorado, Denver, CO); Robert W. Steiner, MD (University of California San Diego Medical Center, San Diego, CA); Giacomo Basadonna, MD (Yale University School of Medicine, New Haven, CT); Karl Brinker, MDCM, FRCP (Dallas Transplant Institute, Dallas, TX); Steven Steinberg, MD (Sharp Memorial Hospital, San Diego, CA); Arthur J. Matas, MD (University of Minnesota, Minneapolis, MN); Anne L. King, MD (Medical College of Virginia, Richmond, VA); Bertram L. Kasiske, MD (Hennepin County Medical Center, Minneapolis, MN); David J. Cohen, MD (New York Presbyterian Hospital-Columbia, New York, NY); David Surer, MD (New York Presbyterian Hospital Rogosin Institute, New York, NY); Sharon Inokuchi, MD, PharmD (California Pacific Medical Center, San Francisco, CA); John D. Pirsch, MD (University of Wisconsin, Madison, WI); Jonathan S. Bromberg, MD, PhD (Mount Sinai Medical Center, New York, NY); Matthew R. Weir, MD (University of Maryland, Baltimore, MD); Stuart M. Greenstein, MD (Montefiore Medical Center, Bronx, NY); Stephen J. Tomlanovich, MD (UCSF Medical Center, San Francisco, CA); Robert Mendez, MD (St. Vincent/NIT, Los Angeles, CA); Lawrence Kahana, MD (Tampa General Hospital, Tampa, FL); Alice K. Henning, MS (The EMMES Corporation, Rockville, MD); M. Roy First, MD, William E. Fitzsimmons, PharmD, Kim Salm, RN, Diane Tolzman, RN (Astellas Pharma US, Deerfield, IL).