- Top of page
- Materials and Methods
We conducted a multicenter randomized study in liver transplantation to compare standard-dose tacrolimus to reduced-dose tacrolimus with mycophenolate mofetil to reduce the occurrence of tacrolimus side effects. Two primary outcomes (censored criteria) were monitored during 48 weeks post-transplantation: occurrence of renal dysfunction or arterial hypertension or diabetes (evaluating benefit) and occurrence of acute graft rejection (evaluating risk). Interim analyses were performed every 40 patients to stop the study in the case of increased risk of graft rejection. One hundred and ninety-five patients (control: 100; experimental: 95) had been included when the study was stopped. Acute graft rejection occurred in 46 (46%) and 28 (30%) patients in control and experimental groups, respectively (HR = 0.59; 95% CI: [0.37–0.94]; p = 0.024). Renal dysfunction or arterial hypertension or diabetes occurred in 80 (80%) and 61 (64%) patients in control and experimental groups, respectively (HR = 0.68; 95% CI: [0.49–0.95]; p = 0.021). Renal dysfunction occurred in 42 (42%) and 23 (24%) patients in control and experimental groups, respectively (HR = 0.49; 95% CI: [0.29–0.81]; p = 0.004). Leucopoenia (p = 0.001), thrombocytopenia (p = 0.017) and diarrhea (p = 0.002) occurred more frequently in the experimental group. Reduced-dose tacrolimus with mycophenolate mofetil reduces the occurrence of renal dysfunction and the risk of graft rejection. This immunosuppressive regimen could replace full-dose tacrolimus in adult liver transplantation.
- Top of page
- Materials and Methods
Long-term immunosuppression using CNIs is essential to patients undergoing liver transplantation. Randomized studies have revealed lower graft rejection rates with tacrolimus than cyclosporine (1–3). According to the manufacturer's recommendations, an initial oral dose of 0.15 mg/kg must be administered to obtain whole blood trough levels between 10 and 20 ng/mL.
With such a regimen, renal dysfunction, arterial hypertension, or diabetes mellitus are common (4–6) and major causes of long-term mortality (7–9). These dose- and blood concentration-dependent effects (10) occur early in the post-transplant period.
Randomized studies have reported that reduced-dose tacrolimus regimens exhibit beneficial effects on post-transplant renal dysfunction. Renal dysfunction may be partially reversed by tapering tacrolimus dosage or substituting tacrolimus with MMF, a T- and B-cell-specific immunosuppressant, with nonimmunological side effects differing from those of CNIs (11). A pre-emptive strategy using antibody induction therapy with interleukin-2 receptor antagonists and delayed reduced-dose tacrolimus has also shown lower nephrotoxicity (12,13), but beneficial effects on post-transplant arterial hypertension and diabetes have not been reported to date.
We tested the hypothesis whether de novo reduced-dose tacrolimus could decrease renal dysfunction, arterial hypertension and diabetes mellitus in first liver transplant patients. Patients on reduced-dose tacrolimus received MMF to reduce acute graft rejection risk.
- Top of page
- Materials and Methods
In total, 195 patients were recruited between May 6, 2003 and December 3, 2007. Although the first two analyses did not reveal any significant between-group difference in acute graft rejection, the third analysis showed a trend toward increased risk in the control group (p = 0.091), which was confirmed by the fourth analysis (p=0.005). The results are summarized in Table 1. As these results were not in line with our hypothesis, the DSMB, meeting on December 19, 2007, did not recommend stopping recruitment. However, the Scientific Committee, meeting on February 28, 2008, decided to stop the study based on the following: (i) slow recruitment rates since late 2005 and (ii) significant between-group differences in terms of acute graft rejection rates.
Table 1. Results of interim analyses on the occurrence of acute graft rejection
| ||Active control||Experimental||OR [95% CI] HR [95% CI]||p-Value|
|First analysis||10/21 (48%)||7/19 (37%)||0.64 [0.18–2.28]||0.492|
| (40 patients)|
| March 9, 2006|| || ||0.86 [0.33–2.25]||0.754|
|Second analysis||23/44 (52%)||15/36 (42%)||0.65 [0.27–1.59]||0.346|
| (80 patients)|
| September 8, 2006|| || ||0.80 [0.42–1.54]||0.507|
|Third analysis||32/62 (52%)||21/58 (36%)||0.53 [0.26–1.11]||0.091|
| (120 patients)|
| February 16, 2007|| || ||0.63 [0.36–1.10]||0.106|
|Fourth analysis||45/85 (53%)||23/75 (31%)||0.39 [0.21–0.75]||0.005|
| (160 patients)|
| December 3, 2007|| || ||0.49 [0.30–0.82]||0.006|
Study flow chart (Figure 1)
Among the 195 patients, 100 were assigned to the control group and 95 to the experimental group. Except for one patient who died prior to receiving randomized treatment, all patients were given the allocated immunosuppression. At week 48, 93 (93%) and 91 (96%) control and experimental patients, respectively, were still alive (p = 0.399). Graft survival rates were 92% and 94% in control and experimental patients, respectively (p = 0.649).
Except for gender (p = 0.040), there was no significant between-group difference in any of the variables. The results are summarized in Table 2. All patients with hepatocellular carcinoma had underlying cirrhosis. Among cirrhotic patients, 45% and 46% were classified child C in control and experimental groups, respectively.
Table 2. Patients’ characteristics at inclusion
| ||Active control (n = 100)||Experimental (n = 95)||All patients (n = 195)|
| Cirrhosis||61 (61%)||56 (59%)||117 (60%)|
| Alcoholic||44 (44%)||39 (41%)||83 (43%)|
| Hepatitis B||1 (1%)||3 (3%)||4 (2%)|
| Hepatitis C||7 (7%)||9 (10%)||16 (8%)|
| Other||9 (9%)||5 (5%)||14 (7%)|
| Hepatocellular carcinoma||32 (32%)||36 (38%)||68 (35%)|
| Other malignancy||1 (1%)||1 (1%)||2 (1%)|
| Benign tumoral disease||2 (2%)||0 (0%)||2 (1%)|
| Chronic cholestatic disease||4 (4%)||2 (2%)||6 (3%)|
|Age (years), median (IQR)||52 (45–57)||51 (44–57)||51 (44–57)|
|Weight (kg), median (IQR)||71 (61–83)||75 (64–83)||73 (63–83)|
|Plasma creatinine (μmol/L), median (IQR)||75 (64–88)||77 (65–95)||75 (64–92)|
|GFR (mL/min), median (IQR)||99 (81–121)||101 (80–133)||100 (81–127)|
|SBP (mmHg), median (IQR)||120 (110–130)||120 (110–130)||120 (110–130)|
|DBP (mmHg), median (IQR)||70 (60–75)||70 (60–80)||70 (60–75)|
|Plasma glucose (mmol/L), median (IQR)||5.6 (5.1–6.4)||5.7 (5.1–6.5)||5.6 (5.1–6.4)|
|Donor age (years), median (IQR)||48 (34–62)||52 (38–63)||48 (37–63)|
|Liver CIT (h), median (IQR)||8.8 (7.0–11.4)||9.3 (6.9–11.3)||9.0 (7.0–11.3)|
Given graft rejection risk, 46 (46%) control and 28 (30%) experimental patients had at least one acute rejection episode. Most of these episodes occurred within 2–3 weeks following transplantation, the between-group difference being apparent within the first month (Figure 2A, p = 0.024). Acute rejection grade was minor, moderate, and severe in 15 and 5, 24 and 18, and 6 and 1 control and experimental patients, respectively. Chronic rejection histological features were present in two experimental patients. No biopsy was performed after suspicion of rejection because of thrombocytopenia in 1 and 2 control and experimental patients, respectively. At week 48, a routine biopsy could be performed in 72 and 69 control and experimental patients, respectively. This biopsy did not allow detecting any new first rejection episode.
Figure 2. (A) Kaplan–Meier distributions of occurrence of acute graft rejection (risk) and (B) Kaplan–Meier distributions of occurrence of renal dysfunction or arterial hypertension or diabetes (benefit).
Download figure to PowerPoint
Regarding adverse reactions, 80 (80%) and 61 (64%) control and experimental patients developed renal dysfunction, arterial hypertension or diabetes mellitus. The difference between groups became apparent after the first month (Figure 2B, p = 0.021). The results are summarized in Table 3.
Table 3. Primary outcomes at 48 weeks
| ||Active control (n = 100)||Experimental (n = 95)||OR [95% CI] HR [95% CI]||p-Value|
|Risk: acute graft rejection||46 (46%)||28 (30%)||0.49 [0.27–0.89]||0.018|
| || ||0.59 [0.37–0.94]||0.024|
|Benefit: long-term side effects||80 (80%)||61 (64%)||0.45 [0.24–0.86]||0.015|
| || ||0.68 [0.49–0.95]||0.021|
| Renal dysfunction||42 (42%)||23 (24%)||0.44 [0.24–0.82]||0.009|
| || ||0.49 [0.29–0.81]||0.004|
| Arterial hypertension||56 (56%)||44 (46%)||0.68 [0.39–1.19]||0.177|
| || ||0.75 [0.51–1.12]||0.156|
| Diabetes||35 (35%)||31 (33%)||0.90 [0.50–1.63]||0.727|
| || ||0.95 [0.59–1.54]||0.836|
There were fewer patients who developed renal dysfunction in the experimental than in the control group (Figure 3A, p = 0.004). Before transplantation, GFR was within normal ranges in both groups (Table 2). At study end, GFR was higher (90 ± 30 mL/min vs. 78 ± 26 mL/min, p = 0.018) in the experimental than the control group. In the experimental group, fewer patients developed arterial hypertension, though the differences did not reach statistical significance (Figure 3B, p = 0.156). Finally, there was no between-group difference in diabetes mellitus (Figure 3C, p = 0.836). Results are summarized in Table 3.
Figure 3. (A) Kaplan–Meier distributions of occurrence of renal dysfunction, (B) Kaplan–Meier distributions of occurrence of arterial hypertension and (C) Kaplan–Meier distributions of occurrence of diabetes.
Download figure to PowerPoint
Tacrolimus daily doses and trough concentrations
As expected, tacrolimus intake was lower in the experimental group (Figure 4A). In this group, mean trough plasma concentrations of tacrolimus were at the upper limit of the designed range throughout the study, while they tended to decrease in the control group, approaching the lower limits of the designed range (Figure 4B).
Figure 4. (A) Mean dose of tacrolimus over the 48 weeks of follow up (mg/day) and (B) mean trough concentration of tacrolimus over the 48 weeks of follow up (ng/mL).
Download figure to PowerPoint
No between-group differences were observed except for diarrhea (p = 0.002), leucopenia (p = 0.001), and thrombocytopenia (p = 0.017), which were more frequent in the experimental group. Results are summarized in Table 4. In the 28 (29%) experimental patients who had leucopenia, the daily dose of MMF was 2, 2.5 and 3 g for 18, 1 and 9 patients, respectively. In the 11 (11%) control patients who had leucopenia, 4 received MMF at a daily dose of 1.5 g for 1 patient and 3 g for 3. In the 13 (14%) experimental patients who had thrombocytopenia, the daily dose of MMF was 2 g or less and 3 g for 7 and 6 patients, respectively. In the 4 (4%) control patients who had thrombocytopenia, none received MMF.
Table 4. Adverse events (other than renal dysfunction, arterial hypertension, or diabetes)
| ||Active control||Experimental|
|(n = 100)||(n = 95)|
|CMV infections||10 (10%)||10 (11%)|
|Extrahepatic infections, excluding CMV infections||42 (42%)||40 (42%)|
| Confusion||17 (17%)||17 (18%)|
| Convulsion||7 (7%)||12 (13%)|
| Shakiness||7 (7%)||5 (5%)|
| Sensorimotor disorder||6 (6%)||5 (5%)|
| Encephalopathy||2 (2%)||2 (2%)|
|Blood system disorders|
| Leucopenia||11 (11%)||28 (29%)|
| Anemia||9 (9%)||13 (14%)|
| Thrombocytopenia||4 (4%)||13 (14%)|
| Aplasia bone marrow||1 (1%)||1 (1%)|
|Diarrhea||7 (7%)||22 (23%)|
| Arrhythmia||6 (6%)||4 (4%)|
| Cardiac failure||2 (2%)||1 (1%)|
| Hypotension||1 (1%)||2 (2%)|
| Pulmonary hypertension||2 (2%)||0 (0%)|
| Myocarditis||0 (0%)||1 (1%)|
| Phlebitis||1 (1%)||0 (0%)|
| Arteritis obliterans||0 (0%)||1 (1%)|
| Pericardial effusion||0 (0%)||1 (1%)|
| Thoracic pain||1 (1%)||1 (1%)|
| Cardiac arrest||0 (0%)||1 (1%)|
|Respiratory failure||7 (7%)||6 (6%)|
|Hyperkaliemia||3 (3%)||7 (7%)|
| New||1 (1%)||3 (3%)|
| Recurrence of hepatocellular carcinoma||1 (1%)||2 (2%)|
|Hypercholesterolemia||2 (2%)||1 (1%)|
Rejection episodes were treated with increased tacrolimus dosage in 29 (63%) and 22 (79%) control and experimental patients, respectively.
Tacrolimus to cyclosporine switch was performed in 3 control (because of neurological complications for 2 and renal failure for 1) and 2 experimental (because of neurological complications) patients, respectively. One experimental patient was switched from tacrolimus to sirolimus (because of hepatocellular carcinoma and hepatitis C infection recurrence).
The evolution of the proportion of patients receiving MMF and steroids (with daily doses) during the study period is reported in Table 5. At week 48, in the experimental group, MMF had been withdrawn in 11 (13%) patients for the following reasons: blood system disorders for 6 patients, cytomegalovirus infection for 2 patients, hepatitis C infection recurrence for 1 patient, diarrhea for 1 patient and unknown reason for 1 patient. In the control group, MMF had been introduced and maintained until week 48 in 19 (21%) patients for the following reasons: tacrolimus adverse event for 12 patients, acute rejection episode for 7 patients. In contrast, steroids exposure was similar in both groups all along the study period.
Table 5. MMF and steroids intakes during the study period
| || ||Active control (n = 100)||Experimental (n = 95)|
|MMF intake at|
| End of week 1||Number of patients||0/99 (0%)||93/94 (99%)|
|Daily dose (mg)||–||2825 ± 494|
| End of week 6||Number of patients||14/96 (15%)||91/94 (97%)|
|Daily dose (mg)||1983 ± 626||2703 ± 548|
| End of week 12||Number of patients||11/95 (12%)||87/94 (93%)|
|Daily dose (mg)||1727 ± 684||1994 ± 448|
| End of week 24||Number of patients||13/93 (14%)||81/92 (88%)|
|Daily dose (mg)||1615 ± 682||1852 ± 434|
| End of week 48||Number of patients||19/92 (21%)||75/86 (87%)|
|Daily dose (mg)||1526 ± 589||1757 ± 470|
|Steroids intake at|
| End of week 1||Number of patients||98/99 (99%)||93/94 (99%)|
|Daily dose (mg)||22 ± 8||20 ± 5|
| End of week 6||Number of patients||96/96 (100%)||93/94 (99%)|
|Daily dose (mg)||16 ± 5||15 ± 6|
| End of week 12||Number of patients||94/95 (99%)||91/94 (97%)|
|Daily dose (mg)||10 ± 5|| 9 ± 3|
| End of week 24||Number of patients||80/93 (86%)||78/92 (85%)|
|Daily dose (mg)|| 6 ± 6|| 5 ± 4|
| End of week 48||Number of patients||26/92 (28%)||20/86 (23%)|
|Daily dose (mg)|| 5 ± 4|| 4 ± 2|
- Top of page
- Materials and Methods
In this study, renal dysfunction, arterial hypertension, or diabetes mellitus occurred in 80% of liver transplant patients treated with standard-dose tacrolimus and low-dose steroids. De novo immunosuppression with reduced-dose tacrolimus under the protection of MMF and low-dose steroids was shown to significantly reduce adverse reaction incidence. Improvements were mainly accounted for by significantly less incident renal dysfunction and a trend toward less incident arterial hypertension. There was no difference in incident diabetes mellitus between groups. In contrast with the initial hypothesis (absence of difference between the two groups on the incidence of acute graft rejection), this strategy of adding MMF to reduced-dose tacrolimus also allowed for a significant reduction in acute rejection episodes.
Renal dysfunction is common in liver transplant patients (7,14,15). In this study, renal dysfunction was considered only if it persisted or appeared after week 9 post-transplant, a delay chosen to exclude patients with altered renal function due to postoperative antibiotic toxicity or tubular necrosis caused by hemodynamic disturbances. Renal dysfunction rates at 1-year post-transplant were reduced by nearly half in the tacrolimus reduced-dose group, a benefit already apparent at the first postoperative month and which was maintained throughout the study period. Renal failure detection was based on plasma creatinine alone as this is the only routinely measured parameter in transplant patient follow-up. In patients with essentially normal renal function, plasma creatinine increases during CNI therapy are closely related to decreasing GFR (16).
The mechanism underlying initial reduced-dose tacrolimus benefits to renal function must still be elucidated. One explanation is that tacrolimus nephrotoxicity is dose dependent, especially within the initial post-transplant period, when the kidney is at risk of ischemia reperfusion due to perioperative events, such as vascular clamping, hemodynamic instability, and multiple drug toxicity. Another explanation is that MMF, which was added to prevent an increased acute graft rejection risk, protected the kidney. Indeed, MMF has been shown to display antifibrotic effects which may counteract CNIs’ harmful effects on endothelial cells (17).
Two recently published randomized trials have evaluated the strategy of preventing renal nephrotoxicity by de novo reduced-dose tacrolimus. Yoshida et al. (12) compared standard induction and maintenance tacrolimus dosage with an MMF plus delayed/reduced-dose tacrolimus protocol. Improved GFR was observed in the investigational arm until month 1 post-transplant, with no significant differences thereafter. In a three-arm study, Neuberger et al. (13) compared standard-dose tacrolimus, MMF plus reduced-dose tacrolimus, and MMF plus anti-IL2R mAb plus delayed reduced-dose tacrolimus protocol. Improved renal function was observed at 1 year post-transplant in patients receiving reduced-dose tacrolimus; however the differences only became significant for delayed reduced-dose tacrolimus. However, this study had very high rates of premature termination (30–45%). Moreover, tacrolimus levels in experimental groups remained close to standard-dose levels. In our study, with the exception of three patients, all surviving patients were followed up until week 48, and the analysis was performed on an intent-to-treat basis. Tacrolimus levels were closely monitored and at each interim analysis, DSMB advised researchers to stick to recommended ranges. This may explain why (i) tacrolimus trough levels were lower in the experimental than the control group, and (ii) the experimental group had a reduced incidence of tacrolimus nephrotoxicity.
In the majority of transplant patients, arterial pressure usually increases to hypertensive levels. In our study, 56% of control patients and 46% of experimental patients developed de novo arterial hypertension, which compares favorably to the 75–85% reported in the scientific literature (5,18–20). In liver transplant patients, arterial hypertension has multiple causes, mainly direct alteration of the vascular endothelial tone (21). However, sodium retention due to renal dysfunction and mineralocorticoid effects of corticosteroids are interacting factors present in both groups, and may explain why the between-group differences did not reach significance.
A comprehensive review of trials in liver transplant recipients receiving CNI-based therapy revealed an overall incidence of diabetes mellitus of approximately 15% (22). De novo diabetes mellitus was found in one-third of our patients. This higher rate observed in our study might be explained by the definition of diabetes we used, i.e. that of the American Diabetes Association and World Health Organization. As the incidence was similar in both groups, this might reflect that the threshold for tacrolimus toxicity on pancreatic beta cells is lower than the dosages used in our study groups. Although emphasis has often been placed on CNIs, there are other risk factors for diabetes mellitus following liver transplantation, such as the diabetogenic effects of corticosteroid treatments and HCV infection (23). Because reduced-doses of tacrolimus could have increased the incidence of acute graft rejection, we gave MMF to the experimental group. MMF has no renal, vascular, or pancreatic toxicity; and has been shown to be protective against acute graft rejection in post-transplant CNI withdrawal strategies. A daily dose of 3 g was given as MMF bioavailability has been shown to be reduced in liver transplant patients as compared to kidney transplant patients, mainly due to impaired glucuroconjugation due to liver dysfunction (24). MMF's excellent efficacy results must be weighed against its adverse reactions. Although MMF administration in the experimental group was associated with gastrointestinal and bone marrow suppressive effects, these were reversible when the dose was reduced, and there were few treatment discontinuations. As MMF is known to prevent acute rejection without tacrolimus’ side effects, it is tempting to imagine a protocol without tacrolimus so as to avoid renal dysfunction, arterial hypertension, or diabetes mellitus. However, in studies, MMF monotherapy in liver transplantation was associated with an unacceptable risk of acute graft rejection and subsequent graft loss (25). This observation suggests that tacrolimus must be added to MMF, although the minimum daily dose has not yet been determined. Administering MMF led to less acute graft rejections, thereby reducing the need for increased corticosteroid and tacrolimus doses to treat these rejection episodes. This could explain why renal dysfunction and arterial hypertension were observed at a lower rate in the experimental group.
Some methodological aspects of our study have to be considered. Firstly, exclusion criteria may have been too restrictive, particularly with the exclusion of patients with pre-existing diabetes mellitus, hypertension or renal dysfunction. Although this may reduce the generalization of study results, it gives credit to the relation between tacrolimus intake and side-effect occurrence. Secondly, the adverse effects of tacrolimus in controls may have been boosted. Indeed, the initial trough levels of tacrolimus, although within the designed range of the study, may have been higher than is aimed for by an experienced high-volume center. This may be why in the control arm the tendency of clinicians was to reduce tacrolimus doses in order to avoid toxicity. Interestingly, in the experimental arm, the tendency was to increase doses in order to avoid graft rejection. A similar tendency has been reported by others (13). However, in our study, a significant difference persisted between groups until week 20 post-transplant. Thirdly, the follow-up was only 1 year, a period of time that might be insufficient to reveal long-term CNI side effects. However, it has been shown that early kidney function is predictive of the long-term development of end-stage renal failure and increased mortality (15). Finally, it should be noted that, in our study, HCV- and HBV-related liver disease represented only 10% of transplantation indications. Therefore, our data may not be easily extrapolated to populations with high rates of HCV- and HBV-related cirrhosis and for which MMF introduction seems to increase the risk of HCV-induced allograft injury (26).
In summary, even though clinical practice has moved away from the target tacrolimus concentrations recommended in the product license on the grounds of toxicity, the beneficial effect of de novo reduced-dose tacrolimus immunosuppression has not been demonstrated yet. In renal transplantation, the Symphony study (27) succeeded in overcoming this restriction. Our study clearly brings new data in this direction showing that de novo reduced-dose tacrolimus, in combination with MMF and steroids, reduced the incidence of renal dysfunction, reduced the rate of acute graft rejection and showed a trend toward reducing the incidence of arterial hypertension. Although this trial did not show a significant effect on the incidence of diabetes mellitus, this ‘renal sparing’ immunosuppression combination could be the protocol of choice in adult liver transplantation.